Jungmin_ML_DL_CV_LLM
Jungmin_ML_DL_CV_LLM
are you doing your own data? style mixing is within generate_video.py and I also agreed with all image is not working well for all part of style mixing. I think...
plz open links...
thank you for replying this comment quickly, are you still working on SAM, right? In here, is there no implementation code? I am wondering, In SAM github, they did not...
thank you so much. I have another question, I want to know how to get style space from w space. and do you try to use w+space? do you know...
upgrade transformer in huggingface
Hi, can you give me advice which line do I modify to apply your idea?
> @john09282922 refer to the section "Simple to implement Input Perturbation in diffusion models" of the readme file Hi, thanks for replying my question, I saw that, but I want...
> @john09282922 you should make the modification on their script [losses.py](https://github.com/ermongroup/ddim/blob/main/functions/losses.py) > > more specifically, line 10 thanks for giving me detail info. I love your code than original ddim...
> @john09282922 e in DDIM is epsilon in DDPM. I am not clear about 'you are using x_start after the idea equation', what does it mean? this is original ddim...
> @john09282922 replace `x = x0 * a.sqrt() + e * (1.0 - a).sqrt()` with `x = x0 * a.sqrt() + (e+gamma*w) * (1.0 - a).sqrt()` where w~N(0, I) >...