How to Install LORA Models for SDXL in Fooocus Under 30 Seconds

In this video, we dive into LORA models inside Foocus, an AI image generation tool that runs Stable Diffusion including SDXL models. As mentioned, it takes only 43 seconds 😉
My other contents at x.com/DebarghaDB
DM me at / debargha_bhattacharjee
Drop an email to [e-mail]

[@donvarsmak5688]

Im having an issue when trying to use downoaled sdxl models:
AttributeError: 'NoneType' object has no attribute 'model'

[@DavidScott-yj5vr]

where is the foocus folder found on a mac?

[@139660713]

But how to install those models?

[@VoodooD0g]

bro

[@nourghafarji]

very usefull thanks

[@windows13zumba-od7lu]

Thank you, I appreciate it very much!!!

[@Debargha_Bhattacharjee]

Glad it helped!

[@zerethins]

How did you install the LORAs? In the video, which folder did you put them in? Is there anything else that needed to be installed?​@@Debargha_Bhattacharjee

[@Debargha_Bhattacharjee]

@@zerethins You have to install the LORA extension first. In newer models of Automatic 1111 the extension comes preinstalled, so that saves the installation part. You have to put the models inside the 'Lora' folder under 'models'. Nothing else is necessary.

[@blacksage81]

I did a six month sentence in Comfy as I wanted to text SDXL models and Auto1111 would offer me wait times of up to an hour, where comfy had it done it about 90seconds(depending on the workflow)

[@asciikat2571]

nice one man

[@___x__x_r___xa__x_____f______]

, super interesting and very well designed video. also. How does one use the already downloaded models from Auto1111 in Fooocus in order not to take more disk space?

[@Debargha_Bhattacharjee]

You can edit user_path_config file to change the target locations for accessing files stored at another location.

[@gruberjens4354]

Is Fooocus only compatible with sdxl loras? SD 1.5 loras too?

[@kanavwastaken]

Yep

[@snasxyyojhael5064]

@@kanavwastaken Oh my gosh no wonder the LORAs I've downloaded never worked! Cause all of them were SD 1.5. I downloaded an SDXL LORA and it actually worked.

[@axelislam8069]

where did u download the SDXL LORA plz help@@snasxyyojhael5064

[@aliasad1618]

i see what you did with the lora pronunciation dude xD

[@youua]

In what language he speaks?

[@HeartSiddharth]

how do i use Fooocus in a discord bot with slash command
can you build a video on it "?

[@Debargha_Bhattacharjee]

You can use Playground AI for similar functions. It also uses SDXL just like Foocus.

[@nicolofais6865]

HI! it was the video I was looking for, in fact I didn't understand why my models, despite posting the images, all looked like relatives or descendants haha Question: can all this be done for Android?
Thank you!

[@Debargha_Bhattacharjee]

I doubt android phones can ever do all these. These functions require powerful GPUs. You can use websites like Leonardo ai, playground ai or something like runpod/rundiffusion etc to run stable diffusion online. That way you can do all these via your phone or tablet.

[@SuraGamingZone]

👍

[@LittleDiamond]

Hello, where are models and lora files when you use Fooocus online? Thanks :)

[@Debargha_Bhattacharjee]

By online do you mean via Runpod or Rundiffusion?

[@DoLundThump]

​@@Debargha_Bhattacharjee he means Google Collab

[@Debargha_Bhattacharjee]

@@DoLundThump I haven't used google collab for this purpose. Sadly, I can't help him in that regard.

[@benjaminbv8046]

How do I do it in Google Colab?

[@Debargha_Bhattacharjee]

As per my knowledge, google collab has restricted the use of ai generation models as of now.

[@marcosangelmedina5845]

I also work on TCS :D hehehe just kidding

[@phalxor]

what?

[@lucfrost1554]

Good video, although, try to speak a bit slower. Without the subs, its hard to get everything you said.

[@Debargha_Bhattacharjee]

Thank you. I have been working on that for some time. I hope it's gotten a little better in my recent videos.

[@lucfrost1554]

@@Debargha_BhattacharjeeThanks you also taking my comment for what it is; a constructive one. Visually its perfect and easy to understand. Keep on the great work.

[@Hobbies-and-Experiments]

That civit is terrible i was there for 1 min and had to leave its got like 20 ppl all speaking at one, none of what you do here is possible now, and it's not free, 100 credits is what you get and i spend none as its a mess.

[@Debargha_Bhattacharjee]

Yeah it has recently gotten like that. I miss the good old days.

[@DoLundThump]

Bro credits are for generation, but downloading models is free

[@DoLundThump]

Bhai mai bhi Indian hu lekin kasam se aap ki english bilkul samajh nahi aati

[@tristanthedarkstar]

You speak and do stuff too fast in this video. Can barely understand your English.

[@Debargha_Bhattacharjee]

Thanks for pointing out. I have been trying to improve that part for quite some time now. Hopefully in my future videos it would be much better.

[@tristanthedarkstar]

@@Debargha_Bhattacharjee thanks for the reply. I am sure you'll improve. Not a criticism--only an observation.

[@Debargha_Bhattacharjee]

@@tristanthedarkstar Thanks for the support.