Open-source large-scale video generative models for text-to-video, image-to-video, speech-to-video and character animation.
- 스타
- 2,479
- 포크
- 0
- 열린 이슈
- 0
- 마지막 커밋
- 2개월 전
핵심 기능
- 01Generates audio-driven avatar video from a reference image and audio input.
- 02Supports continuous streaming generation with block-wise autoregressive processing.
- 03Offers a multi-GPU real-time pipeline and a single-GPU offline mode.
- 04Provides CLI and Gradio web UI inference scripts.
- 05Includes optional FP8 quantization and compilation for faster or lighter inference.
활용 사례
- Building real-time interactive digital human or virtual host applications.
- Generating long-form talking avatar videos from audio.
- Researching streaming video generation and audio-driven animation.
- Running avatar inference on a single 80GB GPU for experimentation.
토픽
자주 묻는 질문
- LiveAvatar은(는) 무엇인가요?
- Open-source framework for real-time, audio-driven avatar video generation that streams with effectively unlimited length, built on a 14B Wan2.2-S2V base model.
- LiveAvatar은(는) 무료로 사용할 수 있나요? 라이선스는 무엇인가요?
- LiveAvatar은(는) Apache-2.0 라이선스의 오픈소스입니다. 상업용 제품에 사용하기 전에 라이선스 조건을 확인하세요.
- LiveAvatar은(는) 아직 유지보수되고 있나요?
- 6개월 이내 업데이트. LiveAvatar의 마지막 커밋은 2026년 8월 24일입니다.
- LiveAvatar은(는) 어디에 사용할 수 있나요?
- Building real-time interactive digital human or virtual host applications. Generating long-form talking avatar videos from audio. Researching streaming video generation and audio-driven animation. Running avatar inference on a single 80GB GPU for experimentation.
- LiveAvatar은(는) 어디서 받을 수 있나요?
- 소스 코드와 설치 방법은 GitHub에 있습니다: https://github.com/Alibaba-Quark/LiveAvatar
요약은 프로젝트 README를 바탕으로 AI가 작성했습니다. GitHub 통계는 매일 갱신됩니다.
관련 프로젝트
Open-source model from Tencent Hunyuan that generates interactive game videos from a reference image and keyboard or mouse inputs.
A node-graph GUI, API and backend for running diffusion models locally to generate images, video, audio and 3D content.
An open-source interactive world model from Skywork AI that renders playable game scenes frame by frame with a causal video model on top of an executable game engine.
Ultra-lightweight on-device text-to-speech model with an 80M-parameter architecture, streaming output at 32kHz, and CPU, CUDA and MPS support.
A unified training engine for multimodal models, supporting vision-language, diffusion, and MoE architectures with FSDP2, Muon, and Liger kernels.