Closed acampbel closed 3 years ago
Thanks Andy, this looks really great!
One small comment, there's now two .prj files, one for the MATLAB project and one for the toolbox. Unless you open up MATLAB and decipher the little icon, it's not obvious which one is which. Is there a naming convention for these two, and should we rename one or both? What do you think?
BTW I never knew CircleCI gives you (some) free GPU build time!
I don’t know of any naming convention, but to me it does seem like a good idea to name them clearly so a contributor knows which is the project and which is the toolbox packaging file.
I wasn’t aware that the gpu support was free. I tried adding a gpu image to my fork and tried a build but it failed saying the gpu images weren’t included in my plan. Have you gotten it to work? That would be really cool to see in action.
On Aug 25, 2020, at 5:01 PM, Justin notifications@github.com<mailto:notifications@github.com> wrote:
Thanks Andy, this looks really great!
One small comment, there's now two .prj files, one for the MATLAB project and one for the toolbox. Unless you open up MATLAB and decipher the little icon, it's not obvious which one is which. Is there a naming convention for these two, and should we rename one or both? What do you think?
BTW I never knew CircleCI gives you (some) free GPU build time!
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/matlab-deep-learning/mtcnn-face-detection/pull/16#issuecomment-680267999, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ALIBBPG5WHGPA2QYKTW7HGTSCQRENANCNFSM4QJ3DI7A.
Ah, you're right the availability of GPU builders seems to be very limited right now, only on their "custom" plan.
Make into a MATLAB project
install
,uninstall
andprojectPaths
since projects takes care of the pathgenerateDAGNetVersions
and register as a startup file for the project to handle non-path related work previously done in installLeverage the circleci orb for streamlined CI
runAllTests
since this is handled through the orb features in the yaml