This repository contains a fine-tuned Wav2Vec2 model that has been trained on a small TIMIT dataset for speech recognition tasks. The model achieved a Word Error Rate (WER) of 30, indicating promising performance on the given small dataset. You can utilize this model for further research and experimentation in the field of speech recognition. In the near future, I plan to expand the collection of fine-tuned Wav2Vec2 models to include various Indian languages. Stay tuned for updates and additions to this repository.