GPT-SoVITS

1 min voice data can also be used to train a good TTS model! (few shot voice cloning)

View on GitHub↗Upstream RVC-Boss/GPT-SoVITS↗

Builder

RVC-Boss

RVC-Boss • individual

Stars

56,229

Using upstream star count

Forks

6,143

Using upstream fork count

Open Issues

Activity Score

0/100

0 commits in 30d

Created

Jan 14, 2024

Project creation date

README Summary

GPT-SoVITS is a few-shot voice cloning system that can train high-quality text-to-speech models using just 1 minute of voice data. It combines GPT-based semantic modeling with SoVITS vocoding to achieve fast voice cloning with minimal training data. The system provides both training and inference capabilities with a user-friendly interface.

AI Dev Skills

Unmapped

Tags

Few-shot LearningVoice CloningText-to-Speech SynthesisAudio GenerationNeural VocodingSpeech ProcessingTransformer ArchitectureGenerative Pre-trained ModelsAudio Feature ExtractionMel-spectrogram ProcessingTransfer LearningPersonalized AICustom TTS Model TrainingSpeech SynthesisTextMedia ProductionAccessibility Voice ServicesSelf-hostedGenerative AIFew-shot Voice CloningGenerative Pre-trained TransformersAudioAudio Content GenerationEducationGamingVoice DubbingOn-premisePersonalized Voice SynthesisContent CreationAccessibilityEntertainmentPython

Recent Activity

Updated 2 months ago

7 Days

30 Days

90 Days

Quality

beta

Quality: medium
Maturity: beta

PM Skills

Scale & Reliability

Languages

Python100.0%

Timeline

Project created: Jan 14, 2024
Forked: Mar 22, 2026
Your last push: 2 months ago
Upstream last push: 2 months ago
Tracked since: Feb 9, 2026

Similar Repos

pgvector cosine similarity · $0

Loading…