AI Jigyasa

Пікірлер

@TairMukashev 4 күн бұрын

Hello! Why can't I change the unet model in this workflow?

@aijigyasa 4 күн бұрын

@@TairMukashev bcuz you didn’t put the model in the Unet folder i think. Go to comfyui/models/unet/ and paste your flux model there..

@TairMukashev 4 күн бұрын

@@aijigyasa thank you so much! this is an amazing workflow

@aijigyasa 4 күн бұрын

@@TairMukashev thank you 😊

@Aitoolstation805 6 күн бұрын

I have Low GPU Machine , Can i use it on online server ? I want to learn from begining do you provide tutorials ? Like How you setup all this workflow ?

@aijigyasa 6 күн бұрын

Yes, you can rent GPU over the internet to use Comfyui and generate images or video. you can find websites like Jarvis labs etc. or use Civitai to create images with the existing workflows shared by the community. if you want I can share a small tutorial on that too.

@aijigyasa 6 күн бұрын

I will share a beginner guide too so you can install everything from scratch. can you share your PC SPECS.

@Aitoolstation805 6 күн бұрын

@@aijigyasa Thankyou so much for your Comment...! My specs are HP Core i5 3rd Gen 500 GB with 16 GB RAM

@aijigyasa 2 күн бұрын

@Aitoolstation805 do you have a GPU Or integrated Graphics?

@Aitoolstation805 2 күн бұрын

@@aijigyasa No i dont have right Now :( But i will buy high GPu Machine to learnt hose properly

8 күн бұрын

Hi, you didn't give the exact URL for the upscale model, only for the main directory. I assume you want to use "4x_NMKD-Superscale-SP_178000_G.pth" here?

@aijigyasa 7 күн бұрын

yes. I just wanted to share the space/website for these upscaling models. you can download 1x,2x,4x according to your liking.

@spoletaev 2 күн бұрын

Look for the NMKD Superscale folder. In it, among the subfolders, find the desired file.

@challefd108 8 күн бұрын

Amazing like always, i hope we get a tutorial for beginners on how to upscale like that step by step

@aijigyasa 7 күн бұрын

Thank you. Do you want a beginner guide for upscaling? like for SDXL OR Flux?

@Aitoolstation805 6 күн бұрын

@@aijigyasa Yes We do , For Flux

@aijigyasa 6 күн бұрын

@@Aitoolstation805 alright, will share a tutorial for flux upscaling for beginner.

@Seany06 10 күн бұрын

damn.... mmmmm ... lmao you''re cracking me up... I think we all must be doing this when creating our bae's

@aijigyasa 10 күн бұрын

@@Seany06 lol. you know it. One should love his/her creation.

@luantv369 11 күн бұрын

i am having issues with compatibility ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. open-clip-torch 2.20.0 requires protobuf<4, but you have protobuf 5.29.2 which is incompatible.

@aijigyasa 10 күн бұрын

@@luantv369 well that’s new. looks like you have to downgrade the protobuf or update the clip torch version. I’ll try to find some fix for this and update you soon.

12 күн бұрын

The workflow is not available on the link provided...

@aijigyasa 11 күн бұрын

It’s been Fixed. Thanks for letting me know.

11 күн бұрын

@@aijigyasa Thanks!

@mythic_wave 17 күн бұрын

how can we add our own prompt?

@aijigyasa 17 күн бұрын

@@mythic_wave under power lora node you can find text encode. Click on grey dot to expand it. You can put your prompt there. Make sure to reduce the value to 0.15-0.25 in redux style node.

@challefd108 18 күн бұрын

One day you will become a big KZbinr that actually has a great content, as long as you are consistent, and i will be watching you i scream at my screen (I KNEW HIM WHEN HE HAD 10 SUBSCRIBERS) 😂, Best of luck. not only educational but also Entertainment, sending support from Algeria

@aijigyasa 18 күн бұрын

@@challefd108 wow. Thank you so much. You made my day. Thanks for the support.going to remember you.

@JacobHalton 21 күн бұрын

Just total ripoffs. Garbage

@aijigyasa 21 күн бұрын

@@JacobHalton i don’t get it. You watched the whole video? What is it you didn’t like?

@laraveber 22 күн бұрын

Thank you very much for the video! But I always get an error in the Pulid Load insightFase node. I don't understand what I'm doing wrong...

@aijigyasa 21 күн бұрын

@@laraveber Watch this video, as I have shown how to install Pulid locally and fix the problem that people normally face. Let me know if you still find any problem. Ikzbin.info/www/bejne/eGKUgZ-bj7WIfrM

@laraveber 21 күн бұрын

@@aijigyasa I carefully studied that video and managed to install all the nodes. But this problem is not related to the installation. It seems to me that the point is that I have a 4060 card and CUDA 8, but I need 16, something I read from the PuLID developer, if I understand correctly...

@aijigyasa 21 күн бұрын

@@laraveber Join the discord discord.gg/U8P6HAE9 and paste the screenshot there of the error you are getting. Let me do the work and find something that can help.

@havemoney 23 күн бұрын

Added WF Face swap to this and now it's great, thanks

@aijigyasa 23 күн бұрын

@@havemoney what face swap? Reactor, instant id?

@havemoney 23 күн бұрын

@@aijigyasa yes, ReActor - Fast Face Swap

@aijigyasa 23 күн бұрын

@ ohh. great. I am creating a discord channel. Interested to join in?

@challefd108 25 күн бұрын

Amazing work, already started learning from you brother

@aijigyasa 25 күн бұрын

@@challefd108 thank you so much.

@havemoney 25 күн бұрын

It’s a pity that I can’t treat you to coffee, Russia today is financially and in many ways informationally cut off from the world (

@aijigyasa 25 күн бұрын

@@havemoney its fine bro. I appreciate the thought. Hope everything will be back to normal in russia.

@havemoney 26 күн бұрын

I see all WF have the pulID 0.9 model, version 0.91 is it worse or not updated?

@aijigyasa 26 күн бұрын

@@havemoney Pulid gives errors with different nodes. I have done lots of fixing with this one. And it’s working fine now. Just Waiting for 1 more update. And will update it.

@havemoney 25 күн бұрын

@@aijigyasa Previously, each launch required a reboot of the interface, your fix helped. I will deal with Olama, I previously used LMStudio. Thanks! Add rtx3060 - 12vram 32 ram

@aijigyasa 25 күн бұрын

@ glad to know it helped. Will add this too. How did you know my PC specs?

@havemoney 25 күн бұрын

@@aijigyasa 3060- 12 and 32 ram a common configuration, I described my own =)

@aijigyasa 25 күн бұрын

@ lol, i just upgraded it to 32. But yeah its pretty common specs.

28 күн бұрын

Very exciting work! I would like to use it, but I can't even install PuLID... Do you have a good tutorial on how I can create the library, what model I need and where to copy it? Thank you!

@aijigyasa 28 күн бұрын

Watch this video, as I have shown how to install Pulid locally and fixed the problem that people normally face Ikzbin.info/www/bejne/eGKUgZ-bj7WIfrM

28 күн бұрын

@@aijigyasa Thank you!

@aijigyasa 28 күн бұрын

Let me know if you still find any problem..

@hainguyenhoang11 Ай бұрын

Thank you so much. My computer only has 8VRAM so it can only run SDXL.

@aijigyasa Ай бұрын

@@hainguyenhoang11 Anytime. You can do a lot with 8GB VRam. To be honest. I like the SDXL more than any other base model. Even flux. Let me know if you want to see some other workflows in SDXL.

@hainguyenhoang11 20 күн бұрын

@@aijigyasa Yes, I am very happy to have you share. I am indeed looking for sharing about SDXL.

@tomaszwaszka3394 Ай бұрын

It's not upscaled. It is reimagined... This is redrawn, but close to the original...

@aijigyasa Ай бұрын

@@tomaszwaszka3394 how would you describe a upscale then? For me the image is twice the pixel. That’s upscaling.

@tomaszwaszka3394 Ай бұрын

@@aijigyasa " twice the pixel" Yes, You are right. The thing is that the details in the photo are not the same

@aijigyasa Ай бұрын

@ i got it. well you can lower the denoise level to 0.15 to get the same details. Use a control net like tile to just enlarge the image. I was upscaling to fill the flat areas with details. Thats it.

@tomaszwaszka3394 Ай бұрын

@@aijigyasa 15:16 The shape of the chair's backrest changes (before, and after upscaling).... Don't get me wrong. Effect are amazing...anyway. If you still don't understand what I mean. Take a photo of a car with an illegible license plate. (You don't know whether there is an "8" or a "B") ("D" or "0")... The program doesn't know either, so it "invents" (reimagine) what could be there

@aijigyasa Ай бұрын

@@tomaszwaszka3394 yeah slightly.

@AndreyJulpa Ай бұрын

Where to put Flux Realism Lora? Not in LORA folder that is for sure

@aijigyasa Ай бұрын

@@AndreyJulpa go to models in comfyui directory and you will found xlabs at last. There is a lora folder in it. Paste it there. Refresh comfyui and use load flux lora and you will see realism lora there.

@adriands8207 28 күн бұрын

ComfyUI\models\xlabs\flux

@SouthbayJay_com Ай бұрын

Thank you for the video and the workflow. I can never get PuLID to work. I tried your fix also but still no luck. Is there a common issues you know of?

@aijigyasa Ай бұрын

@@SouthbayJay_com can you paste the error you are getting when working with pulid. Paste here in the comments. I also shared a fixed for pulid in my flux pulid video. You can try that too.

@baheth3elmy16 Ай бұрын

Nice tutorial, thanks!

@ItsKarstenMusic Ай бұрын

can you make this with image to image with a reference? can you add and make a tutorial please? thank you!

@aijigyasa Ай бұрын

Sure. you want image to image with style transfer. will share soon.

@nosyb Ай бұрын

@@aijigyasa cool :)

@ItsKarstenMusic Ай бұрын

great thanks!

@pong4289pan Ай бұрын

I've tried several times to solve the problem following your suggestions, but I still haven't been able to fix it.

@aijigyasa Ай бұрын

whats the error you are getting when running Pulid?

@aijigyasa Ай бұрын

please check if you have installed the correct python based insightface.

@ApexArtistX Ай бұрын

This is the best tutorial... why they have too make it complex for simple things like this.. you are epic

@aijigyasa Ай бұрын

@@ApexArtistX thank you. Wow.. wasn’t expecting this.

@waxcactu Ай бұрын

not free

@aijigyasa Ай бұрын

it's free. hailuo or Minimax gives you free 100 credits every day. you can create 3 videos every day.

@ohsocialmio3884 Ай бұрын

Sad you suddenly start to use AI voice. This takes away your channel's signature. Remember you made me laugh in one of your previous videos?

@aijigyasa Ай бұрын

@@ohsocialmio3884 this video was sitting there in the corner. I thought why not upload it. It has information. Will be sharing same old videos with workflows.

@ohsocialmio3884 Ай бұрын

@@aijigyasa ahh i see. Watched the Tyson match?

@aijigyasa Ай бұрын

Nahh.. was working. The outcome was joke.

@ohsocialmio3884 Ай бұрын

🤣🤣🤣🤣🤣🤣 14:51 "Wauwwwwwwwwwwww she is crispy now".

@jorgesharp Ай бұрын

the final image its so blury

@aijigyasa Ай бұрын

Increase the steps. Use Lora detailers; put blurry in negative when upscaling. that should fix it

@ohsocialmio3884 Ай бұрын

"She looks exactly like her ... HMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMM" 😭😭😭😭😭😭🤣🤣🤣🤣🤣🤣

@aijigyasa Ай бұрын

You loved the bit. I can tell by looking at your face. 🤣🤣🤣🤣🤣🤣

@ohsocialmio3884 Ай бұрын

@@aijigyasa ah man you made me laugh

@aijigyasa Ай бұрын

@ I am glad to know.

@ohsocialmio3884 Ай бұрын

@@aijigyasa Yes, i have watched your video till the end. If you want and have the time can you make a flux workflow to create ultra realistic professional photos (not the plastic skin look with bad stare) based on uploaded photo with positive and negative prompts: 1. Consistency in character (currently i am working with fooocus and i have a female model i use in front of a black backdrop and when i want to use her i simply drop her photo and generate photos) 2. that fix the face that match the uploaded photo 3. next to normal generated photo an additional upscaled photo using the ComfyUI-Flowty-LDSR (currently one of the best for upscaling)

@aijigyasa Ай бұрын

@ sure will look into it.

@baheth3elmy16 Ай бұрын

Does Not Work. I used your workflow and the image did not expand, only gave me gray padding on both sides.

@aijigyasa Ай бұрын

Use inpainting checkpoint. i have given the link in the description. if it still doesn't work screenshot it and link it here.

@MrDebranjandutta Ай бұрын

i coulld not get it working on windows, but worked like a charm in ubuntu. I wonder if the upscaling nodes are necessary, considering default upscale value is 1.20. As it is it's quite slow, and upscaling just adds to the time

@aijigyasa Ай бұрын

great. Upscaling is optional. I used it to refine some minor details; that's it.

@baheth3elmy16 Ай бұрын

Using your workflow and exact models/loras, I got this error: VAEDecodeTiled.decode() missing 1 required positional argument: 'overlap'

@aijigyasa Ай бұрын

I think the problem is with Ultimate SD upscale. comfyui started rolling out vae decode tile with an overlap option. Wait for them to fix it. Meanwhile, you can turn off the Vae decode option at the bottom of Ultimate upscaler.. That should fix the problem.

@Romeo615Videos 2 ай бұрын

it'd be funny 2-3 years from now yall get sued for using celebrities and that was part of the ai scam hahahahahahhahaha Oh man it'd be a laugh to watch all the people cry and act like they didnt know.... Pretty cool times we live in

@aijigyasa 2 ай бұрын

@@Romeo615Videos this video is for educational purposes only. I used a celebrity face bcuz it’s easy to find high quality image. And it’s public. These images were used for all sorts of work memes or small ads. So i think we will be fine. But thanks for showing concern.

@ismgroov4094 2 ай бұрын

oh my god.... thx as always.

@aijigyasa 2 ай бұрын

No worries. Keep Supporting.

@MrDebranjandutta 2 ай бұрын

@@aijigyasa its failing to build the insightface wheel no matter what I do. was reading somewhere the non-portable version does it without probs (maybe I have a path problem) will try tomm

@aijigyasa 2 ай бұрын

@@MrDebranjandutta try installing insightface with your python version. Steps to Follow: Check the Python version in ComfyUI: 1. Go to the ~/comfyui/python_embedded folder and double-click on python.exe. Note the version of Python (e.g., 3.09, 3.10, 3.11, or 3.12). 2. Now open this link(in description) - github.com/Gourieff/Assets/tree/main/Insightface and download the same version as python. and move that file to comfyui folder where run_nvidia_gpu batch file exist. 3. now right click on it and copy as path. 4. right click anywhere in the folder and open in terminal and paste this command - .\python_embeded\python.exe -m pip install [paste path here] onnxruntime and hit enter. that's it. let me know if it fixes your problem.

@MrDebranjandutta 2 ай бұрын

Good stuff bro, subsd. Which gpu are you using btw? I have a poor mans 3060 can I use the full dev-fp8 model?. Like your goofball style. Keep at it

@aijigyasa 2 ай бұрын

I also have a 3060 with 12gb Vram. You can run flux dev fp8 22gb variant. But it will take time and will slow down your pc movements too. Don’t even think about upscaling with FP8. I will recommend stick with the turbo lora’s or GGUF models. Or generate low res image with full fp8 and upscale with turbo with 8 step or GGUF. and thanks for the compliment 😬

@Bogdan_95 2 ай бұрын

After following your steps, I’m still getting the following error: (AttributeError: newbyteorder was removed from the ndarray class in NumPy 2.0. Use arr.view(arr.dtype.newbyteorder(order)) instead). If you have a solution for this, it would really help me move forward.

@aijigyasa 2 ай бұрын

@@Bogdan_95 can you share the screenshot of your workflow? I want to see which node is giving you this error. Meanwhile try updating your comfyui.

@aijigyasa 2 ай бұрын

@@Bogdan_95 try this fix also.. from folder - ComfyUI_windows_portable .\python_embeded\python.exe -m pip install numpy==1.26.3

@Bogdan_95 2 ай бұрын

@@aijigyasa Sure, I can do that. Where should I send the screenshot? Also, I have other workflows-would they break with the NumPy 2.0 update, or would they still work fine?

@aijigyasa 2 ай бұрын

@@Bogdan_95 i don’t think its going to affects other workflows. Specially SdXl or 1.5 based workflows. This error occurs only in flux, specifically gguf loader or unet checkpoint loader. Have you tried the above methods btw? Here is a thread link. You will find a solution here. I hope this helps. No need for screenshot its a compatibility issue. github.com/city96/ComfyUI-GGUF/issues/7#issuecomment-2392951603

@kallamamran 2 ай бұрын

Doesn't look like 1-step upscale. Rather looks like upscaling two times with 20 step each

@aijigyasa 2 ай бұрын

@@kallamamran sorry for the confusion: but by 1 step i mean just one node. You can’t do 1 step upscale bcuz there is no hyper turbo ultra flux lora right now. Will upload the fast upscaling method with this workflow. Stay tuned.

@germanfa1770 2 ай бұрын

Dude. It was interesting to see your workflow. I smiled when flux only makes women. I hope you will do more complicated and interesting workflow. Goodluck and спасибо.

@aijigyasa 2 ай бұрын

@@germanfa1770 thanks man. And yes will share more interesting workflows. Consider subscribing for future video. Appreciate you.