Ok, I tried to figure this out myself use ChatGPT, I can't. Here is my error. IPAdapterAdvanced Error(s) in loading state_dict for Resampler: size mismatch for proj_in.weight: copying a param with shape torch.Size([768, 1280]) from checkpoint, the shape in current model is torch.Size([768, 1664]).
Hi! Do you know how to make IC light try to set light from the front? I always get the light from the back side.. And it's confusing since the masking only 2 dimensional. lol :D
I keep getting this error: Error occurred when executing FrequencyCombination: operands could not be broadcast together with shapes (1024,3,1024) (64,3,64) Suggestions?
Hello, good afternoon, could you help me? I am having a problem with the IP-Adapter, I am using the following models checkpoints= realisticVisionV51_v51VAE.safetensors ic-light= iclight_sd15_fc_unet_ldm.safetensors clip_vision= model.safetensors IPAdapter= ip-adapter-plus_sd15.safetensors but I get this error Error occurred when executing IPAdapterAdvanced: Error(s) in loading state_dict for Resampler: size mismatch for proj_in.weight: copying a param with shape torch.Size([768, 1280]) from checkpoint, the shape in current model is torch.Size([768, 1024]). How should I do to correct this error? Thank you very much in advance, it is the best video I saw that explains how to correctly use ic-light in comfyui
Really helpful; thank you for your video!
perfect workflow what i looking for !
thanks! happy you find it!
Another step further in terms of control in sd. Thanks very clear explanation
thanks!
the results are quite nice, I'm looking forward for SDXL version.
definitely
Great tutorial, Thanks! How does this work on full body photos? I am getting distorted faces :-( Thanks a lot
thanks! like any sd 1.5, you can use reactor
Thank you! amazing new nodes.
it's really nice!
great... I wonder if it can simulate photo studio with controling lights...
thanks! I think that in the near future it will be possible
can you input image instead of generate in bacgkround ?
Ok, I tried to figure this out myself use ChatGPT, I can't. Here is my error. IPAdapterAdvanced
Error(s) in loading state_dict for Resampler:
size mismatch for proj_in.weight: copying a param with shape torch.Size([768, 1280]) from checkpoint, the shape in current model is torch.Size([768, 1664]).
Amazing. Thank you :)
thx!
Hi! Do you know how to make IC light try to set light from the front? I always get the light from the back side.. And it's confusing since the masking only 2 dimensional. lol :D
Thanks alot
you welcome!
I keep getting this error: Error occurred when executing FrequencyCombination:
operands could not be broadcast together with shapes (1024,3,1024) (64,3,64)
Suggestions?
Hello, good afternoon, could you help me? I am having a problem with the IP-Adapter, I am using the following models
checkpoints= realisticVisionV51_v51VAE.safetensors
ic-light= iclight_sd15_fc_unet_ldm.safetensors
clip_vision= model.safetensors
IPAdapter= ip-adapter-plus_sd15.safetensors
but I get this error
Error occurred when executing IPAdapterAdvanced:
Error(s) in loading state_dict for Resampler:
size mismatch for proj_in.weight: copying a param with shape torch.Size([768, 1280]) from checkpoint, the shape in current model is torch.Size([768, 1024]).
How should I do to correct this error?
Thank you very much in advance, it is the best video I saw that explains how to correctly use ic-light in comfyui
not sure, if u use any resize node , try to replace it with another
I'm not sure why even though I have a good VEA foreground, it seems to still be undergoing modifications, which is not what I intended.
not sure what do you mean
I've tried this but the faces end up getting really messed up. Is there a trick to preserving the face quality?
You can upscale the image that can help also before creating the initial image you can use the IP adapter
you can use reactor as last step
the mask is not created, the image appears to be in the background instead of the foreground
what do you mean?
Too bad there's no SDXL
will have different solutions soon