My video card can't run it but it's nice to learn pinokio exists. Also nice to see an ai content creator acknowledge comfyui is anything but comfy to learn.
Thank you for covering Flux...... i haven't been this excited for AI art tools since I learned about Automatic 1111 a couple or so years back (long before ComfyUI ever existed). 👍 BTW, that scenic nature image backdrop is gorgeous. I love the apparent "golden hour" hues.
I managed to install it based on your video, thank you. However, when I generate anything, even at the smallest image size, after 10 minutes nothing happens, it just works, works... what could be the problem? I have a 3070 video card.
I am using AMD Radeon RX 6600, everything works fine. But, I want to create image, unknown error happened! I see the log terminal, nothing was identified as error. Has any advice?? I am using Win 11, 16GB RAM, RX 6600, i3 12100F.
Take closer look at the logs, though something may not show up as an error, there may be information about why it's doing that. Outside of that I don't know what I could tell you that would help.
Hey, when I install it , it first asks me to install 9 other things like git, cuda etc. When I reach to the installation of cuda, it stucks at 7/9 installation what should i do?
They're having a lot of difficulty getting the Pinokio version running, I would recommend installing Forge WebUI, whether that's through stability matrix or Pinokio. It works a lot better especially on computers with less than 12 GB of VRAM
Not sure what you mean. When you install it locally using the newer version, even the older version, it uses whatever GPU you have installed. As long as it has at least 8 to 12 gigs of VRAM.
I'm getting a "ENOENT: no such file or directory, stat 'C:\pinokio\api\flux-webui.git\{{input.event[0]}}'" error when I click on Open Web UI. Any ideas there?
Yeah I started getting that too. I tried reinstalling and it would load, but something is wrong as it won't generate now. Give it some time, they will do an update hopefully and get it fixed.
Hi, I went through the process and downloaded this on my pc but now I need some guidance on how to train Lora's with this interface, I've wasted days on RUclips trying to find tutorials on how to use this desktop version but found nothing. It seems nobody but us have the computer space to use this version. So, if you don't mind, can you do a tutorial on how to do this please or if it's not possible, do I need to delete this and do all the alternatives I see flooding RUclips?
I just watched and followed your directions. The Flux is installed; however, when I attempt to an image, it errors out. It seems to indicate a "connection" error. I have checked and Pinokio is set to be allowed through the firewall. So the hunt is on for a solution. But great video. Thanks.
Yeah since the update they pushed out since I did this video there's been some issues. I'm experiencing my own issues with extremely slow render times at this point.
@@AIchemywithXerophayze-jt1gg yeah, Win7 was no go.. I had to install a win10, just for this.. Great guide, thanks.. I wonder if you can like upload a picture to it, say a picture with an old house, and get it to recreate the house as new.. maybe you know? (and leave everything as it is in the surroundings)
@@AIchemywithXerophayze-jt1gg what I mean is, i'm using easy diffusion, I have 160 prompts I run on each model, i'd like to be able to put them all in and walk away and let it batch it out.
Hey, I followed this guide but keep getting ' torch.OutOfMemoryError: CUDA out of memory ' - my GPU has 12gb of memory, and I have no idea how to use powershell/python - would you be able to help me?
Yeah I know, shortly after I did the video they did some kind of an update which kind of broke it. I tried to work with him to get it working on my system, and at best I could get it to do a render in about 5 to 7 minutes per image. Check out my latest video because I show you how to get it up and running in Forge WebUI. And it works much better.
That's because the installation is literally clicking install. That's how easy pinokio makes it. But there are issues right now. When I did this video it was working really well. They've done some updates and I can't even get it to run on my system now. I reached out to them to see if we can figure out what's going on. They did an update which broke the initial installation. They updated to version two, but that one is having troubles as well. I went from being able to generate an image in about 30 seconds or less, to it now taking several minutes. But it looks like it's not utilizing my GPU for some reason.
@@AIchemywithXerophayze-jt1gg That was not my experience. First I had to approve a number of new dependencies, plus install a version of VS. Then I got stuck on a screen for downloading GeForce, which I already have. Trying to figure out the difference between my experience and someone else's experience would be easier if I could see that experience.
Check out my last two videos, I go through and point out a couple different models that can be used in the forge web UI. I've moved away from the flux web UI from Pinokio
Yeah it keeps giving me connection errors after it ran for a really long time. I have a AMD Ryzen 5 3600 6-Core Processor with 16GB of Ram. I don't know too much about computers but I am pretty sure mine can run this. What else could be the issue??
I tried working with them for a while to see if we could get it going cuz mine stopped working as well. But ever since Forge WebUI was updated to be able to run Flux, I kind of moved over to that.
@@AIchemywithXerophayze-jt1gg Yeah I seen that right after commenting and have it installed now but it is pretty slow even after the first few runs. Oh well still feels good to have one I can use locally. It is my first time using AI without a browser. It takes like 30 minutes for an image sadly, do you think this is right for the specs I mentioned with using Forge? I feel like it is not right, maybe just having too much already running at one time on my computer aka like 100 browser tabs 😅
@@AIchemywithXerophayze-jt1gg Install Forge WebUI via Stability Matrix? or with pinokio, not that i know how to, i'll ask ai, once you tell me which way to go, i am on a imac 2017 32GB Ram. 4gb radonPro
Sorry for the silly question (I'm at the hospital and I can't play the video with sound), does the video explain if you can run this on an amd gpu? I have a 6900 xt
No I believe this is for Nvidia cards only. You may want to look at using Forge webUI, they recently finished a bunch of updates that allow you to use Flux, and I'm not sure as I don't have an AMD based video card, but I believe it does work with AMD. If somebody else in the comments wants to pitch in on this and help educate us on AMD supported image inference that would be great.
I don't think so. I've been using it all day today I checked my account and I don't see charges anywhere on my account for that. I'm just using the Schnell model.
Yeah they're having some issues with the installation. The first version worked great, then they ran an update and it broke it. Their new version 2 supposedly works, but not sure what's going on with it.
![I Replaced ALL my ADOBE APPS with these [free or cheaper] Alternatives!](http://i.ytimg.com/vi/5EfqHg49kMk/mqdefault.jpg)
My video card can't run it but it's nice to learn pinokio exists. Also nice to see an ai content creator acknowledge comfyui is anything but comfy to learn.
Thanks for providing this. It was helpful to see the steps.
Glad it was helpful!
I am getting link to nvidia dirver page.. so cant lunch. I have 2 cards, but I htink they are old.
Thank you for covering Flux...... i haven't been this excited for AI art tools since I learned about Automatic 1111 a couple or so years back (long before ComfyUI ever existed). 👍
BTW, that scenic nature image backdrop is gorgeous. I love the apparent "golden hour" hues.
I'm loving it too. And it's been so nice having it installed locally now.
I managed to install it based on your video, thank you. However, when I generate anything, even at the smallest image size, after 10 minutes nothing happens, it just works, works... what could be the problem? I have a 3070 video card.
The flux web UI that pinokio came out with had some sort of an update that kind of broke it, people are utilizing forge now. Check out my next video
@@AIchemywithXerophayze-jt1gg thanks!
Thanks, I installed now Forge UI and it works :)
I am using AMD Radeon RX 6600, everything works fine. But, I want to create image, unknown error happened! I see the log terminal, nothing was identified as error. Has any advice?? I am using Win 11, 16GB RAM, RX 6600, i3 12100F.
Take closer look at the logs, though something may not show up as an error, there may be information about why it's doing that. Outside of that I don't know what I could tell you that would help.
Hi, the rendering is taking like forever for me and the image is not generating, it never goes past the 3rd step. Any fix??
I moved to using forge webUI. Works better.
Hey, when I install it , it first asks me to install 9 other things like git, cuda etc. When I reach to the installation of cuda, it stucks at 7/9 installation what should i do?
They're having a lot of difficulty getting the Pinokio version running, I would recommend installing Forge WebUI, whether that's through stability matrix or Pinokio. It works a lot better especially on computers with less than 12 GB of VRAM
What can be done so it can use an available GPU for acceleration? Kind of misses the point without it.
Not sure what you mean. When you install it locally using the newer version, even the older version, it uses whatever GPU you have installed. As long as it has at least 8 to 12 gigs of VRAM.
@@AIchemywithXerophayze-jt1gg it doesnt work with AMD GPUs it seems - tried ZLUDA still no go
@@AIchemywithXerophayze-jt1gg I am using Pinokio 2.1.17 and running flux it does not use the gpu just the cpu. Thats for a top of the line RX7900XTX.
I'm getting a "ENOENT: no such file or directory, stat 'C:\pinokio\api\flux-webui.git\{{input.event[0]}}'" error when I click on Open Web UI. Any ideas there?
Yeah I started getting that too. I tried reinstalling and it would load, but something is wrong as it won't generate now. Give it some time, they will do an update hopefully and get it fixed.
I'm having the same problem as well. Did you resolve the problem?
I have an RTX 3080 (10 gig) with 64 gig of system RAM and it still takes forever to generate an image. Am I doing something wrong?
Try using forge instead. This doesn't work very well any more
Hi, I went through the process and downloaded this on my pc but now I need some guidance on how to train Lora's with this interface, I've wasted days on RUclips trying to find tutorials on how to use this desktop version but found nothing. It seems nobody but us have the computer space to use this version. So, if you don't mind, can you do a tutorial on how to do this please or if it's not possible, do I need to delete this and do all the alternatives I see flooding RUclips?
I never actually trained a Lora before. I know Pinocchio just released a specific package for training them though.
I just watched and followed your directions. The Flux is installed; however, when I attempt to an image, it errors out. It seems to indicate a "connection" error. I have checked and Pinokio is set to be allowed through the firewall. So the hunt is on for a solution. But great video. Thanks.
I am also running the RTX 3080. Also with an i9 13xxx
Yeah since the update they pushed out since I did this video there's been some issues. I'm experiencing my own issues with extremely slow render times at this point.
Hmm.. yes I know, I know.. Maybe I am old style, but it seems I cannot install it on Windows 7.. can this be correct? Only on 10?
I don't think it's supported on Windows 7. I would at least move to Windows 10, 11 preferably
@@AIchemywithXerophayze-jt1gg yeah, Win7 was no go.. I had to install a win10, just for this.. Great guide, thanks.. I wonder if you can like upload a picture to it, say a picture with an old house, and get it to recreate the house as new.. maybe you know? (and leave everything as it is in the surroundings)
why is the image file .txt when trying to download the image?
Not sure mine are webp images.
@@AIchemywithXerophayze-jt1gg when you click the down arrow in the upper right corner it downloads as .txt
And can you do batches of images? Or do you have to do everything one at a time?
I'm pretty sure with the flux web UI or gives you the option of specifying multiple images in a batch.
@@AIchemywithXerophayze-jt1gg what I mean is, i'm using easy diffusion, I have 160 prompts I run on each model, i'd like to be able to put them all in and walk away and let it batch it out.
Hey, I followed this guide but keep getting ' torch.OutOfMemoryError: CUDA out of memory ' - my GPU has 12gb of memory, and I have no idea how to use powershell/python - would you be able to help me?
Yeah I know, shortly after I did the video they did some kind of an update which kind of broke it. I tried to work with him to get it working on my system, and at best I could get it to do a render in about 5 to 7 minutes per image. Check out my latest video because I show you how to get it up and running in Forge WebUI. And it works much better.
EDIT: Thanks for the clarification!
This doesn't show the installation. The actual thing in the title?
what? yes it does lmao
@@ja-no6fx He already had it installed. He literally skipped the installation - it was already done - which is the bit I'm trying to learn.
That's because the installation is literally clicking install. That's how easy pinokio makes it. But there are issues right now. When I did this video it was working really well. They've done some updates and I can't even get it to run on my system now. I reached out to them to see if we can figure out what's going on. They did an update which broke the initial installation. They updated to version two, but that one is having troubles as well. I went from being able to generate an image in about 30 seconds or less, to it now taking several minutes. But it looks like it's not utilizing my GPU for some reason.
@@AIchemywithXerophayze-jt1ggim actually having the same issue, except its not generating an image at all. Seems to get stuck
@@AIchemywithXerophayze-jt1gg That was not my experience. First I had to approve a number of new dependencies, plus install a version of VS. Then I got stuck on a screen for downloading GeForce, which I already have.
Trying to figure out the difference between my experience and someone else's experience would be easier if I could see that experience.
Do you know where the models are downloaded from?
Check out my last two videos, I go through and point out a couple different models that can be used in the forge web UI. I've moved away from the flux web UI from Pinokio
@@AIchemywithXerophayze-jt1gg Just got both models working locally! Had to change the size back to 512 to get shorter times. Thanks for getting back.
Thanks for sharing this!
My pleasure!
Yeah it keeps giving me connection errors after it ran for a really long time. I have a AMD Ryzen 5 3600 6-Core Processor with 16GB of Ram. I don't know too much about computers but I am pretty sure mine can run this. What else could be the issue??
I tried working with them for a while to see if we could get it going cuz mine stopped working as well. But ever since Forge WebUI was updated to be able to run Flux, I kind of moved over to that.
@@AIchemywithXerophayze-jt1gg Yeah I seen that right after commenting and have it installed now but it is pretty slow even after the first few runs. Oh well still feels good to have one I can use locally. It is my first time using AI without a browser.
It takes like 30 minutes for an image sadly, do you think this is right for the specs I mentioned with using Forge? I feel like it is not right, maybe just having too much already running at one time on my computer aka like 100 browser tabs 😅
@@AIchemywithXerophayze-jt1gg Install Forge WebUI via Stability Matrix? or with pinokio, not that i know how to, i'll ask ai, once you tell me which way to go, i am on a imac 2017 32GB Ram. 4gb radonPro
Thank you for the video!! :) I am about to install, but one quick question does Pinokio only run the schnell? Or can we use the other models as well?
It runs that one and another one that is a merged model.
@@AIchemywithXerophayze-jt1gg awesome, thank you. I will give it a try and let you know how it runs for me too. :)
It's supposed to run dev but I can't get past the CLIP error even with VAEs loaded.
Hi Admin! Is Flux AI work for faceswap, like Google Foocus?
There are a lot of new comfy UI workflows that people are working on. I wouldn't doubt it if they have one that does that.
Sorry for the silly question (I'm at the hospital and I can't play the video with sound), does the video explain if you can run this on an amd gpu? I have a 6900 xt
No I believe this is for Nvidia cards only. You may want to look at using Forge webUI, they recently finished a bunch of updates that allow you to use Flux, and I'm not sure as I don't have an AMD based video card, but I believe it does work with AMD. If somebody else in the comments wants to pitch in on this and help educate us on AMD supported image inference that would be great.
@@AIchemywithXerophayze-jt1gg oh I will, never heard of it! Thanks so much bro!
Like your tone of voice and how you explain things.
I appreciate that!
it works with AMD video cards?
Going to be honest, I don't know. I use Nvidia.
Blackforest is now charging for each image. :-(
I don't think so. I've been using it all day today I checked my account and I don't see charges anywhere on my account for that. I'm just using the Schnell model.
@@AIchemywithXerophayze-jt1gg Wonder why it wouldn't let me generate. I had made a bunch of images prior to today, maybe I used up all my free ones?
Thank you for this info, will try it out. *Edit* meh, it error'd, probably because my vRAM is too low (9GB)...
Yeah they're having some issues with the installation. The first version worked great, then they ran an update and it broke it. Their new version 2 supposedly works, but not sure what's going on with it.
Thanks, man!
Thank you, needs a lot of memory to run, M2 with 32GB Ram won’t even generate an image will stick to cloud services for now but thanks again.
Thanks for sharing
cool, thank you
Yeah they're having problems with it right now. Hopefully they do some updates to get it fixed.
Ty
👋
✌️
Thank you for the video but please, next time get faster to the point. I had to watch it on 2.7 speed otherwise I would blow my brains out.
Interesting
He explained it just fine....what do you want from the guy?