Python wrappers for Kaldi Levenshtein's distance and alignment code.

Unofficial fairseq-free PyTorch implementation of UTMOS (v1, 2022), matching the original system.

Kernel sources for https://huggingface.co/kernels-community

High-Quality Voice Cloning TTS for 600+ Languages

Kimi-Audio, an open-source audio foundation model excelling in audio understanding, generation, and conversation

Perceptual Quality Estimator for speech and audio

Open source framework to vibecode and prototype voice agents with Gradium APIs

Machine Learning Systems

A lightweight deep learning training framework implemented from scratch in C++, featuring a PyTorch-style API.

Qwen3-ASR is an open-source series of ASR models developed by the Qwen team at Alibaba Cloud, supporting stable multilingual speech/music/song recognition, language detection and timestamp prediction.

Mount Hugging Face Buckets and repos as local filesystems. No download, no copy, no waiting.

[TACL'26] VoiceBench: Benchmarking LLM-Based Voice Assistants

[ICLR 2025] SOTA discrete acoustic codec models with 40/75 tokens per second for audio language modeling

[NeurIPS' 25] Benchmark for evaluating TTS models on complex prosodic, expressiveness, and linguistic challenges.

[Interspeech 2026] Official Implementation of "ALARM: Audio–Language Alignment for Reasoning Models"

Practical, Colab-friendly notebooks for fine-tuning and running audio AI models

Pure-PyTorch Parakeet TDT inference

Open Source Speech Language Model

Benchmarking Large Language Models using the Eleusis card game

Real-time text-to-speech with Qwen3-TTS

The most powerful local music generation model that outperforms almost all commercial alternatives, supporting Mac, AMD, Intel, and CUDA devices.

Riva Python client API and CLI utils

Soprano: Instant, Ultra-Realistic Text-to-Speech

Pyroomacoustics is a package for audio signal processing for indoor applications. It was developed as a fast prototyping platform for beamforming algorithms in indoor scenarios.

SoTA open-source TTS

A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)

The Hugging Face Course on Transformers for Audio

Omnilingual ASR Open-Source Multilingual SpeechRecognition for 1600+ Languages

SoulX-Podcast is an inference codebase by the Soul AI team for generating high-fidelity podcasts from text.