AIbase
EN
Home
Vakyansh Wav2vec2 Punjabi Pam 10
V

Vakyansh Wav2vec2 Punjabi Pam 10

Developed by Harveenchadha
A Punjabi automatic speech recognition model fine-tuned based on the CLSRIL-23 multilingual pretrained model, supporting 16kHz sampled speech input.
Speech Recognition
Transformers
 OtherOpen Source License:MIT#Punjabi speech recognition#Multilingual pretraining fine-tuning#No language model optimization
Downloads 96
Release Time : 3/2/2022

Model Overview

This model is an automatic speech recognition system optimized for Punjabi, based on the Wav2Vec2 architecture, suitable for converting Punjabi speech into text.

Model Features

Multilingual Pretraining Foundation
Fine-tuned based on the CLSRIL-23 multilingual pretrained model, with robust speech feature extraction capabilities.
Punjabi Optimization
Specifically optimized for Punjabi, suitable for speech recognition tasks in this language.
No Language Model Support
Current results do not use a language model; future improvements can be made by adding a language model to further enhance recognition accuracy.

Model Capabilities

Punjabi speech recognition
16kHz sampling rate speech processing

Use Cases

Speech-to-Text
Punjabi Speech Transcription
Convert Punjabi speech content into text format
Word Error Rate (WER) 33.17
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
English简体中文繁體中文にほんご
© 2025AIbase