baggepinnen / FluxOptTools.jl

Use Optim to train Flux models and visualize loss landscapes
MIT License
59 stars 4 forks source link

Update landscape plot #20

Closed mBarreau closed 1 year ago

mBarreau commented 1 year ago

The current landscape plot is a plot recipe. However, the argument l is already in used for the linewidth size, conflicting thus with the size of the landscape plot. This is not easily possible to draw several times the same landscape so a seed has been added. Here are the proposed modifications:

The tests have been updated and seemingly better looking landscape command has been added to the readme (together with the reference to the original paper proposing this method).

codecov[bot] commented 1 year ago

Codecov Report

Patch coverage: 87.50% and project coverage change: +0.15 :tada:

Comparison is base (02a8f03) 95.40% compared to head (a7e1d42) 95.55%.

Additional details and impacted files ```diff @@ Coverage Diff @@ ## master #20 +/- ## ========================================== + Coverage 95.40% 95.55% +0.15% ========================================== Files 2 2 Lines 87 90 +3 ========================================== + Hits 83 86 +3 Misses 4 4 ``` | Flag | Coverage Δ | | |---|---|---| | unittests | `95.55% <87.50%> (+0.15%)` | :arrow_up: | Flags with carried forward coverage won't be shown. [Click here](https://docs.codecov.io/docs/carryforward-flags?utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Fredrik+Bagge+Carlson#carryforward-flags-in-the-pull-request-comment) to find out more. | [Impacted Files](https://codecov.io/gh/baggepinnen/FluxOptTools.jl/pull/20?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Fredrik+Bagge+Carlson) | Coverage Δ | | |---|---|---| | [src/FluxOptTools.jl](https://codecov.io/gh/baggepinnen/FluxOptTools.jl/pull/20?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Fredrik+Bagge+Carlson#diff-c3JjL0ZsdXhPcHRUb29scy5qbA==) | `91.11% <87.50%> (+0.63%)` | :arrow_up: | Help us with your feedback. Take ten seconds to tell us [how you rate us](https://about.codecov.io/nps?utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Fredrik+Bagge+Carlson). Have a feature suggestion? [Share it here.](https://app.codecov.io/gh/feedback/?utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Fredrik+Bagge+Carlson)

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.