metadata
base_model: microsoft/phi-1_5
library_name: peft
license: mit
tags:
- generated_from_trainer
model-index:
- name: Phi-Medical-QA-LoRA
results: []
Phi-Medical-QA-LoRA
This model is a fine-tuned version of microsoft/phi-1_5 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.9509
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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 3.1312 | 0.1684 | 100 | 2.7874 |
| 2.6198 | 0.3367 | 200 | 2.5280 |
| 2.4484 | 0.5051 | 300 | 2.3807 |
| 2.3069 | 0.6734 | 400 | 2.2949 |
| 2.2195 | 0.8418 | 500 | 2.2151 |
| 2.2088 | 1.0101 | 600 | 2.1664 |
| 2.0887 | 1.1785 | 700 | 2.1273 |
| 2.0318 | 1.3468 | 800 | 2.0870 |
| 2.0239 | 1.5152 | 900 | 2.0519 |
| 1.9709 | 1.6835 | 1000 | 2.0285 |
| 1.9697 | 1.8519 | 1100 | 2.0116 |
| 1.9416 | 2.0202 | 1200 | 1.9881 |
| 1.8815 | 2.1886 | 1300 | 1.9745 |
| 1.8816 | 2.3569 | 1400 | 1.9661 |
| 1.8641 | 2.5253 | 1500 | 1.9577 |
| 1.843 | 2.6936 | 1600 | 1.9526 |
| 1.8638 | 2.8620 | 1700 | 1.9509 |
Framework versions
- PEFT 0.11.1
- Transformers 4.42.4
- Pytorch 1.13.1+cu117
- Datasets 2.19.2
- Tokenizers 0.19.1