Spaces:
Sleeping
Sleeping
Data Folder
This data folder contains datasets and resources used for the Mini-Transformer project.
Structure
- corpus/: Text corpora for training and evaluation.
- raw/: Original, unprocessed datasets.
- processed/: Cleaned and preprocessed data ready for model training.
- README.md: Documentation for the data folder.
Usage
Place your datasets in the appropriate subfolders. Ensure data is properly formatted before use.
Used Datasets:
- small dataset: "ahazeemi/iwslt14-en-fr"
- medium dataset: "Helsinki-NLP/europarl", "en-fr"
- large dataset: "wmt/wmt14", "fr-en"
Notes
- Do not commit sensitive or proprietary data.
- Update this README if the folder structure changes.
- For questions about data formats, refer to the main project documentation.