search

thu-ml / cond-image-leakage

Official implementation for "Identifying and Solving Conditional Image Leakage in Image-to-Video Diffusion Model" (NeurIPS 2024)
Apache License 2.0
292 stars 28 forks source link

Does the 1024x576 model work for portrait images? #2

Closed doogyhatts closed 4 months ago

doogyhatts commented 4 months ago

Just wondering about the usage of the model for the 576x1024 portrait resolution. Or is the 1024x576 model specifically catered to only the landscape format?

zhuhz22 commented 4 months ago

Yes. We finetuned the original DynamiCrafter model on Webvid-2M,which contains portraits,so the model is expected to work for portrait images.You could have a try on examples/DynamiCrafter/inference_CIL_1024.sh .Here are some demos.

doogyhatts commented 4 months ago

Thanks for the reply. I will feedback to Kijai.