whisper-gujarati-base

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

Whisper Gujarati Base

This model is a fine-tuned version of openai/whisper-base on the Gujarati 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 repository.

Usage

In order to evaluate this model on an entire dataset, the evaluation codes available in the whisper-finetune repository can be used.

The same repository also provides the scripts for faster inference using whisper-jax.

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-gujarati-base", chunk_length_s=30, device=device)
>>> transcribe.model.config.forced_decoder_ids = transcribe.tokenizer.get_decoder_prompt_ids(language="gu", task="transcribe")

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

For faster inference of whisper models, the whisper-jax library can be used. Please follow the necessary installation steps as mentioned here, 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-gujarati-base", batch_size=16)
>>> transcribe.model.config.forced_decoder_ids = transcribe.tokenizer.get_decoder_prompt_ids(language="gu", task="transcribe")

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

Training and evaluation data

Training Data:

• ULCA ASR Corpus
• Microsoft Speech Corpus (Indian Languages)
• Google/Fleurs Train+Dev set
• OpenSLR

Evaluation Data:

• Microsoft Speech Corpus (Indian Languages) Test Set
• Google/Fleurs Test Set

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 3.3e-05
• train_batch_size: 80
• eval_batch_size: 88
• seed: 22
• optimizer: adamw_bnb_8bit
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 4000
• training_steps: 7225 (terminated upon convergence. Initially set to 21250 steps)
• mixed_precision_training: True

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.