Update README.md
This commit is contained in:
@@ -6,6 +6,8 @@ A universal swap node that supports ComfyUI native workflow, allowing 4_6G users
|
||||
* Support text encoder or diffusion models, is enable text encoder will need more Ram
|
||||
|
||||
# Example
|
||||
* run minimax H3 5min 0.4 just need 4.5G Vram (要降低Te的占用需要加te模块,或者用comfyUI自带的Vbar,OOM再加TE swap,避免内存占用)
|
||||

|
||||
* run bernini int4 +loras ,512x384x120frames just need 9-10G Vram (if unpack node,notice batch size is wrong 注意官方模板解开后,batch size指向是错的,须改成1)
|
||||

|
||||
* run klein9B Q8 just need 4.8G Vram
|
||||
|
||||
Reference in New Issue
Block a user