🔊 Text-Prompted Generative Audio Model

A guidance language for controlling large language models.

Instruct-tune LLaMA on consumer hardware

A multi-voice TTS system trained with an emphasis on quality

AirLLM 70B inference with single 4GB GPU

An Open Source text-to-speech system built by inverting Whisper.

“让爷康康”是一款手机 AI 应用程序，可以监测不良坐姿并进行语音提示

无需情感标注的情感可控语音合成模型，基于VITS

VITS2: Improving Quality and Efficiency of Single-Stage Text-to-Speech with Adversarial Learning and Architecture Design

Multilingual G2P in 100 languages

Speech Toolkit for Malaysian language, https://malaya-speech.readthedocs.io/

Measurement and processing of binaural impulse responses for personalized surround virtualization on headphones.

Software kit that uses deep learning to generate vocaloid music (melodies and lyrics)

Large-language Model Evaluation framework with Elo Leaderboard and A-B testing

Convert BART models to ONNX with quantization. 3X reduction in size, and upto 3X boost in inference speed

German Alpaca Dataset (Cleaned + Translated)

Audio Super Resolution model to decompress mp3 into wav file with recovered frequencies built on a "ResUNet" style architecture

SovitsTokenizer: A low-bitrate audio tokenizer that converts speech into discrete tokens

Fine-Tune Wav2Vec2 Bert 2.0 for Jyutping Recognition

一些新东西和优化