metadata
license: llama2
base_model: codellama/CodeLlama-34b-Instruct-hf
tags:
- trl
- sft
- unsloth
- generated_from_trainer
datasets:
- generator
model-index:
- name: CodeLlama-34b-Instruct-Shader-hf
results: []
CodeLlama-34b-Instruct-Shader-hf
This model is a fine-tuned version of codellama/CodeLlama-34b-Instruct-hf on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 0.7792
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: 0.0025
- train_batch_size: 4
- eval_batch_size: 1
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: polynomial
- num_epochs: 2
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.3476 | 0.25 | 1000 | 1.3561 |
| 1.3274 | 0.49 | 2000 | 1.1683 |
| 1.1964 | 0.74 | 3000 | 1.0129 |
| 0.8117 | 0.98 | 4000 | 0.9106 |
| 0.5179 | 1.23 | 5000 | 0.8560 |
| 0.7726 | 1.48 | 6000 | 0.8096 |
| 0.7396 | 1.72 | 7000 | 0.7855 |
| 0.6125 | 1.97 | 8000 | 0.7792 |
Framework versions
- Transformers 4.38.2
- Pytorch 2.2.1
- Datasets 2.18.0
- Tokenizers 0.15.2