Tomáš Nekvinda

MOSS-TTS. MOSS‑TTS Family is an open‑source speech and sound generation model family from MOSI.AI and the OpenMOSS team. It is designed for high‑fidelity, high‑expressiveness, and complex real‑world scenarios, covering stable long‑form speech, multi‑speaker dialogue, voice/character design, environmental sound effects, and real‑time streaming TTS.

adversarial-robustness-toolbox. Adversarial Robustness Toolbox (ART) - Python Library for Machine Learning Security - Evasion, Poisoning, Extraction, Inference - Red and Blue Teams

OmniVoice. High-Quality Voice Cloning TTS for 600+ Languages

PeriodWave. The official Implementation of PeriodWave and PeriodWave-Turbo

csm. A Conversational Speech Generation Model

FlashLabs-Chroma. Worlds first open-source real-time end-to-end spoken dialogue model with personalized voice cloning.

Qwen3-TTS. Qwen3-TTS is an open-source series of TTS models developed by the Qwen team at Alibaba Cloud, supporting stable, expressive, and streaming speech generation, free-form voice design, and vivid voice cloning.

Respiro-en. Official implementation of paper: Frame-Wise Breath Detection with Self-Training: An Exploration of Enhancing Breath Naturalness in Text-to-Speech

LinaCodec. A highly compressive and high-quality neural audio codec for speech models.

cro-dl. 🐍 CRo-DL (Český Rozhlas Downloader) - MůjRozhlas.cz 📻 offline

tts. Inworld TTS

VoxCPM. VoxCPM2: Tokenizer-Free TTS for Multilingual Speech Generation, Creative Voice Design, and True-to-Life Cloning

X-Codec-2.0. Codec for paper: LLaSA: Scaling Train-time and Inference-time Compute for LLaMA-based Speech Synthesis

chatterbox. SoTA open-source TTS

Sidon. Training code and dataset cleasing with Sidon

adversarial-attacks-pytorch. PyTorch implementation of adversarial attacks [torchattacks]

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

S3Tokenizer. Reverse Engineering of Supervised Semantic Speech Tokenizer (S3Tokenizer) proposed in CosyVoice

delayed-streams-modeling. Kyutai's Speech-To-Text and Text-To-Speech models based on the Delayed Streams Modeling framework.

BlaGPT. Experimental playground for benchmarking language model (LM) architectures, layers, and tricks on smaller datasets. Designed for flexible experimentation and exploration.

tidy-tunes. Tidy Tunes is an easy-to-use pipeline for mining high-quality audio data for speech generation models. To do so, it chains multiple open source models while minimizing dependencies.

pytorch-nips2017-attack-example. A PyTorch baseline attack example for the NIPS 2017 adversarial competition

ten-vad. Voice Activity Detector (VAD) : low-latency, high-performance and lightweight

parsonaut. Python

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

dia. A TTS model capable of generating ultra-realistic dialogue in one pass.

Orpheus-TTS. Towards Human-Sounding Speech

Zonos. Zonos-v0.1 is a leading open-weight text-to-speech model trained on more than 200k hours of varied multilingual speech, delivering expressiveness and quality on par with—or even surpassing—top TTS providers.

Speech-Articulatory-Coding. Jupyter Notebook

audiobox-aesthetics. Unified automatic quality assessment for speech, music, and sound.

open-r1. Fully open reproduction of DeepSeek-R1

Matcha-TTS. [ICASSP 2024] 🍵 Matcha-TTS: A fast TTS architecture with conditional flow matching

cuml. cuML - RAPIDS Machine Learning Library

awesome-maps. There is more than google: A collection of great online maps 🌍🗺🌎

biblically-accurate-sampler. llm sampler that only allows words that are in the bible

stable-ts. Transcription, forced alignment, and audio indexing with OpenAI's Whisper

audiotools. Object-oriented handling of audio data, with GPU-powered augmentations, and more.

Fast-GeCo. Source code and demo for INTERSPEECH 2024 paper: Noise-robust Speech Separation with Fast Generative Correction

gpt-bert. Official implementation of "GPT or BERT: why not both?"

Transformer-Explainability. [CVPR 2021] Official PyTorch implementation for Transformer Interpretability Beyond Attention Visualization, a novel method to visualize classifications by Transformer based networks.

seed-vc. zero-shot voice conversion & singing voice conversion, with real-time support

F5-TTS. Official code for "F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matching"

Adafruit_TSL2561. Unified sensor driver for Adafruit's TSL2561 breakouts

wavefit-pytorch. PyTorch implementation of WaveFit [2022, Google] which is one of SOTA lightweight/fast speech vocoders.

moshi. Moshi is a speech-text foundation model and full-duplex spoken dialogue framework. It uses Mimi, a state-of-the-art streaming neural audio codec.

demucs. Code for the paper Hybrid Spectrogram and Waveform Source Separation

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

utmos. A toolkit to calculate speech audio quality. Not affiliated with the original authors

SpeechTokenizer. This is the code for the SpeechTokenizer presented in the SpeechTokenizer: Unified Speech Tokenizer for Speech Language Models. Samples are presented on

FreeV. [InterSpeech 24] FreeV: Free Lunch For Vocoders Through Pseudo Inversed Mel Filter

torchdiffeq. Differentiable ODE solvers with full GPU support and O(1)-memory backpropagation.

LibriTTS-P. LibriTTS-P: A Corpus with Speaking Style and Speaker Identity Prompts for Text-to-Speech and Style Captioning

friendly-stable-audio-tools. Refactored / updated version of `stable-audio-tools` which is an open-source code for audio/music generative models originally by Stability AI.

astral. ☀️ Go calculations for the position of the sun and moon.

awesome-discrete-diffusion-models. A curated list for awesome discrete diffusion models resources.

tts-arabic-pytorch. 🎙️ Arabic TTS models (Tacotron2, FastPitch)

arabic_vocalizer. Arabic deep-learning based diacritization models (Shakkala, Shakkelha) in the ONNX format.

snac. Multi-Scale Neural Audio Codec (SNAC) compresses audio into discrete codes at a low bitrate

dataspeech. Python

parler-tts. Inference and training library for high-quality TTS models.

dasp-pytorch. Differentiable audio signal processors in PyTorch

torch-pitch-shift. Pitch-shift audio clips quickly with PyTorch (CUDA supported)! Additional utilities for searching efficient transformations are included.

metavoice-src. Foundational model for human-like, expressive TTS

DeepFilterNet. Noise supression using deep filtering

descript-audio-codec. State-of-the-art audio codec with 90x compression factor. Supports 44.1kHz, 24kHz, and 16kHz mono/stereo audio.

gpt-fast. Simple and efficient pytorch-native transformer text generation in <1000 LOC of python.

audiocraft. Audiocraft is a library for audio processing and generation with deep learning. It features the state-of-the-art EnCodec audio compressor / tokenizer, along with MusicGen, a simple and controllable music generation LM with textual and melodic conditioning.

pyctcdecode. A fast and lightweight python-based CTC beam search decoder for speech recognition.

ControlNet. Let us control diffusion models!

mamba. Mamba SSM architecture

ttts. Train the next generation of TTS systems.

vocos. Vocos: Closing the gap between time-domain and Fourier-based neural vocoders for high-quality audio synthesis

FreeVC. FreeVC: Towards High-Quality Text-Free One-Shot Voice Conversion

national-anthems-clustering. Jupyter Notebook

ProphetNet. A research project for natural language generation, containing the official implementations by MSRA NLC team.

diffusion. Denoising Diffusion Probabilistic Models

vocode-core. 🤖 Build voice-based LLM agents. Modular + open source.

laughter-detection. Python

BigVGAN. BigVGAN with Neural Source-Filter

StarGANv2-VC. StarGANv2-VC: A Diverse, Unsupervised, Non-parallel Framework for Natural-Sounding Voice Conversion

AutoVocoder. Autovocoder: Fast Waveform Generation from a Learned Speech Representation using Differentiable Digital Signal Processing

Knowledge-Distillation-Toolkit. :no_entry: [DEPRECATED] A knowledge distillation toolkit based on PyTorch and PyTorch Lightning.

unilm. Large-scale Self-supervised Pre-training Across Tasks, Languages, and Modalities

panns_inference. Python

fairseq. Facebook AI Research Sequence-to-Sequence Toolkit written in Python.

CLAP. Contrastive Language-Audio Pretraining

tortoise-tts-fast. Fast TorToiSe inference (5x or your money back!)

so-vits-svc-fork. so-vits-svc fork with realtime support, improved interface and more features.

encodec. State-of-the-art deep learning based audio codec supporting both mono 24 kHz audio and stereo 48 kHz audio.

consistency_models. Official repo for consistency models.

Prompt-Engineering-Guide. 🐙 Guides, papers, lessons, notebooks and resources for prompt engineering, context engineering, RAG, and AI Agents.

Large-Audio-Models. Keep track of big models in audio domain, including speech, singing, music etc.

univnet. Unofficial PyTorch Implementation of UnivNet Vocoder (https://arxiv.org/abs/2106.07889)

torch-nansypp. NANSY++: Unified Voice Synthesis with Neural Analysis and Synthesis

SimpleParsing. Simple, Elegant, Typed Argument Parsing with argparse

libebur128. A library implementing the EBU R128 loudness standard.

simple-asgan. Training code and trained checkpoints for ASGAN.

prompt-to-prompt. Jupyter Notebook

HarmoF0. Python

wav2lip-hq. Extension of Wav2Lip repository for processing high-quality videos.

whisper. Robust Speech Recognition via Large-Scale Weak Supervision

denoiser. Real Time Speech Enhancement in the Waveform Domain (Interspeech 2020)We provide a PyTorch implementation of the paper Real Time Speech Enhancement in the Waveform Domain. In which, we present a causal speech enhancement model working on the raw waveform that runs in real-time on a laptop CPU. The proposed model is based on an encoder-decoder architecture with skip-connections. It is optimized on both time and frequency domains, using multiple loss functions. Empirical evidence shows that it is capable of removing various kinds of background noise including stationary and non-stationary noises, as well as room reverb. Additionally, we suggest a set of data augmentation techniques applied directly on the raw waveform which further improve model performance and its generalization abilities.

Awesome-Singing-Voice-Synthesis-and-Singing-Voice-Conversion. A paper and project list about the cutting edge Speech Synthesis, Text-to-Speech (TTS), Singing Voice Synthesis (SVS), Voice Conversion (VC), Singing Voice Conversion (SVC), and related interesting works (such as Music Synthesis, Automatic Music Transcription, Automatic MOS Prediction, SSL-based ASR...etc).

fast_pytorch_kmeans. This is a pytorch implementation of k-means clustering algorithm

TorchPQ. Approximate nearest neighbor search with product quantization on GPU in pytorch and cuda

pyxmeans. Quick implementation of xmeans in python and C

pedalboard. 🎛 🔊 A Python library for audio.

DailyTalk. Official repository of DailyTalk: Spoken Dialogue Dataset for Conversational Text-to-Speech, ICASSP 2023

ai-audio-startups. Community list of startups working with AI in audio and music technology

cuda-samples. Samples for CUDA Developers which demonstrates features in CUDA Toolkit

keops. KErnel OPerationS, on CPUs and GPUs, with autodiff and without memory overflows

cylimiter. A C++/Cython audio limiter for Python.

diffusers. 🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch.

Python-Wrapper-for-World-Vocoder. A Python wrapper for the high-quality vocoder "World"

ParaLip. Parallel and High-Fidelity Text-to-Lip Generation; AAAI 2022 ; Official code

textlesslib. Library for Textless Spoken Language Processing

nvidia-htop. A tool for enriching the output of nvidia-smi.

NLP-progress. Repository to track the progress in Natural Language Processing (NLP), including the datasets and the current state-of-the-art for the most common NLP tasks.

lyra. A Very Low-Bitrate Codec for Speech Compression

TransformerTTS. 🤖💬 Transformer TTS: Implementation of a non-autoregressive Transformer based neural network for text to speech.

vector-quantize-pytorch. Vector (and Scalar) Quantization, in Pytorch

qsubmit. Wrapper for batch engine submission commands (remake of Obo's Perl qsubmit)

torch-imle. Implicit MLE: Backpropagating Through Discrete Exponential Family Distributions

pytorch-normalizing-flows. Normalizing flows in PyTorch. Current intended use is education not production.

multilexnorm2021. MultiLexNorm 2021 competition system from ÚFAL

glow. Code for reproducing results in "Glow: Generative Flow with Invertible 1x1 Convolutions"

Commonsense-Dialogues. A crowdsourced dataset of dialogues grounded in social contexts involving utilization of commonsense.

gecko. Gecko - A Tool for Effective Annotation of Human Conversations

multi-speaker-tacotron. VCTK multi-speaker tacotron for ICASSP 2020

soloist. Python

crepe. CREPE: A Convolutional REpresentation for Pitch Estimation -- pre-trained model (ICASSP 2018)

uda. Unsupervised Data Augmentation (UDA)

Diverse-Reference-Augmentation. Code and Data for our Findings of ACL 2021 paper titled 'Improving Automated Evaluation of Open Domain Dialog via Diverse Reference Augmentation. Varun Gangal *, Harsh Jhamtani *, Eduard Hovy, Taylor Berg-Kirkpatrick'

GEM-metrics. Automatic metrics for GEM tasks

Soft-DTW-Loss. PyTorch implementation of Soft-DTW: a Differentiable Loss Function for Time-Series in CUDA

vits. VITS: Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech

ufal-template-alt. Customized UFAL poster template

lhotse. Tools for handling multimodal data in machine learning projects.

survey. A Survey on Neural Speech Synthesis https://arxiv.org/pdf/2106.15561.pdf

adapters. A Unified Library for Parameter-Efficient and Modular Transfer Learning

x-transformers. A concise but complete full-attention transformer with a set of promising experimental features from various papers

bertviz. BertViz: Visualize Attention in Transformer Models

nopdb. NoPdb: Non-interactive Python Debugger

DialogWAE. Source Code for DialogWAE: Multimodal Response Generation with Conditional Wasserstein Autoencoder (https://arxiv.org/abs/1805.12352)

posterior-collapse-list. A curated list of techniques to avoid posterior collapse

DialogRPT. EMNLP 2020: "Dialogue Response Ranking Training with Large-Scale Human Feedback Data"

entmax. The entmax mapping and its loss, a family of sparse softmax alternatives.

sentence-transformers. State-of-the-Art Embeddings, Retrieval, and Reranking

fed. Code for SIGdial 2020 paper: Unsupervised Evaluation of Interactive Dialog with DialoGPT (https://arxiv.org/abs/2006.12719)

usr. Code for ACL 2020 paper: USR: An Unsupervised and Reference Free Evaluation Metric for Dialog Generation (https://arxiv.org/pdf/2005.00456)

self_dialogue_corpus. The Self-dialogue Corpus - a collection of self-dialogues across music, movies and sports

e2e_dialog_challenge. End-To-End Task-Completion Dialogue Challenge

dstc8-schema-guided-dialogue. The Schema-Guided Dialogue Dataset

airdialogue. Python

chat_corpus. chat corpus collection from various open sources

Topical-Chat. A dataset containing human-human knowledge-grounded open-domain conversations.

Taskmaster. Please see the readme file as well as our 2019 EMNLP paper linked here -->

pytorch-lightning. Pretrain, finetune ANY AI model of ANY size on 1 or 10,000+ GPUs with zero code changes.

sam. SAM: Sharpness-Aware Minimization (PyTorch)

TC-Bot. User Simulation for Task-Completion Dialogues

simmc. With the aim of building next generation virtual assistants that can handle multimodal inputs and perform multimodal actions, we introduce two new datasets (both in the virtual shopping domain), the annotation schema, the core technical tasks, and the baseline models. The code for the baselines and the datasets will be opensourced.

einops. Flexible and powerful tensor operations for readable and reliable code (for pytorch, jax, TF and others)

lambda-networks. Implementation of LambdaNetworks, a new approach to image recognition that reaches SOTA with less compute

lambda-bert. A 🤗-style implementation of BERT using lambda layers instead of self-attention

faiss. A library for efficient similarity search and clustering of dense vectors.

unlikelihood_training. Neural Text Generation with Unlikelihood Training

ToD-BERT. Pre-Trained Models for ToD-BERT

Target-Guided-Conversation. "Target-Guided Open-Domain Conversation" in ACL 2019

ParlAI. A framework for training and evaluating AI models on a variety of openly available dialogue datasets.

dialog-rl. Python

perin. PERIN is Permutation-Invariant Semantic Parser developed for MRP 2020

howdy. 🛡️ Windows Hello™ style facial authentication for Linux

ada-hessian. Easy-to-use AdaHessian optimizer (PyTorch)

SC-GPT. Few-shot Natural Language Generation for Task-Oriented Dialog

ConvLab-2. ConvLab-2: An Open-Source Toolkit for Building, Evaluating, and Diagnosing Dialogue Systems

NeuralDialogPapers. Summary of deep learning models for dialog systems (Tiancheng Zhao LTI, CMU)

BRC. A repository containing the code for the Bistable Recurrent Cell

Cross-Lingual-Voice-Cloning. Tacotron 2 - PyTorch implementation with faster-than-realtime inference modified to enable cross lingual voice cloning.

LeetCode. :pencil2: LeetCode solutions in C++ 11 and Python3

detr. End-to-End Object Detection with Transformers

espnet. End-to-End Speech Processing Toolkit

speedyspeech. Python

WaveRNN. WaveRNN Vocoder + TTS

meta-tasnet. A PyTorch implementation of Meta-TasNet from "Meta-learning Extractors for Music Source Separation

lazy-adam. :angel: Lazy Adam optimizer (pytorch)

datasets-CMU_Wilderness. CMU Wilderness Multilingual Speech Dataset

AdaIN-style. Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization

css10. CSS10: A Collection of Single Speaker Speech Datasets for 10 Languages

tacotron. A TensorFlow implementation of Google's Tacotron speech synthesis with pre-trained model (unofficial)

TTS. :robot: :speech_balloon: Deep learning for Text to Speech (Discussion forum: https://discourse.mozilla.org/c/tts)

Real-Time-Voice-Cloning. Clone a voice in 5 seconds to generate arbitrary speech in real-time

tgen. Statistical NLG for spoken dialogue systems

TinyEKF. Lightweight C/C++ Extended Kalman Filter with Python for prototyping

i2cdevlib. I2C device library collection for AVR/Arduino or other C++-based MCUs

KalmanFilter. This is a Kalman filter used to calculate the angle, rate and bias from from the input of an accelerometer/magnetometer and a gyroscope.

neat-openai-gym. NEAT for Reinforcement Learning on the OpenAI Gym

CEM-RL. Combining Evolutionary Algorithms and deep RL in various ways

ArduinoMotionSensorExample. MPU6050/MPU6500/MPU9150/MPU9250 over I2c for Arduino

codingame-cpp-merge. Automatic merger for C/C++ in the CodinGame IDE

deep-voice-conversion. Deep neural networks for voice conversion (voice style transfer) in Tensorflow