Karel Vesely

seamless_communication. Foundational Models for State-of-the-Art Speech and Text Translation

Edge-Punct-Casing. Python

kaldi-model-server. Simple Kaldi model server for chain (nnet3) models in online recognition mode directly from a local microphone

greek_podcasts_asr. Shell

mirascope. The LLM Anti-Framework

fixwav. Quick utility to fix WAV files with incorrect lengths

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

audiomentations. A Python library for audio data augmentation. Useful for making audio ML models work well in the real world, not just in the lab.

mistral-inference. Official inference library for Mistral models

gpt4all. GPT4All: Run Local LLMs on Any Device. Open-source and available for commercial use.

CQT_toolbox_python. Constant-Q Transform Toolbox for Python/MATLAB

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

GigaSpeech. Large, modern dataset for speech recognition

kaldilm. Python wrapper for kaldi's arpa2fst

mitlm. MIT Language Modeling Toolkit

pocolm. Small language toolkit for creation, interpolation and pruning of ARPA language models

libriheavy. Libriheavy: a 50,000 hours ASR corpus with punctuation casing and context

json. JSON for Modern C++

X-Punctuator. A PyTorch implementation of a punctuation prediction system using (B)LSTM, which automatically adds suitable punctuation into text without punctuation.

k2. FSA/FST algorithms, differentiable, with PyTorch compatibility.

wikiextractor. A tool for extracting plain text from Wikipedia dumps

stk. STK, the Brno speech recognition toolkit.

fdndlp. A speech dereverberation algorithm, also called wpe

keyczar. Easy-to-use crypto toolkit