Hey in this process , i didnt need to install flux models ??😅

The default learning rate seems way to high. Good for 20 images but probably horrible for 50

I get all black output, black image. But i guess workflow works. Any idea what is wrong?

Thank you for the great tutorial. Only one thing happend to my installation. Load Diffusion Model is just crashing with showing me a necessary reconnect and then it just stops. Do you know what could cause the issue? Havent changed anything in the workflow so far.

Why would you not cut out the wait time for the images to generate. Thumbs down for wasting your viewers time.

I wish Krea had the editor function in which you can edit certain parts of the picture.. or even replace the face using another photo like in Midjourney. Also being able to zoom out the image or expand to either direction would be nice features to have. Overall, it's a nice platform but the image generation features are lacking

hello SamplerCustomAdvanced expected scalar type Float but found BFloat16 I can't fix this problem,can you help me please

but can i use images generated using flux dev model commercially??

My Results from this WF are very bad. The face resemblence is zero and redux gives me a mess of a picture. Mind i ask what Cuda-Toolkit, Cudnn and Onnx Version you are running?

magic quill does not make good background replacements. its good for clothes, but not for bg...

Nice - do a demo with biden next.

You’re betting on a dead platform. James Cameron said comfy is the past, 40 years into the past, a node’s platform for AI it’s like going to mars with the Apollo 11.

Error SamplerCustomAdvanced - forward_orig() takes from 7 to 9 positional arguments but 10 were given

How do you import other models? The download feature isn't reliable and the model list doesn't update with new names?

Got error after installing on pinkio ENOENT: no such file or directory, stat 'C:\Users\admin\pinokio\api\echomimic2.git\{{input.event[0]}}'

I enjoyed this! Easy to implement. Where do I tweak to make the outcome look closer to the face( Load Image Face). I am having difficulty making it look like that person by default. Cheers.

Sooooo glad I watched this. I have invested over a year in learning and getting good at midjourney and using the style references. I hope they stay king!

no portable users, only for pinokio version of comfyui, error cannot be fix in portable :(

I have done everything right following the steps but I have a small problem that I can't solve. It does not detect the Flux Pulid model. I have created the folder 'pulid' inside the 'model' folder as you do in the video, but I don't know why it doesn't detect it. Has anyone had a similar problem?

why wouldnt you put the model in the drive folder instead of the api folder? models are supposed to go into Drive. but im not sure if it can use full structure.

Hey, is there plans for fill model + pulid? I tried your inpaint faceswap workflow, it kinda works but results not good if face is small and not facing camera directly (i tried changing denoise level). I also tried comfy ui example for flux fill inpaint - it works pretty nice for same image, but there is no pulid. I tried add it by myself yesterday and it kinda works but resemblence to face is way too low, face is barely recognizable. Maby you know how to make it the right way)

`[email protected]` is not supported because: requires device with capability > (8, 0) but your GPU has capability (7, 5) (too old) bf16 is only supported on A100+ GPUs `cutlassF-pt` is not supported because: bf16 is only supported on A100+ GPUs

To fix this Error SamplerCustomAdvanced forward_orig() takes from 7 to 9 positional arguments but 10 were given go to your E:\pinokio\api\comfyui.git\app\comfy\ldm\flux folder and edit the model file with notepad+ and search for "out = self.forward_orig" line 184 and remove ", transformer_options" at the end of that line and save changes this will fix this errors watch the video to the end to see how I fix this error in the video.

Thanks, for the krea tuto, I've tried their video upscale. Do you have a better solution than using Krea video upscale? You have any ideas Which tool they use to upscale our video and you have a free/better solution we could use? Thanks ❤

Why is there no option to download the final Image ? Is there a way to download the final Image? I dont see any download button.

Great video and workflow unfortunately it does not seem to work with the DDesktop version of Comfyui "load PuLID Flux Model has nothing to choose I placed it in the pulid folder in the models directory in my portable version and it works for that but not for desktop version

Thank you sooooo much for mentioning how to solve that error 🙏🏻❤️

Edit: THIS IS THE ONE THAT WORKS! (Ignore my direct comment below) Thank you @AIMotionStudio I was hoping this one would work but it has errors: SamplerCustomAdvanced forward_orig() takes from 7 to 9 positional arguments but 10 were given

Getting Error ` UltimateSDUpscale view size is not compatible with input tensor's size and stride (at least one dimension spans across two contiguous subspaces). Use .reshape(...) instead. `

ty very clear and helpfull

Thanks a lot!! I will test it !

is there a video for mac?

To fix this Error SamplerCustomAdvanced forward_orig() takes from 7 to 9 positional arguments but 10 were given go to your : github.com/comfyanonymous/ComfyUI/issues/5601 just go to your E:\pinokio\api\comfyui.git\app\comfy\ldm\flux folder and edit the model file with notepad+ and search for "out = self.forward_orig" line 184 and remove ", transformer_options" at the end of that line and save changes this will fix this errors

We were able to fix this error using this: github.com/comfyanonymous/ComfyUI/issues/5601 just go to your E:\pinokio\api\comfyui.git\app\comfy\ldm\flux folder and edit the model file with notepad+ and search for "out = self.forward_orig" line 184 and remove ", transformer_options" at the end of that line and save changes this will fix this errors

Thanks for this video. What graphics card did you use and how long did it take to generate the video? Do you think it works on an RTX 3070 with 8GB of Vram?

well done👍

Can i install magicquill on mac?

You are pathetic!

they key is to build the insightface wheel manually to get it to work

Hello, I hope you can help me because I get an error when I start CMD. I get "git" which is not recognized as an internal or external command, program or executable batch file. I hope you can help me. I have the latest version of comfyui installed on pinokio.

🎉🎉🎉🎉

SamplerCustomAdvanced "Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu! (when checking argument for argument weight in method wrapper_CUDA__native_layer_norm)" The above error occurs when generating consecutively. Thank you. ### Loading: ComfyUI-Manager (V2.51.9) ### ComfyUI Revision: 2862 [15c39ea7] | Released on '2024-11-26' # ComfyUI Error Report ## Error Details - **Node ID:** 48 - **Node Type:** SamplerCustomAdvanced - **Exception Type:** RuntimeError - **Exception Message:** Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu! (when checking argument for argument weight in method wrapper_CUDA__native_layer_norm) ## Stack Trace ``` ## Additional Context (Please add any additional context or steps to reproduce the error here)MODEL"],[123,51,0,62,2,"EVA_CLIP"],[124,53,0,62,3,"FACEANALYSIS"],[125,45,0,62,1,"PULIDFLUX"],[126,54,0,62,4,"IMAGE"],[127,31,0,62,0,"MODEL"],[128,41,0,6,0,"CLIP"],[129,31,0,17,0,"MODEL"],[130,49,0,63,0,"IMAGE"],[132,64,0,65,0,"LATENT"],[133,66,0,64,0,"IMAGE"],[134,10,0,64,1,"VAE"],[135,66,1,65,1,"MASK"],[136,65,0,48,4,"LATENT"],[137,66,0,67,0,"IMAGE"],[138,68,0,6,1,"STRING"],[139,67,2,68,0,"STRING"],[140,69,0,67,1,"FL2MODEL"],[141,66,0,70,0,"IMAGE"],[142,49,0,70,1,"IMAGE"],[143,70,0,71,0,"IMAGE"]],"groups":[{"id":1,"title":"Face Swap Masking","bounding":[726,599,333,422],"color":"#3f789e","font_size":24,"flags":{}},{"id":2,"title":"Flux Models","bounding":[15,34,337,554],"color":"#3f789e","font_size":24,"flags":{}},{"id":3,"title":"Prompts","bounding":[371,41,342,861],"color":"#3f789e","font_size":24,"flags":{}},{"id":4,"title":"Input Image","bounding":[1425,37,338,548],"color":"#3f789e","font_size":24,"flags":{}},{"id":5,"title":"PuLID","bounding":[724,41,329,548],"color":"#3f789e","font_size":24,"flags":{}},{"id":6,"title":"Ksampler","bounding":[1069,41,345,980],"color":"#3f789e","font_size":24,"flags":{}},{"id":7,"title":"Output","bounding":[1427,604,437,552],"color":"#a1309b","font_size":24,"flags":{}}],"config":{},"extra":{"ds":{"scale":0.8390545288824038,"offset":[223.73253580966474,126.80426492827088]}},"version":0.4} ``` ## Additional Context (Please add any additional context or steps to reproduce the error here)

Love your videos so much! Can you make a review video on FlexClip’s AI tools? Really looking forward to that!

how we can add on Cpanel and Worpress

Thanks for the tutorial, I was able to set up everything accordingly. But I am getting this error at 73% SamplerCustomAdvanced expected scalar type Half but found BFloat16. Please help :)

I have 4gb vram will it work?

i actually found out in F5 TTS if the voice you're cloning sounds clear and less distorted it can match almost exactly like the audio of that voice and its personality.

Output images are blurry. I increased steps but didn't help

Thank you!

How do I transfer photos to gallery