Hello 2b
H
Hello 2b
Developed by patrickvonplaten
This model is an automatic speech recognition (ASR) model fine-tuned on the COMMON_VOICE-TR Turkish dataset based on facebook/wav2vec2-xls-r-2b
Downloads 27
Release Time : 3/2/2022
Model Overview
A fine-tuned model for Turkish speech recognition based on the wav2vec2-xls-r-2b architecture
Model Features
Large-scale pre-training foundation
Fine-tuned based on facebook's wav2vec2-xls-r-2b model, with powerful speech feature extraction capabilities
Turkish language optimization
Specially fine-tuned for Turkish, adapting to the phonetic characteristics of Turkish
Multi-GPU training
Utilizes multi-GPU distributed training to improve training efficiency
Model Capabilities
Turkish speech-to-text
Continuous speech recognition
Large-scale speech data processing
Use Cases
Speech transcription
Turkish speech transcription
Convert Turkish speech content into text
Achieved a word error rate of 0.9531 on the COMMON_VOICE-TR test set
Voice assistants
Turkish voice command recognition
Used as the front-end speech recognition module for Turkish voice assistants
Featured Recommended AI Models