PU-GAN
PU-GAN copied to clipboard
liruihui
PU-GAN: a Point Cloud Upsampling Adversarial Network, ICCV, 2019
I am trying to trian the model for up_ratio=8 but I realized there are a lot of error in the code that doesn't make it possible to train for that...
Hi, thank you so much for your hard work for PU-GAN. I encountered an error when running `bash compile.sh` as shown [here](https://github.com/liruihui/PU-GAN/tree/master/Docker). I followed the instruction to install the packages....
I'm sorry to bother you, but I really want to know how you simplified the `get_uniform_loss` function. I didn't see the `imbalance` appeared in your paper and I want to...
I download the pretrained model from Google Drive by link: https://drive.google.com/open?id=12kWoB0-_tflq65RNpJEnNGTTwPXa6IOH Run project in Docker with provided Dockerfile, but got error when restoring model: ``` pu-gan_1 | None pu-gan_1 |...
In the "Common/ops.py", the Line81: net = tf.reshape(net,[tf.shape(net)[0], up_ratio, -1, tf.shape(net)[-1]]) why is tf.shape(net)[-1]
Hi , The description of the scheduler in the paper is : “ After 50k iterations, we gradually reduce both rates by a decay rate of 0.7 per 50k iterations...
Thank you for this great work. Just a simple question, is the pretrain model that you share was train with exactly the code in the github. And is is the...
Hi, I was wondering how can you generate exact number 8192 points from poisson disc sampling as your ground truth ? Because when I run evaluation.py, it seems that **pred...
First of all, thank you for sharing the code. I tried running the code in the docker environment you provided on two separate computers, one with RTX TITAN and the...
How to test the upsampling effect of autonomous driving point cloud data, such as nuscence dataset. How to generate .xyz data, do I need to fuse frame-by-frame point clouds first?
