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 INSTALL Ai-TOOLKIT WAN 2.2 LORA TRAINING! (Patreon)

Published:
2025-08-31 18:17:22
Imported:
2026-06
Tags:

Downloads

Content

Hey everyone! I've created a 1-click installer for the “ULTIMATE WAN 2.2 LORA TRAINING AI‑TOOLKIT WEBUI.” VIDEOS & IMAGES! Just fricking insane! This really is ONE of the best Ai model currently available and Ai-Toolkit makes training a Lora super ultra easy as well, basically the ultimate combination, perfect for complete beginners! With it you can train any LoRA you want for WAN: people, styles, and even full video LoRAs… it’s just nuts!

You can check out the video right here: https://youtu.be/bhBZHPvX7gI

The installer of course automates the entire install process!

LOCAL INSTALL!!! RECOMMENDED TO HAVE AT LEAST 24GB OF VRAM FOR THE WAN TEXT TO IMAGE LORA TRAINING! if it's an issue, definitely use the Runpod installer instead as recommended in the video!

Local install:

Requisites:

– CUDA 12.8 (RTX-50) OR CUDA 12.6 (RTX-40/30/20/GTX)

– Git, Python 3.10 64-bit, Node 18/20 LTS

1. Download the AI-TOOLKIT_AUTO_INSTALL.bat and place it in a folder WITHOUT any spaces in the path!!!

2. Run the bat file and select either option 1 or 2 depending on your GPU

3. Wait for the install to finish

4. ???

5. Profit 😎Use then one of the launcher once you have closed the ui and you want to restart it.

RUNPOD:

1. Create an account if you haven't already: Runpod

2. Click on Pod (on the left side) then click deploy

3. Choose a 5090 GPU (if you intend on training a text to image lora OR a text to video lora with text encoder unloaded and 512 resolutions) otherwise choose the 6000 pro blackwell GPU, and chose the pytorch 2.8.0 then edit the template and choose 100GB for both the container and volume disk, then under expose ports, edit:

8888 into 8888, 8675. Then go into environment variables below, click add and then type:

AI_TOOLKIT_AUTH on the left then type a password on the right

then scroll then and deploy.

4. Go to my pods, wait for everything to finish and then click "connect", then "Connect to HTTP SERVICE port 8888"

5. Then drag and drop the AI-TOOLKIT_AUTO_INSTALL-RUNPOD.sh file on the left side of the UI then click on the "Terminal" icon on the right side on the UI

6. Copy and paste these two lines then press enter:

chmod +x AI-TOOLKIT_AUTO_INSTALL-RUNPOD.sh

./AI-TOOLKIT_AUTO_INSTALL-RUNPOD.sh

7. Wait for everything to be installed

8. Go back to the port 8675 and click on it to launch the webui then enter your password you created previously

IF YOU WANT TO RESTART THE UI AFTER STOPPING AND RESTARTING THE POD

Use the same

chmod +x AI-TOOLKIT_AUTO_INSTALL-RUNPOD.sh

./AI-TOOLKIT_AUTO_INSTALL-RUNPOD.sh

command as you did for the install

As always, supporting me on Patreon allows me to keep creating helpful resources like this for the community. Thank you for your support - now go have some fun😉!

Files

Comments

Lawrence007

Hey I keep getting download failed at around 240 megabytes for the other wan2.2 autoinstaller.

Lance Fanger

This is great, is there an updated workflow for WAN2.2 I2V and image generation?

Constantin Mirea

when opening the launcher it says "the system cannot find the path specified" then closes... any fix for this please?

Jisa

Am I blind? I don't see the file for the local?

Pak26

My training on Runpod seems to be stuck at step 9 out of 3000. I rent a RTX Pro 6000 and kept everything by default like in the video. For Wan 2.1 it went very smooth, no freeze, so i'm kind of wondering if it's normal ? Maybe it's a little bit different for Wan 2.2 and I must be more patient !?

Pak26

After stopping the pod and restarting it, now it generate 2 identical image in the preview and also the Vram jumped to 50gb, before I restart it it was only using 27gb of vram. My guess is that now it train both High and Low and that it freezed the first time because there was not enough vram available !? The only weird thing now is that the console doesnt update but I can see that the steps are going up... so finger cross that everything seems to work now like expected !

Pak26

Just a little update! It reach step 265 then stop again.... :(

Oh Gott

Great! Will there be an updated XY-Plot workflow coming?

Pak26

Quick and last update! After many attempts trying to do a Lora with the same pod and it failed because of many weird problems like described in my other posts, I decided finally to terminate the pod I was using and just start a brand new one. This time, It succeed the first time with no errors or weird glitches. Maybe something went wrong in the installation process of comfy ui or maybe I was just using a defect pods ? Whatever...All good now ! Thanks again AItrepreneur for your awesome tutorial and amazing tools !

Luma Rose

I am waiting for that aswell. lol

Luma Rose

i know the fix for that, just use the offical AI-toolkit pod that is prepaired with everything needed so you dont even need to use console, you only need to open SERVICE port 8675, and then connect, i trained 2 different loras yesterday for WAN2.2 with 2x RTX 6000 PRO here is the link to AI-toolkit officall pod: https://console.runpod.io/explore/0fqzfjy6f3

Holy Buckeye (edited)

Comment edits

2026-09-08 20:00 Anyone else having issues running this throug the Run pod? I did exactly as the video showed, once i paste the 2 lines of code and hit Enter, it starts downloading everything then fails and just stops there??
2025-09-02 18:11

Florin-Mihai Cosma

hello, i try to finish setup, but when i try to connect to htts 8675 port, i reiceve error: Bad gateway Error code 502, what can i do? Thanks

Maranda Renolds

when using your original wan2.2 setup most of my pictures end up pretty blurry like film grain is on strong even when its not any solution?

Pak26

Try loading a Lora of around 1750~2000 steps in both High and Low and reduce the strength to 0.3~0.4 for both.

Pak26

Terminate the pod you use and try another one.

Pak26

Maybe by checking LOW VRAM and some other options like shown in his video, but It rarely go under 24gb vram during training... so i would say, probably not enough.

FractalD

Does anyone run into issues with Runpod jobs randomly hanging and unable to pause and restart? I've lost a few trainings.

Nola

Fantastic video, this one and the WAN 2.1 one. I trained a WAN 2.2 lora yesterday using this flow, all the same settings you mention in your video, and I have to say I'm not blown away by the results, at least not judging by the sample images in the AI Toolkit UI. Are the sample images supposed to look a little blurry and washed out? Further, the face isn't really matching the face of the person I trained it for. AItrepreneur's sample images certainly looked like Margot Robbie toward the final stages (I'm on 2500 of 3000 right now). I know my training dataset is good. I guess I will wait until it's fully finished, but if the face still doesn't match my person after training is complete, I'll be looking for guidance.

BeltaLoada

Yes -- this has been my exprience as well. I can't get a training to complete because they hang and are unable to be resumed. (Well sometimes they can resume, but most of the time the resumes fail) Something is wrong with this install

Jake

Hey! I’m trying to learn how to do realistic text/image to video creation. I want to upload photos of myself and create realistic, brief, video content. I’m struggling to determine if I need to run a Lora trained on myself or go straight to Wan2.2. Looking for a bit of guidance since I am overwhelmed by the amount of content! Thanks!

_i_

Runpod sessions just keep hanging up every 80-120 steps. something is broken

Ian Solo

I must say that after training my first character i am very disappointed . I got much better results with FluxGym. Can we had some steps to a finished job ?

Peter

i just recently used this and since yesterday after clicking update all, the manager is nowhere to be found in the UI. Hence, after using the autoinstall file, I can't restart it. Any help on this? Thanks!

_i_

disregard - I think I had spaced my high-to-low model swap value too high. I set it back to 1 and everything runs smoothly.

FractalD

I found out that if you try to pause or stop a job, it won't actually stop until the current iteration is finished, which could be a while. So, it seems like it's stuck but it's not.

Senpai Ashua

Can u do it for Mimic too? Or a video how to install what u need for working. Because runpod is not my favourite. Quite expensive with time. Or anyone knows how i can install all and what i n eed to install to get wan workflow working on mimicpc comfy?

VinPre

It is best if you first learn how to create Videos in general using Wan 2.2 and after that you can create a Lora of yourself. Using this Post

James

I intend to run locally on my 4090. Appears to be a standalone install. Most of my python direct interaction is under comfuyi standalone, but I also have separate miniconda environments to support ST, etc. Do you suggest setting up a new conda environment before running your standalone bat installer (with your specified python version, and cuda 12.8 is already on my base Win 11 system). Anything I should know before jumping in and trying it?

Morbid

it is unreliable my 10 $ wasted for nothing you cant leave the pc or you wont know if its running or not ostris official is better n reliable

Nils Johan Anti

theres some issue when running this locally where it doesnt set up a venv the correct way and will fetch torchaudio etc. from global python, paste the error code into gemini/qwen etc. (not chatgpt doesnt work) and you will get a guide on getting it working

Vaivij

Hi, I get the following error running L40S on Runpod: Generating baseline samples before training Generating Images: 0%| | 0/10 [00:00 File "/workspace/ai-toolkit/run.py", line 120, in main()main() File "/workspace/ai-toolkit/run.py", line 108, in main File "/workspace/ai-toolkit/run.py", line 108, in main raise eraise e File "/workspace/ai-toolkit/run.py", line 96, in main File "/workspace/ai-toolkit/run.py", line 96, in main job.run()job.run() File "/workspace/ai-toolkit/jobs/ExtensionJob.py", line 22, in run File "/workspace/ai-toolkit/jobs/ExtensionJob.py", line 22, in run process.run()process.run() File "/workspace/ai-toolkit/jobs/process/BaseSDTrainProcess.py", line 2007, in run File "/workspace/ai-toolkit/jobs/process/BaseSDTrainProcess.py", line 2007, in run self.sample(self.step_num)self.sample(self.step_num) File "/workspace/ai-toolkit/extensions_built_in/sd_trainer/DiffusionTrainer.py", line 288, in sample File "/workspace/ai-toolkit/extensions_built_in/sd_trainer/DiffusionTrainer.py", line 288, in sample super().sample(step, is_first)super().sample(step, is_first) File "/workspace/ai-toolkit/jobs/process/BaseSDTrainProcess.py", line 367, in sample File "/workspace/ai-toolkit/jobs/process/BaseSDTrainProcess.py", line 367, in sample self.sd.generate_images(gen_img_config_list, sampler=sample_config.sampler)self.sd.generate_images(gen_img_config_list, sampler=sample_config.sampler) File "/workspace/ai-toolkit/venv/lib/python3.12/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context File "/workspace/ai-toolkit/venv/lib/python3.12/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context return func(*args, **kwargs)return func(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/workspace/ai-toolkit/toolkit/models/base_model.py", line 645, in generate_images File "/workspace/ai-toolkit/toolkit/models/base_model.py", line 645, in generate_images img = self.generate_single_image(img = self.generate_single_image( ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/workspace/ai-toolkit/extensions_built_in/diffusion_models/wan22/wan22_14b_model.py", line 542, in generate_single_image File "/workspace/ai-toolkit/extensions_built_in/diffusion_models/wan22/wan22_14b_model.py", line 542, in generate_single_image output = pipeline(output = pipeline( ^^^^^^^^^^^^^^^^^^ File "/workspace/ai-toolkit/extensions_built_in/diffusion_models/wan22/wan22_pipeline.py", line 242, in __call__ File "/workspace/ai-toolkit/extensions_built_in/diffusion_models/wan22/wan22_pipeline.py", line 242, in __call__ noise_pred = current_model(noise_pred = current_model( ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/workspace/ai-toolkit/venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl File "/workspace/ai-toolkit/venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/workspace/ai-toolkit/venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl File "/workspace/ai-toolkit/venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/workspace/ai-toolkit/venv/lib/python3.12/site-packages/diffusers/models/transformers/transformer_wan.py", line 663, in forward File "/workspace/ai-toolkit/venv/lib/python3.12/site-packages/diffusers/models/transformers/transformer_wan.py", line 663, in forward hidden_states = block(hidden_states, encoder_hidden_states, timestep_proj, rotary_emb)hidden_states = block(hidden_states, encoder_hidden_states, timestep_proj, rotary_emb) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/workspace/ai-toolkit/venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl File "/workspace/ai-toolkit/venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/workspace/ai-toolkit/venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl File "/workspace/ai-toolkit/venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/workspace/ai-toolkit/venv/lib/python3.12/site-packages/diffusers/models/transformers/transformer_wan.py", line 474, in forward File "/workspace/ai-toolkit/venv/lib/python3.12/site-packages/diffusers/models/transformers/transformer_wan.py", line 474, in forward self.scale_shift_table + temb.float()self.scale_shift_table + temb.float() ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^~~~~~~~~~~~~~~~~~~~~~~~~~~ RuntimeErrorRuntimeError: : Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu!Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu!

Onex Sys

having the same issues I guess: import torchaudioimport torchaudio ModuleNotFoundErrorModuleNotFoundError: : No module named 'torchaudio'No module named 'torchaudio' any chance on fixing that? really frustrating .. put quite some work in to preparations and setup .. and once again I am stuck :(

newtoai

Hi. Could you please fix the auto installation for offline use ?