Whisper Kannada Tiny

🚀 Whisper Kannada Tiny

This model is a fine - tuned version of [openai/whisper - tiny](https://huggingface.co/openai/whisper - tiny) on Kannada data from multiple publicly available ASR corpuses. It was fine - tuned as part of the Whisper fine - tuning sprint, offering a powerful solution for Kannada automatic speech recognition.

🚀 Quick Start

This model is a fine - tuned version of [openai/whisper - tiny](https://huggingface.co/openai/whisper - tiny) on the Kannada data available from multiple publicly available ASR corpuses. It has been fine - tuned as a part of the Whisper fine - tuning sprint.

NOTE: The code used to train this model is available for re - use in the [whisper - finetune](https://github.com/vasistalodagala/whisper - finetune) repository.

✨ Features

• Fine - tuned on Kannada data from multiple ASR corpuses.
• Code for training is available for re - use.
• Can be evaluated on an entire dataset using provided evaluation codes.
• Supports faster inference with whisper - jax.

📦 Installation

No specific installation steps are provided in the original document.

💻 Usage Examples

Basic Usage

In order to infer a single audio file using this model, the following code snippet can be used:

>>> 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-kannada-tiny", chunk_length_s=30, device=device)
>>> transcribe.model.config.forced_decoder_ids = transcribe.tokenizer.get_decoder_prompt_ids(language="kn", task="transcribe")

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

Advanced Usage

For faster inference of whisper models, the [whisper - jax](https://github.com/sanchit - gandhi/whisper - jax) library can be used. Please follow the necessary installation steps as mentioned [here](https://github.com/vasistalodagala/whisper - finetune#faster - evaluation - with - whisper - jax), before using the following code snippet:

>>> 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-kannada-tiny", batch_size=16)
>>> transcribe.model.config.forced_decoder_ids = transcribe.tokenizer.get_decoder_prompt_ids(language="kn", task="transcribe")

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

📚 Documentation

Training and evaluation data

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e - 05
• train_batch_size: 88
• eval_batch_size: 88
• seed: 22
• optimizer: adamw_bnb_8bit
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 10000
• training_steps: 15008 (terminated upon convergence. Initially set to 51570 steps)
• mixed_precision_training: True

🔧 Technical Details

The model is a fine - tuned version of [openai/whisper - tiny](https://huggingface.co/openai/whisper - tiny) on Kannada data. It was trained with specific hyperparameters and evaluated on multiple datasets as described above.

📄 License

This model is released under the Apache 2.0 license.

Acknowledgement

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