|
--- |
|
license: mit |
|
library_name: peft |
|
tags: |
|
- generated_from_trainer |
|
base_model: microsoft/phi-2 |
|
model-index: |
|
- name: fine-tuning-Phi2-with-webglm-qa-with-lora_3 |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# fine-tuning-Phi2-with-webglm-qa-with-lora_3 |
|
|
|
This model is a fine-tuned version of [microsoft/phi-2](https://huggingface.co/microsoft/phi-2) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.1155 |
|
|
|
## 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: 5e-05 |
|
- train_batch_size: 2 |
|
- eval_batch_size: 2 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 5 |
|
- total_train_batch_size: 10 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 50 |
|
- training_steps: 500 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:-----:|:----:|:---------------:| |
|
| 8.243 | 0.2 | 10 | 7.8185 | |
|
| 7.4602 | 0.4 | 20 | 6.3280 | |
|
| 4.794 | 0.6 | 30 | 3.1068 | |
|
| 1.6994 | 0.8 | 40 | 0.6354 | |
|
| 0.543 | 1.0 | 50 | 0.5653 | |
|
| 0.4542 | 1.2 | 60 | 0.4874 | |
|
| 0.4449 | 1.39 | 70 | 0.4225 | |
|
| 0.3623 | 1.59 | 80 | 0.3685 | |
|
| 0.278 | 1.79 | 90 | 0.3283 | |
|
| 0.2385 | 1.99 | 100 | 0.2983 | |
|
| 0.2499 | 2.19 | 110 | 0.2748 | |
|
| 0.2113 | 2.39 | 120 | 0.2590 | |
|
| 0.1966 | 2.59 | 130 | 0.2420 | |
|
| 0.217 | 2.79 | 140 | 0.2242 | |
|
| 0.1731 | 2.99 | 150 | 0.2121 | |
|
| 0.1779 | 3.19 | 160 | 0.2033 | |
|
| 0.1687 | 3.39 | 170 | 0.1909 | |
|
| 0.156 | 3.59 | 180 | 0.1833 | |
|
| 0.1464 | 3.78 | 190 | 0.1763 | |
|
| 0.1637 | 3.98 | 200 | 0.1706 | |
|
| 0.1455 | 4.18 | 210 | 0.1649 | |
|
| 0.128 | 4.38 | 220 | 0.1621 | |
|
| 0.1537 | 4.58 | 230 | 0.1562 | |
|
| 0.1193 | 4.78 | 240 | 0.1502 | |
|
| 0.1323 | 4.98 | 250 | 0.1464 | |
|
| 0.1346 | 5.18 | 260 | 0.1440 | |
|
| 0.1049 | 5.38 | 270 | 0.1411 | |
|
| 0.1265 | 5.58 | 280 | 0.1377 | |
|
| 0.13 | 5.78 | 290 | 0.1363 | |
|
| 0.1059 | 5.98 | 300 | 0.1335 | |
|
| 0.1141 | 6.18 | 310 | 0.1300 | |
|
| 0.1097 | 6.37 | 320 | 0.1297 | |
|
| 0.1088 | 6.57 | 330 | 0.1287 | |
|
| 0.106 | 6.77 | 340 | 0.1261 | |
|
| 0.1011 | 6.97 | 350 | 0.1243 | |
|
| 0.0999 | 7.17 | 360 | 0.1235 | |
|
| 0.1081 | 7.37 | 370 | 0.1223 | |
|
| 0.0999 | 7.57 | 380 | 0.1207 | |
|
| 0.1057 | 7.77 | 390 | 0.1203 | |
|
| 0.0937 | 7.97 | 400 | 0.1192 | |
|
| 0.0842 | 8.17 | 410 | 0.1195 | |
|
| 0.0907 | 8.37 | 420 | 0.1182 | |
|
| 0.1109 | 8.57 | 430 | 0.1176 | |
|
| 0.0901 | 8.76 | 440 | 0.1170 | |
|
| 0.1005 | 8.96 | 450 | 0.1162 | |
|
| 0.0961 | 9.16 | 460 | 0.1159 | |
|
| 0.0927 | 9.36 | 470 | 0.1156 | |
|
| 0.0916 | 9.56 | 480 | 0.1158 | |
|
| 0.0908 | 9.76 | 490 | 0.1156 | |
|
| 0.0909 | 9.96 | 500 | 0.1155 | |
|
|
|
|
|
### Framework versions |
|
|
|
- PEFT 0.7.1 |
|
- Transformers 4.36.2 |
|
- Pytorch 2.0.0 |
|
- Datasets 2.15.0 |
|
- Tokenizers 0.15.0 |
