phi-2-QLoRa1-Python
This model is a fine-tuned version of microsoft/phi-2 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 4.3315
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
- gradient_accumulation_steps: 2
- total_train_batch_size: 4
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 4.3884 | 0.1333 | 100 | 4.3373 |
| 4.2606 | 0.2667 | 200 | 4.3355 |
| 4.1367 | 0.4 | 300 | 4.3477 |
| 4.2479 | 0.5333 | 400 | 4.3369 |
| 4.2398 | 0.6667 | 500 | 4.3360 |
| 3.9217 | 0.8 | 600 | 4.3400 |
| 4.0251 | 0.9333 | 700 | 4.3373 |
| 4.3893 | 1.0667 | 800 | 4.3477 |
| 4.4014 | 1.2 | 900 | 4.3426 |
| 3.8868 | 1.3333 | 1000 | 4.3346 |
| 4.3634 | 1.4667 | 1100 | 4.3389 |
| 3.9379 | 1.6 | 1200 | 4.3337 |
| 4.0013 | 1.7333 | 1300 | 4.3340 |
| 4.3155 | 1.8667 | 1400 | 4.3331 |
| 4.5401 | 2.0 | 1500 | 4.3330 |
| 3.8402 | 2.1333 | 1600 | 4.3335 |
| 4.3658 | 2.2667 | 1700 | 4.3336 |
| 3.6929 | 2.4 | 1800 | 4.3326 |
| 4.482 | 2.5333 | 1900 | 4.3321 |
| 4.2755 | 2.6667 | 2000 | 4.3317 |
| 4.1472 | 2.8 | 2100 | 4.3315 |
| 4.1581 | 2.9333 | 2200 | 4.3315 |
Framework versions
- PEFT 0.17.1
- Transformers 4.56.1
- Pytorch 2.8.0+cu126
- Datasets 4.0.0
- Tokenizers 0.22.0