Installation guide for Windows

[mp3pintyo]

Install Ubuntu on WSL2 on Windows 10 https://ubuntu.com/tutorials/install-ubuntu-on-wsl2-on-windows-10#7-enjoy-ubuntu-on-wsl

git lfs install: curl -s https://packagecloud.io/install/repositories/github/git-lfs/script.deb.sh | sudo bash sudo apt-get install git-lfs git lfs install

ffmpeg install: sudo add-apt-repository ppa:mc3man/trusty-media sudo apt-get update sudo apt-get dist-upgrade sudo apt-get install ffmpeg

git clone https://github.com/magic-research/magic-animate.git cd magic-animate pip install nvidia-pyindex pip3 install -r requirements.txt

you create the folder structure you need pretrained_models git lfs clone https://huggingface.co/zcxu-eric/MagicAnimate

pretrained_models/sd-vae-ft-mse https://huggingface.co/stabilityai/sd-vae-ft-mse/resolve/main/diffusion_pytorch_model.safetensors https://huggingface.co/stabilityai/sd-vae-ft-mse/resolve/main/config.json

pretrained_models/stable-diffusion-v1-5 https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/v1-5-pruned-emaonly.safetensors

stable-diffusion-v1-5/tokenizer All file

stable-diffusion-v1-5/text_encoder https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/text_encoder/pytorch_model.bin

stable-diffusion-v1-5/unet https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/unet/config.json https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/unet/diffusion_pytorch_model.bin

stable-diffusion-v1-5/scheduler https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/scheduler/scheduler_config.json

Run: python3 -m demo.gradio_animate or bash scripts/animate.sh

[FurkanGozukara]

can you try with this raw video instead of pre shared motion video

https://github.com/magic-research/magic-animate/assets/19240467/b1900be5-7ed8-4072-bcbb-b959fcdcaae2

[mp3pintyo]

can you try with this raw video instead of pre shared motion video ex2.mp4

But this is not a DensePose video. This definitely won't work.

[FurkanGozukara]

DensePose

ok how we gonna generate DensePose video?

[mp3pintyo]

DensePose

ok how we gonna generate DensePose video?

Whit this: https://github.com/facebookresearch/detectron2

[FurkanGozukara]

DensePose

ok how we gonna generate DensePose video?

Whit this: https://github.com/facebookresearch/detectron2

Thanks opened an issue thread there

Do you know how?

https://github.com/facebookresearch/detectron2/issues/5167

[zcxu-eric]

Hi, we acutally don't have window platform to test the installation and inference. You can refer to https://github.com/magic-research/magic-animate/issues/18 for generating your own DensePose seuqnces. Thanks.

[Rokkonrol]

Is there no way to install it without utilizing WSL2? I've been trying for around 2 hours to get it to work via Anaconda and/or Powershell on Windows 11, but it always come up with errors.

I'm wanting to run it on my cloud PC, but the service I use (Shadow) does not allow nested virtualization, so I can't utilize WSL2. Super frustrating.

[makemefeelgr8]

I tried installing it natively with 0 luck. The anaconda way fails, as there are about 10 libraries missing. As for the pip way pip3 install -r requirements.txt it fails with:

ERROR: Could not find a version that satisfies the requirement nvidia-cudnn-cu11==8.5.0.96 (from versions: 0.0.1.dev5, 8.9.4.19, 8.9.4.25, 8.9.5.29)
ERROR: No matching distribution found for nvidia-cudnn-cu11==8.5.0.96

They dropped windows support in tensorflow (talk about being out of touch), so the only way to fix this would be to install an older version of tensorflow. I'm no python developer, and I haven't found where the package version is specified. Thus, this is all the progress I made.

[Priestru]

urgh

/home/yuuru/magic-animate/magicanimate/pipelines/pipeline_animation.py:43: FutureWarning: Importing `DiffusionPipeline` or `ImagePipelineOutput` from diffusers.pipeline_utils is deprecated. Please import from diffusers.pipelines.pipeline_utils instead.
  from diffusers.pipeline_utils import DiffusionPipeline
Traceback (most recent call last):
  File "/home/yuuru/miniconda3/envs/manimate/lib/python3.8/runpy.py", line 194, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "/home/yuuru/miniconda3/envs/manimate/lib/python3.8/runpy.py", line 87, in _run_code
    exec(code, run_globals)
  File "/home/yuuru/magic-animate/magicanimate/pipelines/animation.py", line 282, in <module>
    run(args)
  File "/home/yuuru/magic-animate/magicanimate/pipelines/animation.py", line 271, in run
    main(args)
  File "/home/yuuru/magic-animate/magicanimate/pipelines/animation.py", line 75, in main
    tokenizer = CLIPTokenizer.from_pretrained(config.pretrained_model_path, subfolder="tokenizer")
  File "/home/yuuru/miniconda3/envs/manimate/lib/python3.8/site-packages/transformers/tokenization_utils_base.py", line 1838, in from_pretrained
    raise EnvironmentError(
OSError: Can't load tokenizer for 'pretrained_models/stable-diffusion-v1-5'. If you were trying to load it from 'https://huggingface.co/models', make sure you don't have a local directory with the same name. Otherwise, make sure 'pretrained_models/stable-diffusion-v1-5' is the correct path to a directory containing all relevant files for a CLIPTokenizer tokenizer.

[mp3pintyo]

urgh File "/home/yuuru/magic-animate/magicanimate/pipelines/animation.py", line 75, in main tokenizer = CLIPTokenizer.from_pretrained(config.pretrained_model_path, subfolder="tokenizer") 'pretrained_models/stable-diffusion-v1-5' is the correct path to a directory containing all relevant files for a CLIPTokenizer tokenizer.

Please read back! I described that you need to download the tokenizer folder.

[Priestru]

Please read back! I described that you need to download the tokenizer folder.

Yeah, I sure did. Trying again from scratch.

Okay, git clones helped. I think it's about different line endings

[FurkanGozukara]

Is there no way to install it without utilizing WSL2? I've been trying for around 2 hours to get it to work via Anaconda and/or Powershell on Windows 11, but it always come up with errors.

I'm wanting to run it on my cloud PC, but the service I use (Shadow) does not allow nested virtualization, so I can't utilize WSL2. Super frustrating.

there is a way

i spent hours to make it work yesterday like more than 6 hours

works with regular venv and python 3.10.x

can be downloaded from here i am preparing a tutorial and working on automatic dense pose generator right now

[FurkanGozukara]

finally released full scripts including auto DensePose maker : https://github.com/magic-research/magic-animate/issues/44

[huiyichanmian]

There is an issue with installing on Windows, you can view this video： https://www.bilibili.com/video/BV1ig4y1f7BQ/?share_source=copy_web&vd_source=842db193ae3e8fc29019a57821e30000

[friendlyFriend4000]

https://github.com/sdbds/magic-animate-for-windows

Here is a one click solution without any paywalls.

I think it's bad taste to use the issue section for promoting oneself economic endeavors because it goes against the spirit of open source. You cannot take open source content and put it behind a paywall in good faith. It doesn't matter if you worked 24hours straight. The project creators have probably worked much longer.

[Augus202312]

urgh 文件“/home/yuuru/magic-animate/magicanimate/pipelines/animation.py”，第 75 行，在主 tokenizer = CLIPTokenizer.from_pretrained(config.pretrained_model_path, subfolder="tokenizer") 'pretrained_models/stable-diffusion-v1 -5' 是包含 CLIPTokenizer 分词器所有相关文件的目录的正确路径。

请回读！我描述过您需要下载 tokenizer 文件夹。

[friendlyFriend4000]

urgh 文件“/home/yuuru/magic-animate/magicanimate/pipelines/animation.py”，第 75 行，在主 tokenizer = CLIPTokenizer.from_pretrained(config.pretrained_model_path, subfolder="tokenizer") 'pretrained_models/stable-diffusion-v1 -5' 是包含 CLIPTokenizer 分词器所有相关文件的目录的正确路径。

请回读！我描述过您需要下载 tokenizer 文件夹。

open an issue here https://github.com/sdbds/magic-animate-for-windows/issues