qinglew
qinglew
The architecture of the network is different from the original when concatenate the features.
I build the pointnet2_ops on Windows 10, and the version of Python is 3.7. When I use the `furthest_point_sample` and `gather_operation` to sample points from a dense point cloud, I...
## Description I compiled this code successfully and got a executable file `mesh_sampling`. After I running the command `./mesh_sampling -h`, it shows that `-n_samples` is the option of the number...
According to your code of flow2flow which is for unpaired image to image translation, you just shuffle the data before training. I think you need to shuffle the target images...
You didn't use the graph-based encoder the paper described.
For each object, all parts can be composed into a complete mesh. Are the meshes are normalized into unit sphere or 3d bounding box?
In line 62 of `optimization/trainer_baseline.py`, the return values of `model.forward(y, x)` is `z, x_bpd` represents the latent code and minus log-det respectively according to class `FlowModel` in `models/architectures/condNF.py`. However, you...
Thank you very much for releasing the code of DPF-Net. I read it carefully and have a doubt for the loss. Here are the details: 1. In `lib/networks/losses.py`, the class...