metadata
license: other
base_model: meta-llama/Meta-Llama-3-8B
tags:
- llama-factory
- full
- generated_from_trainer
model-index:
- name: C020_random_sample_llama3-8b-base_pretrain_20240505_135320
results: []
C020_random_sample_llama3-8b-base_pretrain_20240505_135320
This model is a fine-tuned version of /data/pro-align/progressalign/shared_storage/downloaded_models/llama3-8b-base on the C020_random_sample_data dataset. It achieves the following results on the evaluation set:
- Loss: 1.9418
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: 4
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- total_train_batch_size: 32
- total_eval_batch_size: 64
- 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 |
|---|---|---|---|
| 1.9087 | 0.4032 | 200 | 1.9717 |
| 1.8752 | 0.8065 | 400 | 1.9418 |
| 1.6383 | 1.2097 | 600 | 1.9440 |
| 1.7073 | 1.6129 | 800 | 1.9435 |
| 1.6699 | 2.0161 | 1000 | 1.9428 |
| 1.7212 | 2.4194 | 1200 | 1.9445 |
| 1.7346 | 2.8226 | 1400 | 1.9443 |
| 1.7028 | 3.2258 | 1600 | 1.9448 |
| 1.7383 | 3.6290 | 1800 | 1.9450 |
Framework versions
- Transformers 4.40.1
- Pytorch 2.3.0
- Datasets 2.19.0
- Tokenizers 0.19.1