W

Wav2vec2 Xlsr Nepali

Developed by gagan3012
A Nepali automatic speech recognition model fine-tuned from facebook/wav2vec2-large-xlsr-53, trained on OpenSLR and Common Voice datasets, achieving a test WER of 5.97%.
Downloads 1,950
Release Time : 3/2/2022

Model Overview

This model is an optimized automatic speech recognition (ASR) model for Nepali, capable of converting Nepali speech into text.

Model Features

Low word error rate
Achieves a word error rate (WER) of 5.97% on the OpenSLR ne test set
No language model required
Can be used directly without additional language model support
Multi-dataset training
Fine-tuned using Common Voice and OpenSLR ne datasets

Model Capabilities

Nepali speech recognition
Speech-to-text

Use Cases

Speech transcription
Nepali speech transcription
Convert Nepali speech content into text
Accuracy reaches 94.03% (WER 5.97%)
Voice assistants
Nepali voice command recognition
For developing voice assistant applications supporting Nepali
Featured Recommended AI Models
┬й 2025AIbase