metadata
library_name: peft
license: llama3.2
base_model: meta-llama/Llama-3.2-3B
tags:
- trl
- sft
- generated_from_trainer
datasets:
- generator
model-index:
- name: llama3-3b-summarize-gpt4o-128k
results: []
llama3-3b-summarize-gpt4o-128k
This model is a fine-tuned version of meta-llama/Llama-3.2-3B on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 2.5189
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: 0.0002
- train_batch_size: 4
- eval_batch_size: 2
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- total_eval_batch_size: 16
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.5142 | 1.0 | 219 | 2.4458 |
| 1.4574 | 2.0 | 438 | 2.4565 |
| 1.4298 | 3.0 | 657 | 2.4537 |
| 1.3972 | 4.0 | 876 | 2.4622 |
| 1.3868 | 5.0 | 1095 | 2.4810 |
| 1.3692 | 6.0 | 1314 | 2.5001 |
| 1.3509 | 7.0 | 1533 | 2.5038 |
| 1.3413 | 8.0 | 1752 | 2.5123 |
| 1.3562 | 9.0 | 1971 | 2.5176 |
| 1.3538 | 9.9565 | 2180 | 2.5189 |
Framework versions
- PEFT 0.15.1
- Transformers 4.50.3
- Pytorch 2.6.0+cu124
- Datasets 3.5.0
- Tokenizers 0.21.1