HW_LIANG

decoder structure of C3D, R3D, R21D

3

Based on your repo, it seems that you are using the same decoder structure for all of the backbones(c3d, r3d, r(2+1)d). But in your paper, it seems you used different...

R(2+1)D backbone

1

We noticed in your paper that you got an amazing result with R(2+1)D backbone. Could you public your implementation of R(2+1)D model as well as the best checkpoint? Thanks!

when will you release the training code? Thanks!

ModuleNotFoundError: No module named 'mmdet3d.ops.iou_loss.iou_loss_ext'

3

Hi, I strictly followed the installation instructions. But I always met this problem using the mmdetection v0.13.0 with the files in ../mmdet3d_full. ' File "MUTR3D/mmdetection3d/mmdet3d/ops/iou_loss/sort.py", line 4, in from .iou_loss_ext...

when can the full code be released?

how to control motion magnitude

1

In the train.py file, I noticed for motion control, the motion magnitudes are computed in both RGB(batch["motion"]) and latent space (latent_motion = calculate_latent_motion_score(latents)), but only the latter is used in...

Hello, I was trying to run the demo.py file but I had the error below: : cannot connect to X server I ran it in a docker environment. Do you...

Image generation in stage 1

1

Hey, Congrats on this great work! One point I'm not very clear about stage1. For single-image to 4D generation, will you generate F+V images, or F\timesV images as a matrix...

I was trying to finetune the VAE decoder. I met this error: raise ValueError("Attempting to unscale FP16 gradients."). I set args.mixed_precision as fp16. And here is how I load the...