qwen3_8b_scope_kshot_2gpu
This model is a fine-tuned version of Qwen/Qwen3-8B on the scope_kshot_format dataset. It achieves the following results on the evaluation set:
- Loss: 0.5126
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: 1
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- num_devices: 2
- gradient_accumulation_steps: 8
- total_train_batch_size: 16
- total_eval_batch_size: 2
- 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: 2.0
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.6759 | 0.1360 | 500 | 0.6661 |
| 0.6258 | 0.2719 | 1000 | 0.6303 |
| 0.6131 | 0.4079 | 1500 | 0.6015 |
| 0.5749 | 0.5438 | 2000 | 0.5814 |
| 0.5728 | 0.6798 | 2500 | 0.5645 |
| 0.5627 | 0.8158 | 3000 | 0.5497 |
| 0.5387 | 0.9517 | 3500 | 0.5397 |
| 0.392 | 1.0876 | 4000 | 0.5497 |
| 0.3751 | 1.2235 | 4500 | 0.5453 |
| 0.3874 | 1.3595 | 5000 | 0.5357 |
| 0.3798 | 1.4954 | 5500 | 0.5304 |
| 0.3702 | 1.6314 | 6000 | 0.5244 |
| 0.3706 | 1.7674 | 6500 | 0.5185 |
| 0.3595 | 1.9033 | 7000 | 0.5142 |
Framework versions
- Transformers 4.53.1
- Pytorch 2.5.1+cu121
- Datasets 3.6.0
- Tokenizers 0.21.2
- Downloads last month
- -