search

JeffreyXiang / ivid

PyTorch implementation of the ICCV paper "3D-aware Image Generation using 2D Diffusion Models"
https://jeffreyxiang.github.io/ivid/
MIT License
306 stars 4 forks source link

Inference in our data based on pre-trained model #10

Open Masoud-kamali opened 8 months ago

Masoud-kamali commented 8 months ago

Hi,

Just a quick question, could it be possible to inference in our images using pre-trained models? if so, please describe a little bit.

Thanks

JeffreyXiang commented 7 months ago

To inference images without corresponding depth maps or not belongs to any ImageNet-1k categories is challenging for the ImageNet pretrained models.

You could estimate depth map using a monocular depth estimator and assign the image a closest category inside ImageNet-1k for inference. But the performance is not guaranteed.