DPatch icon indicating copy to clipboard operation
DPatch copied to clipboard

mAP has hardly decreased

Open hoopool opened this issue 5 years ago • 9 comments

I have try to run train.py of the project, and run test.py of the project using the latest patch that saved by train.py (without using the untargetd_DPATCH which download from the url that given at the README.md file). after 100 epoch, it got 0.7095 mAP, but after 9757 epoch, it still got 0.7040 mAP. I just wander it there any parameters that I must modified before run train.py? I only modified max_epoch value in cfgs.exps.darknet19_exp1.py to run more epoch.

than you~

hoopool avatar Sep 26 '19 01:09 hoopool

same problem here. I've tried to modify the threshold values, increase the batch size, adjusted the learning rate, momentum, increase the epoch, etc.. but to no avail. The mAP is still way above the 0.6 mark (usually hovers around 0.7). Is there a way to verify if the patch (416x416 and reduced before attaching to image) is indeed attached to the image prior to running YOLO?

Thanks!

quahez avatar Oct 04 '19 08:10 quahez

same problem here. I've tried to modify the threshold values, increase the batch size, adjusted the learning rate, momentum, increase the epoch, etc.. but to no avail. The mAP is still way above the 0.6 mark (usually hovers around 0.7). Is there a way to verify if the patch (416x416 and reduced before attaching to image) is indeed attached to the image prior to running YOLO?

Thanks!

Did you succeed in training and testing? can you tell me your email, my email is [email protected].

xubo0321 avatar Nov 01 '19 06:11 xubo0321

I have try to run train.py of the project, and run test.py of the project using the latest patch that saved by train.py (without using the untargetd_DPATCH which download from the url that given at the README.md file). after 100 epoch, it got 0.7095 mAP, but after 9757 epoch, it still got 0.7040 mAP. I just wander it there any parameters that I must modified before run train.py? I only modified max_epoch value in cfgs.exps.darknet19_exp1.py to run more epoch.

than you~ Can we discuss the method together, I also have some problems, can you email me or tell me your email, my email is [email protected].

xubo0321 avatar Nov 01 '19 06:11 xubo0321

@hoopool @quahez @xubo0321 I also find the same question, have you finished? can you give me some suggestion? thank you

CNHNLP avatar Mar 10 '20 01:03 CNHNLP

mAP has hardly decreased 我测试了我认为可以调整的参数,但是实验效果依旧不是很好,包括使用了您这边给出的完整版的参数,请问您这边可以给出完整的参数配置么?

chenxiaoyupetter avatar Aug 11 '20 03:08 chenxiaoyupetter

I have the same problem,have you solved it?

xleslie avatar Sep 18 '20 10:09 xleslie

i will help if i can slove the problem

---Original--- From: "xleslie"<[email protected]> Date: Fri, Sep 18, 2020 18:20 PM To: "veralauee/DPatch"<[email protected]>; Cc: "chenxiaoyupetter"<[email protected]>;"Comment"<[email protected]>; Subject: Re: [veralauee/DPatch] mAP has hardly decreased (#3)

I have the same problem,have you solved it?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

chenxiaoyupetter avatar Sep 18 '20 10:09 chenxiaoyupetter

i will help if i can slove the problem ---Original--- From: "xleslie"<[email protected]> Date: Fri, Sep 18, 2020 18:20 PM To: "veralauee/DPatch"<[email protected]>; Cc: "chenxiaoyupetter"<[email protected]>;"Comment"<[email protected]>; Subject: Re: [veralauee/DPatch] mAP has hardly decreased (#3) I have the same problem,have you solved it? — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

Thanks!I will also try to solve the problem,my email is [email protected]

xleslie avatar Sep 18 '20 10:09 xleslie

I have the same problem here.

zxydi1992 avatar Sep 24 '20 05:09 zxydi1992