wav2vec2-base-timit-demo-colab-32-epochs30 Open-Source Speech Recognition Model

Home

Wav2vec2 Base Timit Demo Colab 32 Epochs30

Developed by ying-tina

A speech recognition model fine-tuned from facebook/wav2vec2-base, trained for 30 epochs on the TIMIT dataset

Speech Recognition

Transformers

Open Source License:Apache-2.0 #Speech Recognition #Low Word Error Rate #TIMIT Dataset

Downloads 22

Release Time : 3/2/2022

Model Overview

This model is a fine-tuned version of wav2vec2-base for English speech recognition, suitable for Automatic Speech Recognition (ASR) tasks

Model Features

Efficient Fine-tuning

Efficiently fine-tuned from the wav2vec2-base model with only 30 training epochs

Low Word Error Rate

Achieves a Word Error Rate (WER) of 0.3434 on the evaluation set

Lightweight

Based on the wav2vec2-base architecture, relatively lightweight and suitable for deployment

Model Capabilities

English speech recognition

Audio-to-text conversion

Use Cases

Speech Transcription

Meeting Minutes

Automatically transcribe English meeting recordings into text

Word Error Rate around 34.34%

Voice Notes

Convert English voice notes to text

Training Loss	Epoch	Step	Validation Loss	Wer
3.5243	4.0	500	1.4532	0.9540
0.6178	8.0	1000	0.5490	0.4627
0.223	12.0	1500	0.4513	0.3881
0.1299	16.0	2000	0.4573	0.3698
0.0875	20.0	2500	0.4950	0.3637
0.0613	24.0	3000	0.4327	0.3479
0.0478	28.0	3500	0.4615	0.3434

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 Base Timit Demo Colab 32 Epochs30

Model Overview

Model Features

Model Capabilities

Use Cases

🚀 wav2vec2-base-timit-demo-colab-32-epochs30

🚀 Quick Start

📚 Documentation

Model description

Intended uses & limitations

Training and evaluation data

🔧 Technical Details

Training procedure

Training hyperparameters

Training results

Framework versions

📄 License