smandava98

Choice of densepose

1

Hi. Thanks for the great work. What was the reasoning behind the choice for densepose? I see that the segmentation masks are not always accurate so it leads to some...

Confused on how to use pretrained backbone where you need full details of the image

1

Hi, Thank you for the great work. I am conceptually confused on how to use this backbone. I read through your code on the depth estimation head but it is...

Questions about feature maps/patch size

1

I'm still not entirely sure how ViTs learn various features as I've always intuitively understood convnets better. How can I get feature maps that include the color? For my use...

Is mesh only way to save output of a NeRF?

1

Hi. I am doing some research in 3D graphics and I am attempting to somehow save the output of a NeRF model as a point cloud or voxel grid. Is...

Worse performance than FGT on custom video

5

Hi. I ran the HQ model and I keep getting worse performance than the other inpainting models that E2FGVI is supposed to be better than. E2FGVI for some reason just...

CUDA OOM for ~1-2 min long videos

3

Hi, My videos are ~ 1 min long (30 FPS) which leads to between 1k-2k frames... I am on an A100 GPU but I keep running out of memory. I...

Confused on the AMASS SMPL format

1

Hi @EricGuo5513 ! Thanks for the great dataset. I am struggling to read how you convert SMPL to the 263 dimensional vector that is used with the AMASS dataset. I...

FPS

1

Hi. Great work! I just had one simple question: What was the reasoning behind why motion was downsampled to 20 fps? Is it because of compute constraints or you found...

taking long time to test compared to FGT

2

Hi @hitachinsk! Firstly, great work and am very impressed. I liked the performance of ISVI as compared to FGT however I noticed that (for a video of only 60 frames),...

How do you get different masks when one object is inside the other? Because in that case it seems it uses one object id for both masks.