metadata
license: other
base_model: meta-llama/Meta-Llama-3-8B
tags:
- llama-factory
- full
- generated_from_trainer
model-index:
- name: C016_random_sample_llama3-8b-base_pretrain_20240504_181744
results: []
C016_random_sample_llama3-8b-base_pretrain_20240504_181744
This model is a fine-tuned version of /data/pro-align/progressalign/shared_storage/downloaded_models/llama3-8b-base on the C016_random_sample_data dataset. It achieves the following results on the evaluation set:
- Loss: 2.4196
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 |
|---|---|---|---|
| 2.5472 | 0.1947 | 200 | 2.5262 |
| 2.4431 | 0.3895 | 400 | 2.4733 |
| 2.4163 | 0.5842 | 600 | 2.4443 |
| 2.4462 | 0.7790 | 800 | 2.4281 |
| 2.4353 | 0.9737 | 1000 | 2.4196 |
| 2.2111 | 1.1685 | 1200 | 2.4290 |
| 2.2503 | 1.3632 | 1400 | 2.4281 |
| 2.258 | 1.5579 | 1600 | 2.4271 |
| 2.254 | 1.7527 | 1800 | 2.4266 |
| 2.2508 | 1.9474 | 2000 | 2.4266 |
| 2.2112 | 2.1422 | 2200 | 2.4287 |
| 2.2063 | 2.3369 | 2400 | 2.4293 |
| 2.2544 | 2.5316 | 2600 | 2.4291 |
| 2.2024 | 2.7264 | 2800 | 2.4289 |
| 2.2074 | 2.9211 | 3000 | 2.4288 |
| 2.2268 | 3.1159 | 3200 | 2.4297 |
| 2.1556 | 3.3106 | 3400 | 2.4294 |
| 2.1953 | 3.5054 | 3600 | 2.4296 |
| 2.2002 | 3.7001 | 3800 | 2.4294 |
| 2.2437 | 3.8948 | 4000 | 2.4291 |
Framework versions
- Transformers 4.40.1
- Pytorch 2.3.0
- Datasets 2.19.0
- Tokenizers 0.19.1