README.md

---
library_name: transformers
language: en
license: mit
datasets:
- stanfordnlp/imdb
base_model:
- openai-community/gpt2
---

# Model Card: GPT-2-IMDb

An in-domain GPT-2, pre-trained from scratch on the IMDb dataset text.

## Model Details

### Description

This model is based on the [GPT-2](https://huggingface.co/openai-community/gpt2)
architecture and was pre-trained from scratch (in-domain) using the text in IMDb dataset, excluding its test split.

- **Developed by:** [Cesar Gonzalez-Gutierrez](https://ceguel.es)
- **Funded by:** [ERC](https://erc.europa.eu)
- **Architecture:** GPT-2
- **Language:** English
- **License:** MIT
- **Base model:** [GPT-2](https://huggingface.co/openai-community/gpt2)

### Checkpoints

Intermediate checkpoints from the pre-training process are available and can be accessed using specific tags,
which correspond to training epochs and steps:

| Epoch | Step | Tags | |
|---|---|---|---|
| 1 | 703 | epoch-1 | step-703 |
| 5 | 3515 | epoch-5 | step-3515 |
| 10 | 7031 | epoch-10 | step-7031 |
| 20 | 14063 | epoch-20 | step-14063 |
| 30 | 21095 | epoch-30 | step-21095 |
| 40 | 28126 | epoch-40 | step-28126 |
| 50 | 35158 | epoch-50 | step-35158 |
| 60 | 42190 | epoch-60 | step-42190 |
| 70 | 49221 | epoch-70 | step-49221 |
| 80 | 56240 | epoch-80 | step-56240 |

To load a model from a specific intermediate checkpoint, use the `revision` parameter with the corresponding tag:
```python
from transformers import AutoModelForCausalLM

model = AutoModelForMaskedLM.from_pretrained("<model-name>", revision="<checkpoint-tag>")
```

### Sources

- **Paper:** [Information pending]

## Training Details

For more details on the training procedure, please refer to the base model's documentation:
[Training procedure](https://huggingface.co/openai-community/gpt2#training-procedure).

### Training Data

All texts from IMDb dataset, excluding the test partition.

#### Training Hyperparameters

- **Precision:** fp16
- **Batch size:** 8
- **Gradient accumulation steps:** 12

## Uses

For typical use cases and limitations, please refer to the base model's guidance: 
[Inteded uses & limitations](https://huggingface.co/openai-community/gpt2#intended-uses--limitations).

## Bias, Risks, and Limitations

This model inherits potential risks and limitations from the base model. Refer to:
[Limitations and bias](https://huggingface.co/openai-community/gpt2#limitations-and-bias).

## Environmental Impact

- **Hardware Type:** NVIDIA A100 PCIE 40GB
- **Runtime:** 7 h
- **Cluster Provider:** [Artemisa](https://artemisa.ific.uv.es/web/)
- **Compute Region:** EU
- **Carbon Emitted:** 1.08 kg CO2 eq.

## Citation

**BibTeX:**

[More Information Needed]