metadata
license: other
base_model: facebook/opt-350m
tags:
- trl
- sft
- generated_from_trainer
model-index:
- name: content
results: []
content
This model is a fine-tuned version of facebook/opt-350m on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 3.5972
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: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 4.0419 | 0.2 | 50 | 3.8081 |
| 3.816 | 0.4 | 100 | 3.7579 |
| 3.78 | 0.6 | 150 | 3.7016 |
| 3.753 | 0.8 | 200 | 3.6749 |
| 3.6787 | 1.0 | 250 | 3.6132 |
| 2.987 | 1.2 | 300 | 3.6374 |
| 3.0092 | 1.4 | 350 | 3.6043 |
| 3.0088 | 1.6 | 400 | 3.5676 |
| 2.945 | 1.8 | 450 | 3.5404 |
| 2.9204 | 2.0 | 500 | 3.5082 |
| 2.2216 | 2.2 | 550 | 3.6194 |
| 2.212 | 2.4 | 600 | 3.6117 |
| 2.198 | 2.6 | 650 | 3.6019 |
| 2.1787 | 2.8 | 700 | 3.5973 |
| 2.1878 | 3.0 | 750 | 3.5972 |
Framework versions
- Transformers 4.37.2
- Pytorch 2.1.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2