Laidawang

reduce decoding_t to 1, When you decode more than a dozen frames at the same time, it will explode

[memory.log](https://github.com/lllyasviel/Fooocus/files/13800296/memory.log) I recorded the memory of 20 times outpainting(four directions). Memory usage is indeed rising, although it is not obvious. This is my code, Can somebody else try to reproduce...

inpaint foocus patch is just a lora, Modify the correct lora loading method, just copy the way fooocus loaded. for inpainting head im not sure.

not yet,I read the codes of both and they are basically the same, but I don't have a good way to modify the code. Fooocus redefines a lot of the...

they also seem to use [fooocus_inpaint_head.pth](https://huggingface.co/lllyasviel/fooocus_inpaint/blob/main/fooocus_inpaint_head.pth) I'm not quite sure what this will do, I read the code and maybe an additional patch for unet?  The [inpaint_v26.fooocus.patch](https://huggingface.co/lllyasviel/fooocus_inpaint/blob/main/inpaint_v26.fooocus.patch) is more...

> Actually it seems more like a controlnet, something more like this one: https://huggingface.co/destitech/controlnet-inpaint-dreamer-sdxl.实际上，它看起来更像是一个控制网，更像是这个：https://huggingface.co/destitech/controlnet-inpaint-dreamer-sdxl。 > > They also use a custom sampler for the inpainting, but I agree, it would...

https://github.com/lllyasviel/Fooocus/blob/main/modules/inpaint_worker.py#L187 Another thing worth considering is how to implement this patch for inpaint head model.

> > https://github.com/lllyasviel/Fooocus/blob/main/modules/inpaint_worker.py#L187 Another thing worth considering is how to implement this patch for inpaint head model. > > But the code is just updating the first conv, no? You...

Maybe consider loading it in comfy and saving it as overall weights and then using it in diffusers?

But as I saw in fooocus, the base model will still be used in the second stage, so the most elegant way is to load and unload it freely.