metadata
license: llama2
base_model: codellama/CodeLlama-7b-hf
tags:
- trl
- sft
- generated_from_trainer
model-index:
- name: lora-fine-tuning-llama2-SQL-lora-codellama
results: []
lora-fine-tuning-llama2-SQL-lora-codellama
This model is a fine-tuned version of codellama/CodeLlama-7b-hf on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.5720
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: 4
- seed: 42
- gradient_accumulation_steps: 3
- total_train_batch_size: 12
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.03
- num_epochs: 1
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 2.7803 | 0.06 | 5 | 2.5059 |
| 1.2647 | 0.12 | 10 | 1.1731 |
| 0.8026 | 0.18 | 15 | 0.8834 |
| 0.6482 | 0.24 | 20 | 0.8281 |
| 0.8146 | 0.3 | 25 | 0.7858 |
| 0.7458 | 0.36 | 30 | 0.7275 |
| 0.5377 | 0.42 | 35 | 0.6520 |
| 0.5659 | 0.48 | 40 | 0.6560 |
| 0.6104 | 0.54 | 45 | 0.6101 |
| 0.6253 | 0.6 | 50 | 0.6024 |
| 0.4878 | 0.66 | 55 | 0.5891 |
| 0.4777 | 0.72 | 60 | 0.5830 |
| 0.634 | 0.78 | 65 | 0.5831 |
| 0.5562 | 0.84 | 70 | 0.5771 |
| 0.4696 | 0.9 | 75 | 0.5734 |
| 0.4193 | 0.96 | 80 | 0.5720 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.1