metadata
license: llama3
base_model: meta-llama/Meta-Llama-3-8B-Instruct
tags:
- trl
- sft
- generated_from_trainer
model-index:
- name: Summary_L3_1000steps_1e5rate_SFT
results: []
Summary_L3_1000steps_1e5rate_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: 0.7019
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-05
- 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: 1000
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.7518 | 0.2 | 50 | 0.6955 |
| 0.7657 | 0.4 | 100 | 0.7030 |
| 0.7138 | 0.6 | 150 | 0.6648 |
| 0.6394 | 0.8 | 200 | 0.6382 |
| 0.5783 | 1.0 | 250 | 0.6033 |
| 0.4656 | 1.2 | 300 | 0.5986 |
| 0.4742 | 1.4 | 350 | 0.5881 |
| 0.417 | 1.6 | 400 | 0.5612 |
| 0.3351 | 1.8 | 450 | 0.5599 |
| 0.4481 | 2.0 | 500 | 0.5488 |
| 0.185 | 2.2 | 550 | 0.6115 |
| 0.1621 | 2.4 | 600 | 0.6201 |
| 0.1701 | 2.6 | 650 | 0.6293 |
| 0.1325 | 2.8 | 700 | 0.6154 |
| 0.166 | 3.0 | 750 | 0.6194 |
| 0.0347 | 3.2 | 800 | 0.6931 |
| 0.0422 | 3.4 | 850 | 0.7013 |
| 0.0449 | 3.6 | 900 | 0.7014 |
| 0.0358 | 3.8 | 950 | 0.7020 |
| 0.0422 | 4.0 | 1000 | 0.7019 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.0.0+cu117
- Datasets 2.19.2
- Tokenizers 0.19.1