GitLab

Feature:
    1. added mlu590 bfloat16, single-gpu and multi-gpus inference.
    2. added mlu590 int8 inference.

Co-authored-by: root <root@pt-2ec29e323c56463f8ad6d292b6047107-worker-0.pt-2ec29e323c56463f8ad6d292b6047107.ns-devsft-3460edd0.svc.cluster.local>
Co-authored-by: gushiqiao <975033167@qq.com>

Co-authored-by: root <root@pt-731f15f5c62740f495519352dc90d9c4-worker-0.pt-731f15f5c62740f495519352dc90d9c4.ns-devsft-3460edd0.svc.cluster.local>

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

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

https://github.com/Linboyan-trc/SpargeAttn/tree/02_develop



need install

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

[feat]: refactor to adapt new config system and align precision for qwen-image & qwen-image-edit (#383)

[feat]: refactor to adapt new config system and align precision for
qwen-image & qwen-image-edit

support self-forcing wan2.1 dmd

* update seko talk

* update

* Update configs

* update

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

* add gguf loading functions

* support wan22 distill