metadata
license: other
library_name: peft
tags:
- trl
- sft
- generated_from_trainer
base_model: deepseek-ai/deepseek-coder-1.3b-instruct
model-index:
- name: results
results: []
results
This model is a fine-tuned version of deepseek-ai/deepseek-coder-1.3b-instruct on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.1804
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: 2.5e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.03
- lr_scheduler_warmup_steps: 50
- training_steps: 500
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.8004 | 0.01 | 25 | 1.8163 |
| 1.8977 | 0.01 | 50 | 1.6053 |
| 1.5414 | 0.02 | 75 | 1.4712 |
| 0.9012 | 0.02 | 100 | 1.3793 |
| 1.2741 | 0.03 | 125 | 1.3288 |
| 0.7702 | 0.04 | 150 | 1.2922 |
| 1.3633 | 0.04 | 175 | 1.2580 |
| 0.5385 | 0.05 | 200 | 1.2341 |
| 1.3172 | 0.06 | 225 | 1.2157 |
| 0.5901 | 0.06 | 250 | 1.2088 |
| 1.2612 | 0.07 | 275 | 1.1989 |
| 0.6468 | 0.07 | 300 | 1.1937 |
| 1.2991 | 0.08 | 325 | 1.1891 |
| 0.5557 | 0.09 | 350 | 1.1858 |
| 1.148 | 0.09 | 375 | 1.1838 |
| 0.634 | 0.1 | 400 | 1.1823 |
| 1.2453 | 0.1 | 425 | 1.1812 |
| 0.5921 | 0.11 | 450 | 1.1807 |
| 1.0155 | 0.12 | 475 | 1.1805 |
| 0.6354 | 0.12 | 500 | 1.1804 |
Framework versions
- PEFT 0.7.2.dev0
- Transformers 4.37.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.0