SFT_outputs
This model is a fine-tuned version of /opt/huawei/explorer-env/checkpoint/longvideo_data/DramaSR_cleaned/assets/Qwen3-8B on the sft_training_data dataset. It achieves the following results on the evaluation set:
- Loss: 1.1952
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: 1e-05
- train_batch_size: 8
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 4
- total_train_batch_size: 256
- total_eval_batch_size: 8
- optimizer: Use paged_adamw_8bit 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: 5.0
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.2071 | 1.0 | 96 | 1.1540 |
| 1.1484 | 2.0 | 192 | 1.1297 |
| 1.0432 | 3.0 | 288 | 1.1359 |
| 0.967 | 4.0 | 384 | 1.1620 |
| 0.9125 | 5.0 | 480 | 1.1952 |
Framework versions
- Transformers 4.52.4
- Pytorch 2.8.0+cu128
- Datasets 3.6.0
- Tokenizers 0.21.1
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support