metadata
base_model: yahma/llama-7b-hf
tags:
- generated_from_trainer
model-index:
- name: V0224B1
results: []
V0224B1
This model is a fine-tuned version of yahma/llama-7b-hf on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.7525
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 |
|---|---|---|---|
| 1.1045 | 0.13 | 10 | 1.0762 |
| 0.9886 | 0.26 | 20 | 0.9163 |
| 0.8808 | 0.39 | 30 | 0.8540 |
| 0.8273 | 0.52 | 40 | 0.8241 |
| 0.8082 | 0.65 | 50 | 0.8067 |
| 0.7915 | 0.78 | 60 | 0.7954 |
| 0.7687 | 0.91 | 70 | 0.7883 |
| 0.7644 | 1.04 | 80 | 0.7814 |
| 0.7454 | 1.17 | 90 | 0.7759 |
| 0.7613 | 1.3 | 100 | 0.7717 |
| 0.7512 | 1.43 | 110 | 0.7681 |
| 0.7416 | 1.55 | 120 | 0.7644 |
| 0.7315 | 1.68 | 130 | 0.7613 |
| 0.7434 | 1.81 | 140 | 0.7594 |
| 0.7477 | 1.94 | 150 | 0.7563 |
| 0.7299 | 2.07 | 160 | 0.7555 |
| 0.7148 | 2.2 | 170 | 0.7540 |
| 0.7272 | 2.33 | 180 | 0.7538 |
| 0.7203 | 2.46 | 190 | 0.7532 |
| 0.7216 | 2.59 | 200 | 0.7531 |
| 0.7233 | 2.72 | 210 | 0.7527 |
| 0.7213 | 2.85 | 220 | 0.7526 |
| 0.7234 | 2.98 | 230 | 0.7525 |
Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.2+cu121
- Datasets 2.14.6
- Tokenizers 0.14.1