|
|
--- |
|
|
license: other |
|
|
base_model: yahma/llama-7b-hf |
|
|
tags: |
|
|
- generated_from_trainer |
|
|
model-index: |
|
|
- name: V0224O4 |
|
|
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. --> |
|
|
|
|
|
# V0224O4 |
|
|
|
|
|
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: 0.7661 |
|
|
|
|
|
## 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 | |
|
|
|:-------------:|:-----:|:----:|:---------------:| |
|
|
| 4.1474 | 0.13 | 10 | 1.7099 | |
|
|
| 1.306 | 0.26 | 20 | 1.0662 | |
|
|
| 0.9881 | 0.39 | 30 | 0.9297 | |
|
|
| 0.8907 | 0.52 | 40 | 0.8719 | |
|
|
| 0.8543 | 0.65 | 50 | 0.8475 | |
|
|
| 0.8285 | 0.78 | 60 | 0.8288 | |
|
|
| 0.7987 | 0.91 | 70 | 0.8152 | |
|
|
| 0.7878 | 1.04 | 80 | 0.8042 | |
|
|
| 0.7603 | 1.17 | 90 | 0.7955 | |
|
|
| 0.7743 | 1.3 | 100 | 0.7908 | |
|
|
| 0.7619 | 1.43 | 110 | 0.7851 | |
|
|
| 0.7517 | 1.55 | 120 | 0.7817 | |
|
|
| 0.7411 | 1.68 | 130 | 0.7780 | |
|
|
| 0.752 | 1.81 | 140 | 0.7743 | |
|
|
| 0.7545 | 1.94 | 150 | 0.7711 | |
|
|
| 0.7325 | 2.07 | 160 | 0.7706 | |
|
|
| 0.7139 | 2.2 | 170 | 0.7692 | |
|
|
| 0.7255 | 2.33 | 180 | 0.7676 | |
|
|
| 0.7179 | 2.46 | 190 | 0.7676 | |
|
|
| 0.72 | 2.59 | 200 | 0.7670 | |
|
|
| 0.7222 | 2.72 | 210 | 0.7665 | |
|
|
| 0.7203 | 2.85 | 220 | 0.7660 | |
|
|
| 0.721 | 2.98 | 230 | 0.7661 | |
|
|
|
|
|
|
|
|
### Framework versions |
|
|
|
|
|
- Transformers 4.36.0.dev0 |
|
|
- Pytorch 2.1.2+cu121 |
|
|
- Datasets 2.14.6 |
|
|
- Tokenizers 0.14.1 |
|
|
|
