---
library_name: peft
license: other
base_model: Qwen/Qwen3-30B-A3B
tags:
- llama-factory
- lora
- generated_from_trainer
model-index:
- name: Qwen3-30B-A3B-alpaca-th-52k-dolly-th-15k-wangchan-instruct
  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. -->

# Qwen3-30B-A3B-alpaca-th-52k-dolly-th-15k-wangchan-instruct

This model is a fine-tuned version of [Qwen/Qwen3-30B-A3B](https://huggingface.co/Qwen/Qwen3-30B-A3B) on the alpaca-th-52k, the dolly-th-15k and the wangchan-instruct datasets.
It achieves the following results on the evaluation set:
- Loss: 0.6631

## 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: 2
- eval_batch_size: 2
- seed: 42
- distributed_type: multi-GPU
- num_devices: 64
- gradient_accumulation_steps: 8
- total_train_batch_size: 1024
- total_eval_batch_size: 128
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 3.0

### Training results

| Training Loss | Epoch  | Step | Validation Loss |
|:-------------:|:------:|:----:|:---------------:|
| 0.9351        | 0.1149 | 10   | 0.9997          |
| 0.8087        | 0.2299 | 20   | 0.8204          |
| 0.7724        | 0.3448 | 30   | 0.7787          |
| 0.7386        | 0.4598 | 40   | 0.7544          |
| 0.7351        | 0.5747 | 50   | 0.7382          |
| 0.7431        | 0.6897 | 60   | 0.7254          |
| 0.7183        | 0.8046 | 70   | 0.7151          |
| 0.711         | 0.9195 | 80   | 0.7065          |
| 0.6909        | 1.0345 | 90   | 0.6995          |
| 0.6893        | 1.1494 | 100  | 0.6939          |
| 0.6796        | 1.2644 | 110  | 0.6874          |
| 0.65          | 1.3793 | 120  | 0.6812          |
| 0.6615        | 1.4943 | 130  | 0.6775          |
| 0.6555        | 1.6092 | 140  | 0.6739          |
| 0.6522        | 1.7241 | 150  | 0.6713          |
| 0.6545        | 1.8391 | 160  | 0.6687          |
| 0.648         | 1.9540 | 170  | 0.6668          |
| 0.6285        | 2.0690 | 180  | 0.6663          |
| 0.6652        | 2.1839 | 190  | 0.6655          |
| 0.6307        | 2.2989 | 200  | 0.6647          |
| 0.6383        | 2.4138 | 210  | 0.6641          |
| 0.6394        | 2.5287 | 220  | 0.6636          |
| 0.632         | 2.6437 | 230  | 0.6632          |
| 0.6416        | 2.7586 | 240  | 0.6631          |
| 0.6228        | 2.8736 | 250  | 0.6631          |
| 0.6316        | 2.9885 | 260  | 0.6630          |


### Framework versions

- PEFT 0.15.2
- Transformers 4.52.3
- Pytorch 2.7.0+cu126
- Datasets 3.6.0
- Tokenizers 0.21.1