Its not working for me

[letscreate321]

Phi3Transformer does not support an attention implementation through torch.nn.functional.scaled_dot_product_attention yet. Please request the support for this architecture: huggingface/transformers#28005. If you believe this error is a bug, please open an issue in Transformers GitHub repository and load your model with the argument attn_implementation="eager" meanwhile. Example: model = AutoModel.from_pretrained("openai/whisper-tiny", attn_implementation="eager")

[1038lab]

update the custom node

[deceani]

I get the same error, just installed today, tried manager install, and manual git install, I have latest version of comfyui, all is updated and still doesnt work

[deceani]

I got rid of that error if i used the update_comfyui_and_python_dependencies.bat from the update folder on comfyui. Now i got an torch.OutOfMemoryError: Allocation on device because I dont have enoguh vram on this pc but I will try on another one.

[1038lab]

I get the same error, just installed today, tried manager install, and manual git install, I have latest version of comfyui, all is updated and still doesnt work

what's error message you got,

[1038lab]

I got rid of that error if i used the update_comfyui_and_python_dependencies.bat from the update folder on comfyui. Now i got an torch.OutOfMemoryError: Allocation on device because I dont have enoguh vram on this pc but I will try on another one.

This model is somewhat VRAM-intensive, so it is recommended to limit the input image size to 512 or 768 pixels to reduce VRAM usage. Setting the output resolution to 1024x1024 should work fine, and the results will run as expected. Additionally, there is a significant difference in VRAM consumption when loading one image versus two, which may require further optimization.