feat: add sdxl v-pred suppport (leejet#536)

fix: avoid sd2((non inpaint) crash on v-pred check (leejet#537)

chore: fix CUDA on GitHub Action (leejet#567)

chore: change SD_CUBLAS/SD_USE_CUBLAS to SD_CUDA/SD_USE_CUDA

feat: support more LoRA models (leejet#520)

 feat: support Inpaint models (leejet#511)

fix: fix metal build (leejet#513)

chore: remove rocm5.5 build temporarily

refactor: add some sd vesion helper functions

feat: add sd3.5 medium and skip layer guidance support (leejet#451)

* mmdit-x

* add support for sd3.5 medium

* add skip layer guidance support (mmdit only)

* ignore slg if slg_scale is zero (optimization)

* init out_skip once

* slg support for flux (expermiental)

* warn if version doesn't support slg

* refactor slg cli args

* set default slg_scale to 0 (oops)

* format code

---------

Co-authored-by: leejet <[email protected]>