423A35C7/pytorch3d
1
0
Fork 0
mirror of https://github.com/facebookresearch/pytorch3d.git synced 2025-12-22 23:30:35 +08:00
Code Issues Packages Projects Releases Wiki Activity

Setup website with docusaurus (#11)

Browse Source 
Summary:
Set up landing page, docs page, and html versions of the ipython notebook tutorials.
Pull Request resolved: https://github.com/fairinternal/pytorch3d/pull/11

Reviewed By: gkioxari

Differential Revision: D19730380

Pulled By: nikhilaravi

fbshipit-source-id: 5df8d3f2ac2f8dce4d51f5d14fc336508c2fd0ea
This commit is contained in:
Nikhila Ravi
2020-02-04 17:25:51 -08:00
committed by Facebook Github Bot
parent e290f87ca9
commit 15d3a4557e
48 changed files with 1900 additions and 72 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
docs/notes/batching.md
View File

@@ -1,8 +1,13 @@
---
hide_title: true
sidebar_label: Batching
---
# Batching
In deep learning, every optimization step operates on multiple input examples for robust training. Thus, efficient batching is crucial. For image inputs, batching is straighforward; N images are resized to the same height and width and stacked as a 4 dimensional tensor of shape `N x 3 x H x W`. For meshes, batching is less straighforward.
<img src="../figs/batch_intro.png" alt="batch_intro" align="middle"/>
<img src="assets/batch_intro.png" alt="batch_intro" align="middle"/>
## Batch modes for meshes
@@ -12,13 +17,13 @@ Assume you want to construct a batch containing two meshes, with `mesh1 = (v1: V
* Padded: The padded representation constructs a tensor by padding the extra values. Specifically, `meshes.verts_padded()` returns a tensor of shape `2 x max(V1, V2) x 3` and pads the extra vertices with `0`s. Similarly, `meshes.faces_padded()` returns a tensor of shape `2 x max(F1, F2) x 3` and pads the extra faces with `-1`s.
* Packed: The packed representation concatenates the examples in the batch into a tensor. In particular, `meshes.verts_packed()` returns a tensor of shape `(V1 + V2) x 3`. Similarly, `meshes.faces_packed()` returns a tensor of shape `(F1 + F2) x 3` for the faces. In the packed mode, auxiliary variables are computed that enable efficient conversion between packed and padded or list modes.
<img src="../figs/batch_modes.gif" alt="batch_modes" height="450" align="middle" />
<img src="assets/batch_modes.gif" alt="batch_modes" height="450" align="middle" />
## Use cases for batch modes
The need for different mesh batch modes is inherent to the way pytorch operators are implemented. To fully utilize the optimized pytorch ops, the [Meshes][meshes] data structure allows for efficient conversion between the different batch modes. This is crucial when aiming for a fast and efficient training cycle. An example of this is [Mesh R-CNN][meshrcnn]. Here, in the same forward pass different parts of the network assume different inputs, which are computed by converting between the different batch modes. In particular, [vert_align][vert_align] assumes a *padded* input tensor while immediately after [graph_conv][graphconv] assumes a *packed* input tensor.
<img src="../figs/meshrcnn.png" alt="meshrcnn" width="700" align="middle" />
<img src="assets/meshrcnn.png" alt="meshrcnn" width="700" align="middle" />
[meshes]: https://github.com/facebookresearch/pytorch3d/blob/master/pytorch3d/structures/meshes.py