Patch model patch order slow down generating 10 times

[frankyifei]

My card is 3060 with 12GB vram, I loaded flux Q8_0, with several loras, the recent change to ComfyUI broke the workflow for me, like #151. Your new commit fixed the problem, now the lora works but is just very slow for me.
without the node, the log is

Requested to load Flux
loaded partially 9943.83151171875 9943.367309570312 0

with the node, the speed is 10 times slower, is the model not loaded properly

Requested to load Flux
loaded completely 0.0 12119.509887695312 True

[frankyifei]

It works now, guess it is my WSL environment issue