wav2vec2-xls-r-300m-indonesian Open Source Model - Accurately Achieve Automatic Speech Recognition of Indonesian

Home

Wav2vec2 Xls R 300m Indonesian

Developed by Wikidepia

An automatic speech recognition model fine-tuned on Indonesian speech data based on Facebook's XLS-R-300M model

Speech Recognition

Transformers

OtherOpen Source License:Apache-2.0 #Indonesian speech recognition #Low word error rate #Multi-scenario adaptation

Downloads 4,486

Release Time : 3/2/2022

Model Overview

This model is an automatic speech recognition (ASR) model optimized for Indonesian, based on Facebook's wav2vec2-xls-r-300m architecture, fine-tuned on Common Voice 8.0 and MagicHub Indonesian conversational speech corpus.

Model Features

High-performance Indonesian recognition

Achieves a word error rate (WER) of 5.046% and a character error rate (CER) of 1.699% on the Common Voice 8 test set

Multi-dataset training

Combined training on Common Voice 8.0 and MagicHub Indonesian conversational speech corpus

Robustness evaluation

Performance evaluated on robust speech challenge datasets, demonstrating recognition capabilities under various conditions

Model Capabilities

Indonesian speech recognition

Speech-to-text

Automatic speech transcription

Use Cases

Speech transcription

Voice assistants

Used as the speech recognition component for Indonesian voice assistants

Meeting minutes

Automatically transcribe Indonesian meeting content

Accessibility technology

Real-time caption generation

Generate real-time captions for Indonesian video content

Property	Details
Model Type	Fine - tuned version of `facebook/wav2vec2-xls-r-300m`
Training Data	`mozilla-foundation/common_voice_8_0` and MagicHub Indonesian Conversational Speech Corpus
Tags	automatic - speech - recognition, hf - asr - leaderboard, id, mozilla - foundation/common_voice_8_0, robust - speech - event
Metrics	wer, cer

Task	Dataset	Metric	Value
Automatic Speech Recognition	Common Voice 8 (`mozilla - foundation/common_voice_8_0`, args: id)	Test WER	5.046
Automatic Speech Recognition	Common Voice 8 (`mozilla - foundation/common_voice_8_0`, args: id)	Test CER	1.699
Automatic Speech Recognition	Robust Speech Event - Dev Data (`speech - recognition - community - v2/dev_data`, args: id)	Test WER	41.31
Automatic Speech Recognition	Robust Speech Event - Test Data (`speech - recognition - community - v2/eval_data`, args: id)	Test WER	52.23

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご

Wav2vec2 Xls R 300m Indonesian

Model Overview

Model Features

Model Capabilities

Use Cases

🚀 Wav2Vec2 XLS-R-300M - Indonesian

📄 License

📚 Documentation

Model Information

Model Performance