FluxML / FluxML/FluxJS.jl

compile function not working in case of Conv layers.

Open
#4 1 comment 0 reactions 0 assignees View on GitHub
Dominant language
Julia
Stars
41
Forks
8
PR merge metrics
No merged PRs in 30d

Description

`FluxJS.compile()` doesn't seem to work when Convolution layers are involved:

```
>>> using Flux
>>> using FluxJS
>>> m = Chain(
Conv((2,2), 1=>16, relu),
x -> maxpool(x, (2,2)),
Conv((2,2), 16=>8, relu),
x -> maxpool(x, (2,2)),
x -> reshape(x, :, size(x, 4)),
Dense(288, 10), softmax)
>>> FLuxJS.compile("mnist",m, rand(10))
Error: StackOverflow
```
The error message is too big.

Contributor guide

No contributing guide indexed for this repository

Research direction

Start with the shown FluxJS.compile("mnist", m, rand(10)) reproduction using the Chain with Conv and maxpool layers, and reproduce the StackOverflow error. Trace how the Conv layers are handled during compilation; done means this example compiles without StackOverflow and does not produce an excessively large error message.

Written by the indexing model from the issue text.

Assessment

Tech stack
julia
Domain
machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.