same, is there another repository?

nvm found solution down the comments

@@ANerdyNub hey mate could you please share the solution you found i'm struggling with the same issue thanks

@@loumyguerin8226 ah this was a while back don't really remember but pretty sure I found another repository and replaced the one that's not working with that, try searching for it on Google

@@loumyguerin8226 oh this was a while ago, I think I just replaced the repository that didn't work with another k-diffusion-sd that did work in the installation script

I'll be looking at the negative prompt, apparently it could just be 1 line of code changed, but the ( + ) [ - ] does work, I've been testing that all morning. It might no be to the extreme that other stable-diffusion version due to the amount of steps. 20 to 40 is pretty low compared to 50 to 150 in a fraction of the time on a gpu build.

@@darkhemicdigitalart3433 Any update?

Hey, Do you remember how you added models to this? Kinda new to things and I'm wondering the process.

@@fengroukan9148 , I'm still an hour away from home to write a proper answer but I can write you a step-by-step guide tomorrow, somewhere around my lunch time (portuguese 13h Europe/lisbon time zone). While you wait, though: if i remember, stable diffusion will need to find the files in the snapshots folder of a given release. Models on windows will only get symlinked if you install as an administrator so it takes a bit of patience to get things running.

@@tanguero2k7 I appreciate that, currently just trying to get it to run out of spite at this point via piecemeal upgrades to problems that arent mentioned. Like k diffusion and needing torch.

Sorry, @@fengroukan9148, life happened 🙂and it was totally impossible to spare any time to address the tutorial as I planned. Restoring my SDCO setup from a backup, will get back to you asap.

have you tried?

I have the same issue. Any idea how to fix?

for me it says ModuleNotFoundError: No module named 'pytorch_lightning'

i have been trying aswell and no success so far

Have either of you tried rename swapping the files and calling your new model just model.ckpt. I've only started looking at this in other stable diffusion repos and that seems to be what they are doing, only calling 1 model at a time and reloading stable-diffusion after the model swap

@@darkhemicdigitalart3433 do you have discord? id like to see if we can figure this out id love to help test this

@@darkhemicdigitalart3433 and to answer your question yes i took the .ckpt file and renamed it and it still works

I've experienced that myself when the image resolution is too high or I've batched too high. I'm going to be make a video on how I've been using it so it doesn't crash like that

Glad to hear that.

I'm looking into adding features. So far it's near impossible to change up some code from other stable-diffusion repos so looks like I'm going to try to add them to my repo but I'm not making promises as I'm doing this in my spare time.

the link is bad, download module from another repository, dont need github usr/pass

So did it work for you? I can't get past the TypeError: __init__() got an unexpected keyword argument 'source', installed K-diffusion and Transformers manually and still no luck.

Same here

Hola, pudiste lograrlo?

were you able to do it?

you didn't copy the model from hugging face into the stable diffusion directory, do that and run the install script again, should fix the issue

@@darkhemicdigitalart3433 hello, can I use other models on this version of diffusion? Like using waifu diffusion model instead

@@lenovolenovo6530 I've only tried the 2 stable-diffusion models the one that's 4gb vs 7gb.

@@darkhemicdigitalart3433 hello, thank you for all your work! i have a lil problem when i run ". un_sdco.bat", it says "ModuleNotFoundError: No module named 'pytorch_lightning'". Is there any way i can fix that

you aren't in the stable diffusion directory and are you adding bash -i as well

@@darkhemicdigitalart3433 Apologies, I missed a step in the error. I'm working with Windows and I do the whole Cdrive stable-diffusion-cpuonly-main. un_sdco.bat step of the process, then get this appears on command line in the Anaconda window: call conda activate sdco EnvironmentNameNotFound: Could not find conda environment: sdco You can list all discoverable environments with `conda info --envs`. Then it automatically tries to do the step listed above and produces that error.

Did you cd stable-diffusion-cpuonly-main then bash -i ./run_sdco.bat?

@@darkhemicdigitalart3433 Well, that's embarrassing. Somehow my brain kept cd-ing to the wrong area, then never thought to look at it. Sometimes I hate myself :) Anyway, thanks for the help and great video. Everything is now working on my end. Great Stuff!

when did he install stable diffusion

You don't have enough RAM, and it says you tried to use 54mb of ram so perhaps maybe add ram to your pc

are you try this???

Honestly depends on your setting and cpu. Laptop seems to be about 10 minutes for a 512x512 at 20steps. Desktop is about 4 to 5 minutes

I can't say for sure as I've never run the openvino version. I just installed psensor to check the temps. My temps are around 65 Celsius with it stating 59 min and 72 max. It spikes up once in a while to 72 but drops almost immediately and i'm on a ryzen 7 2700 with a stock dell cooler, which has a max temp of 85 per the internet gods. I hope this info helps.

I haven't managed to get the openvino version working at all, tried for hours on end but i kept running into errors on top of errors. I'm running this one on a ryzen 5 2600 overclocked to 4.2 Ghz with a 240mm AIO and it's staying in the low 60's most of the time.

@@Eminic112 Gotcha, But I am using my GTX 4GB 1650

@@Tony-eo8zz how do you do that? Sorry I'm not very experienced

Ever find an answer? I'm running into the same issue. Says the repository doesn't exist. :/

It runs on my wife's ryzen 5 laptop just fine, it's completely on the cpu, I think the real-esgan is using vulkan.

@@darkhemicdigitalart3433 I just got it working, turns out I had the wrong version of Anaconda, though it works fantastically! Thank you so much for this, this is a blessing to tons of people without the power to run it on their GPU, whether they have a GPU that's not strong enough, or don't have CUDA due to AMD! Amazing work!

@@darkhemicdigitalart3433 There is a couple more questions I had, would it be possible to use the modified models with this program, ones that have been trained differently than the base Stable Diffusion? A way I thought it could be possible would be renaming the other modified model to the same filename as the base Stable Diffusion checkpoint, would that work? Is there any possible way I could get more than one model that I could switch between and run with Webui? Also, I was wondering if it would be possible to still share the link to friends like in the original version of WebUI?

can you describe how you made this work? I installed k_diffusion, but still get the same error when running the command . un_sdco.bat

@@ordfrog For me it was showing a module not found error, I ran the command 'pip install k_diffusion' and after that it started giving me a different module not found error. I installed each module it couldn't find with pip, and I was able to run it. Are you seeing something similar, or different?

@@JadeLion1235 i ran install k_diffusion command from PowerShell and even tried with older versions without getting any other results. What python, torch, anaconda versions do you use and have you used all other files as described and linked? Thanks for your reply!

@@ordfrog I downloaded the latest Anaconda version, python 3.10.9, pytorch 2.0.0 build py3.10_cpu_0 I downloaded both model files and placed them as described in the video before running the install script. Not sure if it would make a difference, but I did not use the anaconda powershell, just the regular anaconda prompt Hope you can get it working!

I'm having the same issues with the k-diffusion. I also installed manually using the pip command mentioned earlier, but still the same problem of the module not found. Has anyone solved this?

use anaconda powershell prompt to execute the install, not windows command prompt

do you think it work for low Ram???

CD to your stable-diffusion directory

@@darkhemicdigitalart3433 I was in the ~/stable-diffusion-cpuonly when I ran it, but tried again to be sure. but still get the same message

@@jazzjazzy did you install anaconda or just used the python that comes with mint?

@@darkhemicdigitalart3433 yeah install Anaconda version 2202.05 with Build py39_0

@@darkhemicdigitalart3433 I have the same problem

I solved it by using "pip install chardet"

Followed the guide, and got this as an output. Not really certain how to fix.