search

showlab / BoxDiff

[ICCV 2023] BoxDiff: Text-to-Image Synthesis with Training-Free Box-Constrained Diffusion
227 stars 11 forks source link

cross_attention.py is deprecated; attention_processor.py should be used instead #13

Open spejamas opened 2 months ago

spejamas commented 2 months ago

ptp_utils.py imports CrossAttention from diffusers/models/cross_attention.py:

from diffusers.models.cross_attention import CrossAttention

But as of July 26 2023, cross_attention.py is deprecated; attention_processor.py should be used instead:

https://github.com/huggingface/diffusers/pull/4299

zjysteven commented 1 month ago

I made a diffuser-compatible pipeline here https://github.com/huggingface/diffusers/tree/main/examples/community#stable-diffusion-boxdiff.