|
|
--- |
|
|
library_name: peft |
|
|
license: other |
|
|
base_model: Qwen/Qwen3-32B |
|
|
tags: |
|
|
- llama-factory |
|
|
- lora |
|
|
- generated_from_trainer |
|
|
model-index: |
|
|
- name: Qwen3-32B-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. --> |
|
|
|
|
|
# Qwen3-32B-alpaca-th-52k-dolly-th-15k-wangchan-instruct |
|
|
|
|
|
This model is a fine-tuned version of [Qwen/Qwen3-32B](https://huggingface.co/Qwen/Qwen3-32B) on the alpaca-th-52k, the dolly-th-15k and the wangchan-instruct datasets. |
|
|
It achieves the following results on the evaluation set: |
|
|
- Loss: 0.6417 |
|
|
|
|
|
## 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: 32 |
|
|
- gradient_accumulation_steps: 8 |
|
|
- total_train_batch_size: 512 |
|
|
- total_eval_batch_size: 64 |
|
|
- 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 | |
|
|
|:-------------:|:------:|:----:|:---------------:| |
|
|
| 0.9564 | 0.0575 | 10 | 1.0507 | |
|
|
| 0.806 | 0.1149 | 20 | 0.8268 | |
|
|
| 0.7551 | 0.1724 | 30 | 0.7598 | |
|
|
| 0.7158 | 0.2299 | 40 | 0.7396 | |
|
|
| 0.7217 | 0.2874 | 50 | 0.7252 | |
|
|
| 0.7078 | 0.3448 | 60 | 0.7130 | |
|
|
| 0.6719 | 0.4023 | 70 | 0.7029 | |
|
|
| 0.6855 | 0.4598 | 80 | 0.6964 | |
|
|
| 0.7328 | 0.5172 | 90 | 0.6907 | |
|
|
| 0.6663 | 0.5747 | 100 | 0.6848 | |
|
|
| 0.7049 | 0.6322 | 110 | 0.6792 | |
|
|
| 0.6772 | 0.6897 | 120 | 0.6751 | |
|
|
| 0.687 | 0.7471 | 130 | 0.6721 | |
|
|
| 0.6786 | 0.8046 | 140 | 0.6700 | |
|
|
| 0.6389 | 0.8621 | 150 | 0.6672 | |
|
|
| 0.6673 | 0.9195 | 160 | 0.6649 | |
|
|
| 0.6711 | 0.9770 | 170 | 0.6633 | |
|
|
| 0.6614 | 1.0345 | 180 | 0.6615 | |
|
|
| 0.6219 | 1.0920 | 190 | 0.6602 | |
|
|
| 0.6542 | 1.1494 | 200 | 0.6587 | |
|
|
| 0.6596 | 1.2069 | 210 | 0.6572 | |
|
|
| 0.6526 | 1.2644 | 220 | 0.6567 | |
|
|
| 0.657 | 1.3218 | 230 | 0.6551 | |
|
|
| 0.6124 | 1.3793 | 240 | 0.6537 | |
|
|
| 0.6489 | 1.4368 | 250 | 0.6526 | |
|
|
| 0.614 | 1.4943 | 260 | 0.6515 | |
|
|
| 0.656 | 1.5517 | 270 | 0.6504 | |
|
|
| 0.6255 | 1.6092 | 280 | 0.6492 | |
|
|
| 0.6419 | 1.6667 | 290 | 0.6486 | |
|
|
| 0.6275 | 1.7241 | 300 | 0.6473 | |
|
|
| 0.6324 | 1.7816 | 310 | 0.6466 | |
|
|
| 0.6334 | 1.8391 | 320 | 0.6461 | |
|
|
| 0.6213 | 1.8966 | 330 | 0.6452 | |
|
|
| 0.6269 | 1.9540 | 340 | 0.6443 | |
|
|
| 0.6408 | 2.0115 | 350 | 0.6437 | |
|
|
| 0.6213 | 2.0690 | 360 | 0.6441 | |
|
|
| 0.6146 | 2.1264 | 370 | 0.6440 | |
|
|
| 0.6572 | 2.1839 | 380 | 0.6438 | |
|
|
| 0.6264 | 2.2414 | 390 | 0.6435 | |
|
|
| 0.6051 | 2.2989 | 400 | 0.6434 | |
|
|
| 0.5983 | 2.3563 | 410 | 0.6429 | |
|
|
| 0.6388 | 2.4138 | 420 | 0.6425 | |
|
|
| 0.6227 | 2.4713 | 430 | 0.6425 | |
|
|
| 0.6335 | 2.5287 | 440 | 0.6421 | |
|
|
| 0.6247 | 2.5862 | 450 | 0.6420 | |
|
|
| 0.6404 | 2.6437 | 460 | 0.6418 | |
|
|
| 0.6218 | 2.7011 | 470 | 0.6418 | |
|
|
| 0.6368 | 2.7586 | 480 | 0.6417 | |
|
|
| 0.6191 | 2.8161 | 490 | 0.6417 | |
|
|
| 0.6234 | 2.8736 | 500 | 0.6417 | |
|
|
| 0.6079 | 2.9310 | 510 | 0.6417 | |
|
|
| 0.6243 | 2.9885 | 520 | 0.6417 | |
|
|
|
|
|
|
|
|
### Framework versions |
|
|
|
|
|
- PEFT 0.15.2 |
|
|
- Transformers 4.52.3 |
|
|
- Pytorch 2.7.0+cu126 |
|
|
- Datasets 3.6.0 |
|
|
- Tokenizers 0.21.1 |
