apache/mxnet

3D Upsampling

Open

#18,493 opened on Jun 5, 2020

View on GitHub
 (15 comments) (3 reactions) (0 assignees)C++ (20,808 stars) (6,711 forks)batch import
C++Feature requestgood first issue

Description

Description

Currently MXNet does not support upsampling on 5D tensors (3D images). There is more and more work being done in the medical field where 3D images are very common. As MXNet supports 3D transposed convolutions, 3D Upsampling would be very helpful as well, since there are instances where upsampling has been shown to be as effective if not more so than a transposed convolutions.

References

PyTorch, TF both support 3D upsampling.

Contributor guide

Tech stack
cpppython
Domain
machine learningai
Issue type
feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
4
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
3-5 days
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
stale
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
mostly clear
Prerequisites
C++Deep learning basicsMXNet internals
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
20
Research direction
Review existing 2D upsampling operator in MXNet (likely in src/operator/nn/upsampling.cc). Study the 3D transposed convolution implementation for pattern reference. Check PyTorch's interpolate function for 3D tensors to align behavior. The main challenge is supporting 5D tensors and adding backward pass if needed.