Data augmentation M1,M2,M3 in the paper

[Kevinz-code]

Hello, I'm very interested in your implementation and I want to implement your code to get a start.
However, I was wondering the meaning of M1, M2, M3 in your paper.
I mean, does it mean:
M1: randomflipping
M2: randomresizecrop
M3: mixup
or
M1: randomflipping
M2: randomfilpping + randomresizecrop
M3: randomfilpping + randomresizecrop + mixup

I'm looking forward to your answer, thanks in advance!

[hellowangqian]

The later is what I mean, thank you for pointing out the confusion. Get Outlook for Android<https://aka.ms/ghei36>

…

________________________________ From: kevin655 <notifications@github.com> Sent: Tuesday, December 8, 2020 6:25:32 AM To: hellowangqian/multi-label-image-classification <multi-label-image-classification@noreply.github.com> Cc: Subscribed <subscribed@noreply.github.com> Subject: [hellowangqian/multi-label-image-classification] Data augmentation M1,M2,M3 in the paper (#3) Hello, I'm very interested in your implementation and I want to implement your code to get a start. However, I was wondering the meaning of M1, M2, M3 in your paper. I mean, does it mean: M1: randomflipping M2: randomresizecrop M3: mixup or M1: randomflipping M2: randomfilpping + randomresizecrop M3: randomfilpping + randomresizecrop + mixup I'm looking forward to your answer, thanks in advance! — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<#3>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABX467VAP3T6WUJMSZI3CH3STXBFZANCNFSM4URQBASA>.

[Kevinz-code]

The later is what I mean, thank you for pointing out the confusion. Get Outlook for Androidhttps://aka.ms/ghei36
…
________________________________ From: kevin655 notifications@github.com Sent: Tuesday, December 8, 2020 6:25:32 AM To: hellowangqian/multi-label-image-classification multi-label-image-classification@noreply.github.com Cc: Subscribed subscribed@noreply.github.com Subject: [hellowangqian/multi-label-image-classification] Data augmentation M1,M2,M3 in the paper (#3) Hello, I'm very interested in your implementation and I want to implement your code to get a start. However, I was wondering the meaning of M1, M2, M3 in your paper. I mean, does it mean: M1: randomflipping M2: randomresizecrop M3: mixup or M1: randomflipping M2: randomfilpping + randomresizecrop M3: randomfilpping + randomresizecrop + mixup I'm looking forward to your answer, thanks in advance! — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<#3>, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ABX467VAP3T6WUJMSZI3CH3STXBFZANCNFSM4URQBASA.

Thank you for your accurately reply,
one thing I also want to know is that when using mixup in COCO, it's better to train more steps before decay learining rate ? (I saw that in your 'resnet101_model1fc.py', the step_size was set 'step_size=5').
Best wishes.