LLaSE-G1: Incentivizing Generalization Capability for LLaMA-based Speech Enhancement

An open source dataset for source separation

SpEx+(tied) source code

Spark-TTS Inference Code

SlamKit is an open source tool kit for efficient training of SpeechLMs. It was used for "Slamming: Training a Speech Language Model on One GPU in a Day"

Research and Production Oriented Speaker Verification, Recognition and Diarization Toolkit

Production-tested AI infrastructure tools for efficient AGI development and community-driven innovation

OSUM: Open Speech Understanding Model, open-sourced by ASLP@NPU.

Unified automatic quality assessment for speech, music, and sound.

Open-source industrial-grade ASR models supporting Mandarin, Chinese dialects and English, achieving a new SOTA on public Mandarin ASR benchmarks, while also offering outstanding singing lyrics rec…

🧑‍🚀 全世界最好的LLM资料总结（数据处理、模型训练、模型部署、o1 模型、小语言模型、视觉语言模型） | Summary of the world's best LLM resources.

Clean, minimal, accessible reproduction of DeepSeek R1-Zero

[Unofficial] PyTorch implementation of "Conformer: Convolution-augmented Transformer for Speech Recognition" (INTERSPEECH 2020)

An unofficial implementation of the Personal VAD speaker-conditioned voice activity detection method. Bachelor's thesis project.

Awesome speech/audio LLMs, representation learning, and codec models

A high-throughput and memory-efficient inference and serving engine for LLMs

Official repository for Mamba-based Segmentation Model for Speaker Diarization

wenet_LLM_from_ASLP

The official Meta Llama 3 GitHub site

Inference code for Llama models

LLaMA-Omni is a low-latency and high-quality end-to-end speech interaction model built upon Llama-3.1-8B-Instruct, aiming to achieve speech capabilities at the GPT-4o level.

real time face swap and one-click video deepfake with only a single image

✨✨VITA-1.5: Towards GPT-4o Level Real-Time Vision and Speech Interaction

🤗 PEFT: State-of-the-art Parameter-Efficient Fine-Tuning.

Code for loralib, an implementation of "LoRA: Low-Rank Adaptation of Large Language Models"

MELD: A Multimodal Multi-Party Dataset for Emotion Recognition in Conversation

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.