metadata
library_name: peft
tags:
- generated_from_trainer
base_model: codellama/CodeLlama-13b-Instruct-hf
model-index:
- name: peft-lora-CodeLlama-13b-huggingface-copilot
results: []
peft-lora-CodeLlama-13b-huggingface-copilot
This model is a fine-tuned version of codellama/CodeLlama-13b-Instruct-hf on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.4166
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: 0.0003
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- training_steps: 2000
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.5101 | 0.05 | 100 | 0.5661 |
| 0.5131 | 0.1 | 200 | 0.5247 |
| 0.5291 | 0.15 | 300 | 0.5053 |
| 0.4597 | 0.2 | 400 | 0.4944 |
| 0.4241 | 0.25 | 500 | 0.4790 |
| 0.4306 | 0.3 | 600 | 0.4694 |
| 0.4228 | 0.35 | 700 | 0.4618 |
| 0.3666 | 0.4 | 800 | 0.4527 |
| 0.4728 | 0.45 | 900 | 0.4438 |
| 0.3866 | 0.5 | 1000 | 0.4371 |
| 0.3695 | 0.55 | 1100 | 0.4326 |
| 0.3303 | 0.6 | 1200 | 0.4295 |
| 0.2989 | 0.65 | 1300 | 0.4258 |
| 0.3495 | 0.7 | 1400 | 0.4223 |
| 0.3437 | 0.75 | 1500 | 0.4203 |
| 0.302 | 0.8 | 1600 | 0.4179 |
| 0.3655 | 0.85 | 1700 | 0.4164 |
| 0.3213 | 0.9 | 1800 | 0.4162 |
| 0.2654 | 0.95 | 1900 | 0.4164 |
| 0.274 | 1.0 | 2000 | 0.4166 |
Framework versions
- PEFT 0.10.1.dev0
- Transformers 4.40.0.dev0
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2