test: Metal test now works

[avik-pal]

No description provided.

[gemini-code-assist]

Summary of Changes

Hello @avik-pal, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request focuses on improving the reliability and consistency of tests related to the Metal backend. It updates the test suite to correctly reflect the passing status of a key weight initializer test on Metal and pins the Metal package dependency to a specific version to prevent potential compatibility issues in the future.

Highlights

• Metal Test Fix: The kaiming_normal weight initializer test is no longer marked as broken for the Metal backend, indicating that it now passes successfully.
• Metal Package Version Pin: The test configuration for the Metal backend now explicitly specifies version 1.9 for the Metal package, ensuring consistent test environments.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request fixes a failing test for the Metal backend by pinning the Metal.jl package to version 1.9. The change to enable the test is correct. However, I've suggested a small improvement to the version pinning in runtests.jl to make it more flexible for future updates, which will improve maintainability.

[gemini-code-assist]

Pinning the Metal package to the exact version 1.9 might be too restrictive and could prevent your test environment from receiving important non-breaking bug fixes or performance improvements in future patch releases. It's generally better to allow for compatible updates.

Consider using a caret ^ for the version specifier, like version="^1.9". This will allow any version from 1.9.0 up to (but not including) 2.0.0, following semantic versioning. This makes your test setup more robust and less likely to require manual updates for minor package releases.

    push!(EXTRA_PKGS, PackageSpec(; name="Metal", version="^1.9"))

[github-actions]

Benchmark Results (Julia v1.11)

Time benchmarks

	main	4e0166f...	main / 4e0166f...
basics/MHA	4.09 ± 0.31 μs	4.12 ± 0.36 μs	0.993 ± 0.12
basics/MHA (first run)	4.34 ± 0.41 μs	4.46 ± 0.61 μs	0.973 ± 0.16
basics/MHA reactant	0.0609 ± 0.0055 ms	0.0661 ± 0.0071 ms	0.922 ± 0.13
basics/MHA reactant (comp + run)	0.14 ± 0.0038 s	0.147 ± 0.0039 s	0.952 ± 0.036
basics/conv	12.6 ± 16 μs	12.2 ± 16 μs	1.03 ± 1.9
basics/conv (first run)	10.4 ± 2.1 μs	10.7 ± 2.1 μs	0.976 ± 0.27
basics/conv reactant	0.0508 ± 0.0019 ms	0.0516 ± 0.0024 ms	0.983 ± 0.059
basics/conv reactant (comp + run)	0.0994 ± 0.0037 s	0.101 ± 0.0042 s	0.984 ± 0.055
basics/dense	0.181 ± 0.01 μs	0.181 ± 0.011 μs	1 ± 0.082
basics/dense (first run)	0.2 ± 0.01 μs	0.2 ± 0.01 μs	1 ± 0.071
basics/dense reactant	0.046 ± 0.0016 ms	0.0505 ± 0.0049 ms	0.91 ± 0.095
basics/dense reactant (comp + run)	0.0816 ± 0.002 s	0.0818 ± 0.0044 s	0.996 ± 0.059
time_to_load	0.886 ± 0.012 s	0.912 ± 0.035 s	0.972 ± 0.039

Memory benchmarks

	main	4e0166f...	main / 4e0166f...
basics/MHA	0.087 k allocs: 6.05 kB	0.087 k allocs: 6.05 kB	1
basics/MHA (first run)	0.087 k allocs: 6.05 kB	0.087 k allocs: 6.05 kB	1
basics/MHA reactant	19 allocs: 0.578 kB	19 allocs: 0.578 kB	1
basics/MHA reactant (comp + run)	18 k allocs: 1.37 MB	18 k allocs: 1.37 MB	1
basics/conv	0.038 k allocs: 5.12 kB	0.038 k allocs: 5.12 kB	1
basics/conv (first run)	0.038 k allocs: 5.12 kB	0.038 k allocs: 5.12 kB	1
basics/conv reactant	15 allocs: 0.438 kB	15 allocs: 0.438 kB	1
basics/conv reactant (comp + run)	6.14 k allocs: 0.813 MB	6.15 k allocs: 0.813 MB	1
basics/dense	2 allocs: 0.109 kB	2 allocs: 0.109 kB	1
basics/dense (first run)	2 allocs: 0.109 kB	2 allocs: 0.109 kB	1
basics/dense reactant	15 allocs: 0.422 kB	15 allocs: 0.422 kB	1
basics/dense reactant (comp + run)	5.91 k allocs: 0.796 MB	5.9 k allocs: 0.796 MB	1
time_to_load	0.159 k allocs: 11.2 kB	0.159 k allocs: 11.2 kB	1