metadata
license: llama3
base_model: meta-llama/Meta-Llama-3-8B-Instruct
tags:
- trl
- sft
- generated_from_trainer
model-index:
- name: Summary_L3_450steps_1e7rate_SFT
results: []
Summary_L3_450steps_1e7rate_SFT
This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.8574
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: 1e-07
- train_batch_size: 2
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 100
- training_steps: 450
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 2.1137 | 0.2 | 50 | 2.1001 |
| 2.0888 | 0.4 | 100 | 2.0502 |
| 1.9952 | 0.6 | 150 | 1.9737 |
| 1.9325 | 0.8 | 200 | 1.9171 |
| 1.8839 | 1.0 | 250 | 1.8812 |
| 1.8548 | 1.2 | 300 | 1.8635 |
| 1.906 | 1.4 | 350 | 1.8578 |
| 1.8425 | 1.6 | 400 | 1.8573 |
| 1.8518 | 1.8 | 450 | 1.8574 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.0.0+cu117
- Datasets 2.19.2
- Tokenizers 0.19.1