AnimateDiff
AnimateDiff copied to clipboard
sdxl webui cannot generate GIF or MP4
Error Info:
……stable-diffusion-webui/extensions/sd-webui-animatediff-master/scripts/animatediff.py", line 71, in postprocess_batch_list
self.prompt_scheduler.save_infotext_img(p)
AttributeError: 'NoneType' object has no attribute 'save_infotext_img'
……extensions/sd-webui-animatediff-master/scripts/animatediff.py", line 83, in postprocess
self.prompt_scheduler.save_infotext_txt(res)
AttributeError: 'NoneType' object has no attribute 'save_infotext_txt'
webui output:only image,no gif or mp4 video.
Same error in runpod And generate only random img like "Batch count" Holly, I just realized I hadn't updated the version :D
Same error in runpod And generate only random img like "Batch count" Holly, I just realized I hadn't updated the version :D
Does it work? I have updated sd-webui-animatediff, the images are still random, cannot generate a video.
Same error in runpod And generate only random img like "Batch count" Holly, I just realized I hadn't updated the version :D
Does it work? I have updated sd-webui-animatediff, the images are still random, cannot generate a video.
same with you, and may be the ckpt problem.
Exception: bad file inside D:\SD\sd-webui\extensions\sd-webui-animatediff\model\mm_sdxl_v10_beta.ckpt: mm_sdxl_nightly,ckpt/byteorder
AttributeError: 'NoneType' object has no attribute 'pop'
then the same error info above.
same with webui
I have the same issue with webui, it just generates a bunch of separate images, not sequential in any way
Exact same issue.
AttributeError: 'NoneType' object has no attribute 'pop'
Then just does a batch of images with different seeds.
EDIT: looking into it more, it's saying won't run the animateddiff models without --disable-safe-unpickle Being that I grabbed them from civitai, first thought was it was corrupt or messed with in some way. But it has the exact same error with the mm_sdxl_v10_beta.ckpt model directly from huggingface.
I solved changed model, because i think not all model are compatible with Animatediff
Same error with the mm_sdxl_v10_beta.ckpy and the animatediffMotion_sdxlV10Beta.ckpt from civit. Just a batch run of images no output.
Tested it with the Nightvisionxl, juggernaughtxl, and colossusproject basemodels
why?
I solved changed model, because i think not all model are compatible with Animatediff
How to do?please help
I solved changed model, because i think not all model are compatible with Animatediff
How to do?please help
I solved changed model, because i think not all model are compatible with Animatediff
Well animatediff supposedly works with SDXL now, hence the animatediffMotion_sdxlV10Beta.ckpt model. I'm just using with the vanilla SDXL base model from hugging face and it's throwing that error every time.
In my case, the problem was caused by the motion model was considered unsafe when verifying the file, so the program is not going to read it, just skip the check by simply adding "--disable-safe-unpickle" into commandline argument, then the problem is solved. ( By the way, I've read an article online saying that this problem can be solved by upgrading the torch and xformers, it didn't work for me though.)
In my case, the problem was caused by the motion model was considered unsafe when verifying the file, so the program is not going to read it, just skip the check by simply adding "--disable-safe-unpickle" into commandline argument, then the problem is solved. ( By the way, I've read an article online saying that this problem can be solved by upgrading the torch and xformers, it didn't work for me though.)
yes I'd mentioned that above re "--disable-safe-unpickle", but turning this off allows execution of malicious code. It's like saying, just turn off your anti-virus software = problem solved. Not really.
In my case, the problem was caused by the motion model was considered unsafe when verifying the file, so the program is not going to read it, just skip the check by simply adding "--disable-safe-unpickle" into commandline argument, then the problem is solved. ( By the way, I've read an article online saying that this problem can be solved by upgrading the torch and xformers, it didn't work for me though.)
yes I'd mentioned that above re "--disable-safe-unpickle", but turning this off allows execution of malicious code. It's like saying, just turn off your anti-virus software = problem solved. Not really.
Sorry, somehow I missed that message. I agree that it is not a solution to add in "--disable-safe-unpickle", is just a temporary measure for animetediff to work, after I disabled the safe check, a lot of problems occurred, like the webui required a total restart after animatediff for a normal t2i to work.
This solves the problem: https://github.com/continue-revolution/sd-webui-animatediff/issues/379