requirements.txt

[SoftologyPro]

Can you provide a requirements.txt with versioned packages?
I am going in circles trying to guess versions and get them to cooperate.

[eniora]

requirements.txt

I modified it and it worked for me. After installing the requirements.txt with "pip install -r requirements.txt", do this if you have an nvidia GPU:

pip install torch==2.7.1+cu128 torchvision==0.22.1+cu128 --index-url https://download.pytorch.org/whl/cu128

[SoftologyPro]

Thanks, after using that requirements.txt these patches are still needed to get it starting...

pip install -r requirements.txt
pip install huggingface-hub==0.25.2
pip install pyrallis==0.3.1
pip install openai-clip==1.0.1
pip uninstall -y diffusers
pip install diffusers==0.33.1
pip uninstall -y transformers
pip install transformers==4.52.4

[eniora]

Glad you got it working. Cheers!
For huggingface-hub I used huggingface-hub==0.23.1 and not >= because version 0.23.1 is also good. I forgot to update that in the file.

A hint: if you're using venv, take a copy of the "site packages" folder located at \venv\Lib, that way you won't need to go through all this pain in the future if you decide to reinstall Instant Style (same applies for any python app).

[SoftologyPro]

Working as in got past the requirements. Not working beyond that.

Firstly line 109 in InstantStyle-Plus\infer_style.py needs to be changed from
checkpoint = torch.load(model_path, map_location="cpu")
to
checkpoint = torch.load(model_path, map_location="cpu", weights_only=False)
That avoids the error

Traceback (most recent call last):
  File "D:\Tests\InstantStyle-Plus\InstantStyle-Plus\infer_style.py", line 109, in <module>
    checkpoint = torch.load(model_path, map_location="cpu")
  File "D:\Tests\InstantStyle-Plus\InstantStyle-Plus\venv\lib\site-packages\torch\serialization.py", line 1524, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint.
        (1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
        (2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
        WeightsUnpickler error: Unsupported global: GLOBAL numpy.core.multiarray.scalar was not an allowed global by default. Please use `torch.serialization.add_safe_globals([numpy.core.multiarray.scalar])` or the `torch.serialization.safe_globals([numpy.core.multiarray.scalar])` context manager to allowlist this global if you trust this class/function.

Once that is fixed, it hits the main roadblock...
ValueError: The provided pretrained_model_name_or_path "./checkpoints/sdxlUnstableDiffusers_v8HeavensWrathVAE" is neither a valid local path nor a valid repo id. Please check the parameter.
The readme has no mention of this model.
I tried downloading this one
https://huggingface.co/tomjackson2023/test/blob/a03630625adb65b5598298e7c233e90c082fbf3c/sdxlUnstableDiffusers_v8HeavensWrathVAE.safetensors
under the /checkpoints/sdxlUnstableDiffusers_v8HeavensWrathVAE/ folder but that gives
OSError: Error no file named model_index.json found in directory ./checkpoints/sdxlUnstableDiffusers_v8HeavensWrathVAE.
So, it is looking for a HF repo name that has the model and model.json etc files.

Any idea what the correct repo path should be?

[SoftologyPro]

Here is a "one click" install.bat to install the required packages etc in Windows.
The only thing missing is the correct sdxlUnstableDiffusers_v8HeavensWrathVAE model download

@echo off

cd
echo *** %time% *** Deleting InstantStyle-Plus directory if it exists
if exist InstantStyle-Plus\. rd /S /Q InstantStyle-Plus

echo *** %time% *** Cloning InstantStyle-Plus repository
git clone https://github.com/instantX-research/InstantStyle-Plus
cd InstantStyle-Plus

echo *** %time% *** Creating venv
python -m venv venv
call venv\scripts\activate.bat

echo *** %time% *** Upgrading pip
python -m pip install -U pip
python -m pip install pip==24.0
pip install wheel==0.45.1
pip install setuptools==65.5.0
pip install ninja==1.11.1.4

echo *** %time% *** Installing requirements

pip install accelerate==0.32.1
pip install safetensors==0.5.3
pip install einops==0.8.1
pip install spaces>=0.19.4
pip install omegaconf==2.3.0
pip install peft==0.13.2
pip install huggingface-hub>=0.23.1
pip install opencv-python==4.11.0.86
pip install gradio==3.48.0
pip install gradio-client==0.6.1
pip install controlnet_aux==0.0.10
pip install gdown==5.2.0
pip install onnxruntime-gpu==1.22.0

pip install pyrallis==0.3.1
pip install diffusers==0.33.1
pip install transformers==4.52.4
pip install git+https://github.com/openai/CLIP.git
pip install timm==0.6.12
pip install huggingface-hub==0.33.0

echo *** %time% *** Installing GPU Torch
pip uninstall -y torch
pip uninstall -y torch
pip uninstall -y torch
pip install torch==2.7.0+cu128 torchvision torchaudio --index-url https://download.pytorch.org/whl/cu128

pip install triton-windows

echo *** %time% *** Cloning CSD_Score
git clone https://github.com/haofanwang/CSD_Score

echo *** %time% *** Downloading models
huggingface-cli download --resume-download h94/IP-Adapter --local-dir checkpoints/IP-Adapter
huggingface-cli download --resume-download TheMistoAI/MistoLine --local-dir checkpoints/MistoLine
huggingface-cli download --resume-download xinsir/controlnet-tile-sdxl-1.0 --local-dir checkpoints/controlnet-tile-sdxl-1.0

echo *** %time% *** Downloading models
cd CSD_Score
md models
cd models
rem ..\..\..\wget https://openaipublic.azureedge.net/clip/models/b8cca3fd41ae0c99ba7e8951adf17d267cdb84cd88be6f7c2e0eca1737a03836/ViT-L-14.pt -O "ViT-L-14.pt" -nc --no-check-certificate
..\..\..\wget https://huggingface.co/PixilAI/models/resolve/0e26fe9adb9ebec32645b0bdc26971a25f399f80/clip_interrogator/ViT-L-14.pt -O "ViT-L-14.pt" -nc --no-check-certificate
..\..\..\wget https://huggingface.co/datasets/Softology-Pro/VoC/resolve/main/CSD_checkpoint.pth -O "checkpoint.pth" -nc --no-check-certificate
cd..
cd..
..\wget https://huggingface.co/datasets/Softology-Pro/VoC/resolve/main/sdxlUnstableDiffusers_v8HeavensWrathVAE.safetensors -O "checkpoints\sdxlUnstableDiffusers_v8HeavensWrathVAE.safetensors" -nc --no-check-certificate

call venv\scripts\deactivate.bat
cd..

echo *** %time% *** Finished InstantStyle-Plus install
echo.
echo *** %time% *** Scroll up and check for errors.  Do not assume it worked.
pause

[zmyhh]

Has the correct download for the sdxlUnstableDiffusers_v8HeavensWrathVAE model been resolved?