AIbase
EN
Home
ASCEND Dataset Model
A

ASCEND Dataset Model

Developed by GleamEyeBeast
A fine-tuned speech recognition model based on facebook/wav2vec2-xls-r-300m, trained on the ASCEND dataset
Speech Recognition
Transformers
Open Source License:Apache-2.0#Speech Recognition#Low Word Error Rate#Multilingual Support
Downloads 22
Release Time : 3/14/2022

Model Overview

This model is a fine-tuned model for Automatic Speech Recognition (ASR) tasks, capable of converting speech into text

Model Features

Fine-tuned from Large-scale Pretrained Model
Fine-tuned from the facebook/wav2vec2-xls-r-300m pretrained model, featuring powerful speech feature extraction capabilities
Optimized Recognition Performance
After 20 training epochs, achieved a Word Error Rate (WER) of 0.9540 on the validation set
Efficient Training Configuration
Utilized mixed-precision training and gradient accumulation techniques to optimize training efficiency

Model Capabilities

Speech to Text
Automatic Speech Recognition
Speech Content Transcription

Use Cases

Speech Transcription
Automatic Meeting Minutes Generation
Automatically convert meeting recordings into text transcripts
Approximately 95.4% accuracy
Voice Command Recognition
Recognize user voice commands and convert them into executable commands
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
EnglishįŽ€äŊ“中文įšéĢ”中文ãĢãģんご
Š 2025AIbase