Lighteval is your all-in-one toolkit for evaluating LLMs across multiple backends

Foundational Model for Speech Recognition Tasks

The most popular spellchecking library.

A fast, local neural text to speech system

Application of MB-iSTFT-VITS components to vits2_pytorch

unofficial vits2-TTS implementation in pytorch

A multilingual ASR model that can recognize ten Turkic languages—Azerbaijani, Bashkir, Chuvash, Kazakh, Kyrgyz, Sakha, Tatar, Turkish, Uyghur, and Uzbek.

An easy-to-run semantic segmentation model based on Unet

Text reading pipeline that combines segmentation and OCR-models.

An easy-to-run OCR model pipeline based on CRNN and CTC loss

Apertium linguistic data for Bashkir

QALD-9-Plus Dataset for Knowledge Graph Question Answering

🤗 The largest hub of ready-to-use datasets for ML models with fast, easy-to-use and efficient data manipulation tools

500 AI Machine learning Deep learning Computer vision NLP Projects with code

Modules to convert numbers to words. 42 --> forty-two

DDSP: Differentiable Digital Signal Processing

🐸STT - a deep learning toolkit for Speech-to-Text, battle-tested in research and production

Deezer source separation library including pretrained models.

Машинное обучение на ФКН ВШЭ