|
|
--- |
|
|
library_name: transformers |
|
|
license: mit |
|
|
base_model: gpt2 |
|
|
tags: |
|
|
- generated_from_trainer |
|
|
model-index: |
|
|
- name: codeparrot-ds |
|
|
results: [] |
|
|
--- |
|
|
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
|
|
# codeparrot-ds |
|
|
|
|
|
This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on an unknown dataset. |
|
|
It achieves the following results on the evaluation set: |
|
|
- Loss: 1.0531 |
|
|
|
|
|
## 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: 32 |
|
|
- eval_batch_size: 32 |
|
|
- seed: 42 |
|
|
- gradient_accumulation_steps: 8 |
|
|
- total_train_batch_size: 256 |
|
|
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments |
|
|
- lr_scheduler_type: cosine |
|
|
- lr_scheduler_warmup_steps: 1000 |
|
|
- num_epochs: 1 |
|
|
- mixed_precision_training: Native AMP |
|
|
|
|
|
### Training results |
|
|
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|
|:-------------:|:------:|:-----:|:---------------:| |
|
|
| 2.5481 | 0.0766 | 5000 | 1.7253 | |
|
|
| 1.6683 | 0.1533 | 10000 | 1.5118 | |
|
|
| 1.5221 | 0.2299 | 15000 | 1.4112 | |
|
|
| 1.444 | 0.3065 | 20000 | 1.3444 | |
|
|
| 1.3844 | 0.3832 | 25000 | 1.2949 | |
|
|
| 1.3335 | 0.4598 | 30000 | 1.2463 | |
|
|
| 1.288 | 0.5365 | 35000 | 1.2043 | |
|
|
| 1.2418 | 0.6131 | 40000 | 1.1618 | |
|
|
| 1.201 | 0.6897 | 45000 | 1.1239 | |
|
|
| 1.1635 | 0.7664 | 50000 | 1.0908 | |
|
|
| 1.1366 | 0.8430 | 55000 | 1.0679 | |
|
|
| 1.1154 | 0.9196 | 60000 | 1.0559 | |
|
|
| 1.108 | 0.9963 | 65000 | 1.0531 | |
|
|
|
|
|
|
|
|
### Framework versions |
|
|
|
|
|
- Transformers 4.57.0 |
|
|
- Pytorch 2.8.0+cu128 |
|
|
- Datasets 4.2.0 |
|
|
- Tokenizers 0.22.1 |
|
|
|
