Back: Catalog

DiffSinger

This project is the official PyTorch application that performs song voice synthesis (SVS) and text reading (TTS) operations using a shallow diffusion mechanism. It is based on the statement presented at the AAAI-2022 conference and includes acceleration support with the PNDM plugin. The system allows for inference from raw inputs and is compatible with vocoder structures with different data sets.

DI

DiffSinger

New64

MoonInTheRiver · Voice & Text-to-Speech

STARS4.9K
FORKS832
COMMITS0
CONTRIBUTORS0
VERSIONpretrain-model
UPDATED3 mo ago
PythonPyTorchSes SenteziAIMIT

Common commercial products in this category: ElevenLabs, Amazon Polly, Google Cloud Text-to-Speech

What can it do?
  • Using text data, you run voice synthesis (TTS) on the pipeline.
  • With the lyrics, you run DiffSinger models that synthesize song sound from MIDI or F0 curtain inputs (SVS).
  • You accelerate diffusion processes with the help of the PNDM plugin.
User reviews
  • • NeutralGitHub Issue

    Hello, I have issue as I try to use another english dataset. And I'm wondering why Inference from packed test set can wo

    “Hello, I have issue as I try to use another english dataset. And I'm wondering why Inference from packed test set can work (`CUDA_VISIBLE_DEVICES=0 python tasks/run.py --config usr/configs/midi/e2e/opencpop/ds100_adj_rel.yaml --exp_name $MY_DS_EXP_NAME --reset --infer`) but…”

  • • NeutralGitHub Issue

    Inference with unseen songs

    “Hi. Since the DiffSinger(PopCS) needs ground-truth f0 information at inference, is it possible to synthesize an unseen song (with phoneme labels, phoneme duration and notes provided) using the DIffSinger(PopCS) model?”

GitHub

Verified against README · Oct 25, 2026 · source

My note

Alternative pages

ElevenLabs alternativesAmazon Polly alternativesGoogle Cloud Text-to-Speech alternatives

More in: Voice & Text-to-Speech