Back to Models

Wav2vec2 Base 960H

Speech Recognition

Wav2Vec2 Base 960h - English ASR with CTC decoding

Integration

main.rs
use xybrid_sdk::{Xybrid, Envelope};

// Load the ASR model
let model = Xybrid::model("wav2vec2-base-960h").load()?;

// Transcribe audio to text
let result = model.run(&Envelope::audio(audio_bytes))?;

println!("{}", result.text.unwrap());

Details

Task
Speech Recognition
Family
Meta
Parameters
95M
Format
onnx
Quantization
fp32
Size
220.3 MB
Model ID
wav2vec2-base-960h