sh-lee-prml/HierSpeechpp

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1238 stars · 152 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Official PyTorch implementation of HierSpeech++, a fast zero-shot speech synthesizer for text-to-speech and voice conversion based on hierarchical variational inference. It includes pre-trained models, a text-to-vec framework, and a speech super-resolution module (16 kHz to 48 kHz).

Use cases

• synthesize speech from text in a zero-shot setting
• clone a voice from a short reference audio prompt
• convert one speaker's voice to another
• upsample 16 kHz speech to 48 kHz
• run a demo TTS model on HuggingFace Spaces

When to choose

• you need fast, robust zero-shot TTS without LLM-based autoregressive inference
• you want voice conversion with high speaker similarity
• you need human-level quality speech synthesis with pre-trained checkpoints

When to avoid

• you need a production-ready TTS service with an API
• you need training scripts that are not yet released
• you work outside Python/PyTorch environments

Facets

library · maturity active

tts speech-recognition machine-learning deep-learning speech-processing artificial-intelligence machine-learning python zero-shot-tts voice-conversion pytorch speech-synthesis variational-autoencoder speech-super-resolution gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="sh-lee-prml/HierSpeechpp")