…IGC-Engineering#190)

* feat: lite avatar handler support multi session

* feat: update readme

* fix cr comments

* support qwen-omni realtime api

* feat: align submodule to main

* fix musetalk import error

* feat: update yaml

* update: update readme

---------

Co-authored-by: yunzong.wf <yunzong.wf@alibaba-inc.com>
Co-authored-by: chenggang.cg@alibaba-inc.com <chenggang.cg@alibaba-inc.com>
Co-authored-by: 泸奕 <ly449033@alibaba-inc.com>

* Update README.md

* Update README.md

* Update readme_en.md

* Update README.md

* add download script for musetalk

* Update avatar_musetalk_processor.py

* Update README.md

fix the config path for docker in readme

* Update readme_en.md

fix a bug in the readme

* fix download warning

* fix mps property bug

* fix musetalk for 50 gpu

* fix: add cudnn lib path to LD_LIBRARY_PATH in case onnxruntime can not use gpu

* fix get value bug

* add cuda12.8 dockerfile support

* fix: lite avatar use shared memory for data transfer between processes

* feat: reset config

---------

Co-authored-by: lovepope <lovepope@gmail.com>
Co-authored-by: sudowind <wfpkueecs@163.com>

* fix download warning

* fix mps property bug

* fix musetalk for 50 gpu

* fix: add cudnn lib path to LD_LIBRARY_PATH in case onnxruntime can not use gpu

* fix get value bug

* add cuda12.8 dockerfile support

* fix: lite avatar use shared memory for data transfer between processes

* feat: reset config

* add run_docker_cuda128.sh

---------

Co-authored-by: lovepope <lovepope@163.com>
Co-authored-by: sudowind <wfpkueecs@163.com>
Co-authored-by: HumanAIGC-Engineering <humanaigc.engineering@gmail.com>

…nAIGC-Engineering#210)

* added the 'connection_ttl' param explicitly to all default configs and
documented it in the faq.

* Update README.md

* Update readme_en.md

* Update README.md

…umanAIGC-Engineering#224)

Supplement missing parameters

*def handle  need to call finish streaming when catch exception

* fix liteavatar bug

* remove unnecessary variable

---------

Co-authored-by: lovepope <lovepope@gmail.com>,bingochaos <523834173@qq.com>,HaveAnApplePie <luyi1019105596@gmail.com>

* [fix] 关闭swagger页面

* [feat] 使用共享内存池来在liteavatar的进程之间传递图像/音频数据，代替pytorch原生的共享内存传递。

* [feat]令fastrtc默认使用H264硬件编码，代替原来的vp8

* [refactor] 修复liteavatar销毁和创建相关的竞态问题，并添加更多日志

---------

Co-authored-by: heguanqiao.hgq <heguanqiao.hgq@alibaba-inc.com>
Co-authored-by: 泸奕 <ly449033@alibaba-inc.com>

* Create docker-compose.yml

* Create turnserver.conf

* Update pre_config_install.sh

* Update post_config_install.sh

* Update docker-compose.yml

* Update README.md

* Update README.md

* Update readme_en.md

…ring#255)

* 如果models 中存在则使用绝对路径加载模型

* Use absolute path after model download completes

* 关闭swagger页面

* [feat] 使用共享内存池来在liteavatar的进程之间传递图像/音频数据，代替pytorch原生的共享内存传递。

* [feat]令fastrtc默认使用H264硬件编码，代替原来的vp8

* [refactor] 修复liteavatar销毁和创建相关的竞态问题，并添加更多日志

* [fix] Enhance H.264 encoding logic with improved fallback handling and encoder selection

* [update] Add unit tests for H.264 encoder fallback logic and error handling