New version: MLJFlux v0.3.0

[JuliaRegistrator]

• Registering package: MLJFlux
• Repository: https://github.com/FluxML/MLJFlux.jl
• Created by: @ablaom
• Version: v0.3.0
• Commit: 37b5f31727524899203abda197171a7a7696f673
• Reviewed by: @ablaom
• Reference: Issue to trigger new releases FluxML/MLJFlux.jl#47 (comment)
• Description: Wrapping deep learning models from the package Flux.jl for use in the MLJ.jl toolbox
• Release notes:

• (mildy breaking) Extend compatibility requirement: Flux = "0.13, 0.14". If using Flux 0.14 or later, it may be necessary to additionally import cuDNN to recover the functionality of previous versions (e.g. if using convolution neural networks in builders) (New version HilbertSpaceFillingCurve: 0.0.3 #232)

• Bump compatibility requirement: Metalhead = "0.8"

[github-actions]

Your new version pull request met all of the guidelines for auto-merging and is scheduled to be merged in the next round.

---

If you want to prevent this pull request from being auto-merged, simply leave a comment. If you want to post a comment without blocking auto-merging, you must include the text [noblock] in your comment. You can edit blocking comments, adding [noblock] to them in order to unblock auto-merging.