metadata
library_name: peft
license: llama3.2
base_model: meta-llama/Llama-3.2-3B-Instruct
tags:
- generated_from_trainer
model-index:
- name: Meta-Llama3.2-3B-Instruct-009
results: []
Meta-Llama3.2-3B-Instruct-009
This model is a fine-tuned version of meta-llama/Llama-3.2-3B-Instruct on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.4892
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: 5e-05
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.8207 | 1.0 | 83 | 0.7515 |
| 0.7171 | 2.0 | 166 | 0.6729 |
| 0.6519 | 3.0 | 249 | 0.6164 |
| 0.6042 | 4.0 | 332 | 0.5788 |
| 0.5645 | 5.0 | 415 | 0.5496 |
| 0.5229 | 6.0 | 498 | 0.5265 |
| 0.5125 | 7.0 | 581 | 0.5096 |
| 0.4959 | 8.0 | 664 | 0.4981 |
| 0.5099 | 9.0 | 747 | 0.4913 |
| 0.5066 | 9.8848 | 820 | 0.4892 |
Framework versions
- PEFT 0.15.1
- Transformers 4.51.3
- Pytorch 2.6.0+cu126
- Datasets 3.5.0
- Tokenizers 0.21.1