File size: 2,343 Bytes
30bc546 17c9be9 30bc546 17c9be9 30bc546 e2df9cf 30bc546 6b95560 30bc546 3d31fe6 36a1820 4358d34 30bc546 58a5fd3 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 |
---
license: apache-2.0
tags:
- bert
- deberta
- text-classification
- fine-tuned
- databricks-dolly
- prompt-category
language: en
datasets:
- databricks/databricks-dolly-15k
base_model:
- microsoft/deberta-v3-base
---
# 🧠 DeBERTa-v3 Base - Prompt Category Classifier (Fine-tuned)
This model is a fine-tuned version of [`microsoft/deberta-v3-base`](https://huggingface.co/microsoft/deberta-v3-base) on the [databricks-dolly-15k](https://huggingface.co/datasets/databricks/databricks-dolly-15k) dataset.
It has been trained to classify the **prompt category** based solely on the **response** text.
## 🗂️ Task
**Text Classification**
**Input**: Response text
**Output**: One of the predefined categories such as:
- `brainstorming`
- `classification`
- `closed_qa`
- `creative_writing`
- `general_qa`
- `information_extraction`
- `open_qa`
- `summarization`
## 📊 Evaluation
The model was evaluated on a balanced version of the dataset. Here are the results:
- **Validation Accuracy**: ~85.5%
- **F1 Score**: ~85.0%
- Best performance on: `creative_writing`, `classification`, `summarization`
- Room for improvement on: `open_qa`
## 🧪 How to Use
```python
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch
model = AutoModelForSequenceClassification.from_pretrained("mariadg/deberta-v3-prompt-recognition")
tokenizer = AutoTokenizer.from_pretrained("mariadg/deberta-v3-prompt-recognition")
text = "The mitochondria is known as the powerhouse of the cell."
inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True)
outputs = model(**inputs)
pred = torch.argmax(outputs.logits, dim=1).item()
print(pred) # Map this index back to label if needed
```
## 📦 Label Mapping
The model outputs a numerical label corresponding to a prompt category. Below is the mapping between label IDs and their respective categories:
- 0: `brainstorming`
- 1: `classification`
- 2: `closed_qa`
- 3: `creative_writing`
- 4: `general_qa`
- 5: `information_extraction`
- 6: `open_qa`
- 7: `summarization`
## 🛠️ Training Details
- **Base model**: `microsoft/deberta-v3-base`
- **Framework**: PyTorch
- **Max length**: 256
- **Batch size**: 16
- **Epochs**: 4
- **Loss function**: `CrossEntropyLoss`
## 🔐 License
Apache 2.0
---
📝 Fine-tuned for research purposes. |