metadata
license: other
base_model: meta-llama/Meta-Llama-3-8B
tags:
- llama-factory
- full
- generated_from_trainer
model-index:
- name: C018_random_sample_llama3-8b-base_pretrain_20240504_182259
results: []
C018_random_sample_llama3-8b-base_pretrain_20240504_182259
This model is a fine-tuned version of /data/pro-align/progressalign/shared_storage/downloaded_models/llama3-8b-base on the C018_random_sample_data dataset. It achieves the following results on the evaluation set:
- Loss: 2.2706
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: 1.5e-05
- train_batch_size: 8
- eval_batch_size: 16
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- total_train_batch_size: 64
- total_eval_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: polynomial
- lr_scheduler_warmup_steps: 20
- num_epochs: 4.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 2.3701 | 0.2186 | 200 | 2.3702 |
| 2.3183 | 0.4372 | 400 | 2.3160 |
| 2.2634 | 0.6557 | 600 | 2.2863 |
| 2.2522 | 0.8743 | 800 | 2.2706 |
| 2.0306 | 1.0929 | 1000 | 2.2777 |
| 2.0095 | 1.3115 | 1200 | 2.2760 |
| 2.0539 | 1.5301 | 1400 | 2.2746 |
| 2.0338 | 1.7486 | 1600 | 2.2743 |
| 2.0648 | 1.9672 | 1800 | 2.2737 |
| 2.0297 | 2.1858 | 2000 | 2.2766 |
| 2.0487 | 2.4044 | 2200 | 2.2767 |
| 2.0329 | 2.6230 | 2400 | 2.2770 |
| 2.0213 | 2.8415 | 2600 | 2.2766 |
| 2.0559 | 3.0601 | 2800 | 2.2771 |
| 2.0543 | 3.2787 | 3000 | 2.2773 |
| 2.0317 | 3.4973 | 3200 | 2.2772 |
| 1.988 | 3.7158 | 3400 | 2.2770 |
| 2.0355 | 3.9344 | 3600 | 2.2772 |
Framework versions
- Transformers 4.40.1
- Pytorch 2.3.0
- Datasets 2.19.0
- Tokenizers 0.19.1