---
license: mit
base_model: gpt2
tags:
- generated_from_trainer
model-index:
- name: codeparrot-ds-v2
  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-v2

This model is trained from scratch from [gpt2](https://huggingface.co/gpt2) on an python code dataset.
It achieves the following results on the evaluation set:
- Loss: 1.0617

## 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: Adam with betas=(0.9,0.999) and epsilon=1e-08
- 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 |
|:-------------:|:------:|:-----:|:---------------:|
| 2.5649        | 0.0766 | 5000  | 1.7424          |
| 1.6811        | 0.1533 | 10000 | 1.5239          |
| 1.5331        | 0.2299 | 15000 | 1.4213          |
| 1.4544        | 0.3065 | 20000 | 1.3544          |
| 1.3946        | 0.3832 | 25000 | 1.3049          |
| 1.3434        | 0.4598 | 30000 | 1.2571          |
| 1.2978        | 0.5365 | 35000 | 1.2146          |
| 1.2515        | 0.6131 | 40000 | 1.1707          |
| 1.2106        | 0.6897 | 45000 | 1.1335          |
| 1.1728        | 0.7664 | 50000 | 1.1002          |
| 1.1457        | 0.8430 | 55000 | 1.0769          |
| 1.1243        | 0.9196 | 60000 | 1.0646          |
| 1.1169        | 0.9963 | 65000 | 1.0617          |


### Framework versions

- Transformers 4.40.2
- Pytorch 2.2.1+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1