FluxML/Flux.jl

Weights shape not validated against kernel, channels

Open

#2,506 opened on Oct 25, 2024

View on GitHub
 (5 comments) (1 reaction) (0 assignees)Julia (4,725 stars) (619 forks)batch import
good first issuehelp wanted

Description

weights = Flux.kaiming_normal()(3, 3, 1)
Conv((3, 3), 1 => 1; pad = (1, 1), init = (_...) -> weights)
# Conv((3,), 3 => 1, pad=1)  # 10 parameters

weights = Flux.kaiming_normal()(3, 3, 1, 1)
Conv((3, 3), 1 => 1; pad = (1, 1), init = (_...) -> weights)
# Conv((3, 3), 1 => 1, pad=1)  # 10 parameters

I wanted to strictly specify the weight init for testing, but encountered this odd result. I think there should be validation to ensure that the weight shape matches the kernel size and input channels, and error if there is a mismatch.

Contributor guide

Tech stack
None
Domain
machine learning
Issue type
feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-3 hours
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
fresh
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
clear
Prerequisites
Understanding of Conv layer weight shape conventions
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
70
Research direction
Look into the Conv layer initialization code in src/layers/conv.jl. Add validation to check that the provided weight tensor (via init) matches the expected shape based on kernel size and input/output channels. The error should be thrown if there is a mismatch, similar to how other shape mismatches are handled.