Yeah I think for the default config you need a good GPU. You can try load the CPU version from the same repository called face-to-many-cpu.json but it might be very slow github.com/aj47/techfren-vids/tree/main/comfyui
I have 2 problems : - cant find the insightface folder and second "Load InstantID Model " node is not working it just stays at "undefined" pls help < / 3
@@techfren That fixed it but when i start the program i got this : Error occurred when executing KSampler (Efficient): Expected query, key, and value to have the same dtype, but got query.dtype: struct c10::Half key.dtype: float and value.dtype: float instead.
@@techfren tried, but it crashes earlier, with this message: Error occurred when executing AIO_Preprocessor: The operator 'aten::upsample_bicubic2d.out' is not currently implemented for the MPS device.
Mine won't download cuda, do you know why? I've pressed the install button for it 3 times but it keeps saying cuda not installed Update: it's installed now, but there isnt a start button to run comfy and im not sure why
Hello, I have an error: Could not allocate tensor with 52428800 bytes. There is not enough GPU video memory available! Could you help me with that? I am using an AMD RX 6800XT. Thanks!
try loading one of these two files instead from the same github: face-to-many-rocm.json or face-to-many-cpu.json github.com/aj47/techfren-vids/tree/main/comfyui
I can’t install comfyui the first time cuda couldn’t install, then I decided to clear the files and download again, but alas it didn’t work. don't know why?
@@techfren oh, I just installed RVC and cuda was installed, maybe I need to install it several times or install another script p.s I was happy early. Do you happen to know why I have red windows in stage 9? ran the file for Nvidia and there is no instaid in stage 12 either
When loading the graph, the following node types were not found: InstantIDModelLoader InstantIDFaceAnalysis ApplyInstantID Nodes that have failed to load will show as red on the graph. ниче не фиксит пробовал все тутор и установку вручную хрен пойми с какой библы он подтянул репозиторий и что в контейнерах даже через кастомы не находится и сториннеми библами ручками чтобы заменить
Установил всю эту парашу, потратил всю ночь, занял весь жёсткий диск... панельки в браузере всё равно горят красным, папки insightface в принципе не было. Я никогда не тратил столько времени впустую...
Theres a part in the video where you stop explaining a step by step, it be nice if you make a resume or conclusion. cos damn its kinda easy, but TO install the nodes left, there's where you put code and get lost. PLssssssss :C
Guys there's a mobile app called "remini" that does video game filter. It has a free trial but filled with ads
playground.com/face also found this seems free
playground sadly asks for 15$ after 4 gens :(
@@thatboyp444 ah damn, knew it was too good to be true
4 gens?
@@nikko9586 4 image generations/creations
Hey! I have been seeing your content lately and finally subscribed. Thank you for your efforts keep up the good work!
thank you!!
Do you have any idea what are the pc spec nesecary to run this?
Since i believe that either my hardware is not enough or something went wrong.
Yeah I think for the default config you need a good GPU. You can try load the CPU version from the same repository called face-to-many-cpu.json but it might be very slow
github.com/aj47/techfren-vids/tree/main/comfyui
other tips, upload smaller images, and process fewer steps
Did you change any of the defaults or promt to get your results? Mine is working but the results are a bit meh
i change them around for different images but generally use : "ps1 graphics, video game, 3d, silent hill, pixelated, psx, gamecube, dark"
Does this work on mac? And which folder should the instant id controlnet model go in MacOS?
I think so if pinokio works, you can see the files location by pressing the files folder in pinokio
Best video ever. Very well explained and the best of all... it worked! haha
Amazing! Well done!!
Btw can i get the discord link to join?@@techfren
@@Juankengas in the description
Thanks. I did not saw It haha@@techfren
Hello, I did the whole procedure but I don't know how to generate the image on PS2, could you help me?
open a thread in my discord help channel
I have 2 problems : - cant find the insightface folder and second "Load InstantID Model " node is not working it just stays at "undefined" pls help < / 3
a lot of people having this issue, did you do step 12?
@@techfren just did it and still no folder :(
@@Radomir- try manually download and add folder and file huggingface.co/InstantX/InstantID/tree/main
@@techfren That fixed it but when i start the program i got this : Error occurred when executing KSampler (Efficient):
Expected query, key, and value to have the same dtype, but got query.dtype: struct c10::Half key.dtype: float and value.dtype: float instead.
how to that node ? even though i follow the step.. i dont think you explain how to do the node
Which node? Click the manager button then install missing nodes button to install missing nodes
@@techfren i clicked on the missing node button to install , but it seem to not load anything on the page to install
so it just blank
hi, how can i use this with a picture without a face on it ? i have a "no face detected error" because im wearing a hood
im not too sure but i think it would involve something like removing instantid front he flow
I can get it to work on my Mac M2 but only in CPU mode which takes a long time, can I also use the GPU? Massive thanks for the step-by-step tutorial!
try changing the sampler property in the KSampler node
@@techfren tried, but it crashes earlier, with this message:
Error occurred when executing AIO_Preprocessor:
The operator 'aten::upsample_bicubic2d.out' is not currently implemented for the MPS device.
I can use the CPU one, but then it takes roughly 16 minutes
@@mister-monkeyman damn, tried all the other sampler types?
@@techfren yes, I can only get it to work if I boost it in CPU mode.
Mine won't download cuda, do you know why?
I've pressed the install button for it 3 times but it keeps saying cuda not installed
Update: it's installed now, but there isnt a start button to run comfy and im not sure why
Send a screenshot on my discord
@@techfren I got frustrated and uninstalled 😭
I still rly want to get it tho I just can’t figure it out
I'm trying to do it in this website now and it doesn't have the video game filter :'(
does anyone know how to automatically load the workflow after installing all of the nodes? or do i have to add each of them manually?
Press load from the right menu then find the .json file
Hello, I have an error:
Could not allocate tensor with 52428800 bytes. There is not enough GPU video memory available!
Could you help me with that? I am using an AMD RX 6800XT.
Thanks!
try loading one of these two files instead from the same github: face-to-many-rocm.json or face-to-many-cpu.json
github.com/aj47/techfren-vids/tree/main/comfyui
@@techfren Thank you for the help, but unfortunately nothing changed :/ My GPU also is still being used, even when using the cpu version
I can’t install comfyui the first time cuda couldn’t install, then I decided to clear the files and download again, but alas it didn’t work. don't know why?
Sorry Im not sure. Do you have an Nvidia GPU?
@@techfren Yes
@@techfren oh, I just installed RVC and cuda was installed, maybe I need to install it several times or install another script
p.s I was happy early. Do you happen to know why I have red windows in stage 9? ran the file for Nvidia and there is no instaid in stage 12 either
@@mr_f1me940 not sure, make sure all the files required are downloaded on the correct paths
When loading the graph, the following node types were not found:
InstantIDModelLoader
InstantIDFaceAnalysis
ApplyInstantID
Nodes that have failed to load will show as red on the graph.
ниче не фиксит
пробовал все тутор и установку вручную
хрен пойми с какой библы он подтянул репозиторий
и что в контейнерах
даже через кастомы не находится и сториннеми библами ручками чтобы заменить
Use that AI so anyone can add themselves to Love Island :D
I've seen models that make 3D files from 2D image
Spectacular tutorial, I did every steps and now I don't know how to generate de image, please help me
Upload a image into the left most upper node then press queue prompt on the right menu
conda is not recognized as an internal or external command, even though pinokio installed it
Damn, maybe try download anaconda manually
this run at amd gpus?
Yeah I think so. I have a rocm workflow in the GitHub
Haven't tried it though so please let me know if it works
@techfren
Sure i found your video too late on my country 😆 tomorrow i will test it
how can I do this with videos?
probably need to add another node that does video, sounds interesting i might look into it
@@techfren if you do please let us know. Would love this possibility!
Установил всю эту парашу, потратил всю ночь, занял весь жёсткий диск... панельки в браузере всё равно горят красным, папки insightface в принципе не было. Я никогда не тратил столько времени впустую...
Can u share the comfyui workflow pls🥺
Link in description
NAH SHIT TOO HARD 😂
😅😅
Check the pinned comment
nice tutorial
Thank you
Theres a part in the video where you stop explaining a step by step, it be nice if you make a resume or conclusion. cos damn its kinda easy, but TO install the nodes left, there's where you put code and get lost. PLssssssss :C
Link bro please
Link to what? There's some in the description
I had a good start but kind of gave up on the multiple repositories needing installation still, missed most of them lol