|
--- |
|
license: other |
|
base_model: yahma/llama-7b-hf |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: V0305P5 |
|
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. --> |
|
|
|
# V0305P5 |
|
|
|
This model is a fine-tuned version of [yahma/llama-7b-hf](https://huggingface.co/yahma/llama-7b-hf) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.0750 |
|
|
|
## 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: 0.0003 |
|
- train_batch_size: 4 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 32 |
|
- total_train_batch_size: 128 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: cosine_with_restarts |
|
- lr_scheduler_warmup_steps: 20 |
|
- num_epochs: 3 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:-----:|:----:|:---------------:| |
|
| 1.3878 | 0.09 | 10 | 0.1658 | |
|
| 0.2801 | 0.17 | 20 | 0.1531 | |
|
| 0.1578 | 0.26 | 30 | 0.1529 | |
|
| 0.1521 | 0.34 | 40 | 0.1514 | |
|
| 0.1515 | 0.43 | 50 | 0.1508 | |
|
| 0.1567 | 0.51 | 60 | 0.1507 | |
|
| 0.1517 | 0.6 | 70 | 0.1491 | |
|
| 0.1531 | 0.68 | 80 | 0.1480 | |
|
| 0.149 | 0.77 | 90 | 0.1474 | |
|
| 0.153 | 0.85 | 100 | 0.1521 | |
|
| 0.1511 | 0.94 | 110 | 0.1264 | |
|
| 0.14 | 1.02 | 120 | 0.1210 | |
|
| 0.1316 | 1.11 | 130 | 0.1209 | |
|
| 0.1198 | 1.19 | 140 | 0.1015 | |
|
| 0.1105 | 1.28 | 150 | 0.0888 | |
|
| 0.1067 | 1.37 | 160 | 0.0930 | |
|
| 0.1043 | 1.45 | 170 | 0.0852 | |
|
| 0.0962 | 1.54 | 180 | 0.0887 | |
|
| 0.0968 | 1.62 | 190 | 0.0805 | |
|
| 0.0957 | 1.71 | 200 | 0.0824 | |
|
| 0.0939 | 1.79 | 210 | 0.0819 | |
|
| 0.0918 | 1.88 | 220 | 0.0803 | |
|
| 0.0865 | 1.96 | 230 | 0.0818 | |
|
| 0.0734 | 2.05 | 240 | 0.0791 | |
|
| 0.0591 | 2.13 | 250 | 0.0743 | |
|
| 0.0614 | 2.22 | 260 | 0.0709 | |
|
| 0.0544 | 2.3 | 270 | 0.0836 | |
|
| 0.0611 | 2.39 | 280 | 0.0747 | |
|
| 0.064 | 2.47 | 290 | 0.0711 | |
|
| 0.0582 | 2.56 | 300 | 0.0721 | |
|
| 0.0646 | 2.65 | 310 | 0.0752 | |
|
| 0.0551 | 2.73 | 320 | 0.0761 | |
|
| 0.0571 | 2.82 | 330 | 0.0753 | |
|
| 0.0562 | 2.9 | 340 | 0.0750 | |
|
| 0.0537 | 2.99 | 350 | 0.0750 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.36.0.dev0 |
|
- Pytorch 2.1.2+cu121 |
|
- Datasets 2.14.6 |
|
- Tokenizers 0.14.1 |
|
|
