Code for the paper "Language Models are Unsupervised Multitask Learners"

The simplest, fastest repository for training/finetuning medium-sized GPTs.

CLUCalcSource-4.3.3 original GPL source code, released in 2005 on www.clucalc.info by Christian Perwass

GPT-3: Language Models are Few-Shot Learners

Code for the paper "Language Models are Unsupervised Multitask Learners"

World.JS is a JavaScript Wrapper for World Vocoder Powered by Emscripten

The latest code to make your own virtual sound synthesizer in Windows. Please see the videos associated with this code. Links in the source files.

A single file, easy to use, audio playback and synthesis framework

Robust Speech Recognition via Large-Scale Weak Supervision

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

This is now the official location of the Merlin project.

Docker Official Images

Automated Arch Linux Install

Placeholder repository to allow filing of general bugs/issues/etc against the Clear Linux OS for Intel Architecture linux distribution