modelId
stringlengths 5
139
| author
stringlengths 2
42
| last_modified
timestamp[us, tz=UTC]date 2020-02-15 11:33:14
2025-08-30 06:27:36
| downloads
int64 0
223M
| likes
int64 0
11.7k
| library_name
stringclasses 527
values | tags
listlengths 1
4.05k
| pipeline_tag
stringclasses 55
values | createdAt
timestamp[us, tz=UTC]date 2022-03-02 23:29:04
2025-08-30 06:27:12
| card
stringlengths 11
1.01M
|
|---|---|---|---|---|---|---|---|---|---|
LarryAIDraw/miniyaemori-lora-nochekaiser
|
LarryAIDraw
| 2023-08-30T05:49:46Z | 0 | 0 | null |
[
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-08-30T05:33:57Z |
---
license: creativeml-openrail-m
---
https://civitai.com/models/136583/mini-yaemori-rental-girlfriend-commission
|
pawankmrs/BioMedLM-ner-small-training
|
pawankmrs
| 2023-08-30T05:44:22Z | 0 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-08-30T05:44:21Z |
---
library_name: peft
---
## Training procedure
### Framework versions
- PEFT 0.5.0
|
LarryAIDraw/IlyaPlatiere-08
|
LarryAIDraw
| 2023-08-30T05:42:03Z | 0 | 0 | null |
[
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-08-30T05:42:03Z |
---
license: creativeml-openrail-m
---
|
Nikolai1902/LL2-13B-FIMFiction-QLORA-GGML
|
Nikolai1902
| 2023-08-30T05:40:58Z | 0 | 0 | null |
[
"license:wtfpl",
"region:us"
] | null | 2023-08-29T23:43:35Z |
---
license: wtfpl
---
This model was a QLoRA of LLaMA 2-13B base finetuned on the FIMFiction archive and then merged with the base model (as most GGML loading apps don't support LoRAs), and quantized for llama.cpp-based frontends. Was trained with 1024 context length
There are two options, depending on the resources you have:
- Q5_K_M: Low quality loss K-quantized 5 bits model. Max RAM consumption is 11.73 GB, recommended if you have 12GB of VRAM to load 40 layers
- Q4_K_S: Compact K-quantized 4 bits. Max RAM consumption is 9.87 GB
This not an instruction tuned model, it was trained on raw text, so treat it like an autocomplete.
Seems sensitive to formatting: I found it's usually better at staying on topic when using double spacing in the prompt.
Only tags excluded from the dataset are eqg and humanized.
|
LarryAIDraw/rem_v1
|
LarryAIDraw
| 2023-08-30T05:40:35Z | 0 | 0 | null |
[
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-08-30T05:40:35Z |
---
license: creativeml-openrail-m
---
|
LarryAIDraw/djeeta_20230722
|
LarryAIDraw
| 2023-08-30T05:40:07Z | 0 | 0 | null |
[
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-08-30T05:40:07Z |
---
license: creativeml-openrail-m
---
|
BM-K/NewsKoT5-small
|
BM-K
| 2023-08-30T05:37:14Z | 109 | 0 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"t5",
"text2text-generation",
"ko",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-06-01T07:36:27Z |
---
language:
- ko
---
# NewsKoT5
The training data for this T5 model consists of Korean news articles (29GB). However, the performance has not been fine-tuned through the use of small batches and a limited number of training steps, so it may not be fully optimized.
## Quick tour
```python
from transformers import AutoTokenizer, T5ForConditionalGeneration
tokenizer = AutoTokenizer.from_pretrained("BM-K/NewsKoT5-small")
model = T5ForConditionalGeneration.from_pretrained("BM-K/NewsKoT5-small")
input_ids = tokenizer("한국형발사체 누리호가 실용급 <extra_id_0> 발사체로서 ‘데뷔’를 성공적으로 <extra_id_1>", return_tensors="pt").input_ids
labels = tokenizer("<extra_id_0> 위성 <extra_id_1> 마쳤다 <extra_id_2>", return_tensors="pt").input_ids
outputs = model(input_ids=input_ids,
labels=labels)
```
## News Summarization Performance (F1-score)
After restoring the model's tokenized output to the original text, Rouge performance was evaluated by comparing it to the reference and hypothesis tokenized using [mecab](https://konlpy.org/ko/v0.4.0/).
- Dacon 한국어 문서 생성요약 AI 경진대회 [Dataset](https://dacon.io/competitions/official/235673/overview/description)
- Training: 29,432
- Validation: 7,358
- Test: 9,182
| | #Param | rouge-1 |rouge-2|rouge-l|
|-------|--------:|--------:|--------:|--------:|
| pko-t5-small | 95M | 51.48 | 33.18 | 44.96 |
| NewsT5-small | 61M | 52.15 | 33.59 | 45.41 |
- AI-Hub 문서요약 텍스트 [Dataset](https://www.aihub.or.kr/aihubdata/data/view.do?currMenu=115&topMenu=100&aihubDataSe=realm&dataSetSn=97)
- Training: 245,626
- Validation: 20,296
- Test: 9,931
| | #Param | rouge-1 |rouge-2|rouge-l|
|-------|--------:|--------:|--------:|--------:|
| pko-t5-small | 95M | 53.44 | 34.03 | 45.36 |
| NewsT5-small | 61M | 53.74 | 34.27 | 45.52 |
- [pko-t5-small](https://github.com/paust-team/pko-t5)
|
BM-K/KoMiniLM
|
BM-K
| 2023-08-30T05:37:05Z | 134 | 2 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"bert",
"feature-extraction",
"arxiv:2002.10957",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] |
feature-extraction
| 2022-05-23T04:26:31Z |
# KoMiniLM
🐣 Korean mini language model
## Overview
Current language models usually consist of hundreds of millions of parameters which brings challenges for fine-tuning and online serving in real-life applications due to latency and capacity constraints. In this project, we release a light weight korean language model to address the aforementioned shortcomings of existing language models.
## Quick tour
```python
from transformers import AutoTokenizer, AutoModel
tokenizer = AutoTokenizer.from_pretrained("BM-K/KoMiniLM") # 23M model
model = AutoModel.from_pretrained("BM-K/KoMiniLM")
inputs = tokenizer("안녕 세상아!", return_tensors="pt")
outputs = model(**inputs)
```
## Update history
** Updates on 2022.06.20 **
- Release KoMiniLM-bert-68M
** Updates on 2022.05.24 **
- Release KoMiniLM-bert-23M
## Pre-training
`Teacher Model`: [KLUE-BERT(base)](https://github.com/KLUE-benchmark/KLUE)
### Object
Self-Attention Distribution and Self-Attention Value-Relation [[Wang et al., 2020]](https://arxiv.org/abs/2002.10957) were distilled from each discrete layer of the teacher model to the student model. Wang et al. distilled in the last layer of the transformer, but that was not the case in this project.
### Data sets
|Data|News comments|News article|
|:----:|:----:|:----:|
|size|10G|10G|
### Config
- **KoMiniLM-23M**
```json
{
"architectures": [
"BertForPreTraining"
],
"attention_probs_dropout_prob": 0.1,
"classifier_dropout": null,
"hidden_act": "gelu",
"hidden_dropout_prob": 0.1,
"hidden_size": 384,
"initializer_range": 0.02,
"intermediate_size": 1536,
"layer_norm_eps": 1e-12,
"max_position_embeddings": 512,
"model_type": "bert",
"num_attention_heads": 12,
"num_hidden_layers": 6,
"output_attentions": true,
"pad_token_id": 0,
"position_embedding_type": "absolute",
"return_dict": false,
"torch_dtype": "float32",
"transformers_version": "4.13.0",
"type_vocab_size": 2,
"use_cache": true,
"vocab_size": 32000
}
```
### Performance on subtasks
- The results of our fine-tuning experiments are an average of 3 runs for each task.
```
cd KoMiniLM-Finetune
bash scripts/run_all_kominilm.sh
```
|| #Param | Average | NSMC<br>(Acc) | Naver NER<br>(F1) | PAWS<br>(Acc) | KorNLI<br>(Acc) | KorSTS<br>(Spearman) | Question Pair<br>(Acc) | KorQuaD<br>(Dev)<br>(EM/F1) |
|:----:|:----:|:----:|:----:|:----:|:----:|:----:|:----:|:----:|:----:|
|KoBERT(KLUE)| 110M | 86.84 | 90.20±0.07 | 87.11±0.05 | 81.36±0.21 | 81.06±0.33 | 82.47±0.14 | 95.03±0.44 | 84.43±0.18 / <br>93.05±0.04 |
|KcBERT| 108M | 78.94 | 89.60±0.10 | 84.34±0.13 | 67.02±0.42| 74.17±0.52 | 76.57±0.51 | 93.97±0.27 | 60.87±0.27 / <br>85.01±0.14 |
|KoBERT(SKT)| 92M | 79.73 | 89.28±0.42 | 87.54±0.04 | 80.93±0.91 | 78.18±0.45 | 75.98±2.81 | 94.37±0.31 | 51.94±0.60 / <br>79.69±0.66 |
|DistilKoBERT| 28M | 74.73 | 88.39±0.08 | 84.22±0.01 | 61.74±0.45 | 70.22±0.14 | 72.11±0.27 | 92.65±0.16 | 52.52±0.48 / <br>76.00±0.71 |
| | | | | | | | | |
|**KoMiniLM<sup>†</sup>**| **68M** | 85.90 | 89.84±0.02 | 85.98±0.09 | 80.78±0.30 | 79.28±0.17 | 81.00±0.07 | 94.89±0.37 | 83.27±0.08 / <br>92.08±0.06 |
|**KoMiniLM<sup>†</sup>**| **23M** | 84.79 | 89.67±0.03 | 84.79±0.09 | 78.67±0.45 | 78.10±0.07 | 78.90±0.11 | 94.81±0.12 | 82.11±0.42 / <br>91.21±0.29 |
- [NSMC](https://github.com/e9t/nsmc) (Naver Sentiment Movie Corpus)
- [Naver NER](https://github.com/naver/nlp-challenge) (NER task on Naver NLP Challenge 2018)
- [PAWS](https://github.com/google-research-datasets/paws) (Korean Paraphrase Adversaries from Word Scrambling)
- [KorNLI/KorSTS](https://github.com/kakaobrain/KorNLUDatasets) (Korean Natural Language Understanding)
- [Question Pair](https://github.com/songys/Question_pair) (Paired Question)
- [KorQuAD](https://korquad.github.io/) (The Korean Question Answering Dataset)
<img src = "https://user-images.githubusercontent.com/55969260/174229747-279122dc-9d27-4da9-a6e7-f9f1fe1651f7.png"> <br>
### User Contributed Examples
-
## Reference
- [KLUE BERT](https://github.com/KLUE-benchmark/KLUE)
- [KcBERT](https://github.com/Beomi/KcBERT)
- [SKT KoBERT](https://github.com/SKTBrain/KoBERT)
- [DistilKoBERT](https://github.com/monologg/DistilKoBERT)
- [lassl](https://github.com/lassl/lassl)
|
BM-K/KoDiffCSE-RoBERTa
|
BM-K
| 2023-08-30T05:36:55Z | 27,758 | 4 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"roberta",
"feature-extraction",
"arxiv:2204.10298",
"arxiv:2004.03289",
"arxiv:2105.09680",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] |
feature-extraction
| 2023-02-28T04:17:36Z |
# KoDiffCSE
Difference-based Contrastive Learning for Korean Sentence Embeddings <br>
- [DiffCSE-[NAACL 2022]](https://arxiv.org/abs/2204.10298) <br>
- [[Github]](https://github.com/voidism/DiffCSE) Official implementation of DiffCSE <br>
<img src=https://user-images.githubusercontent.com/55969260/201829550-9674a3ac-cb9b-4e17-b777-7d96fdf5c633.png>
## Quick tour
```python
import torch
from transformers import AutoModel, AutoTokenizer
def cal_score(a, b):
if len(a.shape) == 1: a = a.unsqueeze(0)
if len(b.shape) == 1: b = b.unsqueeze(0)
a_norm = a / a.norm(dim=1)[:, None]
b_norm = b / b.norm(dim=1)[:, None]
return torch.mm(a_norm, b_norm.transpose(0, 1)) * 100
model = AutoModel.from_pretrained('BM-K/KoDiffCSE-RoBERTa')
tokenizer = AutoTokenizer.from_pretrained('BM-K/KoDiffCSE-RoBERTa')
sentences = ['치타가 들판을 가로 질러 먹이를 쫓는다.',
'치타 한 마리가 먹이 뒤에서 달리고 있다.',
'원숭이 한 마리가 드럼을 연주한다.']
inputs = tokenizer(sentences, padding=True, truncation=True, return_tensors="pt")
embeddings, _ = model(**inputs, return_dict=False)
score01 = cal_score(embeddings[0][0], embeddings[1][0]) # 84.56
# '치타가 들판을 가로 질러 먹이를 쫓는다.' @ '치타 한 마리가 먹이 뒤에서 달리고 있다.'
score02 = cal_score(embeddings[0][0], embeddings[2][0]) # 48.06
# '치타가 들판을 가로 질러 먹이를 쫓는다.' @ '원숭이 한 마리가 드럼을 연주한다.'
```
## Setups
[](https://www.python.org/downloads/release/python-385/)
[](https://pytorch.org/get-started/previous-versions/)
## Encoder Models
Baseline encoders used for korean sentence embedding - [KLUE-PLMs](https://github.com/KLUE-benchmark/KLUE/blob/main/README.md)
| Model | Embedding size | Hidden size | # Layers | # Heads |
|----------------------|----------------|-------------|----------|---------|
| KLUE-BERT-base | 768 | 768 | 12 | 12 |
| KLUE-RoBERTa-base | 768 | 768 | 12 | 12 |
> **Warning** <br>
> Large pre-trained models need a lot of GPU memory to train
## Datasets
The data must exist in the "--path_to_data" folder
- [wiki-corpus](https://github.com/jeongukjae/korean-wikipedia-corpus) (Unsupervised Training)
- [KorSTS](https://github.com/kakaobrain/KorNLUDatasets) (Validation & Testing)
## Training - unsupervised
```
python main.py \
--model klue/roberta-base \
--generator_name klue/roberta-small \
--multi_gpu True \
--train True \
--test False \
--max_len 64 \
--batch_size 256 \
--epochs 1 \
--eval_steps 125 \
--lr 0.00005 \
--masking_ratio 0.15 \
--lambda_weight 0.005 \
--warmup_ratio 0.05 \
--temperature 0.05 \
--path_to_data Dataset/ \
--train_data wiki_corpus_examples.txt \
--valid_data valid_sts.tsv \
--ckpt best_checkpoint.pt
```
```
bash run_diff.sh
```
> **Note** <br>
> Using roberta as an encoder is beneficial for training because the KoBERT model cannot build a small-sized generator.
## Evaluation
```
python main.py \
--model klue/roberta-base \
--generator klue/roberta-small \
--train False \
--test True \
--max_len 64 \
--batch_size 256 \
--path_to_data Dataset/ \
--test_data test_sts.tsv \
--path_to_saved_model output/best_checkpoint.pt
```
## Performance - unsupervised
| Model | Average | Cosine Pearson | Cosine Spearman | Euclidean Pearson | Euclidean Spearman | Manhattan Pearson | Manhattan Spearman | Dot Pearson | Dot Spearman |
|------------------------|:----:|:----:|:----:|:----:|:----:|:----:|:----:|:----:|:----:|
| KoSRoBERTa-base<sup>†</sup> | N/A | N/A | 48.96 | N/A | N/A | N/A | N/A | N/A | N/A |
| KoSRoBERTa-large<sup>†</sup> | N/A | N/A | 51.35 | N/A | N/A | N/A | N/A | N/A | N/A |
| | | | | | | | | | |
| KoSimCSE-BERT | 74.08 | 74.92 | 73.98 | 74.15 | 74.22 | 74.07 | 74.07 | 74.15 | 73.14 |
| KoSimCSE-RoBERTa | 75.27 | 75.93 | 75.00 | 75.28 | 75.01 | 75.17 | 74.83 | 75.95 | 75.01 |
| | | | | | | | | | |
| KoDiffCSE-RoBERTa | 77.17 | 77.73 | 76.96 | 77.21 | 76.89 | 77.11 | 76.81 | 77.74 | 76.97 |
- [Korean-SRoBERTa<sup>†</sup>](https://arxiv.org/abs/2004.03289)
## License
This work is licensed under a <a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/">Creative Commons Attribution-ShareAlike 4.0 International License</a>.
<a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/"><img alt="Creative Commons License" style="border-width:0" src="https://i.creativecommons.org/l/by-sa/4.0/88x31.png" /></a><br />
## References
```bibtex
@inproceedings{chuang2022diffcse,
title={{DiffCSE}: Difference-based Contrastive Learning for Sentence Embeddings},
author={Chuang, Yung-Sung and Dangovski, Rumen and Luo, Hongyin and Zhang, Yang and Chang, Shiyu and Soljacic, Marin and Li, Shang-Wen and Yih, Wen-tau and Kim, Yoon and Glass, James},
booktitle={Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL)},
year={2022}
}
@misc{park2021klue,
title={KLUE: Korean Language Understanding Evaluation},
author={Sungjoon Park and Jihyung Moon and Sungdong Kim and Won Ik Cho and Jiyoon Han and Jangwon Park and Chisung Song and Junseong Kim and Yongsook Song and Taehwan Oh and Joohong Lee and Juhyun Oh and Sungwon Lyu and Younghoon Jeong and Inkwon Lee and Sangwoo Seo and Dongjun Lee and Hyunwoo Kim and Myeonghwa Lee and Seongbo Jang and Seungwon Do and Sunkyoung Kim and Kyungtae Lim and Jongwon Lee and Kyumin Park and Jamin Shin and Seonghyun Kim and Lucy Park and Alice Oh and Jungwoo Ha and Kyunghyun Cho},
year={2021},
eprint={2105.09680},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
@article{ham2020kornli,
title={KorNLI and KorSTS: New Benchmark Datasets for Korean Natural Language Understanding},
author={Ham, Jiyeon and Choe, Yo Joong and Park, Kyubyong and Choi, Ilji and Soh, Hyungjoon},
journal={arXiv preprint arXiv:2004.03289},
year={2020}
}
```
|
BM-K/KoSimCSE-bert
|
BM-K
| 2023-08-30T05:36:44Z | 986 | 5 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"bert",
"feature-extraction",
"korean",
"ko",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] |
feature-extraction
| 2022-05-23T00:54:43Z |
---
language: ko
tags:
- korean
---
https://github.com/BM-K/Sentence-Embedding-is-all-you-need
# Korean-Sentence-Embedding
🍭 Korean sentence embedding repository. You can download the pre-trained models and inference right away, also it provides environments where individuals can train models.
## Quick tour
```python
import torch
from transformers import AutoModel, AutoTokenizer
def cal_score(a, b):
if len(a.shape) == 1: a = a.unsqueeze(0)
if len(b.shape) == 1: b = b.unsqueeze(0)
a_norm = a / a.norm(dim=1)[:, None]
b_norm = b / b.norm(dim=1)[:, None]
return torch.mm(a_norm, b_norm.transpose(0, 1)) * 100
model = AutoModel.from_pretrained('BM-K/KoSimCSE-bert')
AutoTokenizer.from_pretrained('BM-K/KoSimCSE-bert')
sentences = ['치타가 들판을 가로 질러 먹이를 쫓는다.',
'치타 한 마리가 먹이 뒤에서 달리고 있다.',
'원숭이 한 마리가 드럼을 연주한다.']
inputs = tokenizer(sentences, padding=True, truncation=True, return_tensors="pt")
embeddings, _ = model(**inputs, return_dict=False)
score01 = cal_score(embeddings[0][0], embeddings[1][0])
score02 = cal_score(embeddings[0][0], embeddings[2][0])
```
## Performance
- Semantic Textual Similarity test set results <br>
| Model | AVG | Cosine Pearson | Cosine Spearman | Euclidean Pearson | Euclidean Spearman | Manhattan Pearson | Manhattan Spearman | Dot Pearson | Dot Spearman |
|------------------------|:----:|:----:|:----:|:----:|:----:|:----:|:----:|:----:|:----:|
| KoSBERT<sup>†</sup><sub>SKT</sub> | 77.40 | 78.81 | 78.47 | 77.68 | 77.78 | 77.71 | 77.83 | 75.75 | 75.22 |
| KoSBERT | 80.39 | 82.13 | 82.25 | 80.67 | 80.75 | 80.69 | 80.78 | 77.96 | 77.90 |
| KoSRoBERTa | 81.64 | 81.20 | 82.20 | 81.79 | 82.34 | 81.59 | 82.20 | 80.62 | 81.25 |
| | | | | | | | | |
| KoSentenceBART | 77.14 | 79.71 | 78.74 | 78.42 | 78.02 | 78.40 | 78.00 | 74.24 | 72.15 |
| KoSentenceT5 | 77.83 | 80.87 | 79.74 | 80.24 | 79.36 | 80.19 | 79.27 | 72.81 | 70.17 |
| | | | | | | | | |
| KoSimCSE-BERT<sup>†</sup><sub>SKT</sub> | 81.32 | 82.12 | 82.56 | 81.84 | 81.63 | 81.99 | 81.74 | 79.55 | 79.19 |
| KoSimCSE-BERT | 83.37 | 83.22 | 83.58 | 83.24 | 83.60 | 83.15 | 83.54 | 83.13 | 83.49 |
| KoSimCSE-RoBERTa | 83.65 | 83.60 | 83.77 | 83.54 | 83.76 | 83.55 | 83.77 | 83.55 | 83.64 |
| | | | | | | | | | |
| KoSimCSE-BERT-multitask | 85.71 | 85.29 | 86.02 | 85.63 | 86.01 | 85.57 | 85.97 | 85.26 | 85.93 |
| KoSimCSE-RoBERTa-multitask | 85.77 | 85.08 | 86.12 | 85.84 | 86.12 | 85.83 | 86.12 | 85.03 | 85.99 |
|
lu5/swinv2-small-patch4-window8-256-finetuned-eurosat
|
lu5
| 2023-08-30T05:23:13Z | 151 | 0 |
transformers
|
[
"transformers",
"pytorch",
"swinv2",
"image-classification",
"generated_from_trainer",
"dataset:food101",
"base_model:microsoft/swinv2-small-patch4-window8-256",
"base_model:finetune:microsoft/swinv2-small-patch4-window8-256",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-08-25T08:50:23Z |
---
license: apache-2.0
base_model: microsoft/swinv2-small-patch4-window8-256
tags:
- generated_from_trainer
datasets:
- food101
metrics:
- accuracy
model-index:
- name: swinv2-small-patch4-window8-256-finetuned-eurosat
results:
- task:
name: Image Classification
type: image-classification
dataset:
name: food101
type: food101
config: default
split: validation
args: default
metrics:
- name: Accuracy
type: accuracy
value: 0.8846732673267327
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# swinv2-small-patch4-window8-256-finetuned-eurosat
This model is a fine-tuned version of [microsoft/swinv2-small-patch4-window8-256](https://huggingface.co/microsoft/swinv2-small-patch4-window8-256) on the food101 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4063
- Accuracy: 0.8847
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 1.3521 | 1.0 | 592 | 0.7233 | 0.7967 |
| 0.9887 | 2.0 | 1184 | 0.5211 | 0.8519 |
| 0.845 | 3.0 | 1776 | 0.4063 | 0.8847 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3
|
jlpan/starcoder-java2py-snippet2
|
jlpan
| 2023-08-30T05:22:48Z | 0 | 0 | null |
[
"generated_from_trainer",
"base_model:bigcode/starcoder",
"base_model:finetune:bigcode/starcoder",
"license:bigcode-openrail-m",
"region:us"
] | null | 2023-08-30T03:03:11Z |
---
license: bigcode-openrail-m
base_model: bigcode/starcoder
tags:
- generated_from_trainer
model-index:
- name: starcoder-java2py-snippet2
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# starcoder-java2py-snippet2
This model is a fine-tuned version of [bigcode/starcoder](https://huggingface.co/bigcode/starcoder) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1902
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 7e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 10
- training_steps: 150
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 0.1883 | 0.17 | 25 | 0.1970 |
| 0.192 | 0.33 | 50 | 0.1936 |
| 0.1913 | 0.5 | 75 | 0.1918 |
| 0.1885 | 0.67 | 100 | 0.1910 |
| 0.1961 | 0.83 | 125 | 0.1905 |
| 0.1897 | 1.0 | 150 | 0.1902 |
### Framework versions
- Transformers 4.32.0.dev0
- Pytorch 2.0.1+cu117
- Datasets 2.12.0
- Tokenizers 0.13.3
|
AmlanSamanta/ludwig-webinar
|
AmlanSamanta
| 2023-08-30T04:40:12Z | 1 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-08-30T04:40:06Z |
---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: float16
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: float16
### Framework versions
- PEFT 0.5.0
- PEFT 0.5.0
|
gooseboose/fbd1
|
gooseboose
| 2023-08-30T04:19:33Z | 19 | 1 |
diffusers
|
[
"diffusers",
"tensorboard",
"text-to-image",
"license:creativeml-openrail-m",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] |
text-to-image
| 2023-08-30T04:17:27Z |
---
license: creativeml-openrail-m
tags:
- text-to-image
widget:
- text: fbd1
---
### fbd1 Dreambooth model trained by gooseboose with [Hugging Face Dreambooth Training Space](https://huggingface.co/spaces/multimodalart/dreambooth-training) with the v1-5 base model
You run your new concept via `diffusers` [Colab Notebook for Inference](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_dreambooth_inference.ipynb). Don't forget to use the concept prompts!
Sample pictures of:
fbd1 (use that on your prompt)
|
sang-kyung/train_monster_toy1
|
sang-kyung
| 2023-08-30T03:38:19Z | 1 | 0 |
diffusers
|
[
"diffusers",
"stable-diffusion",
"stable-diffusion-diffusers",
"text-to-image",
"lora",
"base_model:stabilityai/stable-diffusion-2-1-base",
"base_model:adapter:stabilityai/stable-diffusion-2-1-base",
"license:creativeml-openrail-m",
"region:us"
] |
text-to-image
| 2023-08-30T03:29:39Z |
---
license: creativeml-openrail-m
base_model: stabilityai/stable-diffusion-2-1-base
instance_prompt: a photo of sks monster toy
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
- diffusers
- lora
inference: true
---
# LoRA DreamBooth - sang-kyung/train_monster_toy1
These are LoRA adaption weights for stabilityai/stable-diffusion-2-1-base. The weights were trained on a photo of sks monster toy using [DreamBooth](https://dreambooth.github.io/). You can find some example images in the following.
LoRA for the text encoder was enabled: False.
|
gigant/distilhubert-audio-course-finetuned-gtzan-v5
|
gigant
| 2023-08-30T03:32:09Z | 159 | 0 |
transformers
|
[
"transformers",
"pytorch",
"hubert",
"audio-classification",
"generated_from_trainer",
"dataset:marsyas/gtzan",
"base_model:ntu-spml/distilhubert",
"base_model:finetune:ntu-spml/distilhubert",
"license:apache-2.0",
"model-index",
"endpoints_compatible",
"region:us"
] |
audio-classification
| 2023-08-29T23:41:25Z |
---
license: apache-2.0
base_model: ntu-spml/distilhubert
tags:
- generated_from_trainer
datasets:
- marsyas/gtzan
metrics:
- accuracy
model-index:
- name: distilhubert-audio-course-finetuned-gtzan-v5
results:
- task:
name: Audio Classification
type: audio-classification
dataset:
name: GTZAN
type: marsyas/gtzan
config: all
split: train
args: all
metrics:
- name: Accuracy
type: accuracy
value: 0.87
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilhubert-audio-course-finetuned-gtzan-v5
This model is a fine-tuned version of [ntu-spml/distilhubert](https://huggingface.co/ntu-spml/distilhubert) on the GTZAN dataset.
It achieves the following results on the evaluation set:
- Loss: 0.9236
- Accuracy: 0.87
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.7
- num_epochs: 50
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 2.2989 | 0.99 | 56 | 2.2882 | 0.11 |
| 2.2716 | 2.0 | 113 | 2.2469 | 0.31 |
| 2.1919 | 2.99 | 169 | 2.1317 | 0.4 |
| 2.0117 | 4.0 | 226 | 1.9244 | 0.53 |
| 1.7966 | 4.99 | 282 | 1.7315 | 0.65 |
| 1.6379 | 6.0 | 339 | 1.5920 | 0.59 |
| 1.4496 | 6.99 | 395 | 1.3539 | 0.71 |
| 1.3264 | 8.0 | 452 | 1.1879 | 0.7 |
| 1.0601 | 8.99 | 508 | 1.1342 | 0.7 |
| 0.9737 | 10.0 | 565 | 0.9209 | 0.79 |
| 0.7915 | 10.99 | 621 | 0.8768 | 0.74 |
| 0.6432 | 12.0 | 678 | 0.8060 | 0.8 |
| 0.5217 | 12.99 | 734 | 0.6562 | 0.85 |
| 0.3335 | 14.0 | 791 | 0.7744 | 0.76 |
| 0.2866 | 14.99 | 847 | 0.6969 | 0.82 |
| 0.1425 | 16.0 | 904 | 0.6378 | 0.82 |
| 0.1278 | 16.99 | 960 | 0.6972 | 0.82 |
| 0.0706 | 18.0 | 1017 | 0.7328 | 0.84 |
| 0.0301 | 18.99 | 1073 | 0.9245 | 0.76 |
| 0.0379 | 20.0 | 1130 | 0.8437 | 0.85 |
| 0.0147 | 20.99 | 1186 | 0.7489 | 0.83 |
| 0.0067 | 22.0 | 1243 | 0.8975 | 0.83 |
| 0.0049 | 22.99 | 1299 | 1.1788 | 0.81 |
| 0.0038 | 24.0 | 1356 | 1.1146 | 0.81 |
| 0.0028 | 24.99 | 1412 | 1.0270 | 0.85 |
| 0.0027 | 26.0 | 1469 | 1.0634 | 0.83 |
| 0.0024 | 26.99 | 1525 | 1.0220 | 0.84 |
| 0.0023 | 28.0 | 1582 | 1.0282 | 0.83 |
| 0.0487 | 28.99 | 1638 | 1.0735 | 0.82 |
| 0.0458 | 30.0 | 1695 | 1.1198 | 0.82 |
| 0.2453 | 30.99 | 1751 | 1.1154 | 0.81 |
| 0.0552 | 32.0 | 1808 | 1.1630 | 0.79 |
| 0.1202 | 32.99 | 1864 | 1.2746 | 0.81 |
| 0.2709 | 34.0 | 1921 | 1.3797 | 0.79 |
| 0.275 | 34.99 | 1977 | 1.5372 | 0.75 |
| 0.1268 | 36.0 | 2034 | 0.8140 | 0.86 |
| 0.1582 | 36.99 | 2090 | 1.4153 | 0.77 |
| 0.0054 | 38.0 | 2147 | 1.3796 | 0.79 |
| 0.0299 | 38.99 | 2203 | 1.3653 | 0.78 |
| 0.0199 | 40.0 | 2260 | 0.9987 | 0.87 |
| 0.0021 | 40.99 | 2316 | 1.0689 | 0.84 |
| 0.0007 | 42.0 | 2373 | 1.0383 | 0.85 |
| 0.0006 | 42.99 | 2429 | 1.0493 | 0.84 |
| 0.0006 | 44.0 | 2486 | 1.0744 | 0.85 |
| 0.0005 | 44.99 | 2542 | 0.9151 | 0.86 |
| 0.0004 | 46.0 | 2599 | 0.8946 | 0.87 |
| 0.01 | 46.99 | 2655 | 0.8960 | 0.88 |
| 0.0073 | 48.0 | 2712 | 0.9485 | 0.87 |
| 0.0004 | 48.99 | 2768 | 0.9247 | 0.87 |
| 0.0004 | 49.56 | 2800 | 0.9236 | 0.87 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.0.1+cu117
- Datasets 2.14.4
- Tokenizers 0.13.3
|
matgu23/zws3
|
matgu23
| 2023-08-30T03:31:32Z | 3 | 0 |
diffusers
|
[
"diffusers",
"safetensors",
"text-to-image",
"stable-diffusion",
"license:creativeml-openrail-m",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] |
text-to-image
| 2023-08-30T03:25:45Z |
---
license: creativeml-openrail-m
tags:
- text-to-image
- stable-diffusion
---
### zws3 Dreambooth model trained by matgu23 with [TheLastBen's fast-DreamBooth](https://colab.research.google.com/github/TheLastBen/fast-stable-diffusion/blob/main/fast-DreamBooth.ipynb) notebook
Test the concept via A1111 Colab [fast-Colab-A1111](https://colab.research.google.com/github/TheLastBen/fast-stable-diffusion/blob/main/fast_stable_diffusion_AUTOMATIC1111.ipynb)
Sample pictures of this concept:
|
abeiler/huggingface-goatLora-goatV10-testData
|
abeiler
| 2023-08-30T03:19:52Z | 0 | 0 | null |
[
"pytorch",
"tensorboard",
"generated_from_trainer",
"region:us"
] | null | 2023-08-30T03:07:07Z |
---
tags:
- generated_from_trainer
model-index:
- name: huggingface-goatLora-goatV10-testData
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# huggingface-goatLora-goatV10-testData
This model is a fine-tuned version of [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) on an unknown dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0003
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 1
- mixed_precision_training: Native AMP
### Training results
### Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0
- Datasets 2.12.0
- Tokenizers 0.13.3
|
Azion/e-commerce-bert-base-multilingual-cased
|
Azion
| 2023-08-30T03:18:08Z | 23 | 0 |
sentence-transformers
|
[
"sentence-transformers",
"pytorch",
"bert",
"feature-extraction",
"sentence-similarity",
"transformers",
"autotrain_compatible",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] |
sentence-similarity
| 2023-08-29T09:36:00Z |
---
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
---
# Azion/e-commerce-bert-base-multilingual-cased
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.
<!--- Describe your model here -->
## Usage (Sentence-Transformers)
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
```
pip install -U sentence-transformers
```
Then you can use the model like this:
```python
from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]
model = SentenceTransformer('Azion/e-commerce-bert-base-multilingual-cased')
embeddings = model.encode(sentences)
print(embeddings)
```
## Usage (HuggingFace Transformers)
Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.
```python
from transformers import AutoTokenizer, AutoModel
import torch
#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
token_embeddings = model_output[0] #First element of model_output contains all token embeddings
input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']
# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('Azion/e-commerce-bert-base-multilingual-cased')
model = AutoModel.from_pretrained('Azion/e-commerce-bert-base-multilingual-cased')
# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
# Compute token embeddings
with torch.no_grad():
model_output = model(**encoded_input)
# Perform pooling. In this case, mean pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
print("Sentence embeddings:")
print(sentence_embeddings)
```
## Evaluation Results
<!--- Describe how your model was evaluated -->
For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=Azion/e-commerce-bert-base-multilingual-cased)
## Training
The model was trained with the parameters:
**DataLoader**:
`torch.utils.data.dataloader.DataLoader` of length 37450 with parameters:
```
{'batch_size': 64, 'sampler': 'torch.utils.data.sampler.RandomSampler', 'batch_sampler': 'torch.utils.data.sampler.BatchSampler'}
```
**Loss**:
`sentence_transformers.losses.CosineSimilarityLoss.CosineSimilarityLoss`
Parameters of the fit()-Method:
```
{
"epochs": 2,
"evaluation_steps": 0,
"evaluator": "sentence_transformers.evaluation.EmbeddingSimilarityEvaluator.EmbeddingSimilarityEvaluator",
"max_grad_norm": 1,
"optimizer_class": "<class 'torch.optim.adamw.AdamW'>",
"optimizer_params": {
"lr": 2e-05
},
"scheduler": "WarmupLinear",
"steps_per_epoch": null,
"warmup_steps": 7490,
"weight_decay": 0.01
}
```
## Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
)
```
## Citing & Authors
<!--- Describe where people can find more information -->
|
andyfriedrich-amd/hipify_plus_model
|
andyfriedrich-amd
| 2023-08-30T03:16:41Z | 119 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"t5",
"text2text-generation",
"generated_from_trainer",
"base_model:google-t5/t5-small",
"base_model:finetune:google-t5/t5-small",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-06-26T14:30:53Z |
---
license: apache-2.0
base_model: t5-small
tags:
- generated_from_trainer
metrics:
- bleu
model-index:
- name: hipify_plus_model
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# hipify_plus_model
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 2.1395
- Bleu: 0.0975
- Gen Len: 19.0
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
### Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:------:|:-------:|
| No log | 1.0 | 8 | 2.3028 | 0.1031 | 19.0 |
| No log | 2.0 | 16 | 2.1395 | 0.0975 | 19.0 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0a0+git794e397
- Datasets 2.14.4
- Tokenizers 0.13.3
|
hecool108/ct-p1
|
hecool108
| 2023-08-30T03:05:55Z | 2 | 0 |
diffusers
|
[
"diffusers",
"safetensors",
"text-generation-inference",
"ctee",
"text-to-image",
"en",
"license:openrail",
"region:us"
] |
text-to-image
| 2023-08-30T02:58:26Z |
---
license: openrail
language:
- en
tags:
- text-generation-inference
- ctee
pipeline_tag: text-to-image
---
|
xiatianC/hsaKJHSJ
|
xiatianC
| 2023-08-30T03:01:56Z | 0 | 0 | null |
[
"license:mit",
"region:us"
] | null | 2023-08-30T03:00:45Z |
---
license: mit
---
# Build Stage
# 使用 golang:alpine 作为构建阶段的基础镜像
FROM golang:alpine AS builder
# 添加 git，以便之后能从GitHub克隆项目
RUN apk --no-cache add git
# 从 GitHub 克隆 go-proxy-bingai 项目到 /workspace/app 目录下
RUN git clone https://github.com/Harry-zklcdc/go-proxy-bingai.git /workspace/app
# 设置工作目录为之前克隆的项目目录
WORKDIR /workspace/app
# 编译 go 项目。-ldflags="-s -w" 是为了减少编译后的二进制大小
RUN go build -ldflags="-s -w" -tags netgo -trimpath -o go-proxy-bingai main.go
# Runtime Stage
# 使用轻量级的 alpine 镜像作为运行时的基础镜像
FROM alpine
# 设置工作目录
WORKDIR /workspace/app
# 从构建阶段复制编译后的二进制文件到运行时镜像中
COPY --from=builder /workspace/app/go-proxy-bingai .
# 设置环境变量，此处为随机字符
ENV Go_Proxy_BingAI_USER_TOKEN_1="sajdhaskdhasjkdhasjkdhjsajkdhasjkd"
# 暴露8080端口
EXPOSE 8080
# 容器启动时运行的命令
CMD ["/workspace/app/go-proxy-bingai"]
|
johnnydevriese/vit-airplanes
|
johnnydevriese
| 2023-08-30T03:01:05Z | 246 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"safetensors",
"vit",
"image-classification",
"generated_from_trainer",
"dataset:image_folder",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2022-04-08T16:45:37Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- image_folder
metrics:
- accuracy
model-index:
- name: vit-airplanes
results:
- task:
name: Image Classification
type: image-classification
dataset:
name: image_folder
type: image_folder
args: default
metrics:
- name: Accuracy
type: accuracy
value: 1.0
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# vit-airplanes
This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the image_folder dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0152
- Accuracy: 1.0
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0002
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 4
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 0.0165 | 2.38 | 100 | 0.0152 | 1.0 |
### Framework versions
- Transformers 4.18.0
- Pytorch 1.10.0+cu111
- Datasets 2.0.0
- Tokenizers 0.11.6
|
jlpan/starcoder-java2py-snippet1
|
jlpan
| 2023-08-30T02:55:43Z | 0 | 0 | null |
[
"generated_from_trainer",
"base_model:bigcode/starcoder",
"base_model:finetune:bigcode/starcoder",
"license:bigcode-openrail-m",
"region:us"
] | null | 2023-08-29T23:29:31Z |
---
license: bigcode-openrail-m
base_model: bigcode/starcoder
tags:
- generated_from_trainer
model-index:
- name: starcoder-java2py-snippet1
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# starcoder-java2py-snippet1
This model is a fine-tuned version of [bigcode/starcoder](https://huggingface.co/bigcode/starcoder) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2013
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 9e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 10
- training_steps: 150
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 3.6996 | 0.17 | 25 | 0.3575 |
| 0.2656 | 0.33 | 50 | 0.2185 |
| 0.2123 | 0.5 | 75 | 0.2070 |
| 0.2077 | 0.67 | 100 | 0.2030 |
| 0.2036 | 0.83 | 125 | 0.2016 |
| 0.2042 | 1.0 | 150 | 0.2013 |
### Framework versions
- Transformers 4.32.0.dev0
- Pytorch 2.0.1+cu117
- Datasets 2.12.0
- Tokenizers 0.13.3
|
uer/chinese_roberta_L-12_H-256
|
uer
| 2023-08-30T02:30:45Z | 121 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"zh",
"dataset:CLUECorpusSmall",
"arxiv:1909.05658",
"arxiv:2212.06385",
"arxiv:1908.08962",
"arxiv:1810.04805",
"arxiv:1907.11692",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-02T23:29:05Z |
---
language: zh
datasets: CLUECorpusSmall
widget:
- text: "北京是[MASK]国的首都。"
---
# Chinese RoBERTa Miniatures
## Model description
This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework.
[Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details.
You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below:
| | H=128 | H=256 | H=512 | H=768 |
| -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: |
| **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] |
| **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] |
| **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] |
| **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] |
| **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] |
| **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] |
Here are scores on the devlopment set of six Chinese tasks:
| Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) |
| -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: |
| RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 |
| RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 |
| RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 |
| RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 |
| RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 |
For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128:
- epochs: 3, 5, 8
- batch sizes: 32, 64
- learning rates: 3e-5, 1e-4, 3e-4
## How to use
You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium):
```python
>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512')
>>> unmasker("中国的首都是[MASK]京。")
[
{'sequence': '[CLS] 中 国 的 首 都 是 北 京 。 [SEP]',
'score': 0.8701988458633423,
'token': 1266,
'token_str': '北'},
{'sequence': '[CLS] 中 国 的 首 都 是 南 京 。 [SEP]',
'score': 0.1194809079170227,
'token': 1298,
'token_str': '南'},
{'sequence': '[CLS] 中 国 的 首 都 是 东 京 。 [SEP]',
'score': 0.0037803512532263994,
'token': 691,
'token_str': '东'},
{'sequence': '[CLS] 中 国 的 首 都 是 普 京 。 [SEP]',
'score': 0.0017127094324678183,
'token': 3249,
'token_str': '普'},
{'sequence': '[CLS] 中 国 的 首 都 是 望 京 。 [SEP]',
'score': 0.001687526935711503,
'token': 3307,
'token_str': '望'}
]
```
Here is how to use this model to get the features of a given text in PyTorch:
```python
from transformers import BertTokenizer, BertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)
```
and in TensorFlow:
```python
from transformers import BertTokenizer, TFBertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)
```
## Training data
[CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall.
## Training procedure
Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes.
Taking the case of RoBERTa-Medium
Stage1:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq128_dataset.pt \
--processes_num 32 --seq_length 128 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \
--learning_rate 1e-4 --batch_size 64 \
--data_processor mlm --target mlm
```
Stage2:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq512_dataset.pt \
--processes_num 32 --seq_length 512 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \
--learning_rate 5e-5 --batch_size 16 \
--data_processor mlm --target mlm
```
Finally, we convert the pre-trained model into Huggingface's format:
```
python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \
--output_model_path pytorch_model.bin \
--layers_num 8 --type mlm
```
### BibTeX entry and citation info
```
@article{devlin2018bert,
title={Bert: Pre-training of deep bidirectional transformers for language understanding},
author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1810.04805},
year={2018}
}
@article{liu2019roberta,
title={Roberta: A robustly optimized bert pretraining approach},
author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin},
journal={arXiv preprint arXiv:1907.11692},
year={2019}
}
@article{turc2019,
title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models},
author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1908.08962v2 },
year={2019}
}
@article{zhao2019uer,
title={UER: An Open-Source Toolkit for Pre-training Models},
author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong},
journal={EMNLP-IJCNLP 2019},
pages={241},
year={2019}
}
@article{zhao2023tencentpretrain,
title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities},
author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others},
journal={ACL 2023},
pages={217},
year={2023}
}
```
[2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128
[2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256
[2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512
[2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768
[4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128
[4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256
[4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512
[4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768
[6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128
[6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256
[6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512
[6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768
[8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128
[8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256
[8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512
[8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768
[10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128
[10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256
[10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512
[10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768
[12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128
[12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256
[12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512
[12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
|
uer/chinese_roberta_L-12_H-128
|
uer
| 2023-08-30T02:30:16Z | 104 | 1 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"zh",
"dataset:CLUECorpusSmall",
"arxiv:1909.05658",
"arxiv:2212.06385",
"arxiv:1908.08962",
"arxiv:1810.04805",
"arxiv:1907.11692",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-02T23:29:05Z |
---
language: zh
datasets: CLUECorpusSmall
widget:
- text: "北京是[MASK]国的首都。"
---
# Chinese RoBERTa Miniatures
## Model description
This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework.
[Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details.
You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below:
| | H=128 | H=256 | H=512 | H=768 |
| -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: |
| **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] |
| **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] |
| **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] |
| **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] |
| **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] |
| **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] |
Here are scores on the devlopment set of six Chinese tasks:
| Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) |
| -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: |
| RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 |
| RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 |
| RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 |
| RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 |
| RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 |
For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128:
- epochs: 3, 5, 8
- batch sizes: 32, 64
- learning rates: 3e-5, 1e-4, 3e-4
## How to use
You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium):
```python
>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512')
>>> unmasker("中国的首都是[MASK]京。")
[
{'sequence': '[CLS] 中 国 的 首 都 是 北 京 。 [SEP]',
'score': 0.8701988458633423,
'token': 1266,
'token_str': '北'},
{'sequence': '[CLS] 中 国 的 首 都 是 南 京 。 [SEP]',
'score': 0.1194809079170227,
'token': 1298,
'token_str': '南'},
{'sequence': '[CLS] 中 国 的 首 都 是 东 京 。 [SEP]',
'score': 0.0037803512532263994,
'token': 691,
'token_str': '东'},
{'sequence': '[CLS] 中 国 的 首 都 是 普 京 。 [SEP]',
'score': 0.0017127094324678183,
'token': 3249,
'token_str': '普'},
{'sequence': '[CLS] 中 国 的 首 都 是 望 京 。 [SEP]',
'score': 0.001687526935711503,
'token': 3307,
'token_str': '望'}
]
```
Here is how to use this model to get the features of a given text in PyTorch:
```python
from transformers import BertTokenizer, BertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)
```
and in TensorFlow:
```python
from transformers import BertTokenizer, TFBertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)
```
## Training data
[CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall.
## Training procedure
Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes.
Taking the case of RoBERTa-Medium
Stage1:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq128_dataset.pt \
--processes_num 32 --seq_length 128 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \
--learning_rate 1e-4 --batch_size 64 \
--data_processor mlm --target mlm
```
Stage2:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq512_dataset.pt \
--processes_num 32 --seq_length 512 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \
--learning_rate 5e-5 --batch_size 16 \
--data_processor mlm --target mlm
```
Finally, we convert the pre-trained model into Huggingface's format:
```
python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \
--output_model_path pytorch_model.bin \
--layers_num 8 --type mlm
```
### BibTeX entry and citation info
```
@article{devlin2018bert,
title={Bert: Pre-training of deep bidirectional transformers for language understanding},
author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1810.04805},
year={2018}
}
@article{liu2019roberta,
title={Roberta: A robustly optimized bert pretraining approach},
author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin},
journal={arXiv preprint arXiv:1907.11692},
year={2019}
}
@article{turc2019,
title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models},
author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1908.08962v2 },
year={2019}
}
@article{zhao2019uer,
title={UER: An Open-Source Toolkit for Pre-training Models},
author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong},
journal={EMNLP-IJCNLP 2019},
pages={241},
year={2019}
}
@article{zhao2023tencentpretrain,
title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities},
author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others},
journal={ACL 2023},
pages={217},
year={2023}
}
```
[2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128
[2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256
[2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512
[2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768
[4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128
[4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256
[4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512
[4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768
[6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128
[6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256
[6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512
[6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768
[8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128
[8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256
[8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512
[8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768
[10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128
[10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256
[10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512
[10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768
[12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128
[12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256
[12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512
[12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
|
uer/chinese_roberta_L-10_H-512
|
uer
| 2023-08-30T02:29:34Z | 111 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"zh",
"dataset:CLUECorpusSmall",
"arxiv:1909.05658",
"arxiv:2212.06385",
"arxiv:1908.08962",
"arxiv:1810.04805",
"arxiv:1907.11692",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-02T23:29:05Z |
---
language: zh
datasets: CLUECorpusSmall
widget:
- text: "北京是[MASK]国的首都。"
---
# Chinese RoBERTa Miniatures
## Model description
This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework.
[Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details.
You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below:
| | H=128 | H=256 | H=512 | H=768 |
| -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: |
| **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] |
| **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] |
| **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] |
| **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] |
| **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] |
| **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] |
Here are scores on the devlopment set of six Chinese tasks:
| Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) |
| -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: |
| RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 |
| RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 |
| RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 |
| RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 |
| RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 |
For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128:
- epochs: 3, 5, 8
- batch sizes: 32, 64
- learning rates: 3e-5, 1e-4, 3e-4
## How to use
You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium):
```python
>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512')
>>> unmasker("中国的首都是[MASK]京。")
[
{'sequence': '[CLS] 中 国 的 首 都 是 北 京 。 [SEP]',
'score': 0.8701988458633423,
'token': 1266,
'token_str': '北'},
{'sequence': '[CLS] 中 国 的 首 都 是 南 京 。 [SEP]',
'score': 0.1194809079170227,
'token': 1298,
'token_str': '南'},
{'sequence': '[CLS] 中 国 的 首 都 是 东 京 。 [SEP]',
'score': 0.0037803512532263994,
'token': 691,
'token_str': '东'},
{'sequence': '[CLS] 中 国 的 首 都 是 普 京 。 [SEP]',
'score': 0.0017127094324678183,
'token': 3249,
'token_str': '普'},
{'sequence': '[CLS] 中 国 的 首 都 是 望 京 。 [SEP]',
'score': 0.001687526935711503,
'token': 3307,
'token_str': '望'}
]
```
Here is how to use this model to get the features of a given text in PyTorch:
```python
from transformers import BertTokenizer, BertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)
```
and in TensorFlow:
```python
from transformers import BertTokenizer, TFBertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)
```
## Training data
[CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall.
## Training procedure
Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes.
Taking the case of RoBERTa-Medium
Stage1:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq128_dataset.pt \
--processes_num 32 --seq_length 128 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \
--learning_rate 1e-4 --batch_size 64 \
--data_processor mlm --target mlm
```
Stage2:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq512_dataset.pt \
--processes_num 32 --seq_length 512 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \
--learning_rate 5e-5 --batch_size 16 \
--data_processor mlm --target mlm
```
Finally, we convert the pre-trained model into Huggingface's format:
```
python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \
--output_model_path pytorch_model.bin \
--layers_num 8 --type mlm
```
### BibTeX entry and citation info
```
@article{devlin2018bert,
title={Bert: Pre-training of deep bidirectional transformers for language understanding},
author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1810.04805},
year={2018}
}
@article{liu2019roberta,
title={Roberta: A robustly optimized bert pretraining approach},
author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin},
journal={arXiv preprint arXiv:1907.11692},
year={2019}
}
@article{turc2019,
title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models},
author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1908.08962v2 },
year={2019}
}
@article{zhao2019uer,
title={UER: An Open-Source Toolkit for Pre-training Models},
author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong},
journal={EMNLP-IJCNLP 2019},
pages={241},
year={2019}
}
@article{zhao2023tencentpretrain,
title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities},
author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others},
journal={ACL 2023},
pages={217},
year={2023}
}
```
[2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128
[2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256
[2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512
[2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768
[4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128
[4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256
[4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512
[4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768
[6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128
[6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256
[6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512
[6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768
[8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128
[8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256
[8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512
[8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768
[10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128
[10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256
[10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512
[10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768
[12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128
[12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256
[12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512
[12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
|
uer/chinese_roberta_L-10_H-256
|
uer
| 2023-08-30T02:29:16Z | 110 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"zh",
"dataset:CLUECorpusSmall",
"arxiv:1909.05658",
"arxiv:2212.06385",
"arxiv:1908.08962",
"arxiv:1810.04805",
"arxiv:1907.11692",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-02T23:29:05Z |
---
language: zh
datasets: CLUECorpusSmall
widget:
- text: "北京是[MASK]国的首都。"
---
# Chinese RoBERTa Miniatures
## Model description
This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework.
[Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details.
You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below:
| | H=128 | H=256 | H=512 | H=768 |
| -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: |
| **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] |
| **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] |
| **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] |
| **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] |
| **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] |
| **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] |
Here are scores on the devlopment set of six Chinese tasks:
| Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) |
| -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: |
| RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 |
| RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 |
| RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 |
| RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 |
| RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 |
For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128:
- epochs: 3, 5, 8
- batch sizes: 32, 64
- learning rates: 3e-5, 1e-4, 3e-4
## How to use
You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium):
```python
>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512')
>>> unmasker("中国的首都是[MASK]京。")
[
{'sequence': '[CLS] 中 国 的 首 都 是 北 京 。 [SEP]',
'score': 0.8701988458633423,
'token': 1266,
'token_str': '北'},
{'sequence': '[CLS] 中 国 的 首 都 是 南 京 。 [SEP]',
'score': 0.1194809079170227,
'token': 1298,
'token_str': '南'},
{'sequence': '[CLS] 中 国 的 首 都 是 东 京 。 [SEP]',
'score': 0.0037803512532263994,
'token': 691,
'token_str': '东'},
{'sequence': '[CLS] 中 国 的 首 都 是 普 京 。 [SEP]',
'score': 0.0017127094324678183,
'token': 3249,
'token_str': '普'},
{'sequence': '[CLS] 中 国 的 首 都 是 望 京 。 [SEP]',
'score': 0.001687526935711503,
'token': 3307,
'token_str': '望'}
]
```
Here is how to use this model to get the features of a given text in PyTorch:
```python
from transformers import BertTokenizer, BertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)
```
and in TensorFlow:
```python
from transformers import BertTokenizer, TFBertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)
```
## Training data
[CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall.
## Training procedure
Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes.
Taking the case of RoBERTa-Medium
Stage1:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq128_dataset.pt \
--processes_num 32 --seq_length 128 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \
--learning_rate 1e-4 --batch_size 64 \
--data_processor mlm --target mlm
```
Stage2:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq512_dataset.pt \
--processes_num 32 --seq_length 512 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \
--learning_rate 5e-5 --batch_size 16 \
--data_processor mlm --target mlm
```
Finally, we convert the pre-trained model into Huggingface's format:
```
python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \
--output_model_path pytorch_model.bin \
--layers_num 8 --type mlm
```
### BibTeX entry and citation info
```
@article{devlin2018bert,
title={Bert: Pre-training of deep bidirectional transformers for language understanding},
author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1810.04805},
year={2018}
}
@article{liu2019roberta,
title={Roberta: A robustly optimized bert pretraining approach},
author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin},
journal={arXiv preprint arXiv:1907.11692},
year={2019}
}
@article{turc2019,
title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models},
author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1908.08962v2 },
year={2019}
}
@article{zhao2019uer,
title={UER: An Open-Source Toolkit for Pre-training Models},
author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong},
journal={EMNLP-IJCNLP 2019},
pages={241},
year={2019}
}
@article{zhao2023tencentpretrain,
title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities},
author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others},
journal={ACL 2023},
pages={217},
year={2023}
}
```
[2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128
[2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256
[2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512
[2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768
[4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128
[4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256
[4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512
[4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768
[6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128
[6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256
[6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512
[6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768
[8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128
[8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256
[8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512
[8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768
[10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128
[10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256
[10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512
[10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768
[12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128
[12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256
[12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512
[12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
|
yumjunstar/trocr-small-printedkorean-deleteunusedchar_noise
|
yumjunstar
| 2023-08-30T02:28:54Z | 51 | 0 |
transformers
|
[
"transformers",
"pytorch",
"vision-encoder-decoder",
"image-text-to-text",
"generated_from_trainer",
"endpoints_compatible",
"region:us"
] |
image-text-to-text
| 2023-08-29T17:52:22Z |
---
tags:
- generated_from_trainer
metrics:
- wer
- accuracy
model-index:
- name: trocr-small-printedkorean-deleteunusedchar_noise
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# trocr-small-printedkorean-deleteunusedchar_noise
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 1.3375
- Cer: 0.2783
- Wer: 0.2975
- Accuracy: 45.6667
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 4e-05
- train_batch_size: 128
- eval_batch_size: 192
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Cer | Wer | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|:--------:|
| 1.711 | 0.43 | 1000 | 1.6485 | 0.3288 | 0.3944 | 30.6667 |
| 1.6849 | 0.85 | 2000 | 1.5361 | 0.3098 | 0.3809 | 32.3333 |
| 1.4933 | 1.28 | 3000 | 1.4302 | 0.2935 | 0.3533 | 34.6667 |
| 1.526 | 1.71 | 4000 | 1.4010 | 0.2922 | 0.3400 | 35.6667 |
| 1.3422 | 2.13 | 5000 | 1.3883 | 0.2846 | 0.3331 | 36.0 |
| 1.333 | 2.56 | 6000 | 1.3790 | 0.2871 | 0.3308 | 34.0 |
| 1.3295 | 2.99 | 7000 | 1.3644 | 0.2876 | 0.3294 | 35.6667 |
| 1.3294 | 3.42 | 8000 | 1.3588 | 0.2824 | 0.3202 | 36.6667 |
| 1.3578 | 3.84 | 9000 | 1.3502 | 0.2823 | 0.3162 | 40.6667 |
| 1.3029 | 4.27 | 10000 | 1.3514 | 0.2879 | 0.3228 | 37.0 |
| 1.2777 | 4.7 | 11000 | 1.3507 | 0.2813 | 0.3168 | 38.3333 |
| 1.1781 | 5.12 | 12000 | 1.3507 | 0.2791 | 0.3150 | 40.3333 |
| 1.3025 | 5.55 | 13000 | 1.3459 | 0.2818 | 0.3099 | 41.6667 |
| 1.2024 | 5.98 | 14000 | 1.3401 | 0.2801 | 0.3061 | 41.6667 |
| 1.1792 | 6.4 | 15000 | 1.3412 | 0.2763 | 0.3015 | 44.6667 |
| 1.1586 | 6.83 | 16000 | 1.3410 | 0.2799 | 0.3064 | 43.3333 |
| 1.2098 | 7.26 | 17000 | 1.3439 | 0.2777 | 0.3030 | 43.6667 |
| 1.2122 | 7.69 | 18000 | 1.3418 | 0.2816 | 0.3050 | 43.3333 |
| 1.1323 | 8.11 | 19000 | 1.3409 | 0.2767 | 0.2981 | 45.3333 |
| 1.2215 | 8.54 | 20000 | 1.3386 | 0.2781 | 0.3004 | 44.0 |
| 1.2068 | 8.97 | 21000 | 1.3375 | 0.2762 | 0.2972 | 45.0 |
| 1.0847 | 9.39 | 22000 | 1.3366 | 0.2765 | 0.2969 | 46.0 |
| 1.1791 | 9.82 | 23000 | 1.3375 | 0.2783 | 0.2975 | 45.6667 |
### Framework versions
- Transformers 4.28.0
- Pytorch 1.13.1+cu116
- Datasets 2.14.4
- Tokenizers 0.13.3
|
uer/chinese_roberta_L-8_H-512
|
uer
| 2023-08-30T02:28:15Z | 146 | 3 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"zh",
"dataset:CLUECorpusSmall",
"arxiv:1909.05658",
"arxiv:2212.06385",
"arxiv:1908.08962",
"arxiv:1810.04805",
"arxiv:1907.11692",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-02T23:29:05Z |
---
language: zh
datasets: CLUECorpusSmall
widget:
- text: "北京是[MASK]国的首都。"
---
# Chinese RoBERTa Miniatures
## Model description
This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework.
[Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details.
You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below:
| | H=128 | H=256 | H=512 | H=768 |
| -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: |
| **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] |
| **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] |
| **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] |
| **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] |
| **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] |
| **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] |
Here are scores on the devlopment set of six Chinese tasks:
| Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) |
| -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: |
| RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 |
| RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 |
| RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 |
| RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 |
| RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 |
For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128:
- epochs: 3, 5, 8
- batch sizes: 32, 64
- learning rates: 3e-5, 1e-4, 3e-4
## How to use
You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium):
```python
>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512')
>>> unmasker("中国的首都是[MASK]京。")
[
{'sequence': '[CLS] 中 国 的 首 都 是 北 京 。 [SEP]',
'score': 0.8701988458633423,
'token': 1266,
'token_str': '北'},
{'sequence': '[CLS] 中 国 的 首 都 是 南 京 。 [SEP]',
'score': 0.1194809079170227,
'token': 1298,
'token_str': '南'},
{'sequence': '[CLS] 中 国 的 首 都 是 东 京 。 [SEP]',
'score': 0.0037803512532263994,
'token': 691,
'token_str': '东'},
{'sequence': '[CLS] 中 国 的 首 都 是 普 京 。 [SEP]',
'score': 0.0017127094324678183,
'token': 3249,
'token_str': '普'},
{'sequence': '[CLS] 中 国 的 首 都 是 望 京 。 [SEP]',
'score': 0.001687526935711503,
'token': 3307,
'token_str': '望'}
]
```
Here is how to use this model to get the features of a given text in PyTorch:
```python
from transformers import BertTokenizer, BertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)
```
and in TensorFlow:
```python
from transformers import BertTokenizer, TFBertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)
```
## Training data
[CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall.
## Training procedure
Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes.
Taking the case of RoBERTa-Medium
Stage1:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq128_dataset.pt \
--processes_num 32 --seq_length 128 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \
--learning_rate 1e-4 --batch_size 64 \
--data_processor mlm --target mlm
```
Stage2:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq512_dataset.pt \
--processes_num 32 --seq_length 512 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \
--learning_rate 5e-5 --batch_size 16 \
--data_processor mlm --target mlm
```
Finally, we convert the pre-trained model into Huggingface's format:
```
python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \
--output_model_path pytorch_model.bin \
--layers_num 8 --type mlm
```
### BibTeX entry and citation info
```
@article{devlin2018bert,
title={Bert: Pre-training of deep bidirectional transformers for language understanding},
author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1810.04805},
year={2018}
}
@article{liu2019roberta,
title={Roberta: A robustly optimized bert pretraining approach},
author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin},
journal={arXiv preprint arXiv:1907.11692},
year={2019}
}
@article{turc2019,
title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models},
author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1908.08962v2 },
year={2019}
}
@article{zhao2019uer,
title={UER: An Open-Source Toolkit for Pre-training Models},
author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong},
journal={EMNLP-IJCNLP 2019},
pages={241},
year={2019}
}
@article{zhao2023tencentpretrain,
title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities},
author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others},
journal={ACL 2023},
pages={217},
year={2023}
}
```
[2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128
[2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256
[2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512
[2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768
[4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128
[4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256
[4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512
[4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768
[6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128
[6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256
[6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512
[6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768
[8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128
[8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256
[8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512
[8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768
[10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128
[10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256
[10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512
[10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768
[12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128
[12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256
[12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512
[12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
|
ellen-0221/dl-reddit-misinformation-classifier
|
ellen-0221
| 2023-08-30T02:27:53Z | 0 | 0 |
keras
|
[
"keras",
"tf-keras",
"region:us"
] | null | 2023-08-29T13:33:47Z |
---
library_name: keras
---
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
| Hyperparameters | Value |
| :-- | :-- |
| name | Adam |
| learning_rate | 0.0010000000474974513 |
| decay | 0.0 |
| beta_1 | 0.8999999761581421 |
| beta_2 | 0.9990000128746033 |
| epsilon | 1e-07 |
| amsgrad | False |
| training_precision | float32 |
|
uer/chinese_roberta_L-8_H-256
|
uer
| 2023-08-30T02:27:44Z | 111 | 1 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"zh",
"dataset:CLUECorpusSmall",
"arxiv:1909.05658",
"arxiv:2212.06385",
"arxiv:1908.08962",
"arxiv:1810.04805",
"arxiv:1907.11692",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-02T23:29:05Z |
---
language: zh
datasets: CLUECorpusSmall
widget:
- text: "北京是[MASK]国的首都。"
---
# Chinese RoBERTa Miniatures
## Model description
This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework.
[Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details.
You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below:
| | H=128 | H=256 | H=512 | H=768 |
| -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: |
| **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] |
| **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] |
| **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] |
| **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] |
| **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] |
| **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] |
Here are scores on the devlopment set of six Chinese tasks:
| Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) |
| -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: |
| RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 |
| RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 |
| RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 |
| RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 |
| RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 |
For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128:
- epochs: 3, 5, 8
- batch sizes: 32, 64
- learning rates: 3e-5, 1e-4, 3e-4
## How to use
You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium):
```python
>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512')
>>> unmasker("中国的首都是[MASK]京。")
[
{'sequence': '[CLS] 中 国 的 首 都 是 北 京 。 [SEP]',
'score': 0.8701988458633423,
'token': 1266,
'token_str': '北'},
{'sequence': '[CLS] 中 国 的 首 都 是 南 京 。 [SEP]',
'score': 0.1194809079170227,
'token': 1298,
'token_str': '南'},
{'sequence': '[CLS] 中 国 的 首 都 是 东 京 。 [SEP]',
'score': 0.0037803512532263994,
'token': 691,
'token_str': '东'},
{'sequence': '[CLS] 中 国 的 首 都 是 普 京 。 [SEP]',
'score': 0.0017127094324678183,
'token': 3249,
'token_str': '普'},
{'sequence': '[CLS] 中 国 的 首 都 是 望 京 。 [SEP]',
'score': 0.001687526935711503,
'token': 3307,
'token_str': '望'}
]
```
Here is how to use this model to get the features of a given text in PyTorch:
```python
from transformers import BertTokenizer, BertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)
```
and in TensorFlow:
```python
from transformers import BertTokenizer, TFBertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)
```
## Training data
[CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall.
## Training procedure
Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes.
Taking the case of RoBERTa-Medium
Stage1:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq128_dataset.pt \
--processes_num 32 --seq_length 128 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \
--learning_rate 1e-4 --batch_size 64 \
--data_processor mlm --target mlm
```
Stage2:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq512_dataset.pt \
--processes_num 32 --seq_length 512 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \
--learning_rate 5e-5 --batch_size 16 \
--data_processor mlm --target mlm
```
Finally, we convert the pre-trained model into Huggingface's format:
```
python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \
--output_model_path pytorch_model.bin \
--layers_num 8 --type mlm
```
### BibTeX entry and citation info
```
@article{devlin2018bert,
title={Bert: Pre-training of deep bidirectional transformers for language understanding},
author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1810.04805},
year={2018}
}
@article{liu2019roberta,
title={Roberta: A robustly optimized bert pretraining approach},
author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin},
journal={arXiv preprint arXiv:1907.11692},
year={2019}
}
@article{turc2019,
title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models},
author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1908.08962v2 },
year={2019}
}
@article{zhao2019uer,
title={UER: An Open-Source Toolkit for Pre-training Models},
author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong},
journal={EMNLP-IJCNLP 2019},
pages={241},
year={2019}
}
@article{zhao2023tencentpretrain,
title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities},
author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others},
journal={ACL 2023},
pages={217},
year={2023}
}
```
[2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128
[2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256
[2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512
[2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768
[4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128
[4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256
[4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512
[4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768
[6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128
[6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256
[6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512
[6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768
[8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128
[8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256
[8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512
[8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768
[10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128
[10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256
[10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512
[10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768
[12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128
[12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256
[12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512
[12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
|
uer/chinese_roberta_L-6_H-768
|
uer
| 2023-08-30T02:27:10Z | 125 | 2 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"zh",
"dataset:CLUECorpusSmall",
"arxiv:1909.05658",
"arxiv:2212.06385",
"arxiv:1908.08962",
"arxiv:1810.04805",
"arxiv:1907.11692",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-02T23:29:05Z |
---
language: zh
datasets: CLUECorpusSmall
widget:
- text: "北京是[MASK]国的首都。"
---
# Chinese RoBERTa Miniatures
## Model description
This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework.
[Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details.
You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below:
| | H=128 | H=256 | H=512 | H=768 |
| -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: |
| **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] |
| **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] |
| **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] |
| **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] |
| **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] |
| **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] |
Here are scores on the devlopment set of six Chinese tasks:
| Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) |
| -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: |
| RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 |
| RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 |
| RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 |
| RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 |
| RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 |
For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128:
- epochs: 3, 5, 8
- batch sizes: 32, 64
- learning rates: 3e-5, 1e-4, 3e-4
## How to use
You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium):
```python
>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512')
>>> unmasker("中国的首都是[MASK]京。")
[
{'sequence': '[CLS] 中 国 的 首 都 是 北 京 。 [SEP]',
'score': 0.8701988458633423,
'token': 1266,
'token_str': '北'},
{'sequence': '[CLS] 中 国 的 首 都 是 南 京 。 [SEP]',
'score': 0.1194809079170227,
'token': 1298,
'token_str': '南'},
{'sequence': '[CLS] 中 国 的 首 都 是 东 京 。 [SEP]',
'score': 0.0037803512532263994,
'token': 691,
'token_str': '东'},
{'sequence': '[CLS] 中 国 的 首 都 是 普 京 。 [SEP]',
'score': 0.0017127094324678183,
'token': 3249,
'token_str': '普'},
{'sequence': '[CLS] 中 国 的 首 都 是 望 京 。 [SEP]',
'score': 0.001687526935711503,
'token': 3307,
'token_str': '望'}
]
```
Here is how to use this model to get the features of a given text in PyTorch:
```python
from transformers import BertTokenizer, BertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)
```
and in TensorFlow:
```python
from transformers import BertTokenizer, TFBertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)
```
## Training data
[CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall.
## Training procedure
Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes.
Taking the case of RoBERTa-Medium
Stage1:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq128_dataset.pt \
--processes_num 32 --seq_length 128 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \
--learning_rate 1e-4 --batch_size 64 \
--data_processor mlm --target mlm
```
Stage2:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq512_dataset.pt \
--processes_num 32 --seq_length 512 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \
--learning_rate 5e-5 --batch_size 16 \
--data_processor mlm --target mlm
```
Finally, we convert the pre-trained model into Huggingface's format:
```
python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \
--output_model_path pytorch_model.bin \
--layers_num 8 --type mlm
```
### BibTeX entry and citation info
```
@article{devlin2018bert,
title={Bert: Pre-training of deep bidirectional transformers for language understanding},
author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1810.04805},
year={2018}
}
@article{liu2019roberta,
title={Roberta: A robustly optimized bert pretraining approach},
author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin},
journal={arXiv preprint arXiv:1907.11692},
year={2019}
}
@article{turc2019,
title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models},
author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1908.08962v2 },
year={2019}
}
@article{zhao2019uer,
title={UER: An Open-Source Toolkit for Pre-training Models},
author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong},
journal={EMNLP-IJCNLP 2019},
pages={241},
year={2019}
}
@article{zhao2023tencentpretrain,
title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities},
author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others},
journal={ACL 2023},
pages={217},
year={2023}
}
```
[2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128
[2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256
[2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512
[2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768
[4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128
[4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256
[4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512
[4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768
[6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128
[6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256
[6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512
[6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768
[8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128
[8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256
[8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512
[8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768
[10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128
[10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256
[10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512
[10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768
[12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128
[12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256
[12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512
[12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
|
uer/chinese_roberta_L-6_H-512
|
uer
| 2023-08-30T02:26:53Z | 111 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"zh",
"dataset:CLUECorpusSmall",
"arxiv:1909.05658",
"arxiv:2212.06385",
"arxiv:1908.08962",
"arxiv:1810.04805",
"arxiv:1907.11692",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-02T23:29:05Z |
---
language: zh
datasets: CLUECorpusSmall
widget:
- text: "北京是[MASK]国的首都。"
---
# Chinese RoBERTa Miniatures
## Model description
This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework.
[Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details.
You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below:
| | H=128 | H=256 | H=512 | H=768 |
| -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: |
| **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] |
| **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] |
| **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] |
| **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] |
| **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] |
| **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] |
Here are scores on the devlopment set of six Chinese tasks:
| Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) |
| -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: |
| RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 |
| RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 |
| RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 |
| RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 |
| RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 |
For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128:
- epochs: 3, 5, 8
- batch sizes: 32, 64
- learning rates: 3e-5, 1e-4, 3e-4
## How to use
You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium):
```python
>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512')
>>> unmasker("中国的首都是[MASK]京。")
[
{'sequence': '[CLS] 中 国 的 首 都 是 北 京 。 [SEP]',
'score': 0.8701988458633423,
'token': 1266,
'token_str': '北'},
{'sequence': '[CLS] 中 国 的 首 都 是 南 京 。 [SEP]',
'score': 0.1194809079170227,
'token': 1298,
'token_str': '南'},
{'sequence': '[CLS] 中 国 的 首 都 是 东 京 。 [SEP]',
'score': 0.0037803512532263994,
'token': 691,
'token_str': '东'},
{'sequence': '[CLS] 中 国 的 首 都 是 普 京 。 [SEP]',
'score': 0.0017127094324678183,
'token': 3249,
'token_str': '普'},
{'sequence': '[CLS] 中 国 的 首 都 是 望 京 。 [SEP]',
'score': 0.001687526935711503,
'token': 3307,
'token_str': '望'}
]
```
Here is how to use this model to get the features of a given text in PyTorch:
```python
from transformers import BertTokenizer, BertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)
```
and in TensorFlow:
```python
from transformers import BertTokenizer, TFBertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)
```
## Training data
[CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall.
## Training procedure
Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes.
Taking the case of RoBERTa-Medium
Stage1:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq128_dataset.pt \
--processes_num 32 --seq_length 128 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \
--learning_rate 1e-4 --batch_size 64 \
--data_processor mlm --target mlm
```
Stage2:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq512_dataset.pt \
--processes_num 32 --seq_length 512 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \
--learning_rate 5e-5 --batch_size 16 \
--data_processor mlm --target mlm
```
Finally, we convert the pre-trained model into Huggingface's format:
```
python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \
--output_model_path pytorch_model.bin \
--layers_num 8 --type mlm
```
### BibTeX entry and citation info
```
@article{devlin2018bert,
title={Bert: Pre-training of deep bidirectional transformers for language understanding},
author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1810.04805},
year={2018}
}
@article{liu2019roberta,
title={Roberta: A robustly optimized bert pretraining approach},
author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin},
journal={arXiv preprint arXiv:1907.11692},
year={2019}
}
@article{turc2019,
title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models},
author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1908.08962v2 },
year={2019}
}
@article{zhao2019uer,
title={UER: An Open-Source Toolkit for Pre-training Models},
author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong},
journal={EMNLP-IJCNLP 2019},
pages={241},
year={2019}
}
@article{zhao2023tencentpretrain,
title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities},
author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others},
journal={ACL 2023},
pages={217},
year={2023}
}
```
[2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128
[2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256
[2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512
[2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768
[4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128
[4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256
[4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512
[4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768
[6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128
[6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256
[6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512
[6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768
[8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128
[8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256
[8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512
[8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768
[10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128
[10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256
[10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512
[10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768
[12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128
[12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256
[12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512
[12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
|
fuguizhen/q-FrozenLake-v1-4x4-noSlippery
|
fuguizhen
| 2023-08-30T02:26:51Z | 0 | 0 | null |
[
"FrozenLake-v1-4x4-no_slippery",
"q-learning",
"reinforcement-learning",
"custom-implementation",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-08-30T02:26:49Z |
---
tags:
- FrozenLake-v1-4x4-no_slippery
- q-learning
- reinforcement-learning
- custom-implementation
model-index:
- name: q-FrozenLake-v1-4x4-noSlippery
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: FrozenLake-v1-4x4-no_slippery
type: FrozenLake-v1-4x4-no_slippery
metrics:
- type: mean_reward
value: 1.00 +/- 0.00
name: mean_reward
verified: false
---
# **Q-Learning** Agent playing1 **FrozenLake-v1**
This is a trained model of a **Q-Learning** agent playing **FrozenLake-v1** .
## Usage
model = load_from_hub(repo_id="fuguizhen/q-FrozenLake-v1-4x4-noSlippery", filename="q-learning.pkl")
# Don't forget to check if you need to add additional attributes (is_slippery=False etc)
env = gym.make(model["env_id"])
|
uer/chinese_roberta_L-6_H-256
|
uer
| 2023-08-30T02:26:30Z | 108 | 1 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"zh",
"dataset:CLUECorpusSmall",
"arxiv:1909.05658",
"arxiv:2212.06385",
"arxiv:1908.08962",
"arxiv:1810.04805",
"arxiv:1907.11692",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-02T23:29:05Z |
---
language: zh
datasets: CLUECorpusSmall
widget:
- text: "北京是[MASK]国的首都。"
---
# Chinese RoBERTa Miniatures
## Model description
This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework.
[Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details.
You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below:
| | H=128 | H=256 | H=512 | H=768 |
| -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: |
| **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] |
| **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] |
| **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] |
| **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] |
| **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] |
| **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] |
Here are scores on the devlopment set of six Chinese tasks:
| Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) |
| -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: |
| RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 |
| RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 |
| RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 |
| RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 |
| RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 |
For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128:
- epochs: 3, 5, 8
- batch sizes: 32, 64
- learning rates: 3e-5, 1e-4, 3e-4
## How to use
You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium):
```python
>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512')
>>> unmasker("中国的首都是[MASK]京。")
[
{'sequence': '[CLS] 中 国 的 首 都 是 北 京 。 [SEP]',
'score': 0.8701988458633423,
'token': 1266,
'token_str': '北'},
{'sequence': '[CLS] 中 国 的 首 都 是 南 京 。 [SEP]',
'score': 0.1194809079170227,
'token': 1298,
'token_str': '南'},
{'sequence': '[CLS] 中 国 的 首 都 是 东 京 。 [SEP]',
'score': 0.0037803512532263994,
'token': 691,
'token_str': '东'},
{'sequence': '[CLS] 中 国 的 首 都 是 普 京 。 [SEP]',
'score': 0.0017127094324678183,
'token': 3249,
'token_str': '普'},
{'sequence': '[CLS] 中 国 的 首 都 是 望 京 。 [SEP]',
'score': 0.001687526935711503,
'token': 3307,
'token_str': '望'}
]
```
Here is how to use this model to get the features of a given text in PyTorch:
```python
from transformers import BertTokenizer, BertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)
```
and in TensorFlow:
```python
from transformers import BertTokenizer, TFBertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)
```
## Training data
[CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall.
## Training procedure
Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes.
Taking the case of RoBERTa-Medium
Stage1:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq128_dataset.pt \
--processes_num 32 --seq_length 128 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \
--learning_rate 1e-4 --batch_size 64 \
--data_processor mlm --target mlm
```
Stage2:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq512_dataset.pt \
--processes_num 32 --seq_length 512 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \
--learning_rate 5e-5 --batch_size 16 \
--data_processor mlm --target mlm
```
Finally, we convert the pre-trained model into Huggingface's format:
```
python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \
--output_model_path pytorch_model.bin \
--layers_num 8 --type mlm
```
### BibTeX entry and citation info
```
@article{devlin2018bert,
title={Bert: Pre-training of deep bidirectional transformers for language understanding},
author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1810.04805},
year={2018}
}
@article{liu2019roberta,
title={Roberta: A robustly optimized bert pretraining approach},
author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin},
journal={arXiv preprint arXiv:1907.11692},
year={2019}
}
@article{turc2019,
title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models},
author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1908.08962v2 },
year={2019}
}
@article{zhao2019uer,
title={UER: An Open-Source Toolkit for Pre-training Models},
author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong},
journal={EMNLP-IJCNLP 2019},
pages={241},
year={2019}
}
@article{zhao2023tencentpretrain,
title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities},
author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others},
journal={ACL 2023},
pages={217},
year={2023}
}
```
[2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128
[2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256
[2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512
[2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768
[4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128
[4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256
[4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512
[4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768
[6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128
[6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256
[6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512
[6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768
[8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128
[8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256
[8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512
[8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768
[10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128
[10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256
[10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512
[10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768
[12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128
[12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256
[12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512
[12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
|
uer/chinese_roberta_L-6_H-128
|
uer
| 2023-08-30T02:26:10Z | 119 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"zh",
"dataset:CLUECorpusSmall",
"arxiv:1909.05658",
"arxiv:2212.06385",
"arxiv:1908.08962",
"arxiv:1810.04805",
"arxiv:1907.11692",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-02T23:29:05Z |
---
language: zh
datasets: CLUECorpusSmall
widget:
- text: "北京是[MASK]国的首都。"
---
# Chinese RoBERTa Miniatures
## Model description
This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework.
[Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details.
You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below:
| | H=128 | H=256 | H=512 | H=768 |
| -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: |
| **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] |
| **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] |
| **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] |
| **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] |
| **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] |
| **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] |
Here are scores on the devlopment set of six Chinese tasks:
| Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) |
| -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: |
| RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 |
| RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 |
| RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 |
| RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 |
| RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 |
For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128:
- epochs: 3, 5, 8
- batch sizes: 32, 64
- learning rates: 3e-5, 1e-4, 3e-4
## How to use
You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium):
```python
>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512')
>>> unmasker("中国的首都是[MASK]京。")
[
{'sequence': '[CLS] 中 国 的 首 都 是 北 京 。 [SEP]',
'score': 0.8701988458633423,
'token': 1266,
'token_str': '北'},
{'sequence': '[CLS] 中 国 的 首 都 是 南 京 。 [SEP]',
'score': 0.1194809079170227,
'token': 1298,
'token_str': '南'},
{'sequence': '[CLS] 中 国 的 首 都 是 东 京 。 [SEP]',
'score': 0.0037803512532263994,
'token': 691,
'token_str': '东'},
{'sequence': '[CLS] 中 国 的 首 都 是 普 京 。 [SEP]',
'score': 0.0017127094324678183,
'token': 3249,
'token_str': '普'},
{'sequence': '[CLS] 中 国 的 首 都 是 望 京 。 [SEP]',
'score': 0.001687526935711503,
'token': 3307,
'token_str': '望'}
]
```
Here is how to use this model to get the features of a given text in PyTorch:
```python
from transformers import BertTokenizer, BertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)
```
and in TensorFlow:
```python
from transformers import BertTokenizer, TFBertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)
```
## Training data
[CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall.
## Training procedure
Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes.
Taking the case of RoBERTa-Medium
Stage1:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq128_dataset.pt \
--processes_num 32 --seq_length 128 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \
--learning_rate 1e-4 --batch_size 64 \
--data_processor mlm --target mlm
```
Stage2:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq512_dataset.pt \
--processes_num 32 --seq_length 512 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \
--learning_rate 5e-5 --batch_size 16 \
--data_processor mlm --target mlm
```
Finally, we convert the pre-trained model into Huggingface's format:
```
python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \
--output_model_path pytorch_model.bin \
--layers_num 8 --type mlm
```
### BibTeX entry and citation info
```
@article{devlin2018bert,
title={Bert: Pre-training of deep bidirectional transformers for language understanding},
author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1810.04805},
year={2018}
}
@article{liu2019roberta,
title={Roberta: A robustly optimized bert pretraining approach},
author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin},
journal={arXiv preprint arXiv:1907.11692},
year={2019}
}
@article{turc2019,
title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models},
author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1908.08962v2 },
year={2019}
}
@article{zhao2019uer,
title={UER: An Open-Source Toolkit for Pre-training Models},
author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong},
journal={EMNLP-IJCNLP 2019},
pages={241},
year={2019}
}
@article{zhao2023tencentpretrain,
title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities},
author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others},
journal={ACL 2023},
pages={217},
year={2023}
}
```
[2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128
[2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256
[2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512
[2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768
[4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128
[4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256
[4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512
[4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768
[6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128
[6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256
[6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512
[6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768
[8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128
[8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256
[8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512
[8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768
[10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128
[10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256
[10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512
[10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768
[12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128
[12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256
[12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512
[12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
|
uer/chinese_roberta_L-2_H-256
|
uer
| 2023-08-30T02:23:47Z | 113 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"zh",
"dataset:CLUECorpusSmall",
"arxiv:1909.05658",
"arxiv:2212.06385",
"arxiv:1908.08962",
"arxiv:1810.04805",
"arxiv:1907.11692",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-02T23:29:05Z |
---
language: zh
datasets: CLUECorpusSmall
widget:
- text: "北京是[MASK]国的首都。"
---
# Chinese RoBERTa Miniatures
## Model description
This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework.
[Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details.
You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below:
| | H=128 | H=256 | H=512 | H=768 |
| -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: |
| **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] |
| **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] |
| **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] |
| **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] |
| **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] |
| **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] |
Here are scores on the devlopment set of six Chinese tasks:
| Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) |
| -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: |
| RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 |
| RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 |
| RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 |
| RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 |
| RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 |
For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128:
- epochs: 3, 5, 8
- batch sizes: 32, 64
- learning rates: 3e-5, 1e-4, 3e-4
## How to use
You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium):
```python
>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512')
>>> unmasker("中国的首都是[MASK]京。")
[
{'sequence': '[CLS] 中 国 的 首 都 是 北 京 。 [SEP]',
'score': 0.8701988458633423,
'token': 1266,
'token_str': '北'},
{'sequence': '[CLS] 中 国 的 首 都 是 南 京 。 [SEP]',
'score': 0.1194809079170227,
'token': 1298,
'token_str': '南'},
{'sequence': '[CLS] 中 国 的 首 都 是 东 京 。 [SEP]',
'score': 0.0037803512532263994,
'token': 691,
'token_str': '东'},
{'sequence': '[CLS] 中 国 的 首 都 是 普 京 。 [SEP]',
'score': 0.0017127094324678183,
'token': 3249,
'token_str': '普'},
{'sequence': '[CLS] 中 国 的 首 都 是 望 京 。 [SEP]',
'score': 0.001687526935711503,
'token': 3307,
'token_str': '望'}
]
```
Here is how to use this model to get the features of a given text in PyTorch:
```python
from transformers import BertTokenizer, BertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)
```
and in TensorFlow:
```python
from transformers import BertTokenizer, TFBertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)
```
## Training data
[CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall.
## Training procedure
Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes.
Taking the case of RoBERTa-Medium
Stage1:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq128_dataset.pt \
--processes_num 32 --seq_length 128 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \
--learning_rate 1e-4 --batch_size 64 \
--data_processor mlm --target mlm
```
Stage2:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq512_dataset.pt \
--processes_num 32 --seq_length 512 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \
--learning_rate 5e-5 --batch_size 16 \
--data_processor mlm --target mlm
```
Finally, we convert the pre-trained model into Huggingface's format:
```
python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \
--output_model_path pytorch_model.bin \
--layers_num 8 --type mlm
```
### BibTeX entry and citation info
```
@article{devlin2018bert,
title={Bert: Pre-training of deep bidirectional transformers for language understanding},
author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1810.04805},
year={2018}
}
@article{liu2019roberta,
title={Roberta: A robustly optimized bert pretraining approach},
author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin},
journal={arXiv preprint arXiv:1907.11692},
year={2019}
}
@article{turc2019,
title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models},
author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1908.08962v2 },
year={2019}
}
@article{zhao2019uer,
title={UER: An Open-Source Toolkit for Pre-training Models},
author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong},
journal={EMNLP-IJCNLP 2019},
pages={241},
year={2019}
}
@article{zhao2023tencentpretrain,
title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities},
author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others},
journal={ACL 2023},
pages={217},
year={2023}
}
```
[2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128
[2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256
[2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512
[2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768
[4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128
[4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256
[4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512
[4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768
[6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128
[6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256
[6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512
[6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768
[8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128
[8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256
[8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512
[8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768
[10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128
[10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256
[10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512
[10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768
[12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128
[12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256
[12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512
[12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
|
uer/chinese_roberta_L-2_H-128
|
uer
| 2023-08-30T02:23:10Z | 1,633 | 11 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"zh",
"dataset:CLUECorpusSmall",
"arxiv:1909.05658",
"arxiv:2212.06385",
"arxiv:1908.08962",
"arxiv:1810.04805",
"arxiv:1907.11692",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-02T23:29:05Z |
---
language: zh
datasets: CLUECorpusSmall
widget:
- text: "北京是[MASK]国的首都。"
---
# Chinese RoBERTa Miniatures
## Model description
This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework.
[Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details.
You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below:
| | H=128 | H=256 | H=512 | H=768 |
| -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: |
| **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] |
| **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] |
| **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] |
| **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] |
| **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] |
| **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] |
Here are scores on the devlopment set of six Chinese tasks:
| Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) |
| -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: |
| RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 |
| RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 |
| RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 |
| RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 |
| RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 |
For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128:
- epochs: 3, 5, 8
- batch sizes: 32, 64
- learning rates: 3e-5, 1e-4, 3e-4
## How to use
You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium):
```python
>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512')
>>> unmasker("中国的首都是[MASK]京。")
[
{'sequence': '[CLS] 中 国 的 首 都 是 北 京 。 [SEP]',
'score': 0.8701988458633423,
'token': 1266,
'token_str': '北'},
{'sequence': '[CLS] 中 国 的 首 都 是 南 京 。 [SEP]',
'score': 0.1194809079170227,
'token': 1298,
'token_str': '南'},
{'sequence': '[CLS] 中 国 的 首 都 是 东 京 。 [SEP]',
'score': 0.0037803512532263994,
'token': 691,
'token_str': '东'},
{'sequence': '[CLS] 中 国 的 首 都 是 普 京 。 [SEP]',
'score': 0.0017127094324678183,
'token': 3249,
'token_str': '普'},
{'sequence': '[CLS] 中 国 的 首 都 是 望 京 。 [SEP]',
'score': 0.001687526935711503,
'token': 3307,
'token_str': '望'}
]
```
Here is how to use this model to get the features of a given text in PyTorch:
```python
from transformers import BertTokenizer, BertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)
```
and in TensorFlow:
```python
from transformers import BertTokenizer, TFBertModel
tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512')
model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)
```
## Training data
[CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall.
## Training procedure
Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes.
Taking the case of RoBERTa-Medium
Stage1:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq128_dataset.pt \
--processes_num 32 --seq_length 128 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \
--learning_rate 1e-4 --batch_size 64 \
--data_processor mlm --target mlm
```
Stage2:
```
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--vocab_path models/google_zh_vocab.txt \
--dataset_path cluecorpussmall_seq512_dataset.pt \
--processes_num 32 --seq_length 512 \
--dynamic_masking --data_processor mlm
```
```
python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \
--vocab_path models/google_zh_vocab.txt \
--pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \
--learning_rate 5e-5 --batch_size 16 \
--data_processor mlm --target mlm
```
Finally, we convert the pre-trained model into Huggingface's format:
```
python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \
--output_model_path pytorch_model.bin \
--layers_num 8 --type mlm
```
### BibTeX entry and citation info
```
@article{devlin2018bert,
title={Bert: Pre-training of deep bidirectional transformers for language understanding},
author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1810.04805},
year={2018}
}
@article{liu2019roberta,
title={Roberta: A robustly optimized bert pretraining approach},
author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin},
journal={arXiv preprint arXiv:1907.11692},
year={2019}
}
@article{turc2019,
title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models},
author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1908.08962v2 },
year={2019}
}
@article{zhao2019uer,
title={UER: An Open-Source Toolkit for Pre-training Models},
author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong},
journal={EMNLP-IJCNLP 2019},
pages={241},
year={2019}
}
@article{zhao2023tencentpretrain,
title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities},
author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others},
journal={ACL 2023},
pages={217},
year={2023}
}
```
[2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128
[2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256
[2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512
[2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768
[4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128
[4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256
[4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512
[4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768
[6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128
[6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256
[6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512
[6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768
[8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128
[8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256
[8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512
[8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768
[10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128
[10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256
[10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512
[10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768
[12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128
[12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256
[12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512
[12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
|
akashmaggon/lamini70m
|
akashmaggon
| 2023-08-30T02:16:52Z | 170 | 0 |
transformers
|
[
"transformers",
"pytorch",
"gpt_neox",
"text-generation",
"generated_from_trainer",
"base_model:EleutherAI/pythia-70m",
"base_model:finetune:EleutherAI/pythia-70m",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-08-30T00:55:54Z |
---
license: apache-2.0
base_model: EleutherAI/pythia-70m
tags:
- generated_from_trainer
model-index:
- name: lamini70m
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# lamini70m
This model is a fine-tuned version of [EleutherAI/pythia-70m](https://huggingface.co/EleutherAI/pythia-70m) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6972
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 1.2201 | 1.0 | 43 | 0.7105 |
| 0.6949 | 2.0 | 86 | 0.6984 |
| 0.6667 | 3.0 | 129 | 0.6972 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3
|
TheRains/cv9-special-batch8-lr3-small
|
TheRains
| 2023-08-30T02:14:30Z | 118 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"whisper",
"automatic-speech-recognition",
"whisper-event",
"generated_from_trainer",
"id",
"dataset:mozilla-foundation/common_voice_9_0",
"base_model:openai/whisper-small",
"base_model:finetune:openai/whisper-small",
"license:apache-2.0",
"model-index",
"endpoints_compatible",
"region:us"
] |
automatic-speech-recognition
| 2023-08-29T23:55:10Z |
---
language:
- id
license: apache-2.0
base_model: openai/whisper-small
tags:
- whisper-event
- generated_from_trainer
datasets:
- mozilla-foundation/common_voice_9_0
metrics:
- wer
model-index:
- name: Whisper Small Indonesian
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: mozilla-foundation/common_voice_9_0 id
type: mozilla-foundation/common_voice_9_0
config: id
split: test
args: id
metrics:
- name: Wer
type: wer
value: 99.20864964343225
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# Whisper Small Indonesian
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the mozilla-foundation/common_voice_9_0 id dataset.
It achieves the following results on the evaluation set:
- Loss: 2.8029
- Wer: 99.2086
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.001
- train_batch_size: 8
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 5000
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 3.9757 | 0.97 | 1000 | 3.3365 | 126.4090 |
| 3.4153 | 1.94 | 2000 | 2.9701 | 105.5855 |
| 2.9747 | 2.9 | 3000 | 2.8029 | 99.2086 |
| 2.6552 | 3.87 | 4000 | 2.6929 | 102.4891 |
| 1.9795 | 4.84 | 5000 | 2.7744 | 104.8263 |
### Framework versions
- Transformers 4.31.0.dev0
- Pytorch 2.0.1+cu117
- Datasets 2.13.1
- Tokenizers 0.13.3
|
jason-lee08/my_model
|
jason-lee08
| 2023-08-30T01:21:53Z | 186 | 0 |
transformers
|
[
"transformers",
"safetensors",
"gpt_neo",
"text-generation",
"generated_from_trainer",
"base_model:roneneldan/TinyStories-33M",
"base_model:finetune:roneneldan/TinyStories-33M",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-08-30T01:20:23Z |
---
base_model: roneneldan/TinyStories-33M
tags:
- generated_from_trainer
model-index:
- name: my_model
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# my_model
This model is a fine-tuned version of [roneneldan/TinyStories-33M](https://huggingface.co/roneneldan/TinyStories-33M) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.0283
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 3e-05
- train_batch_size: 32
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 16
- total_train_batch_size: 512
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 8
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| No log | 0.11 | 2 | 1.1872 |
| No log | 0.22 | 4 | 1.0699 |
| No log | 0.33 | 6 | 1.0373 |
| No log | 0.44 | 8 | 1.0283 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.0.1
- Datasets 2.12.0
- Tokenizers 0.13.2
|
FantasticGNU/AnomalyGPT
|
FantasticGNU
| 2023-08-30T00:57:31Z | 0 | 24 | null |
[
"en",
"zh",
"arxiv:2308.15366",
"region:us"
] | null | 2023-08-28T12:04:52Z |
---
language:
- en
- zh
---
# AnomalyGPT: Detecting Industrial Anomalies using Large Vision-Language Models
<p align="left">
🌐 <a href="https://anomalygpt.github.io" target="_blank">Project Page</a> • 🤗 <a href="https://huggingface.co/spaces/FantasticGNU/AnomalyGPT" target="_blank">Online Demo</a> • 📃 <a href="https://arxiv.org/abs/2308.15366" target="_blank">Paper</a> • 🤖 <a href="https://huggingface.co/FantasticGNU/AnomalyGPT" target="_blank">Model</a> • 📹 <a href="https://www.youtube.com/watch?v=lcxBfy0YnNA" target="_blank">Video</a>
</p>
Zhaopeng Gu, Bingke Zhu, Guibo Zhu, Yingying Chen, Ming Tang, Jinqiao Wang
****
<span id='all_catelogue'/>
## Catalogue:
* <a href='#introduction'>1. Introduction</a>
* <a href='#environment'>2. Running AnomalyGPT Demo</a>
* <a href='#install_environment'>2.1 Environment Installation</a>
* <a href='#download_imagebind_model'>2.2 Prepare ImageBind Checkpoint</a>
* <a href='#download_vicuna_model'>2.3 Prepare Vicuna Checkpoint</a>
* <a href='#download_anomalygpt'>2.4 Prepare Delta Weights of AnomalyGPT</a>
* <a href='#running_demo'>2.5 Deploying Demo</a>
* <a href='#train_anomalygpt'>3. Train Your Own AnomalyGPT</a>
* <a href='#data_preparation'>3.1 Data Preparation</a>
* <a href='#training_configurations'>3.2 Training Configurations</a>
* <a href='#model_training'>3.3 Training AnoamlyGPT</a>
* <a href='#examples'>4. Examples</a>
<!-- * <a href='#results'>5. Results</a> -->
* <a href='#license'>License</a>
* <a href='#citation'>Citation</a>
* <a href='#acknowledgments'>Acknowledgments</a>
****
<span id='introduction'/>
### 1. Introduction: <a href='#all_catelogue'>[Back to Top]</a>
!()[./images/compare.png]
**AnomalyGPT** is the first Large Vision-Language Model (LVLM) based Industrial Anomaly Detection (IAD) method that can detect anomalies in industrial images without the need for manually specified thresholds. Existing IAD methods can only provide anomaly scores and need manually threshold setting, while existing LVLMs cannot detect anomalies in the image. AnomalyGPT can not only indicate the presence and location of anomaly but also provide information about the image.
!()[./images/AnomalyGPT.png]
We leverage a pre-trained image encoder and a Large Language Model (LLM) to align IAD images and their corresponding textual descriptions via simulated anomaly data. We employ a lightweight, visual-textual feature-matching-based image decoder to obtain localization result, and design a prompt learner to provide fine-grained semantic to LLM and fine-tune the LVLM using prompt embeddings. Our method can also detect anomalies for previously unseen items with few normal sample provided.
****
<span id='environment'/>
### 2. Running AnomalyGPT Demo <a href='#all_catelogue'>[Back to Top]</a>
<span id='install_environment'/>
#### 2.1 Environment Installation
Clone the repository locally:
```
git clone https://github.com/CASIA-IVA-Lab/AnomalyGPT.git
```
Install the required packages:
```
pip install -r requirements.txt
```
<span id='download_imagebind_model'/>
#### 2.2 Prepare ImageBind Checkpoint:
You can download the pre-trained ImageBind model using [this link](https://dl.fbaipublicfiles.com/imagebind/imagebind_huge.pth). After downloading, put the downloaded file (imagebind_huge.pth) in [[./pretrained_ckpt/imagebind_ckpt/]](./pretrained_ckpt/imagebind_ckpt/) directory.
<span id='download_vicuna_model'/>
#### 2.3 Prepare Vicuna Checkpoint:
To prepare the pre-trained Vicuna model, please follow the instructions provided [[here]](./pretrained_ckpt#1-prepare-vicuna-checkpoint).
<span id='download_anomalygpt'/>
#### 2.4 Prepare Delta Weights of AnomalyGPT:
We use the pre-trained parameters from [PandaGPT](https://github.com/yxuansu/PandaGPT) to initialize our model. You can get the weights of PandaGPT trained with different strategies in the table below. In our experiments and online demo, we use the Vicuna-7B and `openllmplayground/pandagpt_7b_max_len_1024` due to the limitation of computation resource. Better results are expected if switching to Vicuna-13B.
| **Base Language Model** | **Maximum Sequence Length** | **Huggingface Delta Weights Address** |
| :---------------------: | :-------------------------: | :----------------------------------------------------------: |
| Vicuna-7B (version 0) | 512 | [openllmplayground/pandagpt_7b_max_len_512](https://huggingface.co/openllmplayground/pandagpt_7b_max_len_512) |
| Vicuna-7B (version 0) | 1024 | [openllmplayground/pandagpt_7b_max_len_1024](https://huggingface.co/openllmplayground/pandagpt_7b_max_len_1024) |
| Vicuna-13B (version 0) | 256 | [openllmplayground/pandagpt_13b_max_len_256](https://huggingface.co/openllmplayground/pandagpt_13b_max_len_256) |
| Vicuna-13B (version 0) | 400 | [openllmplayground/pandagpt_13b_max_len_400](https://huggingface.co/openllmplayground/pandagpt_13b_max_len_400) |
Please put the downloaded 7B/13B delta weights file (pytorch_model.pt) in the [./pretrained_ckpt/pandagpt_ckpt/7b/](./pretrained_ckpt/pandagpt_ckpt/7b/) or [./pretrained_ckpt/pandagpt_ckpt/13b/](./pretrained_ckpt/pandagpt_ckpt/13b/) directory.
After that, you can download AnomalyGPT weights from the table below.
| Setup and Datasets | Weights Address |
| :---------------------------------------------------------: | :-------------------------------: |
| Unsupervised on MVTec-AD | [AnomalyGPT/train_mvtec](https://huggingface.co/FantasticGNU/AnomalyGPT/blob/main/train_mvtec/pytorch_model.pt) |
| Unsupervised on VisA | [AnomalyGPT/train_visa](https://huggingface.co/FantasticGNU/AnomalyGPT/blob/main/train_visa/pytorch_model.pt) |
| Supervised on MVTec-AD, VisA, MVTec-LOCO-AD and CrackForest | [AnomalyGPT/train_supervised](https://huggingface.co/FantasticGNU/AnomalyGPT/blob/main/train_supervised/pytorch_model.pt) |
After downloading, put the AnomalyGPT weights in the [./code/ckpt/](./code/ckpt/) directory.
In our [online demo](https://huggingface.co/spaces/FantasticGNU/AnomalyGPT), we use the supervised setting as our default model to attain an enhanced user experience. You can also try other weights locally.
<span id='running_demo'/>
#### 2.5. Deploying Demo
Upon completion of previous steps, you can run the demo locally as
```bash
cd ./code/
python web_demo.py
```
****
<span id='train_anomalygpt'/>
### 3. Train Your Own AnomalyGPT <a href='#all_catelogue'>[Back to Top]</a>
**Prerequisites:** Before training the model, making sure the environment is properly installed and the checkpoints of ImageBind, Vicuna and PandaGPT are downloaded.
<span id='data_preparation'/>
#### 3.1 Data Preparation:
You can download MVTec-AD dataset from [[this link]](https://www.mvtec.com/company/research/datasets/mvtec-ad/downloads) and VisA from [[this link]](https://github.com/amazon-science/spot-diff). You can also download pre-training data of PandaGPT from [[here]](https://huggingface.co/datasets/openllmplayground/pandagpt_visual_instruction_dataset/tree/main). After downloading, put the data in the [[./data]](./data/) directory.
The directory of [[./data]](./data/) should look like:
```
data
|---pandagpt4_visual_instruction_data.json
|---images
|-----|-- ...
|---mvtec_anomaly_detection
|-----|-- bottle
|-----|-----|----- ground_truth
|-----|-----|----- test
|-----|-----|----- train
|-----|-- capsules
|-----|-- ...
|----VisA
|-----|-- split_csv
|-----|-----|--- 1cls.csv
|-----|-----|--- ...
|-----|-- candle
|-----|-----|--- Data
|-----|-----|-----|----- Images
|-----|-----|-----|--------|------ Anomaly
|-----|-----|-----|--------|------ Normal
|-----|-----|-----|----- Masks
|-----|-----|-----|--------|------ Anomaly
|-----|-----|--- image_anno.csv
|-----|-- capsules
|-----|-----|----- ...
```
<span id='training_configurations'/>
#### 3.2 Training Configurations
The table below show the training hyperparameters used in our experiments. The hyperparameters are selected based on the constrain of our computational resources, i.e. 2 x RTX3090 GPUs.
| **Base Language Model** | **Epoch Number** | **Batch Size** | **Learning Rate** | **Maximum Length** |
| :---------------------: | :--------------: | :------------: | :---------------: | :----------------: |
| Vicuna-7B | 50 | 16 | 1e-3 | 1024 |
<span id='model_training'/>
#### 3.3 Training AnomalyGPT
To train AnomalyGPT on MVTec-AD dataset, please run the following commands:
```yaml
cd ./code
bash ./scripts/train_mvtec.sh
```
The key arguments of the training script are as follows:
* `--data_path`: The data path for the json file `pandagpt4_visual_instruction_data.json`.
* `--image_root_path`: The root path for training images of PandaGPT.
* `--imagebind_ckpt_path`: The path of ImageBind checkpoint.
* `--vicuna_ckpt_path`: The directory that saves the pre-trained Vicuna checkpoints.
* `--max_tgt_len`: The maximum sequence length of training instances.
* `--save_path`: The directory which saves the trained delta weights. This directory will be automatically created.
* `--log_path`: The directory which saves the log. This directory will be automatically created.
Note that the epoch number can be set in the `epochs` argument at [./code/config/openllama_peft.yaml](./code/config/openllama_peft.yaml) file and the learning rate can be set in [./code/dsconfig/openllama_peft_stage_1.json](./code/dsconfig/openllama_peft_stage_1.json)
****
<span id='examples'/>
### 4. Examples
<h4 align='center'>An image of concrete with crack. </h4>
****
<h4 align='center'>A crack capsule. </h4>
****
<h4 align='center'>An image of a cut hazelnut. </h4>
****
<h4 align='center'>A damaged bottle. </h4>
****
<h4 align='center'>A photo of normal carpet. </h4>
****
<h4 align='center'>A photo of a piece of wood with defect. </h4>
****
<h4 align='center'>A piece of normal fabric. </h4>
****
<span id='license'/>
### License
AnomalyGPT is licensed under the [CC BY-NC-SA 4.0 license](./LICENSE).
****
<span id='citation'/>
### Citation:
If you found AnomalyGPT useful in your research or applications, please kindly cite using the following BibTeX:
```
@article{gu2023anomalyagpt,
title={AnomalyGPT: Detecting Industrial Anomalies using Large Vision-Language Models},
author={Gu, Zhaopeng and Zhu, Bingke and Zhu, Guibo and Chen, Yingying and Tang, Ming and Wang, Jinqiao},
journal={arXiv preprint arXiv:2308.15366},
year={2023}
}
```
****
<span id='acknowledgments'/>
### Acknowledgments:
We borrow some codes and the pre-trained weights from [PandaGPT](https://github.com/yxuansu/PandaGPT). Thanks for their wonderful work!
|
johaanm/llama-test-plannerV5.5
|
johaanm
| 2023-08-30T00:54:58Z | 0 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-08-30T00:54:54Z |
---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float16
### Framework versions
- PEFT 0.4.0
|
touchtech/fashion-images-gender-age-vit-huge-patch14-224-in21k
|
touchtech
| 2023-08-30T00:54:37Z | 13 | 0 |
transformers
|
[
"transformers",
"pytorch",
"vit",
"image-classification",
"vision",
"generated_from_trainer",
"dataset:imagefolder",
"base_model:google/vit-huge-patch14-224-in21k",
"base_model:finetune:google/vit-huge-patch14-224-in21k",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-08-29T19:35:45Z |
---
license: apache-2.0
base_model: google/vit-huge-patch14-224-in21k
tags:
- image-classification
- vision
- generated_from_trainer
datasets:
- imagefolder
metrics:
- accuracy
model-index:
- name: fashion-images-gender-age-vit-huge-patch14-224-in21k
results:
- task:
name: Image Classification
type: image-classification
dataset:
name: touchtech/fashion-images-gender-age
type: imagefolder
config: default
split: train
args: default
metrics:
- name: Accuracy
type: accuracy
value: 0.9894736842105263
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# fashion-images-gender-age-vit-huge-patch14-224-in21k
This model is a fine-tuned version of [google/vit-huge-patch14-224-in21k](https://huggingface.co/google/vit-huge-patch14-224-in21k) on the touchtech/fashion-images-gender-age dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0346
- Accuracy: 0.9895
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 1337
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5.0
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 0.1848 | 1.0 | 2422 | 0.0777 | 0.9798 |
| 0.1055 | 2.0 | 4844 | 0.0708 | 0.9819 |
| 0.0949 | 3.0 | 7266 | 0.0440 | 0.9877 |
| 0.08 | 4.0 | 9688 | 0.0373 | 0.9883 |
| 0.063 | 5.0 | 12110 | 0.0346 | 0.9895 |
### Framework versions
- Transformers 4.33.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3
|
nisten/bigdoc-c7b-instruct-bf16
|
nisten
| 2023-08-30T00:51:24Z | 0 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-08-30T00:51:18Z |
---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: QuantizationMethod.BITS_AND_BYTES
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: bfloat16
### Framework versions
- PEFT 0.6.0.dev0
|
touchtech/fashion-images-perspectives-vit-huge-patch14-224-in21k
|
touchtech
| 2023-08-30T00:48:28Z | 6 | 0 |
transformers
|
[
"transformers",
"pytorch",
"vit",
"image-classification",
"vision",
"generated_from_trainer",
"dataset:imagefolder",
"base_model:google/vit-huge-patch14-224-in21k",
"base_model:finetune:google/vit-huge-patch14-224-in21k",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-08-29T19:30:32Z |
---
license: apache-2.0
base_model: google/vit-huge-patch14-224-in21k
tags:
- image-classification
- vision
- generated_from_trainer
datasets:
- imagefolder
metrics:
- accuracy
model-index:
- name: fashion-images-perspectives-vit-huge-patch14-224-in21k
results:
- task:
name: Image Classification
type: image-classification
dataset:
name: touchtech/fashion-images-perspectives
type: imagefolder
config: default
split: train
args: default
metrics:
- name: Accuracy
type: accuracy
value: 0.9149976711690732
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# fashion-images-perspectives-vit-huge-patch14-224-in21k
This model is a fine-tuned version of [google/vit-huge-patch14-224-in21k](https://huggingface.co/google/vit-huge-patch14-224-in21k) on the touchtech/fashion-images-perspectives dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2604
- Accuracy: 0.9150
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 1337
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5.0
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 0.5932 | 1.0 | 3042 | 0.3863 | 0.8889 |
| 0.4286 | 2.0 | 6084 | 0.4201 | 0.8642 |
| 0.3628 | 3.0 | 9126 | 0.2820 | 0.9101 |
| 0.3183 | 4.0 | 12168 | 0.2604 | 0.9150 |
| 0.2648 | 5.0 | 15210 | 0.2710 | 0.9150 |
### Framework versions
- Transformers 4.33.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3
|
yoo11/taskA_distilbert_model
|
yoo11
| 2023-08-30T00:23:08Z | 61 | 0 |
transformers
|
[
"transformers",
"tf",
"distilbert",
"text-classification",
"generated_from_keras_callback",
"base_model:distilbert/distilbert-base-uncased",
"base_model:finetune:distilbert/distilbert-base-uncased",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-08-30T00:15:17Z |
---
license: apache-2.0
base_model: distilbert-base-uncased
tags:
- generated_from_keras_callback
model-index:
- name: yoo11/taskA_distilbert_model
results: []
---
<!-- This model card has been generated automatically according to the information Keras had access to. You should
probably proofread and complete it, then remove this comment. -->
# yoo11/taskA_distilbert_model
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Train Loss: 0.1501
- Validation Loss: 0.4055
- Train Accuracy: 0.863
- Epoch: 2
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- optimizer: {'name': 'Adam', 'weight_decay': None, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': True, 'is_legacy_optimizer': False, 'learning_rate': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': 4375, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}}, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False}
- training_precision: float32
### Training results
| Train Loss | Validation Loss | Train Accuracy | Epoch |
|:----------:|:---------------:|:--------------:|:-----:|
| 0.3800 | 0.3247 | 0.861 | 0 |
| 0.2582 | 0.3318 | 0.861 | 1 |
| 0.1501 | 0.4055 | 0.863 | 2 |
### Framework versions
- Transformers 4.32.1
- TensorFlow 2.12.0
- Datasets 2.14.4
- Tokenizers 0.13.3
|
botp/chatglm2-6b-Paul
|
botp
| 2023-08-30T00:19:21Z | 14 | 1 |
transformers
|
[
"transformers",
"pytorch",
"chatglm",
"feature-extraction",
"Taiwan",
"Traditional Chinese",
"text-generation",
"custom_code",
"zh",
"license:apache-2.0",
"region:us"
] |
text-generation
| 2023-08-27T12:55:45Z |
---
license: apache-2.0
language:
- zh
tags:
- Taiwan
- Traditional Chinese
widget:
- text: 你聽過台灣總統大選嗎？
example_title: 台灣總統大選
library_name: transformers
pipeline_tag: text-generation
---
# Very Few Dataset Lora Fine-Tuning Verification Model
## 讓大模型學習您教給他的知識 aka 洗腦
##### 語料：47筆
##### 硬體：A6000 48G
##### 訓練時間：28分鐘
|
touchtech/fashion-images-pack-types-vit-large-patch16-224-in21k
|
touchtech
| 2023-08-30T00:12:28Z | 218 | 1 |
transformers
|
[
"transformers",
"pytorch",
"vit",
"image-classification",
"vision",
"generated_from_trainer",
"dataset:imagefolder",
"base_model:google/vit-large-patch16-224-in21k",
"base_model:finetune:google/vit-large-patch16-224-in21k",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-08-29T22:27:47Z |
---
license: apache-2.0
base_model: google/vit-large-patch16-224-in21k
tags:
- image-classification
- vision
- generated_from_trainer
datasets:
- imagefolder
metrics:
- accuracy
model-index:
- name: fashion-images-pack-types-vit-large-patch16-224-in21k
results:
- task:
name: Image Classification
type: image-classification
dataset:
name: touchtech/fashion-images-pack-types
type: imagefolder
config: default
split: train
args: default
metrics:
- name: Accuracy
type: accuracy
value: 0.9894336432797971
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# fashion-images-pack-types-vit-large-patch16-224-in21k
This model is a fine-tuned version of [google/vit-large-patch16-224-in21k](https://huggingface.co/google/vit-large-patch16-224-in21k) on the touchtech/fashion-images-pack-types dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0343
- Accuracy: 0.9894
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 1337
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5.0
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 0.1556 | 1.0 | 1676 | 0.0490 | 0.9861 |
| 0.1185 | 2.0 | 3352 | 0.0343 | 0.9894 |
| 0.0815 | 3.0 | 5028 | 0.0537 | 0.9882 |
| 0.0503 | 4.0 | 6704 | 0.0374 | 0.9915 |
| 0.0447 | 5.0 | 8380 | 0.0362 | 0.9915 |
### Framework versions
- Transformers 4.33.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3
|
victornica/mini_molformer_gsf
|
victornica
| 2023-08-29T23:49:34Z | 150 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"gpt2",
"text-generation",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-08-29T19:20:13Z |
---
license: mit
tags:
- generated_from_trainer
model-index:
- name: mini_molformer_gsf
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# mini_molformer_gsf
This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6773
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0006
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:-----:|:---------------:|
| 1.8056 | 0.1 | 1000 | 1.0889 |
| 1.0277 | 0.19 | 2000 | 0.9550 |
| 0.9453 | 0.29 | 3000 | 0.9052 |
| 0.9026 | 0.39 | 4000 | 0.8735 |
| 0.8751 | 0.48 | 5000 | 0.8492 |
| 0.8538 | 0.58 | 6000 | 0.8306 |
| 0.8381 | 0.68 | 7000 | 0.8125 |
| 0.8225 | 0.77 | 8000 | 0.8030 |
| 0.8124 | 0.87 | 9000 | 0.7906 |
| 0.8 | 0.97 | 10000 | 0.7795 |
| 0.7895 | 1.07 | 11000 | 0.7724 |
| 0.78 | 1.16 | 12000 | 0.7617 |
| 0.7729 | 1.26 | 13000 | 0.7570 |
| 0.766 | 1.36 | 14000 | 0.7477 |
| 0.7577 | 1.45 | 15000 | 0.7425 |
| 0.7524 | 1.55 | 16000 | 0.7354 |
| 0.7454 | 1.65 | 17000 | 0.7286 |
| 0.739 | 1.74 | 18000 | 0.7220 |
| 0.7328 | 1.84 | 19000 | 0.7169 |
| 0.7276 | 1.94 | 20000 | 0.7095 |
| 0.7184 | 2.03 | 21000 | 0.7040 |
| 0.71 | 2.13 | 22000 | 0.6994 |
| 0.7064 | 2.23 | 23000 | 0.6948 |
| 0.7017 | 2.32 | 24000 | 0.6904 |
| 0.6982 | 2.42 | 25000 | 0.6866 |
| 0.6939 | 2.52 | 26000 | 0.6834 |
| 0.6909 | 2.61 | 27000 | 0.6811 |
| 0.6896 | 2.71 | 28000 | 0.6790 |
| 0.6881 | 2.81 | 29000 | 0.6778 |
| 0.6873 | 2.91 | 30000 | 0.6773 |
### Framework versions
- Transformers 4.30.2
- Pytorch 2.0.0
- Datasets 2.1.0
- Tokenizers 0.13.3
|
ashtrevi/distilbert-base-uncased-finetuned-squad-finetuned-thread
|
ashtrevi
| 2023-08-29T23:37:23Z | 123 | 0 |
transformers
|
[
"transformers",
"pytorch",
"distilbert",
"question-answering",
"generated_from_trainer",
"base_model:ashtrevi/distilbert-base-uncased-finetuned-squad",
"base_model:finetune:ashtrevi/distilbert-base-uncased-finetuned-squad",
"endpoints_compatible",
"region:us"
] |
question-answering
| 2023-08-29T23:37:09Z |
---
base_model: ashtrevi/distilbert-base-uncased-finetuned-squad
tags:
- generated_from_trainer
model-index:
- name: distilbert-base-uncased-finetuned-squad-finetuned-thread
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-base-uncased-finetuned-squad-finetuned-thread
This model is a fine-tuned version of [ashtrevi/distilbert-base-uncased-finetuned-squad](https://huggingface.co/ashtrevi/distilbert-base-uncased-finetuned-squad) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 2.4546
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-06
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 4.219 | 0.6 | 100 | 2.7337 |
| 3.3141 | 1.19 | 200 | 2.5648 |
| 3.1199 | 1.79 | 300 | 2.4972 |
| 3.1278 | 2.38 | 400 | 2.4569 |
| 2.9679 | 2.98 | 500 | 2.4546 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3
|
lsoni/bert-finetuned-ner-synonym-replacement
|
lsoni
| 2023-08-29T23:27:58Z | 114 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bert",
"token-classification",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
token-classification
| 2023-08-29T20:04:52Z |
---
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: bert-finetuned-ner-synonym-replacement
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bert-finetuned-ner-synonym-replacement
This model is a fine-tuned version of [bert-base-cased](https://huggingface.co/bert-base-cased) on the combined training dataset(tweetner7(train_2021)+augmented dataset(train_2021) using synonym replacment technique.
It achieves the following results on the evaluation set:
- Loss: 0.4556
- Precision: 0.6825
- Recall: 0.6704
- F1: 0.6764
- Accuracy: 0.8792
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
| 0.5491 | 1.0 | 624 | 0.4161 | 0.7217 | 0.6308 | 0.6732 | 0.8813 |
| 0.3284 | 2.0 | 1248 | 0.4195 | 0.7077 | 0.6485 | 0.6768 | 0.8820 |
| 0.2522 | 3.0 | 1872 | 0.4556 | 0.6825 | 0.6704 | 0.6764 | 0.8792 |
### Framework versions
- Transformers 4.20.1
- Pytorch 1.12.1
- Datasets 2.10.1
- Tokenizers 0.12.1
|
ychenNLP/nllb-200-distilled-1.3B-easyproject
|
ychenNLP
| 2023-08-29T23:20:18Z | 100 | 2 |
transformers
|
[
"transformers",
"pytorch",
"m2m_100",
"text2text-generation",
"nllb",
"translation",
"ace",
"acm",
"acq",
"aeb",
"af",
"ajp",
"ak",
"als",
"am",
"apc",
"ar",
"ars",
"ary",
"arz",
"as",
"ast",
"awa",
"ayr",
"azb",
"azj",
"ba",
"bm",
"ban",
"be",
"bem",
"bn",
"bho",
"bjn",
"bo",
"bs",
"bug",
"bg",
"ca",
"ceb",
"cs",
"cjk",
"ckb",
"crh",
"cy",
"da",
"de",
"dik",
"dyu",
"dz",
"el",
"en",
"eo",
"et",
"eu",
"ee",
"fo",
"fj",
"fi",
"fon",
"fr",
"fur",
"fuv",
"gaz",
"gd",
"ga",
"gl",
"gn",
"gu",
"ht",
"ha",
"he",
"hi",
"hne",
"hr",
"hu",
"hy",
"ig",
"ilo",
"id",
"is",
"it",
"jv",
"ja",
"kab",
"kac",
"kam",
"kn",
"ks",
"ka",
"kk",
"kbp",
"kea",
"khk",
"km",
"ki",
"rw",
"ky",
"kmb",
"kmr",
"knc",
"kg",
"ko",
"lo",
"lij",
"li",
"ln",
"lt",
"lmo",
"ltg",
"lb",
"lua",
"lg",
"luo",
"lus",
"lvs",
"mag",
"mai",
"ml",
"mar",
"min",
"mk",
"mt",
"mni",
"mos",
"mi",
"my",
"nl",
"nn",
"nb",
"npi",
"nso",
"nus",
"ny",
"oc",
"ory",
"pag",
"pa",
"pap",
"pbt",
"pes",
"plt",
"pl",
"pt",
"prs",
"quy",
"ro",
"rn",
"ru",
"sg",
"sa",
"sat",
"scn",
"shn",
"si",
"sk",
"sl",
"sm",
"sn",
"sd",
"so",
"st",
"es",
"sc",
"sr",
"ss",
"su",
"sv",
"swh",
"szl",
"ta",
"taq",
"tt",
"te",
"tg",
"tl",
"th",
"ti",
"tpi",
"tn",
"ts",
"tk",
"tum",
"tr",
"tw",
"tzm",
"ug",
"uk",
"umb",
"ur",
"uzn",
"vec",
"vi",
"war",
"wo",
"xh",
"ydd",
"yo",
"yue",
"zh",
"zsm",
"zu",
"dataset:flores-200",
"arxiv:2211.15613",
"license:cc-by-nc-4.0",
"autotrain_compatible",
"region:us"
] |
translation
| 2023-05-03T06:21:41Z |
---
language:
- ace
- acm
- acq
- aeb
- af
- ajp
- ak
- als
- am
- apc
- ar
- ars
- ary
- arz
- as
- ast
- awa
- ayr
- azb
- azj
- ba
- bm
- ban
- be
- bem
- bn
- bho
- bjn
- bo
- bs
- bug
- bg
- ca
- ceb
- cs
- cjk
- ckb
- crh
- cy
- da
- de
- dik
- dyu
- dz
- el
- en
- eo
- et
- eu
- ee
- fo
- fj
- fi
- fon
- fr
- fur
- fuv
- gaz
- gd
- ga
- gl
- gn
- gu
- ht
- ha
- he
- hi
- hne
- hr
- hu
- hy
- ig
- ilo
- id
- is
- it
- jv
- ja
- kab
- kac
- kam
- kn
- ks
- ka
- kk
- kbp
- kea
- khk
- km
- ki
- rw
- ky
- kmb
- kmr
- knc
- kg
- ko
- lo
- lij
- li
- ln
- lt
- lmo
- ltg
- lb
- lua
- lg
- luo
- lus
- lvs
- mag
- mai
- ml
- mar
- min
- mk
- mt
- mni
- mos
- mi
- my
- nl
- nn
- nb
- npi
- nso
- nus
- ny
- oc
- ory
- pag
- pa
- pap
- pbt
- pes
- plt
- pl
- pt
- prs
- quy
- ro
- rn
- ru
- sg
- sa
- sat
- scn
- shn
- si
- sk
- sl
- sm
- sn
- sd
- so
- st
- es
- sc
- sr
- ss
- su
- sv
- swh
- szl
- ta
- taq
- tt
- te
- tg
- tl
- th
- ti
- tpi
- tn
- ts
- tk
- tum
- tr
- tw
- tzm
- ug
- uk
- umb
- ur
- uzn
- vec
- vi
- war
- wo
- xh
- ydd
- yo
- yue
- zh
- zsm
- zu
language_details: "ace_Arab, ace_Latn, acm_Arab, acq_Arab, aeb_Arab, afr_Latn, ajp_Arab, aka_Latn, amh_Ethi, apc_Arab, arb_Arab, ars_Arab, ary_Arab, arz_Arab, asm_Beng, ast_Latn, awa_Deva, ayr_Latn, azb_Arab, azj_Latn, bak_Cyrl, bam_Latn, ban_Latn,bel_Cyrl, bem_Latn, ben_Beng, bho_Deva, bjn_Arab, bjn_Latn, bod_Tibt, bos_Latn, bug_Latn, bul_Cyrl, cat_Latn, ceb_Latn, ces_Latn, cjk_Latn, ckb_Arab, crh_Latn, cym_Latn, dan_Latn, deu_Latn, dik_Latn, dyu_Latn, dzo_Tibt, ell_Grek, eng_Latn, epo_Latn, est_Latn, eus_Latn, ewe_Latn, fao_Latn, pes_Arab, fij_Latn, fin_Latn, fon_Latn, fra_Latn, fur_Latn, fuv_Latn, gla_Latn, gle_Latn, glg_Latn, grn_Latn, guj_Gujr, hat_Latn, hau_Latn, heb_Hebr, hin_Deva, hne_Deva, hrv_Latn, hun_Latn, hye_Armn, ibo_Latn, ilo_Latn, ind_Latn, isl_Latn, ita_Latn, jav_Latn, jpn_Jpan, kab_Latn, kac_Latn, kam_Latn, kan_Knda, kas_Arab, kas_Deva, kat_Geor, knc_Arab, knc_Latn, kaz_Cyrl, kbp_Latn, kea_Latn, khm_Khmr, kik_Latn, kin_Latn, kir_Cyrl, kmb_Latn, kon_Latn, kor_Hang, kmr_Latn, lao_Laoo, lvs_Latn, lij_Latn, lim_Latn, lin_Latn, lit_Latn, lmo_Latn, ltg_Latn, ltz_Latn, lua_Latn, lug_Latn, luo_Latn, lus_Latn, mag_Deva, mai_Deva, mal_Mlym, mar_Deva, min_Latn, mkd_Cyrl, plt_Latn, mlt_Latn, mni_Beng, khk_Cyrl, mos_Latn, mri_Latn, zsm_Latn, mya_Mymr, nld_Latn, nno_Latn, nob_Latn, npi_Deva, nso_Latn, nus_Latn, nya_Latn, oci_Latn, gaz_Latn, ory_Orya, pag_Latn, pan_Guru, pap_Latn, pol_Latn, por_Latn, prs_Arab, pbt_Arab, quy_Latn, ron_Latn, run_Latn, rus_Cyrl, sag_Latn, san_Deva, sat_Beng, scn_Latn, shn_Mymr, sin_Sinh, slk_Latn, slv_Latn, smo_Latn, sna_Latn, snd_Arab, som_Latn, sot_Latn, spa_Latn, als_Latn, srd_Latn, srp_Cyrl, ssw_Latn, sun_Latn, swe_Latn, swh_Latn, szl_Latn, tam_Taml, tat_Cyrl, tel_Telu, tgk_Cyrl, tgl_Latn, tha_Thai, tir_Ethi, taq_Latn, taq_Tfng, tpi_Latn, tsn_Latn, tso_Latn, tuk_Latn, tum_Latn, tur_Latn, twi_Latn, tzm_Tfng, uig_Arab, ukr_Cyrl, umb_Latn, urd_Arab, uzn_Latn, vec_Latn, vie_Latn, war_Latn, wol_Latn, xho_Latn, ydd_Hebr, yor_Latn, yue_Hant, zho_Hans, zho_Hant, zul_Latn"
tags:
- nllb
- translation
license: "cc-by-nc-4.0"
datasets:
- flores-200
metrics:
- bleu
- spbleu
- chrf++
inference: false
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
This model is a fine-tuned version of [facebook/nllb-200-3.3B](https://huggingface.co/facebook/nllb-200-3.3B) on the [EasyProject](https://github.com/edchengg/easyproject) dataset.
### Framework versions
- Transformers 4.29.2
- Pytorch 1.11.0+cu113
- Datasets 2.8.0
- Tokenizers 0.13.2
- Paper link: [Frustratingly Easy Label Projection for Cross-lingual Transfer](https://arxiv.org/abs/2211.15613)
- Github link: https://github.com/edchengg/easyproject
- Please use the transformers==4.29.2 library as Huggingface recently fixed a bug in [NLLB tokenizer](https://github.com/huggingface/transformers/pull/22313)
# Code
```python
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
import torch
tokenizer = AutoTokenizer.from_pretrained(
"facebook/nllb-200-distilled-600M", src_lang="eng_Latn")
print("Loading model")
model = AutoModelForSeq2SeqLM.from_pretrained("ychenNLP/nllb-200-distilled-1.3b-easyproject")
model.cuda()
input_chunks = ["A translator always risks inadvertently introducing source-language words, grammar, or syntax into the target-language rendering."]
print("Start translation...")
output_result = []
batch_size = 1
for idx in tqdm(range(0, len(input_chunks), batch_size)):
start_idx = idx
end_idx = idx + batch_size
inputs = tokenizer(input_chunks[start_idx: end_idx], padding=True, truncation=True, max_length=128, return_tensors="pt").to('cuda')
with torch.no_grad():
translated_tokens = model.generate(**inputs, forced_bos_token_id=tokenizer.lang_code_to_id["zho_Hans"],
max_length=128, num_beams=5, num_return_sequences=1, early_stopping=True)
output = tokenizer.batch_decode(translated_tokens, skip_special_tokens=True)
output_result.extend(output)
print(output_result)
```
## Citation
```
@inproceedings{chen2023easyproject,
title={Frustratingly Easy Label Projection for Cross-lingual Transfer},
author={Chen, Yang and Jiang, Chao and Ritter, Alan and Xu, Wei},
booktitle={Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Findings)},
year={2023}
}
```
|
DioulaD/birdi-finetuned-ner-address
|
DioulaD
| 2023-08-29T23:16:53Z | 110 | 0 |
transformers
|
[
"transformers",
"pytorch",
"camembert",
"token-classification",
"generated_from_trainer",
"fr",
"base_model:almanach/camembert-base-legacy",
"base_model:finetune:almanach/camembert-base-legacy",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
token-classification
| 2023-08-29T20:44:23Z |
---
base_model: camembert/camembert-base
tags:
- generated_from_trainer
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: birdi-finetuned-ner-address
results: []
language:
- fr
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# birdi-finetuned-ner-address
This model is a fine-tuned version of [camembert/camembert-base](https://huggingface.co/camembert/camembert-base) on the None dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Framework versions
- Transformers 4.33.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3
|
Allenpai/AlpacaLoraRec
|
Allenpai
| 2023-08-29T23:01:27Z | 0 | 0 | null |
[
"region:us"
] | null | 2023-08-06T15:50:40Z |
The following bitsandbytes quantization config was used during training:
load_in_8bit: True
Framework versions
PEFT 0.4.0.dev0
|
Merozed/Merozed
|
Merozed
| 2023-08-29T22:52:36Z | 193 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"vit",
"image-classification",
"huggingpics",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-08-29T22:52:09Z |
---
tags:
- image-classification
- pytorch
- huggingpics
metrics:
- accuracy
- Precision
- Recall
- F1Score
model-index:
- name: FER VIT eval
results:
- task:
name: Image Classification
type: image-classification
metrics:
- name: Accuracy
type: accuracy
value: 0.4703616797924042
- name: Precision
type: Precision
value: 0.3884884715080261
- name: Recall
type: Recall
value: 0.39113321900367737
- name: F1Score
type: F1Score
value: 0.3669418394565582
---
# FER VIT eval
Autogenerated by HuggingPics🤗🖼️
Create your own image classifier for **anything** by running [the demo on Google Colab](https://colab.research.google.com/github/nateraw/huggingpics/blob/main/HuggingPics.ipynb).
Report any issues with the demo at the [github repo](https://github.com/nateraw/huggingpics).
## Example Images
|
jlpan/starcoder-php2py-program2
|
jlpan
| 2023-08-29T22:48:23Z | 0 | 0 | null |
[
"generated_from_trainer",
"base_model:bigcode/starcoder",
"base_model:finetune:bigcode/starcoder",
"license:bigcode-openrail-m",
"region:us"
] | null | 2023-08-29T22:09:39Z |
---
license: bigcode-openrail-m
base_model: bigcode/starcoder
tags:
- generated_from_trainer
model-index:
- name: starcoder-php2py-program2
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# starcoder-php2py-program2
This model is a fine-tuned version of [bigcode/starcoder](https://huggingface.co/bigcode/starcoder) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1172
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 5
- total_train_batch_size: 20
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 10
- training_steps: 100
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 0.1276 | 0.25 | 25 | 0.1184 |
| 0.1196 | 0.5 | 50 | 0.1178 |
| 0.1181 | 1.0 | 75 | 0.1173 |
| 0.1243 | 1.25 | 100 | 0.1172 |
### Framework versions
- Transformers 4.32.0.dev0
- Pytorch 2.0.1+cu117
- Datasets 2.12.0
- Tokenizers 0.13.3
|
ronit01/my_awesome_qa_model
|
ronit01
| 2023-08-29T22:35:16Z | 112 | 0 |
transformers
|
[
"transformers",
"pytorch",
"distilbert",
"question-answering",
"generated_from_trainer",
"base_model:distilbert/distilbert-base-uncased",
"base_model:finetune:distilbert/distilbert-base-uncased",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
question-answering
| 2023-08-27T19:54:52Z |
---
license: apache-2.0
base_model: distilbert-base-uncased
tags:
- generated_from_trainer
model-index:
- name: my_awesome_qa_model
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# my_awesome_qa_model
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5952
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| No log | 1.0 | 26 | 2.1901 |
| No log | 2.0 | 52 | 0.7552 |
| No log | 3.0 | 78 | 0.5952 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3
|
thomasavare/distilbert-ft-test4
|
thomasavare
| 2023-08-29T22:30:54Z | 62 | 0 |
transformers
|
[
"transformers",
"tf",
"distilbert",
"text-classification",
"generated_from_keras_callback",
"base_model:distilbert/distilbert-base-uncased",
"base_model:finetune:distilbert/distilbert-base-uncased",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-08-29T22:30:19Z |
---
license: apache-2.0
base_model: distilbert-base-uncased
tags:
- generated_from_keras_callback
model-index:
- name: distilbert-ft-test4
results: []
---
<!-- This model card has been generated automatically according to the information Keras had access to. You should
probably proofread and complete it, then remove this comment. -->
# distilbert-ft-test4
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset.
It achieves the following results on the evaluation set:
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- optimizer: {'name': 'Adam', 'weight_decay': None, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': True, 'is_legacy_optimizer': False, 'learning_rate': 1e-05, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False}
- training_precision: float32
### Training results
### Framework versions
- Transformers 4.32.1
- TensorFlow 2.12.0
- Datasets 2.14.4
- Tokenizers 0.13.3
|
OmkarB/SQL-GQL-Finetuned-2
|
OmkarB
| 2023-08-29T22:30:38Z | 2 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-08-28T22:50:31Z |
---
library_name: peft
---
## Training procedure
### Framework versions
- PEFT 0.6.0.dev0
|
aldobremen/ppo-Huggy
|
aldobremen
| 2023-08-29T22:30:18Z | 1 | 0 |
ml-agents
|
[
"ml-agents",
"tensorboard",
"onnx",
"Huggy",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-Huggy",
"region:us"
] |
reinforcement-learning
| 2023-08-29T22:30:05Z |
---
library_name: ml-agents
tags:
- Huggy
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-Huggy
---
# **ppo** Agent playing **Huggy**
This is a trained model of a **ppo** agent playing **Huggy**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: aldobremen/ppo-Huggy
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
ThuyNT03/xlm-roberta-base-New_VietNam-aug_replace_BERT-1
|
ThuyNT03
| 2023-08-29T22:27:56Z | 103 | 0 |
transformers
|
[
"transformers",
"pytorch",
"xlm-roberta",
"text-classification",
"generated_from_trainer",
"base_model:FacebookAI/xlm-roberta-base",
"base_model:finetune:FacebookAI/xlm-roberta-base",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-08-29T22:23:13Z |
---
license: mit
base_model: xlm-roberta-base
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
model-index:
- name: xlm-roberta-base-New_VietNam-aug_replace_BERT-1
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# xlm-roberta-base-New_VietNam-aug_replace_BERT-1
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.8418
- Accuracy: 0.7
- F1: 0.7047
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 8
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 0.6469 | 1.0 | 87 | 1.0507 | 0.71 | 0.7162 |
| 0.422 | 2.0 | 174 | 0.9811 | 0.73 | 0.7318 |
| 0.3235 | 3.0 | 261 | 0.9616 | 0.72 | 0.7224 |
| 0.2039 | 4.0 | 348 | 1.2767 | 0.68 | 0.6865 |
| 0.1606 | 5.0 | 435 | 1.5652 | 0.71 | 0.7122 |
| 0.1419 | 6.0 | 522 | 1.7107 | 0.71 | 0.7145 |
| 0.1158 | 7.0 | 609 | 1.7809 | 0.7 | 0.7047 |
| 0.1024 | 8.0 | 696 | 1.8418 | 0.7 | 0.7047 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.0.0
- Datasets 2.14.4
- Tokenizers 0.13.3
|
dt-and-vanilla-ardt/dt-combo_train_walker2d_v2-2908_2140-66
|
dt-and-vanilla-ardt
| 2023-08-29T22:27:20Z | 31 | 0 |
transformers
|
[
"transformers",
"pytorch",
"decision_transformer",
"generated_from_trainer",
"endpoints_compatible",
"region:us"
] | null | 2023-08-29T20:41:56Z |
---
tags:
- generated_from_trainer
model-index:
- name: dt-combo_train_walker2d_v2-2908_2140-66
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# dt-combo_train_walker2d_v2-2908_2140-66
This model is a fine-tuned version of [](https://huggingface.co/) on the None dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 64
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- training_steps: 10000
### Training results
### Framework versions
- Transformers 4.29.2
- Pytorch 2.1.0.dev20230727+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3
|
ThuyNT03/xlm-roberta-base-New_VietNam-aug_replace_w2v-1
|
ThuyNT03
| 2023-08-29T22:18:34Z | 103 | 0 |
transformers
|
[
"transformers",
"pytorch",
"xlm-roberta",
"text-classification",
"generated_from_trainer",
"base_model:FacebookAI/xlm-roberta-base",
"base_model:finetune:FacebookAI/xlm-roberta-base",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-08-29T21:53:52Z |
---
license: mit
base_model: xlm-roberta-base
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
model-index:
- name: xlm-roberta-base-New_VietNam-aug_replace_w2v-1
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# xlm-roberta-base-New_VietNam-aug_replace_w2v-1
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.0760
- Accuracy: 0.73
- F1: 0.7316
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 8
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 1.0605 | 1.0 | 84 | 0.9570 | 0.55 | 0.4830 |
| 0.7614 | 2.0 | 168 | 0.7075 | 0.68 | 0.6682 |
| 0.5718 | 3.0 | 252 | 0.7353 | 0.74 | 0.7413 |
| 0.4409 | 4.0 | 336 | 0.7226 | 0.73 | 0.7346 |
| 0.3768 | 5.0 | 420 | 0.7989 | 0.76 | 0.7612 |
| 0.2887 | 6.0 | 504 | 0.8930 | 0.74 | 0.7424 |
| 0.2218 | 7.0 | 588 | 1.0373 | 0.74 | 0.7429 |
| 0.1798 | 8.0 | 672 | 1.0760 | 0.73 | 0.7316 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.0.0
- Datasets 2.14.4
- Tokenizers 0.13.3
|
touchtech/fashion-images-pack-types-vit-huge-patch14-224-in21k
|
touchtech
| 2023-08-29T22:14:21Z | 8 | 0 |
transformers
|
[
"transformers",
"pytorch",
"vit",
"image-classification",
"vision",
"generated_from_trainer",
"dataset:imagefolder",
"base_model:google/vit-huge-patch14-224-in21k",
"base_model:finetune:google/vit-huge-patch14-224-in21k",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2023-08-29T19:35:55Z |
---
license: apache-2.0
base_model: google/vit-huge-patch14-224-in21k
tags:
- image-classification
- vision
- generated_from_trainer
datasets:
- imagefolder
metrics:
- accuracy
model-index:
- name: fashion-images-pack-types-vit-huge-patch14-224-in21k
results:
- task:
name: Image Classification
type: image-classification
dataset:
name: touchtech/fashion-images-pack-types
type: imagefolder
config: default
split: train
args: default
metrics:
- name: Accuracy
type: accuracy
value: 0.989010989010989
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# fashion-images-pack-types-vit-huge-patch14-224-in21k
This model is a fine-tuned version of [google/vit-huge-patch14-224-in21k](https://huggingface.co/google/vit-huge-patch14-224-in21k) on the touchtech/fashion-images-pack-types dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0436
- Accuracy: 0.9890
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 1337
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5.0
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 0.2292 | 1.0 | 1676 | 0.1293 | 0.9755 |
| 0.1376 | 2.0 | 3352 | 0.0769 | 0.9827 |
| 0.1122 | 3.0 | 5028 | 0.0565 | 0.9852 |
| 0.0759 | 4.0 | 6704 | 0.0501 | 0.9873 |
| 0.0678 | 5.0 | 8380 | 0.0436 | 0.9890 |
### Framework versions
- Transformers 4.33.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3
|
mgmeskill/new-Soccer
|
mgmeskill
| 2023-08-29T22:14:18Z | 4 | 0 |
ml-agents
|
[
"ml-agents",
"tensorboard",
"onnx",
"SoccerTwos",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-SoccerTwos",
"region:us"
] |
reinforcement-learning
| 2023-08-29T22:12:52Z |
---
library_name: ml-agents
tags:
- SoccerTwos
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-SoccerTwos
---
# **poca** Agent playing **SoccerTwos**
This is a trained model of a **poca** agent playing **SoccerTwos**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: mgmeskill/new-Soccer
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
ThuyNT03/xlm-roberta-base-New_VietNam-aug_replace_synonym-1
|
ThuyNT03
| 2023-08-29T21:53:51Z | 103 | 0 |
transformers
|
[
"transformers",
"pytorch",
"xlm-roberta",
"text-classification",
"generated_from_trainer",
"base_model:FacebookAI/xlm-roberta-base",
"base_model:finetune:FacebookAI/xlm-roberta-base",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-08-29T21:47:59Z |
---
license: mit
base_model: xlm-roberta-base
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
model-index:
- name: xlm-roberta-base-New_VietNam-aug_replace_synonym-1
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# xlm-roberta-base-New_VietNam-aug_replace_synonym-1
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.3309
- Accuracy: 0.63
- F1: 0.6393
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 8
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 0.3738 | 1.0 | 86 | 1.0460 | 0.69 | 0.6960 |
| 0.2198 | 2.0 | 172 | 1.6273 | 0.67 | 0.6759 |
| 0.1523 | 3.0 | 258 | 2.0692 | 0.66 | 0.6651 |
| 0.1015 | 4.0 | 344 | 2.0397 | 0.68 | 0.6891 |
| 0.073 | 5.0 | 430 | 2.0420 | 0.69 | 0.7014 |
| 0.0509 | 6.0 | 516 | 2.3116 | 0.68 | 0.6907 |
| 0.0549 | 7.0 | 602 | 2.3421 | 0.65 | 0.6560 |
| 0.0549 | 8.0 | 688 | 2.3309 | 0.63 | 0.6393 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.0.0
- Datasets 2.14.4
- Tokenizers 0.13.3
|
ThuyNT03/xlm-roberta-base-New_VietNam-aug_insert_tfidf-1
|
ThuyNT03
| 2023-08-29T21:42:58Z | 103 | 0 |
transformers
|
[
"transformers",
"pytorch",
"xlm-roberta",
"text-classification",
"generated_from_trainer",
"base_model:FacebookAI/xlm-roberta-base",
"base_model:finetune:FacebookAI/xlm-roberta-base",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-08-29T21:37:55Z |
---
license: mit
base_model: xlm-roberta-base
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
model-index:
- name: xlm-roberta-base-New_VietNam-aug_insert_tfidf-1
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# xlm-roberta-base-New_VietNam-aug_insert_tfidf-1
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.2725
- Accuracy: 0.65
- F1: 0.6607
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 8
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 0.4357 | 1.0 | 87 | 1.0211 | 0.7 | 0.7077 |
| 0.2982 | 2.0 | 174 | 1.0731 | 0.66 | 0.6688 |
| 0.2301 | 3.0 | 261 | 1.3254 | 0.69 | 0.6975 |
| 0.21 | 4.0 | 348 | 1.7521 | 0.67 | 0.6774 |
| 0.1363 | 5.0 | 435 | 2.0581 | 0.67 | 0.6781 |
| 0.1229 | 6.0 | 522 | 2.1290 | 0.67 | 0.6769 |
| 0.0911 | 7.0 | 609 | 2.2635 | 0.64 | 0.6505 |
| 0.0834 | 8.0 | 696 | 2.2725 | 0.65 | 0.6607 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.0.0
- Datasets 2.14.4
- Tokenizers 0.13.3
|
pkulium/opt-1.3b-lora-nm
|
pkulium
| 2023-08-29T21:37:56Z | 2 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-08-29T17:17:47Z |
---
library_name: peft
---
## Training procedure
### Framework versions
- PEFT 0.6.0.dev0
|
Infernaught/codealpaca51295
|
Infernaught
| 2023-08-29T21:24:15Z | 4 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-08-29T21:22:56Z |
---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: float16
### Framework versions
- PEFT 0.4.0
|
Infernaught/codealpaca_jz
|
Infernaught
| 2023-08-29T20:59:30Z | 0 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-08-29T20:59:03Z |
---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: float16
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: float16
### Framework versions
- PEFT 0.4.0
- PEFT 0.4.0
|
Aladin77/Taxi-v3-tests
|
Aladin77
| 2023-08-29T20:57:08Z | 0 | 0 | null |
[
"Taxi-v3",
"q-learning",
"reinforcement-learning",
"custom-implementation",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-08-29T20:57:07Z |
---
tags:
- Taxi-v3
- q-learning
- reinforcement-learning
- custom-implementation
model-index:
- name: Taxi-v3-tests
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: Taxi-v3
type: Taxi-v3
metrics:
- type: mean_reward
value: 7.56 +/- 2.71
name: mean_reward
verified: false
---
# **Q-Learning** Agent playing1 **Taxi-v3**
This is a trained model of a **Q-Learning** agent playing **Taxi-v3** .
## Usage
```python
model = load_from_hub(repo_id="Aladin77/Taxi-v3-tests", filename="q-learning.pkl")
# Don't forget to check if you need to add additional attributes (is_slippery=False etc)
env = gym.make(model["env_id"])
```
|
TheLastBen/AI_Hands_XL
|
TheLastBen
| 2023-08-29T20:50:09Z | 14 | 1 |
diffusers
|
[
"diffusers",
"text-to-image",
"stable-diffusion",
"lora",
"base_model:stabilityai/stable-diffusion-xl-base-1.0",
"base_model:adapter:stabilityai/stable-diffusion-xl-base-1.0",
"license:creativeml-openrail-m",
"region:us"
] |
text-to-image
| 2023-08-29T19:41:24Z |
---
license: creativeml-openrail-m
tags:
- text-to-image
- stable-diffusion
- lora
- diffusers
base_model: stabilityai/stable-diffusion-xl-base-1.0
instance_prompt: showing hands
widget:
- text: showing hands
---
### AI Hands
#### SDXL LoRA by TheLastBen
#### Prompts to start with :
--Name-- showing hands
---
Trained using https://github.com/TheLastBen/fast-stable-diffusion SDXL trainer.
ComfyUI seems to give better results than A1111, but that's just me.
#### Sample pictures:
.png)
.png)
.jpg)
.png)
.png)
.png)
.png)
.png)
|
ThuyNT03/xlm-roberta-base-New_VietNam-aug_swap-1
|
ThuyNT03
| 2023-08-29T20:40:08Z | 103 | 0 |
transformers
|
[
"transformers",
"pytorch",
"xlm-roberta",
"text-classification",
"generated_from_trainer",
"base_model:FacebookAI/xlm-roberta-base",
"base_model:finetune:FacebookAI/xlm-roberta-base",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-08-29T20:01:08Z |
---
license: mit
base_model: xlm-roberta-base
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
model-index:
- name: xlm-roberta-base-New_VietNam-aug_swap-1
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# xlm-roberta-base-New_VietNam-aug_swap-1
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.0905
- Accuracy: 0.71
- F1: 0.7188
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 8
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 1.0296 | 1.0 | 86 | 0.8380 | 0.63 | 0.5245 |
| 0.7549 | 2.0 | 172 | 0.7621 | 0.67 | 0.6630 |
| 0.5459 | 3.0 | 258 | 0.8527 | 0.67 | 0.6701 |
| 0.4274 | 4.0 | 344 | 0.8442 | 0.71 | 0.7131 |
| 0.2929 | 5.0 | 430 | 0.9154 | 0.69 | 0.6995 |
| 0.2086 | 6.0 | 516 | 1.0227 | 0.71 | 0.7191 |
| 0.1671 | 7.0 | 602 | 1.0829 | 0.71 | 0.7191 |
| 0.1421 | 8.0 | 688 | 1.0905 | 0.71 | 0.7188 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.0.0
- Datasets 2.14.4
- Tokenizers 0.13.3
|
SHENMU007/neunit_BASE_V13.5.5
|
SHENMU007
| 2023-08-29T20:38:22Z | 78 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"speecht5",
"text-to-audio",
"1.1.0",
"generated_from_trainer",
"zh",
"dataset:facebook/voxpopuli",
"base_model:microsoft/speecht5_tts",
"base_model:finetune:microsoft/speecht5_tts",
"license:mit",
"endpoints_compatible",
"region:us"
] |
text-to-audio
| 2023-08-29T18:03:11Z |
---
language:
- zh
license: mit
base_model: microsoft/speecht5_tts
tags:
- 1.1.0
- generated_from_trainer
datasets:
- facebook/voxpopuli
model-index:
- name: SpeechT5 TTS Dutch neunit
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# SpeechT5 TTS Dutch neunit
This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the VoxPopuli dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 4000
### Training results
### Framework versions
- Transformers 4.31.0.dev0
- Pytorch 2.0.1+cu117
- Datasets 2.12.0
- Tokenizers 0.13.3
|
ThuyNT03/xlm-roberta-base-New_VietNam-train-1
|
ThuyNT03
| 2023-08-29T20:35:06Z | 103 | 0 |
transformers
|
[
"transformers",
"pytorch",
"xlm-roberta",
"text-classification",
"generated_from_trainer",
"base_model:FacebookAI/xlm-roberta-base",
"base_model:finetune:FacebookAI/xlm-roberta-base",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-08-29T19:56:17Z |
---
license: mit
base_model: xlm-roberta-base
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
model-index:
- name: xlm-roberta-base-New_VietNam-train-1
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# xlm-roberta-base-New_VietNam-train-1
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.9382
- Accuracy: 0.71
- F1: 0.7149
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 8
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 1.0757 | 1.0 | 44 | 0.9819 | 0.54 | 0.4764 |
| 0.9029 | 2.0 | 88 | 0.7825 | 0.66 | 0.6195 |
| 0.7055 | 3.0 | 132 | 0.7774 | 0.67 | 0.6585 |
| 0.5871 | 4.0 | 176 | 0.8691 | 0.68 | 0.6717 |
| 0.4873 | 5.0 | 220 | 0.7771 | 0.73 | 0.7331 |
| 0.3833 | 6.0 | 264 | 0.8852 | 0.67 | 0.6744 |
| 0.3449 | 7.0 | 308 | 0.9458 | 0.66 | 0.6694 |
| 0.3099 | 8.0 | 352 | 0.9382 | 0.71 | 0.7149 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.0.0
- Datasets 2.14.4
- Tokenizers 0.13.3
|
jlpan/starcoder-php2py-snippet1
|
jlpan
| 2023-08-29T20:34:53Z | 5 | 0 |
peft
|
[
"peft",
"generated_from_trainer",
"base_model:bigcode/starcoder",
"base_model:adapter:bigcode/starcoder",
"license:bigcode-openrail-m",
"region:us"
] | null | 2023-08-29T19:05:40Z |
---
license: bigcode-openrail-m
base_model: bigcode/starcoder
tags:
- generated_from_trainer
model-index:
- name: starcoder-php2py-snippet1
results: []
library_name: peft
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# starcoder-php2py-snippet1
This model is a fine-tuned version of [bigcode/starcoder](https://huggingface.co/bigcode/starcoder) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2024
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 9e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 5
- total_train_batch_size: 80
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 10
- training_steps: 100
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 3.6678 | 0.25 | 25 | 0.4089 |
| 0.277 | 0.5 | 50 | 0.2208 |
| 0.2063 | 0.75 | 75 | 0.2040 |
| 0.2039 | 1.0 | 100 | 0.2024 |
### Framework versions
- PEFT 0.5.0.dev0
- PEFT 0.5.0.dev0
- Transformers 4.32.0.dev0
- Pytorch 2.0.1+cu117
- Datasets 2.12.0
- Tokenizers 0.13.3
|
dvaughan79/q-FrozenLake-v1-4x4-noSlippery
|
dvaughan79
| 2023-08-29T20:31:36Z | 0 | 0 | null |
[
"FrozenLake-v1-4x4-no_slippery",
"q-learning",
"reinforcement-learning",
"custom-implementation",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-08-29T20:31:34Z |
---
tags:
- FrozenLake-v1-4x4-no_slippery
- q-learning
- reinforcement-learning
- custom-implementation
model-index:
- name: q-FrozenLake-v1-4x4-noSlippery
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: FrozenLake-v1-4x4-no_slippery
type: FrozenLake-v1-4x4-no_slippery
metrics:
- type: mean_reward
value: 1.00 +/- 0.00
name: mean_reward
verified: false
---
# **Q-Learning** Agent playing1 **FrozenLake-v1**
This is a trained model of a **Q-Learning** agent playing **FrozenLake-v1** .
## Usage
```python
model = load_from_hub(repo_id="dvaughan79/q-FrozenLake-v1-4x4-noSlippery", filename="q-learning.pkl")
# Don't forget to check if you need to add additional attributes (is_slippery=False etc)
env = gym.make(model["env_id"])
```
|
StefanoCaloni/ppo-Huggy
|
StefanoCaloni
| 2023-08-29T20:23:36Z | 0 | 0 |
ml-agents
|
[
"ml-agents",
"tensorboard",
"onnx",
"Huggy",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-Huggy",
"region:us"
] |
reinforcement-learning
| 2023-08-29T20:23:34Z |
---
library_name: ml-agents
tags:
- Huggy
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-Huggy
---
# **ppo** Agent playing **Huggy**
This is a trained model of a **ppo** agent playing **Huggy**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: StefanoCaloni/ppo-Huggy
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
hakanssonjesper/gpt2-test
|
hakanssonjesper
| 2023-08-29T19:56:19Z | 4 | 0 |
transformers
|
[
"transformers",
"llm-rs",
"ggml",
"text-generation",
"arxiv:1910.09700",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-08-29T17:25:33Z |
---
tags:
- llm-rs
- ggml
pipeline_tag: text-generation
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
- **Developed by:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Data Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Data Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed]
|
swer/sdxl-holding-deserteagle
|
swer
| 2023-08-29T19:54:20Z | 2 | 1 |
diffusers
|
[
"diffusers",
"text-to-image",
"autotrain",
"base_model:stabilityai/stable-diffusion-xl-base-1.0",
"base_model:finetune:stabilityai/stable-diffusion-xl-base-1.0",
"region:us"
] |
text-to-image
| 2023-08-29T19:54:14Z |
---
base_model: stabilityai/stable-diffusion-xl-base-1.0
instance_prompt: holding desert eagle
tags:
- text-to-image
- diffusers
- autotrain
inference: true
---
# DreamBooth trained by AutoTrain
Text encoder was not trained.
|
vwxyzjn/starcoderbase-triviaqa
|
vwxyzjn
| 2023-08-29T19:52:36Z | 17 | 0 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"gpt_bigcode",
"text-generation",
"trl",
"rlhf",
"en",
"dataset:trivia_qa",
"license:bigscience-openrail-m",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-08-17T16:45:54Z |
---
license: bigscience-openrail-m
datasets:
- trivia_qa
language:
- en
tags:
- trl
- transformers
- rlhf
---
# starcoderbase-triviaqa
This model is baesed on https://huggingface.co/bigcode/starcoderbase and is fine-tuned on the TriviaQA dataset using reinforcement learning via TRL's `TextEnvironment` (https://github.com/huggingface/trl/pull/424).
### Out of Scope Use
- Replacing human expertise
## Bias, Risks, and Limitations
- May generate answers that are incorrect or misleading.
- May copy answers from the training data verbatim.
- May generate language that is hateful or promotes discrimination ([example](https://huggingface.co/trl-lib/llama-7b-se-rl-peft/discussions/7#64376083369f6f907f5bfe4c)).
- May generate language that is offensive to direct or indirect users or to people or groups mentioned.
### Recommendations
- Answers should be validated through the use of external sources.
- Disparities between the data contributors and the direct and indirect users of the technology should inform developers in assessing what constitutes an appropriate use case.
- Further research is needed to attribute model generations to sources in the training data, especially in cases where the model copies answers from the training data.
|
rishitunu/ecc_segformer_main
|
rishitunu
| 2023-08-29T19:50:01Z | 16 | 0 |
transformers
|
[
"transformers",
"pytorch",
"segformer",
"image-segmentation",
"vision",
"generated_from_trainer",
"base_model:nvidia/mit-b5",
"base_model:finetune:nvidia/mit-b5",
"license:other",
"endpoints_compatible",
"region:us"
] |
image-segmentation
| 2023-08-28T14:08:33Z |
---
license: other
base_model: nvidia/mit-b5
tags:
- image-segmentation
- vision
- generated_from_trainer
model-index:
- name: ecc_segformer_main
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# ecc_segformer_main
This model is a fine-tuned version of [nvidia/mit-b5](https://huggingface.co/nvidia/mit-b5) on the rishitunu/ecc_crackdetector_dataset_main dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1918
- Mean Iou: 0.2329
- Mean Accuracy: 0.4658
- Overall Accuracy: 0.4658
- Accuracy Background: nan
- Accuracy Crack: 0.4658
- Iou Background: 0.0
- Iou Crack: 0.4658
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 6e-05
- train_batch_size: 2
- eval_batch_size: 2
- seed: 1337
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: polynomial
- training_steps: 10000
### Training results
| Training Loss | Epoch | Step | Validation Loss | Mean Iou | Mean Accuracy | Overall Accuracy | Accuracy Background | Accuracy Crack | Iou Background | Iou Crack |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:-------------:|:----------------:|:-------------------:|:--------------:|:--------------:|:---------:|
| 0.1069 | 1.0 | 172 | 0.1376 | 0.1660 | 0.3320 | 0.3320 | nan | 0.3320 | 0.0 | 0.3320 |
| 0.0682 | 2.0 | 344 | 0.1327 | 0.2298 | 0.4596 | 0.4596 | nan | 0.4596 | 0.0 | 0.4596 |
| 0.0666 | 3.0 | 516 | 0.2478 | 0.1200 | 0.2401 | 0.2401 | nan | 0.2401 | 0.0 | 0.2401 |
| 0.0639 | 4.0 | 688 | 0.1732 | 0.1538 | 0.3076 | 0.3076 | nan | 0.3076 | 0.0 | 0.3076 |
| 0.0624 | 5.0 | 860 | 0.1027 | 0.2334 | 0.4668 | 0.4668 | nan | 0.4668 | 0.0 | 0.4668 |
| 0.0557 | 6.0 | 1032 | 0.1003 | 0.1851 | 0.3703 | 0.3703 | nan | 0.3703 | 0.0 | 0.3703 |
| 0.0563 | 7.0 | 1204 | 0.1512 | 0.2007 | 0.4014 | 0.4014 | nan | 0.4014 | 0.0 | 0.4014 |
| 0.054 | 8.0 | 1376 | 0.1000 | 0.2401 | 0.4802 | 0.4802 | nan | 0.4802 | 0.0 | 0.4802 |
| 0.0546 | 9.0 | 1548 | 0.0933 | 0.2238 | 0.4475 | 0.4475 | nan | 0.4475 | 0.0 | 0.4475 |
| 0.0498 | 10.0 | 1720 | 0.0964 | 0.2303 | 0.4606 | 0.4606 | nan | 0.4606 | 0.0 | 0.4606 |
| 0.0515 | 11.0 | 1892 | 0.1107 | 0.2258 | 0.4516 | 0.4516 | nan | 0.4516 | 0.0 | 0.4516 |
| 0.0453 | 12.0 | 2064 | 0.0961 | 0.2557 | 0.5115 | 0.5115 | nan | 0.5115 | 0.0 | 0.5115 |
| 0.0431 | 13.0 | 2236 | 0.1027 | 0.2396 | 0.4792 | 0.4792 | nan | 0.4792 | 0.0 | 0.4792 |
| 0.0418 | 14.0 | 2408 | 0.1027 | 0.2521 | 0.5042 | 0.5042 | nan | 0.5042 | 0.0 | 0.5042 |
| 0.0426 | 15.0 | 2580 | 0.1059 | 0.2561 | 0.5123 | 0.5123 | nan | 0.5123 | 0.0 | 0.5123 |
| 0.0377 | 16.0 | 2752 | 0.1193 | 0.2281 | 0.4561 | 0.4561 | nan | 0.4561 | 0.0 | 0.4561 |
| 0.0369 | 17.0 | 2924 | 0.1161 | 0.2486 | 0.4972 | 0.4972 | nan | 0.4972 | 0.0 | 0.4972 |
| 0.036 | 18.0 | 3096 | 0.1058 | 0.2515 | 0.5029 | 0.5029 | nan | 0.5029 | 0.0 | 0.5029 |
| 0.034 | 19.0 | 3268 | 0.1176 | 0.2434 | 0.4868 | 0.4868 | nan | 0.4868 | 0.0 | 0.4868 |
| 0.0337 | 20.0 | 3440 | 0.1162 | 0.2254 | 0.4509 | 0.4509 | nan | 0.4509 | 0.0 | 0.4509 |
| 0.0281 | 21.0 | 3612 | 0.1203 | 0.2213 | 0.4426 | 0.4426 | nan | 0.4426 | 0.0 | 0.4426 |
| 0.0354 | 22.0 | 3784 | 0.1266 | 0.2384 | 0.4768 | 0.4768 | nan | 0.4768 | 0.0 | 0.4768 |
| 0.0323 | 23.0 | 3956 | 0.1223 | 0.2409 | 0.4818 | 0.4818 | nan | 0.4818 | 0.0 | 0.4818 |
| 0.0299 | 24.0 | 4128 | 0.1356 | 0.2195 | 0.4390 | 0.4390 | nan | 0.4390 | 0.0 | 0.4390 |
| 0.0294 | 25.0 | 4300 | 0.1285 | 0.2318 | 0.4636 | 0.4636 | nan | 0.4636 | 0.0 | 0.4636 |
| 0.0295 | 26.0 | 4472 | 0.1274 | 0.2559 | 0.5119 | 0.5119 | nan | 0.5119 | 0.0 | 0.5119 |
| 0.0252 | 27.0 | 4644 | 0.1387 | 0.2413 | 0.4827 | 0.4827 | nan | 0.4827 | 0.0 | 0.4827 |
| 0.029 | 28.0 | 4816 | 0.1468 | 0.2236 | 0.4472 | 0.4472 | nan | 0.4472 | 0.0 | 0.4472 |
| 0.0218 | 29.0 | 4988 | 0.1448 | 0.2433 | 0.4866 | 0.4866 | nan | 0.4866 | 0.0 | 0.4866 |
| 0.0275 | 30.0 | 5160 | 0.1478 | 0.2318 | 0.4635 | 0.4635 | nan | 0.4635 | 0.0 | 0.4635 |
| 0.0233 | 31.0 | 5332 | 0.1377 | 0.2502 | 0.5005 | 0.5005 | nan | 0.5005 | 0.0 | 0.5005 |
| 0.0252 | 32.0 | 5504 | 0.1458 | 0.2399 | 0.4797 | 0.4797 | nan | 0.4797 | 0.0 | 0.4797 |
| 0.0245 | 33.0 | 5676 | 0.1431 | 0.2480 | 0.4960 | 0.4960 | nan | 0.4960 | 0.0 | 0.4960 |
| 0.0225 | 34.0 | 5848 | 0.1562 | 0.2439 | 0.4879 | 0.4879 | nan | 0.4879 | 0.0 | 0.4879 |
| 0.0242 | 35.0 | 6020 | 0.1633 | 0.2323 | 0.4646 | 0.4646 | nan | 0.4646 | 0.0 | 0.4646 |
| 0.0213 | 36.0 | 6192 | 0.1666 | 0.2274 | 0.4549 | 0.4549 | nan | 0.4549 | 0.0 | 0.4549 |
| 0.0256 | 37.0 | 6364 | 0.1665 | 0.2340 | 0.4680 | 0.4680 | nan | 0.4680 | 0.0 | 0.4680 |
| 0.0237 | 38.0 | 6536 | 0.1658 | 0.2410 | 0.4819 | 0.4819 | nan | 0.4819 | 0.0 | 0.4819 |
| 0.0192 | 39.0 | 6708 | 0.1705 | 0.2286 | 0.4572 | 0.4572 | nan | 0.4572 | 0.0 | 0.4572 |
| 0.0198 | 40.0 | 6880 | 0.1688 | 0.2322 | 0.4644 | 0.4644 | nan | 0.4644 | 0.0 | 0.4644 |
| 0.0214 | 41.0 | 7052 | 0.1717 | 0.2315 | 0.4630 | 0.4630 | nan | 0.4630 | 0.0 | 0.4630 |
| 0.0197 | 42.0 | 7224 | 0.1764 | 0.2338 | 0.4677 | 0.4677 | nan | 0.4677 | 0.0 | 0.4677 |
| 0.0187 | 43.0 | 7396 | 0.1764 | 0.2437 | 0.4874 | 0.4874 | nan | 0.4874 | 0.0 | 0.4874 |
| 0.0212 | 44.0 | 7568 | 0.1874 | 0.2259 | 0.4519 | 0.4519 | nan | 0.4519 | 0.0 | 0.4519 |
| 0.0188 | 45.0 | 7740 | 0.1854 | 0.2362 | 0.4725 | 0.4725 | nan | 0.4725 | 0.0 | 0.4725 |
| 0.0188 | 46.0 | 7912 | 0.1772 | 0.2320 | 0.4641 | 0.4641 | nan | 0.4641 | 0.0 | 0.4641 |
| 0.0228 | 47.0 | 8084 | 0.1783 | 0.2385 | 0.4770 | 0.4770 | nan | 0.4770 | 0.0 | 0.4770 |
| 0.0199 | 48.0 | 8256 | 0.1850 | 0.2317 | 0.4634 | 0.4634 | nan | 0.4634 | 0.0 | 0.4634 |
| 0.0202 | 49.0 | 8428 | 0.1872 | 0.2336 | 0.4672 | 0.4672 | nan | 0.4672 | 0.0 | 0.4672 |
| 0.0181 | 50.0 | 8600 | 0.1803 | 0.2405 | 0.4810 | 0.4810 | nan | 0.4810 | 0.0 | 0.4810 |
| 0.0157 | 51.0 | 8772 | 0.1874 | 0.2349 | 0.4697 | 0.4697 | nan | 0.4697 | 0.0 | 0.4697 |
| 0.0162 | 52.0 | 8944 | 0.1889 | 0.2332 | 0.4665 | 0.4665 | nan | 0.4665 | 0.0 | 0.4665 |
| 0.0178 | 53.0 | 9116 | 0.1948 | 0.2357 | 0.4715 | 0.4715 | nan | 0.4715 | 0.0 | 0.4715 |
| 0.0166 | 54.0 | 9288 | 0.1911 | 0.2333 | 0.4666 | 0.4666 | nan | 0.4666 | 0.0 | 0.4666 |
| 0.0193 | 55.0 | 9460 | 0.1959 | 0.2306 | 0.4611 | 0.4611 | nan | 0.4611 | 0.0 | 0.4611 |
| 0.0199 | 56.0 | 9632 | 0.1999 | 0.2330 | 0.4659 | 0.4659 | nan | 0.4659 | 0.0 | 0.4659 |
| 0.0177 | 57.0 | 9804 | 0.1943 | 0.2319 | 0.4639 | 0.4639 | nan | 0.4639 | 0.0 | 0.4639 |
| 0.019 | 58.0 | 9976 | 0.1926 | 0.2327 | 0.4653 | 0.4653 | nan | 0.4653 | 0.0 | 0.4653 |
| 0.0187 | 58.14 | 10000 | 0.1918 | 0.2329 | 0.4658 | 0.4658 | nan | 0.4658 | 0.0 | 0.4658 |
### Framework versions
- Transformers 4.32.0.dev0
- Pytorch 2.0.1+cpu
- Datasets 2.14.4
- Tokenizers 0.13.3
|
rnkVikcdkam/ppo-Huggy
|
rnkVikcdkam
| 2023-08-29T19:43:27Z | 13 | 0 |
ml-agents
|
[
"ml-agents",
"tensorboard",
"onnx",
"Huggy",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-Huggy",
"region:us"
] |
reinforcement-learning
| 2023-08-29T19:43:22Z |
---
library_name: ml-agents
tags:
- Huggy
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-Huggy
---
# **ppo** Agent playing **Huggy**
This is a trained model of a **ppo** agent playing **Huggy**
using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
- A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your
browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction
- A *longer tutorial* to understand how works ML-Agents:
https://huggingface.co/learn/deep-rl-course/unit5/introduction
### Resume the training
```bash
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser**
1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity
2. Step 1: Find your model_id: rnkVikcdkam/ppo-Huggy
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
KennethTM/gpt2-medium-danish-review-response
|
KennethTM
| 2023-08-29T19:34:51Z | 153 | 0 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"gpt2",
"text-generation",
"da",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-08-29T19:14:18Z |
---
language:
- da
pipeline_tag: text-generation
widget:
- text: "### Bruger:\nAnders\n\n### Anmeldelse:\nUmuligt at komme igennem på telefonen.\n\n### Svar:\nKære Anders\n"
---
# What is this?
A fine-tuned GPT-2 model (medium version, ~354.8 M parameters) for generating responses to customer reviews in Danish.
# How to use
The model is based on the [gpt2-medium-danish model](https://huggingface.co/KennethTM/gpt2-medium-danish) and performs better than the smaller version ([gpt2-small-danish-review-response](https://huggingface.co/KennethTM/gpt2-small-danish-review-response)). Supervised fine-tuning is applied to adapt the model to generate responses to customer reviews in Danish. A prompting template is applied to the examples used for training (see the example below).
Test the model using the pipeline from the [🤗 Transformers](https://github.com/huggingface/transformers) library:
```python
from transformers import pipeline
generator = pipeline("text-generation", model = "KennethTM/gpt2-medium-danish-review-response")
def prompt_template(user, review):
return f"### Bruger:\n{user}\n\n### Anmeldelse:\n{review}\n\n### Svar:\nKære {user}\n"
prompt = prompt_template(user = "Anders", review = "Umuligt at komme igennem på telefonen.")
text = generator(prompt)
print(text[0]["generated_text"])
```
Or load it using the Auto* classes:
```python
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("KennethTM/gpt2-medium-danish-review-response")
model = AutoModelForCausalLM.from_pretrained("KennethTM/gpt2-medium-danish-review-response")
```
|
ThuyNT03/xlm-roberta-base-New_VietNam-aug_delete
|
ThuyNT03
| 2023-08-29T19:34:25Z | 103 | 0 |
transformers
|
[
"transformers",
"pytorch",
"xlm-roberta",
"text-classification",
"generated_from_trainer",
"base_model:FacebookAI/xlm-roberta-base",
"base_model:finetune:FacebookAI/xlm-roberta-base",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-08-29T18:53:46Z |
---
license: mit
base_model: xlm-roberta-base
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
model-index:
- name: xlm-roberta-base-New_VietNam-aug_delete
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# xlm-roberta-base-New_VietNam-aug_delete
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.8915
- Accuracy: 0.7
- F1: 0.7032
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 1.0805 | 1.0 | 44 | 0.9603 | 0.61 | 0.5214 |
| 0.8688 | 2.0 | 88 | 0.7871 | 0.68 | 0.6492 |
| 0.7285 | 3.0 | 132 | 0.7653 | 0.71 | 0.7134 |
| 0.6103 | 4.0 | 176 | 0.7112 | 0.72 | 0.7197 |
| 0.5336 | 5.0 | 220 | 0.7593 | 0.7 | 0.7057 |
| 0.4561 | 6.0 | 264 | 0.7167 | 0.72 | 0.7210 |
| 0.4174 | 7.0 | 308 | 0.7432 | 0.73 | 0.7332 |
| 0.3602 | 8.0 | 352 | 0.8390 | 0.7 | 0.7032 |
| 0.2965 | 9.0 | 396 | 0.8819 | 0.69 | 0.6978 |
| 0.2937 | 10.0 | 440 | 0.8915 | 0.7 | 0.7032 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.0.0
- Datasets 2.14.4
- Tokenizers 0.13.3
|
1warden2/BARTCnn_AWSBlogs
|
1warden2
| 2023-08-29T19:33:02Z | 120 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bart",
"text2text-generation",
"summarization",
"aws",
"blogposts",
"en",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
summarization
| 2023-08-29T19:15:50Z |
---
license: apache-2.0
language:
- en
pipeline_tag: summarization
tags:
- summarization
- aws
- blogposts
---
|
Muniks/ludwig-llama2
|
Muniks
| 2023-08-29T19:19:43Z | 1 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-08-29T19:19:24Z |
---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: float16
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: float16
### Framework versions
- PEFT 0.5.0
- PEFT 0.5.0
|
nikinetrahutama/afx-ai-llama-chat-model-18
|
nikinetrahutama
| 2023-08-29T19:16:49Z | 0 | 0 |
peft
|
[
"peft",
"pytorch",
"llama",
"region:us"
] | null | 2023-08-29T18:59:35Z |
---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: bfloat16
### Framework versions
- PEFT 0.6.0.dev0
|
dkqjrm/20230830015435
|
dkqjrm
| 2023-08-29T19:00:52Z | 107 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"text-classification",
"generated_from_trainer",
"dataset:super_glue",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-08-29T16:54:55Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- super_glue
metrics:
- accuracy
model-index:
- name: '20230830015435'
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# 20230830015435
This model is a fine-tuned version of [bert-large-cased](https://huggingface.co/bert-large-cased) on the super_glue dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3304
- Accuracy: 0.6538
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.009
- train_batch_size: 16
- eval_batch_size: 8
- seed: 44
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 80.0
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| No log | 1.0 | 35 | 0.5691 | 0.6058 |
| No log | 2.0 | 70 | 0.5906 | 0.5962 |
| No log | 3.0 | 105 | 0.5703 | 0.625 |
| No log | 4.0 | 140 | 0.4604 | 0.6154 |
| No log | 5.0 | 175 | 0.4834 | 0.5962 |
| No log | 6.0 | 210 | 0.4973 | 0.4135 |
| No log | 7.0 | 245 | 0.9046 | 0.3654 |
| No log | 8.0 | 280 | 0.3669 | 0.6346 |
| No log | 9.0 | 315 | 0.3828 | 0.4231 |
| No log | 10.0 | 350 | 0.4207 | 0.5769 |
| No log | 11.0 | 385 | 0.7596 | 0.3654 |
| No log | 12.0 | 420 | 0.9833 | 0.6346 |
| No log | 13.0 | 455 | 0.3754 | 0.6346 |
| No log | 14.0 | 490 | 0.4325 | 0.6346 |
| 0.6565 | 15.0 | 525 | 0.4163 | 0.6346 |
| 0.6565 | 16.0 | 560 | 0.7707 | 0.3654 |
| 0.6565 | 17.0 | 595 | 0.4262 | 0.6346 |
| 0.6565 | 18.0 | 630 | 0.3547 | 0.5 |
| 0.6565 | 19.0 | 665 | 0.3355 | 0.6346 |
| 0.6565 | 20.0 | 700 | 0.3787 | 0.4423 |
| 0.6565 | 21.0 | 735 | 0.3718 | 0.6346 |
| 0.6565 | 22.0 | 770 | 0.4742 | 0.3846 |
| 0.6565 | 23.0 | 805 | 0.3361 | 0.6827 |
| 0.6565 | 24.0 | 840 | 0.4078 | 0.6346 |
| 0.6565 | 25.0 | 875 | 0.3701 | 0.6346 |
| 0.6565 | 26.0 | 910 | 0.3726 | 0.6346 |
| 0.6565 | 27.0 | 945 | 0.7422 | 0.6346 |
| 0.6565 | 28.0 | 980 | 0.6071 | 0.6346 |
| 0.5512 | 29.0 | 1015 | 0.4255 | 0.4038 |
| 0.5512 | 30.0 | 1050 | 0.3393 | 0.6346 |
| 0.5512 | 31.0 | 1085 | 0.3556 | 0.6346 |
| 0.5512 | 32.0 | 1120 | 0.4493 | 0.3846 |
| 0.5512 | 33.0 | 1155 | 0.4296 | 0.4231 |
| 0.5512 | 34.0 | 1190 | 0.3491 | 0.625 |
| 0.5512 | 35.0 | 1225 | 0.3334 | 0.625 |
| 0.5512 | 36.0 | 1260 | 0.3552 | 0.6346 |
| 0.5512 | 37.0 | 1295 | 0.3297 | 0.6538 |
| 0.5512 | 38.0 | 1330 | 0.3939 | 0.4231 |
| 0.5512 | 39.0 | 1365 | 0.4693 | 0.3846 |
| 0.5512 | 40.0 | 1400 | 0.4533 | 0.3942 |
| 0.5512 | 41.0 | 1435 | 0.3350 | 0.6346 |
| 0.5512 | 42.0 | 1470 | 0.7957 | 0.3654 |
| 0.4482 | 43.0 | 1505 | 0.3327 | 0.6346 |
| 0.4482 | 44.0 | 1540 | 0.3572 | 0.6346 |
| 0.4482 | 45.0 | 1575 | 0.3303 | 0.6346 |
| 0.4482 | 46.0 | 1610 | 0.3398 | 0.6058 |
| 0.4482 | 47.0 | 1645 | 0.3778 | 0.4135 |
| 0.4482 | 48.0 | 1680 | 0.3528 | 0.5962 |
| 0.4482 | 49.0 | 1715 | 0.3447 | 0.6346 |
| 0.4482 | 50.0 | 1750 | 0.3419 | 0.5673 |
| 0.4482 | 51.0 | 1785 | 0.4567 | 0.3846 |
| 0.4482 | 52.0 | 1820 | 0.3662 | 0.4712 |
| 0.4482 | 53.0 | 1855 | 0.3298 | 0.6827 |
| 0.4482 | 54.0 | 1890 | 0.3821 | 0.4423 |
| 0.4482 | 55.0 | 1925 | 0.3533 | 0.6346 |
| 0.4482 | 56.0 | 1960 | 0.3340 | 0.6346 |
| 0.4482 | 57.0 | 1995 | 0.3546 | 0.5288 |
| 0.3923 | 58.0 | 2030 | 0.3321 | 0.6442 |
| 0.3923 | 59.0 | 2065 | 0.3440 | 0.5481 |
| 0.3923 | 60.0 | 2100 | 0.3326 | 0.6442 |
| 0.3923 | 61.0 | 2135 | 0.3378 | 0.6635 |
| 0.3923 | 62.0 | 2170 | 0.3346 | 0.625 |
| 0.3923 | 63.0 | 2205 | 0.3460 | 0.5 |
| 0.3923 | 64.0 | 2240 | 0.3517 | 0.6346 |
| 0.3923 | 65.0 | 2275 | 0.3301 | 0.6346 |
| 0.3923 | 66.0 | 2310 | 0.3306 | 0.6635 |
| 0.3923 | 67.0 | 2345 | 0.3362 | 0.6346 |
| 0.3923 | 68.0 | 2380 | 0.3382 | 0.5865 |
| 0.3923 | 69.0 | 2415 | 0.3314 | 0.6731 |
| 0.3923 | 70.0 | 2450 | 0.3283 | 0.6538 |
| 0.3923 | 71.0 | 2485 | 0.3304 | 0.6635 |
| 0.3611 | 72.0 | 2520 | 0.3445 | 0.5385 |
| 0.3611 | 73.0 | 2555 | 0.3291 | 0.6731 |
| 0.3611 | 74.0 | 2590 | 0.3414 | 0.5577 |
| 0.3611 | 75.0 | 2625 | 0.3334 | 0.6058 |
| 0.3611 | 76.0 | 2660 | 0.3304 | 0.6731 |
| 0.3611 | 77.0 | 2695 | 0.3415 | 0.5481 |
| 0.3611 | 78.0 | 2730 | 0.3314 | 0.6346 |
| 0.3611 | 79.0 | 2765 | 0.3315 | 0.6346 |
| 0.3611 | 80.0 | 2800 | 0.3304 | 0.6538 |
### Framework versions
- Transformers 4.26.1
- Pytorch 2.0.1+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3
|
dkqjrm/20230830020815
|
dkqjrm
| 2023-08-29T18:53:45Z | 105 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"text-classification",
"generated_from_trainer",
"dataset:super_glue",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-08-29T17:08:35Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- super_glue
metrics:
- accuracy
model-index:
- name: '20230830020815'
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# 20230830020815
This model is a fine-tuned version of [bert-large-cased](https://huggingface.co/bert-large-cased) on the super_glue dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6170
- Accuracy: 0.5577
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.009
- train_batch_size: 16
- eval_batch_size: 8
- seed: 44
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 80.0
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| No log | 1.0 | 35 | 0.8914 | 0.5865 |
| No log | 2.0 | 70 | 2.1528 | 0.4135 |
| No log | 3.0 | 105 | 1.5770 | 0.4038 |
| No log | 4.0 | 140 | 0.8308 | 0.625 |
| No log | 5.0 | 175 | 0.7444 | 0.4135 |
| No log | 6.0 | 210 | 1.6216 | 0.3654 |
| No log | 7.0 | 245 | 0.6072 | 0.6058 |
| No log | 8.0 | 280 | 0.9229 | 0.3654 |
| No log | 9.0 | 315 | 1.2563 | 0.6346 |
| No log | 10.0 | 350 | 1.9376 | 0.3654 |
| No log | 11.0 | 385 | 0.8642 | 0.4231 |
| No log | 12.0 | 420 | 0.8720 | 0.4135 |
| No log | 13.0 | 455 | 0.7846 | 0.625 |
| No log | 14.0 | 490 | 1.3697 | 0.6346 |
| 1.2144 | 15.0 | 525 | 1.2052 | 0.3654 |
| 1.2144 | 16.0 | 560 | 0.7440 | 0.5962 |
| 1.2144 | 17.0 | 595 | 0.8147 | 0.5288 |
| 1.2144 | 18.0 | 630 | 0.8679 | 0.6346 |
| 1.2144 | 19.0 | 665 | 0.8125 | 0.6346 |
| 1.2144 | 20.0 | 700 | 0.7545 | 0.5962 |
| 1.2144 | 21.0 | 735 | 0.5959 | 0.6346 |
| 1.2144 | 22.0 | 770 | 0.7990 | 0.3654 |
| 1.2144 | 23.0 | 805 | 0.8083 | 0.6346 |
| 1.2144 | 24.0 | 840 | 0.6932 | 0.6346 |
| 1.2144 | 25.0 | 875 | 0.6382 | 0.6346 |
| 1.2144 | 26.0 | 910 | 0.6023 | 0.625 |
| 1.2144 | 27.0 | 945 | 0.6502 | 0.625 |
| 1.2144 | 28.0 | 980 | 0.6646 | 0.4231 |
| 0.8752 | 29.0 | 1015 | 0.6646 | 0.5288 |
| 0.8752 | 30.0 | 1050 | 0.6106 | 0.5769 |
| 0.8752 | 31.0 | 1085 | 0.8355 | 0.375 |
| 0.8752 | 32.0 | 1120 | 0.6060 | 0.6058 |
| 0.8752 | 33.0 | 1155 | 0.7944 | 0.375 |
| 0.8752 | 34.0 | 1190 | 0.6461 | 0.6058 |
| 0.8752 | 35.0 | 1225 | 0.6320 | 0.5096 |
| 0.8752 | 36.0 | 1260 | 0.6189 | 0.6154 |
| 0.8752 | 37.0 | 1295 | 0.6007 | 0.625 |
| 0.8752 | 38.0 | 1330 | 0.6415 | 0.5096 |
| 0.8752 | 39.0 | 1365 | 0.6386 | 0.6346 |
| 0.8752 | 40.0 | 1400 | 0.6051 | 0.5962 |
| 0.8752 | 41.0 | 1435 | 0.7365 | 0.3942 |
| 0.8752 | 42.0 | 1470 | 0.7951 | 0.3942 |
| 0.7496 | 43.0 | 1505 | 0.6346 | 0.5385 |
| 0.7496 | 44.0 | 1540 | 0.6475 | 0.4712 |
| 0.7496 | 45.0 | 1575 | 0.7517 | 0.375 |
| 0.7496 | 46.0 | 1610 | 0.6727 | 0.4327 |
| 0.7496 | 47.0 | 1645 | 0.6718 | 0.4712 |
| 0.7496 | 48.0 | 1680 | 0.6113 | 0.5577 |
| 0.7496 | 49.0 | 1715 | 0.6150 | 0.6346 |
| 0.7496 | 50.0 | 1750 | 0.6207 | 0.6346 |
| 0.7496 | 51.0 | 1785 | 0.7305 | 0.375 |
| 0.7496 | 52.0 | 1820 | 0.5944 | 0.6346 |
| 0.7496 | 53.0 | 1855 | 0.6348 | 0.4808 |
| 0.7496 | 54.0 | 1890 | 0.6641 | 0.4808 |
| 0.7496 | 55.0 | 1925 | 0.6014 | 0.6154 |
| 0.7496 | 56.0 | 1960 | 0.6118 | 0.6442 |
| 0.7496 | 57.0 | 1995 | 0.5951 | 0.625 |
| 0.6833 | 58.0 | 2030 | 0.6069 | 0.5769 |
| 0.6833 | 59.0 | 2065 | 0.6264 | 0.5865 |
| 0.6833 | 60.0 | 2100 | 0.6055 | 0.6346 |
| 0.6833 | 61.0 | 2135 | 0.6010 | 0.6346 |
| 0.6833 | 62.0 | 2170 | 0.5987 | 0.6154 |
| 0.6833 | 63.0 | 2205 | 0.6271 | 0.5192 |
| 0.6833 | 64.0 | 2240 | 0.6102 | 0.6346 |
| 0.6833 | 65.0 | 2275 | 0.6039 | 0.6058 |
| 0.6833 | 66.0 | 2310 | 0.6465 | 0.4808 |
| 0.6833 | 67.0 | 2345 | 0.6219 | 0.5481 |
| 0.6833 | 68.0 | 2380 | 0.6189 | 0.5481 |
| 0.6833 | 69.0 | 2415 | 0.5961 | 0.5865 |
| 0.6833 | 70.0 | 2450 | 0.5996 | 0.6058 |
| 0.6833 | 71.0 | 2485 | 0.6017 | 0.6058 |
| 0.6514 | 72.0 | 2520 | 0.6183 | 0.5577 |
| 0.6514 | 73.0 | 2555 | 0.6026 | 0.5962 |
| 0.6514 | 74.0 | 2590 | 0.6205 | 0.4808 |
| 0.6514 | 75.0 | 2625 | 0.6070 | 0.5769 |
| 0.6514 | 76.0 | 2660 | 0.6173 | 0.4904 |
| 0.6514 | 77.0 | 2695 | 0.6138 | 0.5385 |
| 0.6514 | 78.0 | 2730 | 0.6165 | 0.5192 |
| 0.6514 | 79.0 | 2765 | 0.6213 | 0.5192 |
| 0.6514 | 80.0 | 2800 | 0.6170 | 0.5577 |
### Framework versions
- Transformers 4.26.1
- Pytorch 2.0.1+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3
|
dt-and-vanilla-ardt/dt-robust_train_walker2d_v3-2908_1830-99
|
dt-and-vanilla-ardt
| 2023-08-29T18:52:00Z | 31 | 0 |
transformers
|
[
"transformers",
"pytorch",
"decision_transformer",
"generated_from_trainer",
"endpoints_compatible",
"region:us"
] | null | 2023-08-29T17:31:43Z |
---
tags:
- generated_from_trainer
model-index:
- name: dt-robust_train_walker2d_v3-2908_1830-99
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# dt-robust_train_walker2d_v3-2908_1830-99
This model is a fine-tuned version of [](https://huggingface.co/) on the None dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 64
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- training_steps: 10000
### Training results
### Framework versions
- Transformers 4.29.2
- Pytorch 2.1.0.dev20230727+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3
|
BugHunter1/distilhubert-finetuned-gtzan
|
BugHunter1
| 2023-08-29T18:48:19Z | 159 | 0 |
transformers
|
[
"transformers",
"pytorch",
"hubert",
"audio-classification",
"generated_from_trainer",
"dataset:marsyas/gtzan",
"base_model:ntu-spml/distilhubert",
"base_model:finetune:ntu-spml/distilhubert",
"license:apache-2.0",
"model-index",
"endpoints_compatible",
"region:us"
] |
audio-classification
| 2023-08-29T18:48:13Z |
---
license: apache-2.0
base_model: ntu-spml/distilhubert
tags:
- generated_from_trainer
datasets:
- marsyas/gtzan
metrics:
- accuracy
model-index:
- name: distilhubert-finetuned-gtzan
results:
- task:
name: Audio Classification
type: audio-classification
dataset:
name: GTZAN
type: marsyas/gtzan
config: all
split: train
args: all
metrics:
- name: Accuracy
type: accuracy
value: 0.87
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilhubert-finetuned-gtzan
This model is a fine-tuned version of [ntu-spml/distilhubert](https://huggingface.co/ntu-spml/distilhubert) on the GTZAN dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5653
- Accuracy: 0.87
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 3e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 1.0686 | 1.0 | 112 | 1.1059 | 0.65 |
| 0.7149 | 2.0 | 225 | 0.8356 | 0.74 |
| 0.7336 | 3.0 | 337 | 0.7207 | 0.82 |
| 0.5571 | 4.0 | 450 | 0.7318 | 0.8 |
| 0.3585 | 5.0 | 562 | 0.6034 | 0.83 |
| 0.2655 | 6.0 | 675 | 0.6394 | 0.81 |
| 0.1647 | 7.0 | 787 | 0.6114 | 0.81 |
| 0.1216 | 8.0 | 900 | 0.5841 | 0.85 |
| 0.1731 | 9.0 | 1012 | 0.5620 | 0.87 |
| 0.1316 | 9.96 | 1120 | 0.5653 | 0.87 |
### Framework versions
- Transformers 4.33.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3
|
Musha-the-Yusha/Llama-2-7b-chat-finetune
|
Musha-the-Yusha
| 2023-08-29T18:40:37Z | 3 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-08-29T18:30:56Z |
---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float16
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float16
### Framework versions
- PEFT 0.4.0
- PEFT 0.4.0
- PEFT 0.4.0
|
Fduv/DeciCoder-FineTuned-CodeAlpaca
|
Fduv
| 2023-08-29T18:37:12Z | 2 | 0 |
peft
|
[
"peft",
"code",
"en",
"dataset:lucasmccabe-lmi/CodeAlpaca-20k",
"license:apache-2.0",
"region:us"
] | null | 2023-08-29T18:27:42Z |
---
library_name: peft
license: apache-2.0
datasets:
- lucasmccabe-lmi/CodeAlpaca-20k
language:
- en
tags:
- code
---
# Model Card for DeciCoder 1B - CodeAlpaca20k Fine Tuned
This model is a fine tuned version of DeciCoder 1B (https://huggingface.co/Deci/DeciCoder-1b), fine tune for instructions based on Code Alpaca Dataset.
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: float16
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: False
- load_in_4bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
- llm_int8_has_fp16_weight: False
- bnb_4bit_quant_type: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: float16
### Framework versions
- PEFT 0.5.0
## Citation
Thanks for DeciCoder-1b team for making this model open sourced.
```
@misc{DeciFoundationModels,
title = {DeciCoder},
author = {DeciAI Research Team},
year = {2023}
url={[https://huggingface.co/deci/decicoder-1b](https://huggingface.co/deci/decicoder-1b)},
}
```
|
trieudemo11/llama_7b_attrb_cate_b6_l320_low_2_KEEP
|
trieudemo11
| 2023-08-29T18:34:49Z | 2 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-08-29T18:34:30Z |
---
library_name: peft
---
## Training procedure
### Framework versions
- PEFT 0.6.0.dev0
- PEFT 0.6.0.dev0
- PEFT 0.6.0.dev0
- PEFT 0.6.0.dev0
- PEFT 0.6.0.dev0
- PEFT 0.6.0.dev0
- PEFT 0.6.0.dev0
- PEFT 0.6.0.dev0
- PEFT 0.6.0.dev0
- PEFT 0.6.0.dev0
- PEFT 0.6.0.dev0
rsions
- PEFT 0.6.0.dev0
|
fusi0n/Based-7b-GGUF
|
fusi0n
| 2023-08-29T18:31:50Z | 6 | 0 | null |
[
"gguf",
"endpoints_compatible",
"region:us"
] | null | 2023-08-29T15:57:32Z |
GGUF conversion of based-7b from ehartford
original model
https://huggingface.co/ehartford/based-7b
|
mke10/distilbert-base-uncased-finetuned-ner
|
mke10
| 2023-08-29T18:20:53Z | 117 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"tensorboard",
"distilbert",
"token-classification",
"generated_from_keras_callback",
"base_model:distilbert/distilbert-base-uncased",
"base_model:finetune:distilbert/distilbert-base-uncased",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
token-classification
| 2023-08-29T11:48:54Z |
---
license: apache-2.0
base_model: distilbert-base-uncased
tags:
- generated_from_keras_callback
model-index:
- name: mke10/distilbert-base-uncased-finetuned-ner
results: []
---
<!-- This model card has been generated automatically according to the information Keras had access to. You should
probably proofread and complete it, then remove this comment. -->
# mke10/distilbert-base-uncased-finetuned-ner
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Train Loss: 0.0339
- Validation Loss: 0.0603
- Train Precision: 0.9245
- Train Recall: 0.9356
- Train F1: 0.9300
- Train Accuracy: 0.9834
- Epoch: 2
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': 2631, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.01}
- training_precision: float32
### Training results
| Train Loss | Validation Loss | Train Precision | Train Recall | Train F1 | Train Accuracy | Epoch |
|:----------:|:---------------:|:---------------:|:------------:|:--------:|:--------------:|:-----:|
| 0.1985 | 0.0727 | 0.8930 | 0.9177 | 0.9052 | 0.9786 | 0 |
| 0.0535 | 0.0605 | 0.9255 | 0.9323 | 0.9289 | 0.9832 | 1 |
| 0.0339 | 0.0603 | 0.9245 | 0.9356 | 0.9300 | 0.9834 | 2 |
### Framework versions
- Transformers 4.32.1
- TensorFlow 2.12.0
- Datasets 2.14.4
- Tokenizers 0.13.3
|
stablediffusionapi/3AmvPjC7O5hWtaFsl8bf6dkGx
|
stablediffusionapi
| 2023-08-29T18:13:30Z | 17 | 0 |
diffusers
|
[
"diffusers",
"stablediffusionapi.com",
"stable-diffusion-api",
"text-to-image",
"ultra-realistic",
"license:creativeml-openrail-m",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] |
text-to-image
| 2023-05-18T09:52:38Z |
---
license: creativeml-openrail-m
tags:
- stablediffusionapi.com
- stable-diffusion-api
- text-to-image
- ultra-realistic
pinned: true
---
# API Inference
## Get API Key
Get API key from [Stable Diffusion API](http://stablediffusionapi.com/), No Payment needed.
Replace Key in below code, change **model_id** to "3AmvPjC7O5hWtaFsl8bf6dkGx"
Coding in PHP/Node/Java etc? Have a look at docs for more code examples: [View docs](https://stablediffusionapi.com/docs)
Model link: [View model](https://stablediffusionapi.com/models/3AmvPjC7O5hWtaFsl8bf6dkGx)
Credits: [View credits](https://civitai.com/?query=model_search)
View all models: [View Models](https://stablediffusionapi.com/models)
import requests
import json
url = "https://stablediffusionapi.com/api/v3/dreambooth"
payload = json.dumps({
"key": "",
"model_id": "3AmvPjC7O5hWtaFsl8bf6dkGx",
"prompt": "actual 8K portrait photo of gareth person, portrait, happy colors, bright eyes, clear eyes, warm smile, smooth soft skin, big dreamy eyes, beautiful intricate colored hair, symmetrical, anime wide eyes, soft lighting, detailed face, by makoto shinkai, stanley artgerm lau, wlop, rossdraws, concept art, digital painting, looking into camera",
"negative_prompt": "painting, extra fingers, mutated hands, poorly drawn hands, poorly drawn face, deformed, ugly, blurry, bad anatomy, bad proportions, extra limbs, cloned face, skinny, glitchy, double torso, extra arms, extra hands, mangled fingers, missing lips, ugly face, distorted face, extra legs, anime",
"width": "512",
"height": "512",
"samples": "1",
"num_inference_steps": "30",
"safety_checker": "no",
"enhance_prompt": "yes",
"seed": None,
"guidance_scale": 7.5,
"multi_lingual": "no",
"panorama": "no",
"self_attention": "no",
"upscale": "no",
"embeddings": "embeddings_model_id",
"lora": "lora_model_id",
"webhook": None,
"track_id": None
})
headers = {
'Content-Type': 'application/json'
}
response = requests.request("POST", url, headers=headers, data=payload)
print(response.text)
> Use this coupon code to get 25% off **DMGG0RBN**
|
Subsets and Splits
Filtered Qwen2.5 Distill Models
Identifies specific configurations of models by filtering cards that contain 'distill', 'qwen2.5', '7b' while excluding certain base models and incorrect model ID patterns, uncovering unique model variants.
Filtered Model Cards Count
Finds the count of entries with specific card details that include 'distill', 'qwen2.5', '7b' but exclude certain base models, revealing valuable insights about the dataset's content distribution.
Filtered Distill Qwen 7B Models
Filters for specific card entries containing 'distill', 'qwen', and '7b', excluding certain strings and patterns, to identify relevant model configurations.
Filtered Qwen-7b Model Cards
The query performs a detailed filtering based on specific keywords and excludes certain entries, which could be useful for identifying a specific subset of cards but does not provide deeper insights or trends.
Filtered Qwen 7B Model Cards
The query filters for specific terms related to "distilled" or "distill", "qwen", and "7b" in the 'card' column but excludes certain base models, providing a limited set of entries for further inspection.
Qwen 7B Distilled Models
The query provides a basic filtering of records to find specific card names that include keywords related to distilled Qwen 7b models, excluding a particular base model, which gives limited insight but helps in focusing on relevant entries.
Qwen 7B Distilled Model Cards
The query filters data based on specific keywords in the modelId and card fields, providing limited insight primarily useful for locating specific entries rather than revealing broad patterns or trends.
Qwen 7B Distilled Models
Finds all entries containing the terms 'distilled', 'qwen', and '7b' in a case-insensitive manner, providing a filtered set of records but without deeper analysis.
Distilled Qwen 7B Models
The query filters for specific model IDs containing 'distilled', 'qwen', and '7b', providing a basic retrieval of relevant entries but without deeper analysis or insight.
Filtered Model Cards with Distill Qwen2.
Filters and retrieves records containing specific keywords in the card description while excluding certain phrases, providing a basic count of relevant entries.
Filtered Model Cards with Distill Qwen 7
The query filters specific variations of card descriptions containing 'distill', 'qwen', and '7b' while excluding a particular base model, providing limited but specific data retrieval.
Distill Qwen 7B Model Cards
The query filters and retrieves rows where the 'card' column contains specific keywords ('distill', 'qwen', and '7b'), providing a basic filter result that can help in identifying specific entries.