Wav2vec2_10 July Open-source German Automatic Speech Recognition Model - Precise Recognition of German Speech Content

Wav2vec2 10july

Developed by sourabharsh

This is a German automatic speech recognition model based on the XLSR Wav2Vec2 architecture, fine-tuned on the Common Voice German dataset.

Downloads 24

Release Time : 3/2/2022

Model Overview

This model is used for German speech recognition tasks, capable of converting German speech to text with high accuracy.

High Accuracy

Achieves a word error rate of 10.55% and a character error rate of 2.81% on the Common Voice German test set.

Based on XLSR Wav2Vec2

Fine-tuned using the advanced XLSR Wav2Vec2 architecture, suitable for German speech recognition.

Open Source License

Licensed under Apache-2.0, allowing both commercial and research use.

German speech recognition

Speech-to-text

Automatic speech recognition

Speech Transcription

German Meeting Minutes

Automatically transcribe German meeting recordings into text

Accuracy up to 89.45% (based on WER 10.55%)

German Voice Assistant

Provide speech recognition capabilities for German voice assistants

Accessibility Applications

Real-time Caption Generation

Generate real-time captions for German video content

Property	Details
Model Type	Audio model for automatic speech recognition
Training Data	Common Voice
Evaluation Metrics	Word Error Rate (WER), Character Error Rate (CER)
License	Apache - 2.0

Task	Dataset	Metric	Value
Speech Recognition (automatic - speech - recognition)	Common Voice de	Test WER	10.55
Speech Recognition (automatic - speech - recognition)	Common Voice de	Test CER	2.81

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base