Whisper Hindi Small

🚀 Whisper Hindi Small

This model is a fine - tuned version of openai/whisper-small on Hindi data from multiple public ASR corpuses, contributing to the Whisper fine - tuning sprint.

🚀 Quick Start

This model is a fine - tuned variant of openai/whisper-small based on Hindi data from multiple publicly accessible ASR corpuses. It was developed as part of the Whisper fine - tuning sprint.

NOTE: The code for training this model can be reused from the whisper-finetune repository.

✨ Features

• Fine - tuned on multiple Hindi ASR corpuses.
• Code for training and evaluation is publicly available for re - use.
• Supports faster inference with whisper - jax.

📦 Installation

The installation and evaluation codes can be found in the whisper-finetune repository.

💻 Usage Examples

Basic Usage

To infer a single audio file using this model, you can use the following code snippet:

>>> import torch
>>> from transformers import pipeline

>>> # path to the audio file to be transcribed
>>> audio = "/path/to/audio.format"
>>> device = "cuda:0" if torch.cuda.is_available() else "cpu"

>>> transcribe = pipeline(task="automatic-speech-recognition", model="vasista22/whisper-hindi-small", chunk_length_s=30, device=device)
>>> transcribe.model.config.forced_decoder_ids = transcribe.tokenizer.get_decoder_prompt_ids(language="hi", task="transcribe")

>>> print('Transcription: ', transcribe(audio)["text"])

Advanced Usage

For faster inference using whisper - jax, follow the installation steps here and then use the following code:

>>> import jax.numpy as jnp
>>> from whisper_jax import FlaxWhisperForConditionalGeneration, FlaxWhisperPipline

>>> # path to the audio file to be transcribed
>>> audio = "/path/to/audio.format"

>>> transcribe = FlaxWhisperPipline("vasista22/whisper-hindi-small", batch_size=16)
>>> transcribe.model.config.forced_decoder_ids = transcribe.tokenizer.get_decoder_prompt_ids(language="hi", task="transcribe")

>>> print('Transcription: ', transcribe(audio)["text"])

📚 Documentation

Training and evaluation data

• Training Data:
  • GramVaani ASR Corpus
  • ULCA ASR Corpus
  • Shrutilipi ASR Corpus
  • Google/Fleurs Train+Dev set
• Evaluation Data:
  • GramVaani ASR Corpus Test Set
  • Google/Fleurs Test Set

Training hyperparameters

The following hyperparameters were used during training:

🔧 Technical Details

This model is a fine - tuned version of openai/whisper-small on Hindi data. The fine - tuning was carried out as part of the Whisper fine - tuning sprint.

📄 License

This model is licensed under the Apache - 2.0 license.

Acknowledgement

This work was conducted at Speech Lab, IIT Madras. The compute resources were funded by the "Bhashini: National Language translation Mission" project of the Ministry of Electronics and Information Technology (MeitY), Government of India.