18 projects
pymss
Python package for music source separation.
g2ps-torch
g2pS PyTorch backend (GPU + fp16)
g2ps-onnx
g2pS ONNX backend (GPU-capable)
g2ps
Mandarin polyphone disambiguation, numpy-only inference
pypinyin-g2pS
pypinyin adapter for g2pS (numpy/onnx/torch backends)
pymss-core
Core model, configuration, and checkpoint package for music source separation.
funasr-mlx
Native MLX speech inference (SenseVoice, Paraformer, Whisper, Qwen3-ASR, GLM, MoSS, Fun-ASR-Nano) with W4A16 quantization, an OpenAI-compatible server and pseudo-streaming, for Apple Silicon
gamemlx
Pure MLX inference package for GAME (vocal MIDI extraction) on Apple Silicon
gpt-sovits-cpufast
CPU-focused inference fork of GPT-SoVITS
moss-tts-lite
Minimal MOSS-TTS-v1.5 inference stack: 4 dependencies, fast CUDA-graph decode by default (bitwise-exact) + native tier + W4 GPTQ tiers, 8GB-card support
ops-registry-cli
Unified offline search across the CANN operator repositories (ops-blas/math/nn/cv/...): which operator, which repo, which arch
cann-recipes-cli
Search the CANN recipes repositories (infer/train/...): models, modules, ops, docs in one offline FTS5 index
pto-isa-cli
Query the PTO-ISA virtual instruction set documentation offline (FTS5, zh/en)
asc-devkit-docs-cli
Query the Ascend asc-devkit repository: FTS5 docs search plus source code lookup
windextts
Windows-priority, zero-JIT-compile, pure-torch accelerated inference for IndexTTS-2.5
npu-docs-cli
Portable CLI for querying offline AscendC, CCE, Runtime, and PTO-ISA documentation
usvc
Unified SVC framework: shared preprocessing, pluggable backends, single training/inference pipeline.
nemoasr2pytorch
Pure-PyTorch inference port of NeMo ASR (VAD + Parakeet-TDT + Nemotron RNNT) for Windows/WSL
