Home
Creators
Search Recent Random
Posts
Search DMs Popular Hash Lookup Tags Random
Importer
Import FAQ
Register Favorites Keys Review DMs Logout
Community
Telegram ThePornDude
Home Creators Posts Import Register Favorites Logout

1 Click Installer for Automatic1111 SD Web UI, SDXL, ControlNet, All ControlNet Models, TensorRT (RTX Accelerator) For RunPod / Any Linux System (Patreon)

Published:
2024-06-06 13:09:00
Edited:
2024-07-29 09:21:10
Imported:
2026-06
Tags:

Downloads

Content

Join discord and tell me your discord username to get a special rank : SECourses Discord

Patreon exclusive posts index

6 June 2024 Update

Updated to Torch 2.3.1 and xFormers 26

Working flawlessly with huge speed

New IP Adapter SDXL model added ControlNet downloads to be able to do style transfer

TensorRT repo updated and works perfect with LoRAs as well

If you get AttributeError: 'NoneType' object has no attribute 'lowvram' error when changing model, try several times changing and it gets automatically fixed

This error is latest Automatic1111 error

install_latest_auto_1111 .sh and file will install the latest versions of the following extensions

sd-webui-reactor (Reactor), adetailer (After Detailer), sd-webui-controlnet (ControlNet) and Deforum

Tutorial for how to run scripts : https://youtu.be/8Qf4x3-DFf4

How to download models from CivitAI or Hugging Face or upload them full tutorial : https://youtu.be/X5WVZ0NMaTg

If you are going to use TensorRT first go to Settings, User Interface, Quicksettings list, enable sd_net and reload UI before generating TensorRT model : sd_unet_enable_TensorRT.png

TensorRT example config : tensor_RT_settings.png

Latest installer files : runpod_auto1111_v19.zip

Download the zip file and extract into any folder.

These scripts are prepared for RunPod but they would work on any Linux system. It can be local or Cloud. Just change the paths accurately.

The install commands and full instructions are provided at the runpod_instructions_READ .txt inside the attached zip file.

If you register RunPod with my link I appreicate : https://runpod.io?ref=1aka98lq

Login your pod you can use this link : https://runpod.io?ref=1aka98lq

If you don't know how to setup a Pod here written tutorial : 

https://medium.com/@furkangozukara/how-to-deploy-a-pod-on-runpod-and-verify-it-is-working-20e47031c0b5

Full ControlNet Tutorial : https://youtu.be/3E5fhFQUVLo

How To Install On RunPod

Select your GPU. I think best price / performance is RTX 3090

Select RunPod template runpod/stable-diffusion:web-ui-10.2.1

You can use either community cloud, private or network storage

Customize deployment

Make container disk 20 GB and Make volume disk 150 GB - or any size you want according to your needs

Wait until you can click Connect and Connect to Jupyter Lab [Port 8888]

Enter inside stable-diffusion-webui folder

Upload relauncher.py and overwrite file. Then restart your pod. This is mandatory only 1 time.

This is important to kill auto started Automatic1111 Web UI instance and start our own from terminal.

Then upload everything into workspace folder.

If you want to install everything and download everything run below command

Install all command is suggest one

export HF_HOME="/workspace"

cd /workspace

chmod +x install_all.sh

./install_all.sh

After installation to re-run the web UI execute below command

fuser -k 3000/tcp

cd /workspace/stable-diffusion-webui

python relauncher.py

If you use above command and install all, you don't need any of below ones since all will be installed.

Now we have different installers

The first one is install_latest_auto_1111.sh

This file will update automatic1111 to the latest version and its controlnet extension to latest version including insightface library - but will not download controlnet models yet

Then it will automatically start latest version of automatic1111 SD web ui with latest version ControlNet and the following extensions

sd-webui-reactor (Reactor), adetailer (After Detailer), sd-webui-controlnet (ControlNet) and Deforum

This file will download 275 Amazing Fooocus styles csv that you can use in Automatic1111 SD Web UI

To run it copy paste below command

export HF_HOME="/workspace"

cd /workspace

chmod +x install_latest_auto_1111.sh

./install_latest_auto_1111.sh

If you want to install only TensorRT Extension

Execute below installer script command

It will update Automatic1111 to the latest version and install TensorRT repo that supports LoRAs as well : https://github.com/Yomisana/Stable-Diffusion-WebUI-TensorRT

TensorRT not yet supporting ControlNet but they have a development branch : https://github.com/NVIDIA/Stable-Diffusion-WebUI-TensorRT/tree/controlnet_v2 

This file will download 275 Amazing Fooocus styles csv that you can use in Automatic1111 SD Web UI

The tutorial of TensorRT here : https://youtu.be/kvxX6NrPtEk

And another tutorial of TensorRT : https://youtu.be/eKnMVXVjVoU

Script to execute for install

export HF_HOME="/workspace"

cd /workspace

chmod +x install_tensorRT.sh

./install_tensorRT.sh

Let's say you want to download all controlnet models

Then you need to run below command

export HF_HOME="/workspace"

cd /workspace

python control_net_downloader.py

python download_ip_adapter_and_instantid.py

This above command will download huge amount of models 40GB+

control_net_downloader.py file will download all of the available ControlNet models that I know total number of 52.

The downloader will download into the accurate folder with accurate naming.

If you find new ones let me know

python download_ip_adapter_and_instantid.py will download ip-adapter-faceid-plusv2_sd15_lora, ip-adapter-faceid-plusv2_sdxl_lora, ip-adapter_instant_id_sdxl and control_instant_id_sdxl into the accurate folders with accurate naming. Accurate naming is crucial.

I will hopefully update scripts if there be new model releases.

Using ip-adapter-faceid and instant_id_sdxl are not straight forward. 

Please read the following threads:

Ip Adapter Face ID of ControlNet : https://github.com/Mikubill/sd-webui-controlnet/discussions/2442

InstantID of ControlNet : https://github.com/Mikubill/sd-webui-controlnet/discussions/2589

However I find that, these both models are working very poor compared to my direct implementation of their pipelines into a single Gradio app with 1 click installer.

You can download them from

1 Click Automatic Installer & A Very Advanced GUI For InstantId For Windows, RunPod, Linux & Kaggle Notebook:

https://www.patreon.com/posts/1-click-very-gui-97769887

IP-Adapter-FaceID-PlusV2 - 0 Shot Face Transfer - Auto Installer & Gradio App

https://www.patreon.com/posts/ip-adapter-0-app-95759342

Both of my apps has instructions and installers for both Windows, RunPod (Linux) & Kaggle notebooks

after this operation i suggest you to restart your web ui with following command

fuser -k 3000/tcp

cd /workspace/stable-diffusion-webui

python relauncher.py

You can also run download_models.py file

This file will download the following models

Realistic_Vision_V6 (SD 1.5 based model - one of the very best)

sdxl-vae-fp16-fix - fixed SDXL VAE in FP16 so you can run in Automatic1111 in FP16

RealVisXL_V4.0 - one of the most realistic SDXL model - training works fine

Hyper_Realism_V3 - which is the best SD 1.5 realism model that I find

To download any custom model you want, execute below command

cd /workspace/stable-diffusion-webui/models/Stable-diffusion

Then use below format

wget "download_link_of_model"

Files

Comments

indievish

Hi - is the dreambooth workflow same as before or are there any different setting/special considerations for sdxl + automatic 1111?

indievish

ok thanks - so best way right now is SDXL + Lora - does it produce better results compared to SD1.5+Dreambooth ?

SECourses: FLUX, Tutorials, Guides, Resources, Training, Scripts

depends on your purpose. if you are aiming realistic face i think my dreambooth workflow still best. but for all other things probably SDXL lora better. such as styling or having better clothing. SDXL lora is still good and many times better than SD 1.5 LoRAs

Ish Hodaszi

Hello. Thanks for the guides. Is there any way to use the refiner in A1111 currently?

Wicked Luxx (edited)

Comment edits

2026-09-07 10:00 Hello. Can I use LoRAs and extensions from SD on SDXL?
2023-08-10 16:19

Philipp Capetian

Hi. Thanks for the great ressources. However after following the steps whenever connecting through Runpod with the service port, the webUI is stalling at the loading step (the orange squares freeze)

SECourses: FLUX, Tutorials, Guides, Resources, Training, Scripts

hello i tested again and working great. so do this. get a new pod. dont use rtx 4090 (it is not working full performance). use rtx 3090 or similar pod. upload relauncher.py into web ui folder and overwrite relauncher.py. restart pod (only 1 time necessary). then use the auto installer. after that use the start command. dont forget to connect with connect button on pods section and use Connect to HTTP Service [Port 3001] .

Rafał Ryniak

"RuntimeError: Detected that PyTorch and torchvision were compiled with different CUDA versions. PyTorch has CUDA Version=11.7 and torchvision has CUDA Version=11.8. Please reinstall the torchvision that matches your PyTorch install."

Rafał Ryniak

maybe you just need to change the access path

Rafał Ryniak

started, offline work was selected in the settings, again reply from space ;)))

jaehyun AI

It's great to see you here. I learned a lot of the skills I need to create AI-related content from your videos, so I apologize for not subscribing sooner. Good luck with your future endeavors.

IOVANE Tchavtchanidze

hello, reference_only is not working for SDXL do you know any other alternative i can use?

Sana (edited)

Comment edits

2026-09-07 10:00 /workspace/stable-diffusion-webui/models/V 100%[=======================================================================================>] 319.14M 109MB/s in 2.9s 2023-10-05 07:45:59 (109 MB/s) - ‘/workspace/stable-diffusion-webui/models/VAE/sd_xl_base_1.0.safetensors’ saved [334641164/334641164] Traceback (most recent call last): File "/workspace/stable-diffusion-webui/relauncher.py", line 26, in relaunch_process() File "/workspace/stable-diffusion-webui/relauncher.py", line 9, in relaunch_process while (n<1): NameError: name 'n' is not defined root@8210406ec097:/workspace# Getting this error
2023-10-05 07:50

Joel Maynard

does this tensort support sdxl lora?

Joel Maynard

ERROR:root:Failed to add Constant /out/out.0/Constant_2_output_0 Finished refit. Dumping result to disk. Traceback (most recent call last): File "/home/venv/lib/python3.10/site-packages/gradio/routes.py", line 488, in run_predict output = await app.get_blocks().process_api( File "/home/venv/lib/python3.10/site-packages/gradio/blocks.py", line 1431, in process_api result = await self.call_function( File "/home/venv/lib/python3.10/site-packages/gradio/blocks.py", line 1103, in call_function prediction = await anyio.to_thread.run_sync( File "/home/venv/lib/python3.10/site-packages/anyio/to_thread.py", line 33, in run_sync return await get_asynclib().run_sync_in_worker_thread( File "/home/venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 877, in run_sync_in_worker_thread return await future File "/home/venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 807, in run result = context.run(func, *args) File "/home/venv/lib/python3.10/site-packages/gradio/utils.py", line 707, in wrapper response = f(*args, **kwargs) File "/home/stable-diffusion-webui/extensions/Stable-Diffusion-WebUI-TensorRT/ui_trt.py", line 262, in export_lora_to_trt engine.refit(onnx_base_path, onnx_lora_path, dump_refit_path=trt_lora_path) File "/home/stable-diffusion-webui/extensions/Stable-Diffusion-WebUI-TensorRT/utilities.py", line 277, in refit save_file( File "/home/venv/lib/python3.10/site-packages/safetensors/numpy.py", line 74, in save_file if not _is_little_endian(tensor): File "/home/venv/lib/python3.10/site-packages/safetensors/numpy.py", line 167, in _is_little_endian byteorder = tensor.dtype.byteorder AttributeError: 'NoneType' object has no attribute 'dtype'

Joel Maynard

I am getting this error for sdxl lora

Brandon Waters

Relauncher.py is just for runpod right? if we are installing tensor rt on our local linux machine, we just need to install the 1_click and the install_tensorRT.sh script? Any particular order we need to install them in?

Vasiliy Bulanov

Hi! Can you give an advise about running Tensors with custom Lora? TensorRT Lora tab model loading seems to have no effect. And stylasation is disappearing while using it. I'm developing an indie product and Tensors can really lower my buget on large scale. I would apreceate if you can give me a hint:)

Shavel

Hello my Runpod has worked for weeks and suddenly it is giving me this error for 2 days. Happens with creating new pods also. 2023-11-18T17:48:54Z create pod network 2023-11-18T17:48:54Z create container runpod/stable-diffusion:web-ui-10.2.1 2023-11-18T17:49:02Z pending image pull runpod/stable-diffusion:web-ui-10.2.1 2023-11-18T17:49:09Z error pulling image: Error response from daemon: Get "https://registry-1.docker.io/v2/": dial tcp: lookup registry-1.docker.io on 127.0.0.53:53: read udp 127.0.0.1:43926->127.0.0.53:53: i/o timeout Have you seen this before?

Nenad Kuzmanovic

You are very dedicated to your work... I really appreciate and respect that. I think you are by far the best AI content tutorial creator on the internet right now. Don't change anything in your work... I learned a lot from you and our early patreon hangout we had when I was at begining of this journey. Since then I have learned a lot, especially training models that are not the face of a real person. Right now I'm blown away by the results I'm getting in Kohya training only OUT blocks in LORA. I assure you that IS THE ONLY WAY to properly teach a style... There is no other way, cause every single tutorial on YT that explain how to make Style model is WRONG. In their way you just getting Concept and not style.. A whole unexplored universe is hidden behind those settings. I feel like I'm just starting to learn Latent Diffusion. If you want, I can send you the results of those trainings, with dataset images, etc... As a sign of gratitude, because you were my first teacher..

Rylan Czuczman

Hey thanks for these details - is it possible to use ReActor with this installation please? Struggling to find anything online related to RunPod

Rylan Czuczman

I am however getting the following error for any kind of generation (the instructions don't fix it in the error code) : 💾 🗃️ 🖼️ 🎨️ 📐 NansException: A tensor with all NaNs was produced in Unet. This could be either because there's not enough precision to represent the picture, or because your video card does not support half type. Try setting the "Upcast cross attention layer to float32" option in Settings > Stable Diffusion or using the --no-half commandline argument to fix this. Use --disable-nan-check commandline argument to disable this check.

Jorgcamacho

Hi, I think there is no way to update controlnet to use it for the XL version.

Jorgcamacho

is it because of the tensor that controlnet cannot be used for XL?

Felix Rockwell

Hello, if I try to load my file I get this error message on Runpod Stable Diffusion: NansException: A tensor with all NaNs was produced in Unet. This could be either because there's not enough precision to represent the picture, or because your video card does not support half type. Try setting the "Upcast cross attention layer to float32" option in Settings > Stable Diffusion or using the --no-half commandline argument to fix this. Use --disable-nan-check commandline argument to disable this check. Do you know what should I do now?

Pallavi Chauhan

Hi, I followed your Tensor RT installation tutorial for RunPOD. I am not getting Tensor RT option in my stable diffusion webUI. Any clue what might be the problem? driver info shows ths - NVIDIA-SMI 520.56.06

Arcon Septim

Can you please explain how to do clean reinstall so we don't lose model data

Ton Nattapong

Hi, I'm reaching out to seek assistance for a persistent error I've been encountering with a CUDA-related process. Here is the error message: Traceback (most recent call last): File "/workspace/venv/lib/python3.10/site-packages/gradio/routes.py", line 488, in run_predict output = await app.get_blocks().process_api( File "/workspace/venv/lib/python3.10/site-packages/gradio/blocks.py", line 1431, in process_api result = await self.call_function( File "/workspace/venv/lib/python3.10/site-packages/gradio/blocks.py", line 1103, in call_function prediction = await anyio.to_thread.run_sync( File "/workspace/venv/lib/python3.10/site-packages/anyio/to_thread.py", line 33, in run_sync return await get_asynclib().run_sync_in_worker_thread( File "/workspace/venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 877, in run_sync_in_worker_thread return await future File "/workspace/venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 807, in run result = context.run(func, *args) File "/workspace/venv/lib/python3.10/site-packages/gradio/utils.py", line 707, in wrapper response = f(*args, **kwargs) File "/workspace/stable-diffusion-webui/modules/call_queue.py", line 77, in f devices.torch_gc() File "/workspace/stable-diffusion-webui/modules/devices.py", line 61, in torch_gc torch.cuda.empty_cache() File "/workspace/venv/lib/python3.10/site-packages/torch/cuda/memory.py", line 133, in empty_cache torch._C._cuda_emptyCache() RuntimeError: CUDA error: an illegal memory access was encountered CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1. Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.

Ton Nattapong

and this error *** Error running process_batch: /workspace/stable-diffusion-webui/extensions/Stable-Diffusion-WebUI-TensorRT/scripts/trt.py8it/s] Traceback (most recent call last): File "/workspace/stable-diffusion-webui/modules/scripts.py", line 742, in process_batch script.process_batch(p, *script_args, **kwargs) File "/workspace/stable-diffusion-webui/extensions/Stable-Diffusion-WebUI-TensorRT/scripts/trt.py", line 302, in process_batch if self.idx != sd_unet.current_unet.profile_idx: AttributeError: 'NoneType' object has no attribute 'profile_idx'

Dmitry

Hello, thank you very much for your amazing work helping to solve a lot of problems. I used the installation through your script, everything works fine, but recently an update for controlnet was released, instant_ID appeared in it https://www.youtube.com/watch ?v=pd4EY5udcF8 Allows you to create art based on photos that are not inferior in quality to Laura. And instant_id doesn't work. Could you please see what the problem might be and how this problem can be solved?

Dmitry

I managed to solve the problem via "pip install insightface" in the venv folder

Franco Acosta Diaz

Hi, I really like the content of your patreon, thank you very much, I have already downloaded Kohya in Runpod, now I have this. I would like to know if with this command: export HF_HOME="/workspace" cd /workspace chmod +x install_all.sh ./install_all.sh I download everything and all is done and finish. Or do I still have to download the rest with these other commands: export HF_HOME="/workspace" cd /workspace chmod +x install_latest_auto_1111.sh ./install_latest_auto_1111.sh .... ETC.

Franco Acosta Diaz

Hello, I have another question. About the models that are installed, should I choose one of them to use as a base model for training (in this case realistic). But if so, what happens with SDXL, I don't understand?

Jorgcamacho

Hello, in this last update controlnet still does not work in XL, right?

Jorgcamacho

but I have been testing them and I have not been able to generate the hands well.

100Sachen80

Hi Mate, do you plan to update the Windows Version as well? Maybe there is also sth. new to try?

Jorgcamacho

i have a doubt, tensor rt does not work with controlnet during inpaint, right? before i could see the error message but i did not take a screenshot of it.

Dmitry

Hello, you wrote in the instructions that if I run the command: "export HF_HOME="/workspace" cd /workspace chmod +x install_all.sh ./install_all.sh" Then everything that is needed will be installed, but I did not find a line in the script that would run the file: "install_latest_auto_1111.sh" Is this how it should be and is the code with the automatiс1111 update in the py files? I got acquainted with the code from the moment I started watching your videos, do not judge strictly if I do not understand something)

Dmitry

Hello, I have not installed 1111 for a week, now I have installed it and noticed that the preprocessor "instant_id_face_keypoints" no longer works Could you check why and how it can be fixed? Is it possible that this is related to upgrading A 1111 to version 1.9? I tried switching to version 1.8, but it didn't work This is what I get when I try to process an image through a preprocessor: from pydantic import Field, ValidationInfo, field_validator, model_validator ImportError: cannot import name 'ValidationInfo' from 'pydantic' (/workspace/venv/lib/python3.10/site-packages/pydantic/__init__.cpython-310-x86_64-linux-gnu.so)

Art

Hi Furkhan... how to get the same installation (with adetailer and other extensions) on MassedCompute? If I'm not wrong, I read somewhere (that now I can't reach out) on your patreon that there is a command to one-click-install all the useful extensions on the A1111 base version of MC (without having to install manually one by one and restart the app)...

Dmitry

Yes, I use contorl net. However, for other people, everything works as it should. Perhaps the fact is that in your installation script controlnet is updated first and then A 1111 is updated. In addition, some other processors may not work with this installation either, please check if this is related to the update order in the script

linabrooksai (edited)

Comment edits

2026-09-07 10:00 Hello, unfortunately have error with checkpoint window and vae window (it does not appear when I make checkpoint/vae changes), also I have lowvram eror and it does not work when I reload or etc
2024-04-21 22:26

Dmitry

Hello, something happened. After updating A1111(used only the script from the file "install_latest_auto_1111.sh") - A1111 stopped running. Please see what this might be related to. Perhaps you should make a script that installs A1111 from the repository and creates your own venv, rather than using the one that was in the template? UPD Looks like I solved the problem. Added to file "install_latest_auto_1111.sh" after the line "44" pip install -r requirements.txt

Shepard2k

Greetings, when i use your Autoinstaller I'm getting this error: " Cloning BLIP into W:\Auto1111_V8\stable-diffusion-webui\repositories\BLIP... Cloning into 'W:\Auto1111_V8\stable-diffusion-webui\repositories\BLIP'... remote: Enumerating objects: 277, done. remote: Counting objects: 100% (183/183), done. remote: Compressing objects: 100% (46/46), done. remote: Total 277 (delta 145), reused 137 (delta 137), pack-reused 94 (from 1) Receiving objects: 100% (277/277), 7.04 MiB | 23.24 MiB/s, done. Resolving deltas: 100% (152/152), done. Traceback (most recent call last): File "W:\Auto1111_V8\stable-diffusion-webui\launch.py", line 48, in main() File "W:\Auto1111_V8\stable-diffusion-webui\launch.py", line 39, in main prepare_environment() File "W:\Auto1111_V8\stable-diffusion-webui\modules\launch_utils.py", line 422, in prepare_environment if not requirements_met(requirements_file): File "W:\Auto1111_V8\stable-diffusion-webui\modules\launch_utils.py", line 290, in requirements_met import packaging.version ModuleNotFoundError: No module named 'packaging' any idea how to fix this?

Shepard2k

oh I commented on the wrong post, It's a local install here is the video with the error at the end. I reinstalled everything following your last video. Despite everything thank you for your quick response and your hard work overall https://youtu.be/56UI5utd6SY?si=_-KSRk0IMO6nwJP4

Shepard2k

yes it's mountable drive, i added it as a safe.directory to git but i guess that isn't enough. now I did a test installation on a normal drive and it worked, but i don't understand why to be honest. you think it's possible to install everything and then move the folder into my mountable drive? guess i try

Shepard2k

worked ;) thank you