whatsapp-to-train

Turn your whatsapp conversations into training data for a large language model

Usage

Export your whatsapp conversation from the app, and put it in the data directory, with the filename _chat.txt (this is the default export).
Run python3 src/clean.py to clean the data and save it to data/cleaned.txt.
Then, to make your huggingface dataset, run python3 src/make_dataset.py. This will save the dataset to data/chat_dataset.