metadata
library_name: peft
license: other
base_model: Qwen/Qwen3-32B
tags:
- llama-factory
- lora
- generated_from_trainer
model-index:
- name: Qwen3-32B-medqa-seed-4201
results: []
Qwen3-32B-medqa-seed-4201
This model is a fine-tuned version of Qwen/Qwen3-32B on the medqa dataset. It achieves the following results on the evaluation set:
- Loss: 0.0247
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: 4201
- 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 |
|---|---|---|---|
| 1.4047 | 0.1163 | 10 | 1.1082 |
| 0.0377 | 0.2326 | 20 | 0.0421 |
| 0.0309 | 0.3488 | 30 | 0.0327 |
| 0.0276 | 0.4651 | 40 | 0.0300 |
| 0.0307 | 0.5814 | 50 | 0.0288 |
| 0.0266 | 0.6977 | 60 | 0.0282 |
| 0.0293 | 0.8140 | 70 | 0.0276 |
| 0.0254 | 0.9302 | 80 | 0.0269 |
| 0.0215 | 1.0465 | 90 | 0.0265 |
| 0.0176 | 1.1628 | 100 | 0.0262 |
| 0.0227 | 1.2791 | 110 | 0.0257 |
| 0.0198 | 1.3953 | 120 | 0.0254 |
| 0.0245 | 1.5116 | 130 | 0.0253 |
| 0.0217 | 1.6279 | 140 | 0.0252 |
| 0.0209 | 1.7442 | 150 | 0.0249 |
| 0.0227 | 1.8605 | 160 | 0.0247 |
| 0.0195 | 1.9767 | 170 | 0.0244 |
| 0.0174 | 2.0930 | 180 | 0.0246 |
| 0.018 | 2.2093 | 190 | 0.0247 |
| 0.0176 | 2.3256 | 200 | 0.0248 |
| 0.0149 | 2.4419 | 210 | 0.0247 |
| 0.0176 | 2.5581 | 220 | 0.0247 |
| 0.016 | 2.6744 | 230 | 0.0247 |
| 0.0202 | 2.7907 | 240 | 0.0247 |
| 0.0144 | 2.9070 | 250 | 0.0247 |
Framework versions
- PEFT 0.15.2
- Transformers 4.52.3
- Pytorch 2.7.0+cu126
- Datasets 3.6.0
- Tokenizers 0.21.1