[Bug]: EinopsError: Error while processing rearrange-reduction pattern "(b f) d c -> (b d) f c". Input tensor shape: torch.Size([2, 4096, 320]). Additional info: {'f': 16}. Shape mismatch, can't divide axis of length 2 in chunks of 16

[tail611]

Is there an existing issue for this?

• [X] I have searched the existing issues and checked the recent builds/commits of both this extension and the webui

Have you read FAQ on README?

• [X] I have updated WebUI and this extension to the latest version

What happened?

“I want to create a video using Stable Diffusion, but after installing and trying to use AnimateDiff, I get the following error message.”

EinopsError: Error while processing rearrange-reduction pattern "(b f) d c -> (b d) f c". Input tensor shape: torch.Size([2, 4096, 320]). Additional info: {'f': 16}. Shape mismatch, can't divide axis of length 2 in chunks of 16

Steps to reproduce the problem

1. Go to .... AmimativeDiff
2. Press .... Generate
3. ... Error massage

What should have happened?

“The video should be generated.”

Commit where the problem happens

webui: extension: version: v1.6.0  •  python: 3.10.6  •  torch: 2.1.2+cu121  •  xformers: N/A  •  gradio: 3.41.2  •  checkpoint: 7c819b6d13

What browsers do you use to access the UI ?

Microsoft Edge

Command Line Arguments

no

Console logs

sorry i don't know

Additional information

No response

[tail611]

please help me!!ㅠㅠ “I already tried installing the latest version of the web UI, but the same issue occurred, so I downgraded the version.”

[zappityzap]

Try using the motion modules from here: https://huggingface.co/conrevo/AnimateDiff-A1111/tree/main

[tail611]

    p{margin-top:0;margin-bottom:0}

Dear Developer,I am writing to express my sincere gratitude for your support. Despite my efforts to resolve the errors with the provided files and placing them in my animatediff folder, I am still encountering issues. I understand your time is valuable, but I am eager to use AnimateDiff. Could you please assist me one more time? I truly appreciate the continuous efforts for the success and development of Stable Diffusion WebUI. Thank you.Best regards,

Jinwoo from Korea--------- 원본 메일 ---------

보낸사람: zappityzap ***@***.***>
받는사람: continue-revolution/sd-webui-animatediff ***@***.***>
  참조: tail611 ***@***.***>,Manual ***@***.***>
날짜: 24.08.02 23:24 GMT +0900

제목: Re: [continue-revolution/sd-webui-animatediff] [Bug]: EinopsError: Error while processing rearrange-reduction pattern "(b f) d c -> (b d) f c". Input tensor shape: torch.Size([2, 4096, 320]). Additional info: {'f': 16}. Shape mismatch, can't divide axis of length 2 in chunks of 16 (Issue #546)

Try using the motion modules from here: https://huggingface.co/conrevo/AnimateDiff-A1111/tree/main

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: @.***>

[tail611]

    p{margin-top:0;margin-bottom:0}

Dear DeveloperI hope this email finds you well.I am writing to seek further assistance with the installation of AnimateDiff on my system. Despite following the provided instructions and attempting multiple solutions, I am still encountering issues.Here is a summary of the steps I have taken and the problems I am facing:File Placement:I have downloaded the files from the provided link and placed them in the corresponding directories within the animatediff folder.Screenshots of the folder structure are attached for your reference.CUDA Downgrade Issue:I downgraded CUDA from 12.1 to 11.8, but the WebUI still recognizes the version as 12.1.This persists despite ensuring that all CUDA 12.1 files were removed and only CUDA 11.8 files are installed.Screenshot of the installed programs showing CUDA 11.8 is attached.Persistent Errors:I continue to encounter CUDA-related errors and issues with memory allocation when attempting to use AnimateDiff.These errors affect the functionality of the WebUI, causing other features to stop working until a restart is performed.Steps Taken:I uninstalled and reinstalled torch, torchvision, and torchaudio with the correct CUDA version.I have also considered reinstalling the WebUI, but previous attempts led to the WebUI failing to launch.Here are some logs and error messages from my attempts:2024-08-03 14:24:20,662 - AnimateDiff - INFO - AnimateDiff process start.        | 1440/1700 [1:33:40<16:28,  3.80s/it]2024-08-03 14:24:20,662 - AnimateDiff - INFO - Loading motion module animatediffMotion_v15V2.ckpt from F:\stable-diffusion-webui\extensions\sd-webui-animatediff\model\animatediffMotion_v15V2.ckpt2024-08-03 14:24:22,191 - AnimateDiff - INFO - Guessed animatediffMotion_v15V2.ckpt architecture: MotionModuleType.AnimateDiffV2...RuntimeError: CUDA error: device-side assert triggeredCompile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.I would greatly appreciate any further guidance or recommendations you could provide to resolve these issues. I am eager to use AnimateDiff and your help has been invaluable thus far.Thank you for your continued support and understanding.Best regards,

Jinwoo South Korea--------- 원본 메일 ---------

보낸사람: zappityzap ***@***.***>
받는사람: continue-revolution/sd-webui-animatediff ***@***.***>
  참조: tail611 ***@***.***>,Manual ***@***.***>
날짜: 24.08.02 23:24 GMT +0900

제목: Re: [continue-revolution/sd-webui-animatediff] [Bug]: EinopsError: Error while processing rearrange-reduction pattern "(b f) d c -> (b d) f c". Input tensor shape: torch.Size([2, 4096, 320]). Additional info: {'f': 16}. Shape mismatch, can't divide axis of length 2 in chunks of 16 (Issue #546)

Try using the motion modules from here: https://huggingface.co/conrevo/AnimateDiff-A1111/tree/main

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: @.***>