AI Art Generation Handbook/List of Issues and possible Solutions

Here is the list of the problems that might be faced during installations and/or during normal usage of Stable Diffusion

P.S 1 : Press "Ctrl" + "F" in your browser to see if you facing certain issues for fast searchig

P.S: If there are other error / problems that you faced during the installing / generating images, you can add your own answer here. Or put it in the Discussion tabs.

General Use

(1) CUDA Out of Memory

	Descriptions
Affected Web-UI / System	Automatic 1111 / Windows
Error message	Depending on WebUI versions: `RuntimeError:CUDA out of memory` `RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cpu and cuda:0! (when checking argument for argument index in method wrapper__index_select)`
Description	When running Stable Diffusion to generate the first image, generated images would not show at all and show "Error" at where the image supposedly be appeared. When open the command shell window, error message will come out
Workaround /Solution	(i) Ensure the hardware requirements of VRAM (Not RAM) is 4GB and above. Optimally, recommended to buy any graphic card that supported 6GB VRAM and above. If you decide want to train the model using Dreambooth , suggest to use Google Collabs or buying the graphic card that supported 12GB VRAM or above.\| (ii) In the webui-user.bat file, right click to edit the files using notepad. At the line where set " `COMMANDLINE_ARGS =` " , add in these parameters " `--xformers` " and " `--medvram` " and " `--opt-split-attention` " to reduce further the VRAM needed BUT it will added the processing time.

(2) Images does not show up on image view panels after generated images

	Descriptions
Affected Web-UI / System	Automatic 1111 / Windows
Error message	`AttributeError: 'Blocks' object has no attribute 'temp_file_sets'` on the command shell window
Description	This error message appeared during the update of Automatic 1111 (Jan 2023 update) with Dreambooth installed successfully where the image generated are not shown on the image preview gallery on the right panel. Further troubleshooting revealed that there are error message (shown above) that are suspected to be root cause of this issues
Workaround /Solution	At the moment, to review your imgages , need to go to C:Your_Installed_Locations\stable-diffusion-webui\outputs to manually review the generated AI images. Pending for Automatic 1111 to fix the issues

(3) Web-UI did not detect base model for AI art generations

	Descriptions
Affected Web-UI / System	Automatic 1111 / Windows
Error message	`No checkpoints found. When searching for checkpoints, looked at:` `- file C:\Your_Installed_Locations\stable-diffusion-webui\model.ckpt` `- directory C:\Your_Installed_Locations\stable-diffusion-webui\models\Stable-diffusion` `Can't run without a checkpoint. Find and place a .ckpt file into any of those locations. The program will exit.` on the command shell window
Description	This error message appeared during the installation of Stable Diffusions for the first time on the system itself.
Workaround /Solution	Download the models from the link here https://huggingface.co/runwayml/stable-diffusion-v1-5/tree/main and copy the models (.ckpt) files to the `C:\Your_Installed_Locations\stable-diffusion-webui\models\Stable-diffusion`

(4) Unable to generate images - TypeError: 'NoneType' object is not callable

	Descriptions
Affected Web-UI / System	SD.Next with SDXL Model/ Windows
Error message	`TypeError: 'NoneType' object is not callable`
Description	This error message appeared when want to generate AI images with the images unable to be generated
Workaround /Solution	Disable extensions named `a1111-sd-webui-lycoris` by untick it on the left side and update to latest patch whenever available (Aug 2023)

SDXL related

(1) Web-UI generates only black images (For SDXL models)

	Descriptions
Affected Web-UI / System	SD.Next with SDXL Model/ Windows
Error message	`Error loading model weights:` `C:\Your_installed_directory\SD.Next\automatic\models\Stable-diffusion\sd_xl_base_1.0.safe` `tensors Error(s) in loading state_dict for LatentDiffusion:` `size mismatch for model.diffusion_model.input_blocks.4.1.proj_in.weight: copying a` `param with shape torch.Size([640, 640]) from checkpoint, the shape in current model is` `torch.Size([640, 640, 1, 1]).` `size mismatch for` `model.diffusion_model.input_blocks.4.1.transformer_blocks.0.attn2.to_k.weight: copying a param` `with shape torch.Size([640, 2048]) from checkpoint, the shape in current model is` `torch.Size([640, 768]).`
Description	When you want to try the latest Stable Diffusion SDXL model, it will just generate black images only
Workaround /Solution	On the tab , click on `Settings` top tab , `User Interface` at the right side , scroll down to the `Quicksettings list` . Add in by typing `sd_model_checkpoint` , `sd_model_refiner` , `diffuser pipeline` and `sd_backend` . Scroll back to top to click on `Apply Settings` .and then click `Restart Server` button . (But highly suggest to close all of terminal windows ) and re-click webui.bat Once clicked, you should be re-directed back to the main page .

(2) Web-UI generates only cartoon abstract art / line art image although you try to generate photographs like image

	Descriptions
Affected Web-UI / System	Auto1111 with SDXL Model, SD.Next with SDXL Model/ Windows
Error message	N/A
Description	When you want to try the latest Stable Diffusion SDXL model, it will just generate cartoony or at worst images
Workaround /Solution	(i) Generate the image more than 512512px size (See this link > AI Art Generation Handbook/Differing Resolution for SDXL ) . For most optimum result, choose 1024 1024 px images (ii) Use negative prompts such as `cartoon , anime , drawings` to enhance the photorealism (if you want to achieve it)

ControlNet

(1) ControlNet Canny is generating a cartoony images

	Descriptions
Affected Web-UI / System	Auto1111 with SDXL Model, ControlNet Canny XL
Error message	`RuntimeError: You have not selected any ControlNet Model.`
Description	When you want to try to generate a Canny preprocessed but it is generating cartoonish images
Workaround /Solution	(1) Lower down the weight seems helpful (more testing needed)

(2) mat1 and mat2 shapes cannot be multiplied (154x2048 and 768x320)

	Descriptions
Affected Web-UI / System	Auto1111 with SDXL Model, ControlNet_sd15
Error message	`mat1 and mat2 shapes cannot be multiplied (154x2048 and 768x320)`
Description	Images unable to be generated and the error message RuntimeError: mat1 and mat2 shapes cannot be multiplied (154x2048 and 768x320) is shown on the webui and terminal
Workaround /Solution	(i) Ensure the Stable Diffusion Checkpoint(on top left) and ControlNet model (beside Preprocessor) should be same. (Meaning if base model is SDXL , ensure ControlNet is also same SDXL)

Captioning

(1) When using webp images , error message with O with slash above O will appear

	Descriptions
Affected Web-UI / System	Automatic 1111 / Windows
Error message	`File "C:\Your_Installed_Locations\Python\Python310\lib\encodings\cp1252.py", line 19, in encode return codecs.charmap_encode(input,self.errors,encoding_table)[0]` `UnicodeEncodeError: 'charmap' codec can't encode character '\u014d' in position 118: character maps to <undefined>` on the command shell window
Description	If you captioning on webp image format, it might appeared from time to time
Workaround /Solution	(i) Try to convert the images to png and retry captioning (ii) If errors persists, use other images instead (iii) Wait for Automatic1111 to fix the issues

Installation

(1) Pytorch / Torch Vision error when installing Auto1111

	Descriptions
Affected Web-UI / System	Automatic 1111 / Windows
Error message	`RuntimeError: Couldn't install torchstderr: ERROR: Could not find a version that satisfies the requirement torch==x.xx.x +cu yyy (from versions: none)` `ERROR: No matching distribution found for torch==x.xx.x +cu yyy` on the command shell window
Description	This error message appeared during/after the installation of Stable Diffusions for the first time on the system itself. The Web-UI unable to load and the error message closes itself. Futher troubleshooting leads to Pytorch / Torch Vision are unable to caught up to the latest version of Pyton.
Workaround / Solution	(i) Type the cmd command at the Your_Installation_Locations and type `python --version` .Check back on the software pre-requisites and see if the Python version number is higher than the recommended version. If higher, recommended to download and re-install the same version as of recommended Python version (3.10.9) or can try to dial back a few versions backward. (ii) If error persists after step (i), delete the venv folder aka `C:\Your_Installed_Locations\stable-diffusion-webui\venv` and restart the Web-UI (click webui-user.bat) to re-generate the Python Enviroment due to conflict merging Note: Pytorch development usually are not able to caught onto the latest version of Python.

Extension

(1) Extension lists is not loading

	Descriptions
Affected Web-UI / System	Automatic 1111 / Windows
Error message	`ValueError: unknown url type: ''` `[object Object]` on theweb-ui `TypeError: unhashable type: 'dict'` on the command shell window
Description	When want to add extension, the lists is empty, when it is supposedly full
Workaround / Solution	(i) Restart the whole web-ui (ii) If error persisted, check for values in Extension -> Available -> Extension index URL . If the URL field is empty , copy this field below and click Load From button : `https://raw.githubusercontent.com/wiki/AUTOMATIC1111/stable-diffusion-webui/Extensions-index.md`

(2) Extension not updating

	Descriptions
Affected Web-UI / System	Automatic 1111 / Windows
Error message	`Update: Unknown` on the web-ui shown
Description	This error is applicable if you installed dreambooth
Workaround / Solution	(i) Delete all the extensions installed in ControlnetDreamboothStableDiffusion\stable-diffusion-webui\extensions (ii) Restart the web-ui (iii) Reinstall all of previous extension but DO NOT install dreambooth

Dreambooth

As of this moment, training using Dreambooth extension for A1111 are not encouraged to be used as it is bug breaking. Recommend to use kohya-ss

(1) Dreambooth installation error / Dreambooth not appearing on the main page tabs

	Descriptions
Affected Web-UI / System	Automatic 1111 / Windows
Error message	Depending on the Automatic1111 versions, you may unable to proceed with Stable Diffusion with the following error messages `ImportError: DLL load failed while importing _ufuncs: %1 is not a valid Win32 application` `Error loading script: api.py` `Error loading script: dreambooth.py` `Error loading script: main.py`
Description	Depending on the Automatic1111 version, several symptoms may appeared: (i) Web-UI is unable to be opened at all and shows ImportError message in Command Windows (ii) After installing Dreambooth in the Automatic1111, there are no tabs "Dreambooth" appeared between Training and Setting tab.
Workaround /Solution	Re-install the whole Stable Diffusion from scratch [ See this link for step by step details ] and make sure that folder name (aka "`Your_Installed_Locations`") should NOT HAVE ANY EMPTY SPACES in the folder naming NG E.g: "Stable Diffusion Dreambooth" is (Unless you intend to generate images only) OK E.g: "`StableDiffusionDreambooth`" or "`Stable_Diffusion_Dreambooth`" is or delete the venv folder aka `C:\Your_Installed_Locations\stable-diffusion-webui\venv` and restart the Web-UI (click webui-user.bat) to re-generate the Python Enviroment due to conflict merging or pending for Automatic1111 to provide fix in future versions

(2) File Not found when installing Dreambooth

	Descriptions
Affected Web-UI / System	Automatic 1111 / Windows
Error message	`stderr: ERROR: Could not install packages due to an OSError: [Errno 2] No such file or directory: 'C:\\Your_Installed_Locations\\stable-diffusion-webui\\venv\\Lib\\site-packages\\tensorflow\\include\\external\\llvm-project\\mlir\\_virtual_includes\\BufferizableOpInterfaceIncGen\\mlir\\Dialect\\Bufferization\\IR\\BufferizableOpInterface.cpp.inc'HINT: This error might have occurred since this system does not have Windows Long Path support enabled. You can find information on how to enable this at https://pip.pypa.io/warnings/enable-long-paths`
Description	When installing Dreambooth for Stable Diffusion, the above messages may occurs
Workaround /Solution	(i) Create a .reg files on any where at the PC system (ii) Paste following codes inside the .reg file `Windows Registry Editor Version 5.00` `[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\FileSystem]` `"LongPathsEnabled"=dword:00000001` (iii) Double click the .reg file to execute (iv) Re-install the Web-UI from scratch

(3) Invalid training data directory.

	Descriptions
Affected Web-UI / System	Automatic 1111 / Windows
Error message	`Invalid training data directory.`
Description	When want to click Training button in Dreambooth tab for Stable Diffusion, the above messages may occurs
Workaround /Solution	Make sure before training, the button "Create Model" in Create tab must be clicked before proceed to train the model

(4) Pip Installation Error

	Descriptions
Affected Web-UI / System	Automatic 1111 / Windows
Error message	`Fatal error in launcher: Unable to create process using '"Your Installation Directory\venv\Scripts\python.exe" .`
Description	When installing the latest version of dreambooth which requires Torch 2, you encounter the following error while following instructions from this Github link
Workaround /Solution	It meant that your pip version is not upgraded to Torch 2 installation requirements To resolve this, type this before start the command listed on the link `python -m ensurepip --upgrade` `python -m pip install --upgrade pip`

ComfyUI

	Descriptions
Affected Web-UI / System	ComfyUI / Windows
Error message	When launching, this error message appeared at the bottom of the terminal screen `OSError: [Errno 10048] error while attempting to bind on address ('127.0.0.1', 8188): only one usage of each socket address (protocol/network address/port) is normally permitted`
Description	You encounter this message when you want to turn on ComfyUI for the first time.
Workaround /Solution	Check the web browser to see if other local UI that also uses address `127.0.0.1`is open . If open, close the tab , and re-run again.