GitLab

* Init deploy: not ok

* Test data_manager & task_manager

* pipeline is no need for worker

* Update worker text_encoder

* deploy: submit task

* add apis

* Test pipelineRunner

* Fix pipeline

* Tidy worker & test PipelineWorker ok

* Tidy code

* Fix multi_stage for wan2.1 t2v & i2v

* api query task, get result & report subtasks failed when workers stop

* Add model list functionality to Pipeline and API

* Add task cancel and task resume  to API

* Add RabbitMQ queue manager

* update local task manager atomic

* support postgreSQL task manager, add lifespan async init

* worker print -> logger

* Add S3 data manager, delete temp objects after finished.

* fix worker

* release fetch queue msg when closed, run stuck worker in another thread, stop worker when process down.

* DiTWorker run with thread & tidy logger print

* Init monitor without test

* fix monitor

* github OAuth and jwt token access & static demo html page

* Add user to task, ok for local task manager & update demo ui

* sql task manager support users

* task list with pages

* merge main fix

* Add proxy for auth request

* support wan audio

* worker ping subtask and ping life, fix rabbitmq async get,

* s3 data manager with async api & tidy monitor config

* fix merge main & update req.txt & fix html view video error

* Fix distributed worker

* LImit user visit freq

* Tidy

* Fix only rank save

* Fix audio input

* Fix worker fetch None

* index.html abs path to rel path

* Fix dist worker stuck

* support publish output video to rtmp & graceful stop running dit step or segment step

* Add VAReader

* Enhance VAReader with torch dist

* Fix audio stream input

* fix merge refractor main, support stream input_audio and output_video

* fix audio read with prev frames & fix end take frames & tidy worker end

* split audio model to 4 workers & fix audio end frame

* fix ping subtask with queue

* Fix audio worker put block & add whep, whip without test ok

* Tidy va recorder & va reader log, thread canel within 30s

* Fix dist worker stuck: broadcast stop signal

* Tidy

* record task active_elapse & subtask status_elapse

* Design prometheus metrics

* Tidy prometheus metrics

* Fix merge main

* send sigint to ffmpeg process

* Fix gstreamer pull audio by whep & Dockerfile for gstreamer & check params when submitting

* Fix merge main

* Query task with more info & va_reader buffer size = 1

* Fix va_recorder

* Add config for prev_frames

* update frontend

* update frontend

* update frontend

* update frontend
merge

* update frontend & partial backend

* Different rank for va_recorder and va_reader

* Fix mem leak: only one rank publish video, other rank should pop gen vids

* fix task category

* va_reader pre-alloc tensor & va_recorder send frames all & fix dist cancel infer

* Fix prev_frame_length

* Tidy

* Tidy

* update frontend & backend

* Fix lint error

* recover some files

* Tidy

* lint code

---------
Co-authored-by: liuliang1 <liuliang1@sensetime.com>
Co-authored-by: unknown <qinxinyi@sensetime.com>

* [Feat] Make prev_frame_length parameter configurable via config

* Add newline at end of file in wan_audio_runner.py

---------
Co-authored-by: Yang Yong(雍洋) <yongyang1030@163.com>

* [Fix] Fix A14B load quant model bug and tae bug

* [Fix] Fix A14B load quant model bug and tae bug

* update seko talk

* update

* Update configs

* update

---------
Co-authored-by: gushiqiao <975033167@qq.com>

* feature: qwen-image support cpu offload (block) and refactor transfomrer model

* bugfix

* bugfix import

---------
Co-authored-by: Yang Yong(雍洋) <yongyang1030@163.com>

* feature: qwen-image support cpu offload (block) and refactor transfomrer model

* bugfix

* update seko file names

* update

* [Update] update seko-models configs

* Add seko scripts

---------
Co-authored-by: gushiqiao <975033167@qq.com>

* [Feat] Support vae bf16 and add 4090 best config

* [Feat] Support vae bf16 and add 4090 best config

* [Feat] Support vae bf16 and add 4090 best config

* [Feat] Support vae bf16 and add 4090 best config

* update dtype

---------
Co-authored-by: helloyongyang <yongyang1030@163.com>

* update docs

* update docs

* [Feat] audio driven for wan2.2-r2v-5b

* refactor: reduce vae args dim &&  timestep init

* update audio pre_infer

* update runner

* update scripts

* [Fix] move ref timestep padding to Scheduler

* fix wan i2v

---------
Co-authored-by: helloyongyang <yongyang1030@163.com>

* [Fix] Fix audio model offload bug.

* Fix

* Fix

* Fix

* Fix

* Fix

* Fix

* bugfix:adapt to  5B dit model, derive attention_head_dim from config[dim]

* [Fix] Wan2.2 Vae Encode refactor: drop args parameter and use self.cpu_offload

* [Fix] fix bug

* Fix bug

* Update run_wan_vace.sh