search

ThibaultGROUEIX / AtlasNet

This repository contains the source codes for the paper "AtlasNet: A Papier-Mâché Approach to Learning 3D Surface Generation ". The network is able to synthesize a mesh (point cloud + connectivity) from a low-resolution point cloud, or from an image.
http://imagine.enpc.fr/~groueixt/atlasnet/
MIT License
678 stars 120 forks source link

What's the meaning of conv1d with the same input and output size #48

Closed BrandoZhang closed 3 years ago

BrandoZhang commented 4 years ago

Hi, thanks for your great work! When I studied your code, I found the following: https://github.com/ThibaultGROUEIX/AtlasNet/blob/8dd00e940dae208209ae97b6694e647eb5b1a06b/auxiliary/model.py#L132 I am wondering what's the meaning of 1d convolution that takes the same size of input and output.

ThibaultGROUEIX commented 4 years ago

Dear @BrandoZhang ,

it's like a fully connected layer with self.bottleneck_size input neurons and self.bottleneck_size output neurons. You can check out pytorch doc on conv1d.

Best, Thibault