metadata
license: llama2
library_name: peft
tags:
- generated_from_trainer
base_model: codellama/CodeLlama-13b-hf
model-index:
- name: codeLlama13B-StaproCoder
results: []
codeLlama13B-StaproCoder
This model is a fine-tuned version of codellama/CodeLlama-13b-hf on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.1794
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: 0.0005
- 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 |
|---|---|---|---|
| 0.2404 | 0.05 | 100 | 0.2456 |
| 0.289 | 0.1 | 200 | 0.2349 |
| 0.236 | 0.15 | 300 | 0.2281 |
| 0.1928 | 0.2 | 400 | 0.2200 |
| 0.1999 | 0.25 | 500 | 0.2217 |
| 0.1697 | 0.3 | 600 | 0.2134 |
| 0.1825 | 0.35 | 700 | 0.2129 |
| 0.1904 | 0.4 | 800 | 0.2093 |
| 0.159 | 0.45 | 900 | 0.2054 |
| 0.1448 | 0.5 | 1000 | 0.1994 |
| 0.1993 | 0.55 | 1100 | 0.1954 |
| 0.2424 | 0.6 | 1200 | 0.1896 |
| 0.1385 | 0.65 | 1300 | 0.1891 |
| 0.1427 | 0.7 | 1400 | 0.1860 |
| 0.2021 | 0.75 | 1500 | 0.1836 |
| 0.1884 | 0.8 | 1600 | 0.1824 |
| 0.26 | 0.85 | 1700 | 0.1809 |
| 0.1641 | 0.9 | 1800 | 0.1799 |
| 0.1503 | 0.95 | 1900 | 0.1794 |
| 0.1613 | 1.0 | 2000 | 0.1794 |
Framework versions
- PEFT 0.9.0
- Transformers 4.39.0.dev0
- Pytorch 2.2.1+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2