lshqqytiger
fresh install - missing python dependencies - joblib
Fresh install issue. Stable Diffusion WebUI AMDGPU Forge on StabilityMatrix 2.15.0 (Win11, portable mode disabled).
I added modules manually according to: https://github.com/lshqqytiger/stable-diffusion-webui-amdgpu-forge/blob/main/requirements_versions.txt
Launch Trace:
WARNING: ZLUDA works best with SD.Next. Please consider migrating to SD.Next.
fatal: No names found, cannot describe anything.
Python 3.10.17 (main, May 30 2025, 05:32:15) [MSC v.1943 64 bit (AMD64)]
Version: f2.0.1v1.10.1-1.10.1
Commit hash: 96ad3a736bd616460a86dc43cd5672acd9017e6f
ROCm: agents=['gfx1150']
ROCm: version=6.2, using agent gfx1150
ZLUDA support: experimental
ZLUDA load: path='C:\Users\d\AppData\Roaming\StabilityMatrix\Packages\stable-diffusion-webui-amdgpu-forge.zluda' nightly=False
Launching Web UI with arguments: --skip-install --use-zluda --gradio-allowed-path 'C:\Users\d\AppData\Roaming\StabilityMatrix\Images'
Total VRAM 26989 MB, total RAM 61049 MB
pytorch version: 2.7.0+cu118
Set vram state to: NORMAL_VRAM
Device: cuda:0 AMD Radeon(TM) 890M Graphics [ZLUDA] : native
VAE dtype preferences: [torch.bfloat16, torch.float32] -> torch.bfloat16
CUDA Using Stream: False
W1005 22:29:03.730000 23200 venv\Lib\site-packages\torch\distributed\elastic\multiprocessing\redirects.py:29] NOTE: Redirects are currently not supported in Windows or MacOs.
The cache for model files in Transformers v4.22.0 has been updated. Migrating your old cache. This is a one-time only operation. You can interrupt this and resume the migration later on by calling transformers.utils.move_cache().
0it [00:00, ?it/s]
Using pytorch cross attention
Using pytorch attention for VAE
C:\Users\d\AppData\Roaming\StabilityMatrix\Packages\stable-diffusion-webui-amdgpu-forge\venv\lib\site-packages\torch\onnx_internal\registration.py:159: OnnxExporterWarning: Symbolic function 'aten::scaled_dot_product_attention' already registered for opset 14. Replacing the existing function with new function. This is unexpected. Please report it on https://github.com/pytorch/pytorch/issues.
warnings.warn(
ONNX failed to initialize: module 'optimum.onnxruntime.modeling_diffusion' has no attribute 'ORTPipelinePart'
ControlNet preprocessor location: C:\Users\d\AppData\Roaming\StabilityMatrix\Packages\stable-diffusion-webui-amdgpu-forge\models\ControlNetPreprocessor
*** Error loading script: soft_inpainting.py
Traceback (most recent call last):
File "C:\Users\d\AppData\Roaming\StabilityMatrix\Packages\stable-diffusion-webui-amdgpu-forge\modules\scripts.py", line 525, in load_scripts
script_module = script_loading.load_module(scriptfile.path)
File "C:\Users\d\AppData\Roaming\StabilityMatrix\Packages\stable-diffusion-webui-amdgpu-forge\modules\script_loading.py", line 13, in load_module
module_spec.loader.exec_module(module)
File "
2025-10-05 22:29:23,716 - ControlNet - INFO - ControlNet UI callback registered. Model selected: {'checkpoint_info': {'filename': 'C:\Users\d\AppData\Roaming\StabilityMatrix\Packages\stable-diffusion-webui-amdgpu-forge\models\Stable-diffusion\sd\epicrealismXL_vxviiCrystalclear (1).safetensors', 'hash': 'c0c5b134'}, 'additional_modules': [], 'unet_storage_dtype': None} Using online LoRAs in FP16: False Running on local URL: http://127.0.0.1:7860
To create a public link, set share=True in launch().
Startup time: 33.5s (prepare environment: 0.8s, launcher: 0.5s, import torch: 18.0s, initialize shared: 5.2s, other imports: 0.4s, list SD models: 0.1s, load scripts: 4.5s, initialize extra networks: 0.1s, create ui: 2.3s, gradio launch: 1.2s).
Environment vars changed: {'stream': False, 'inference_memory': 1024.0, 'pin_shared_memory': False}
[GPU Setting] You will use 96.21% GPU memory (25965.00 MB) to load weights, and use 3.79% GPU memory (1024.00 MB) to do matrix computation.
That sounds like horribly easy solution but how about you just install joblib module ? Or maybe delete --skip-install command and it will download it at startup ?
I was hoping to get a completion of the requirements file so that the installation would run smoothly on a fresh system.
The thing is the versions of packages might be different depending on which backend you use or how old GPU u use. This is why they are not forced by .txt file and usually they are installed automatically. For example i can't give you my versions since i am using torch 2.2.1 and you are using torch 2.7.0 but maybe someone else has a list of compatible packages versions.
I tried installing it on a relatively fresh system, and many dependencies (including the ‘joblib’ module) did not install automatically.
I tried installing it on a relatively fresh system, and many dependencies (including the ‘joblib’ module) did not install automatically.
Its because you use --skip-install