AIbase
EN
Home
Trillsson3 Ft Keyword Spotting 12
T

Trillsson3 Ft Keyword Spotting 12

Developed by vumichien
An audio classification model based on the TRillsson3 architecture, fine-tuned for keyword detection tasks, achieving 91.5% accuracy on the evaluation set.
Audio Classification
Transformers
#Keyword detection#High-accuracy audio classification#Non-semantic speech processing
Downloads 32
Release Time : 10/23/2022

Model Overview

This model is a fine-tuned version of TRillsson3 on the superb dataset, primarily used for keyword detection tasks.

Model Features

High accuracy
Achieves 91.5% accuracy on the evaluation set.
Based on TRillsson3
Fine-tuned from the non-semantic speech representation model TRillsson3.
Efficient training
Utilizes Adam optimizer and linear learning rate scheduler for efficient training.

Model Capabilities

Audio classification
Keyword detection
Speech feature extraction

Use Cases

Voice interaction
Wake word detection
Detects device wake words.
High-accuracy keyword detection.
Voice command recognition
Recognizes simple voice 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