Some problems in calculating mAP in eval.py

[Cynthia-324]

Hi,thank you for your sharing~ When I use this project training the VOCdataset, I can use the demo.py to test the pictures, but when using eval.py calculating mAP, I get very low mAP（No boxes to NMS, just 0.03）,is there something wrong with eval.py?

[y78h11b09]

I have the same problem, I got the hightest mAp just 0.19 after trained 186 epoches

[qtw1998]

I have the same problem, I got the hightest mAp just 0.19 after trained 186 epoches

you can have a check now, the codes have been updated.

[y78h11b09]

Thank you very much. Yesternight, I also found the code updated. and I found the eval code before updated had a problem, which is red, i.e, for label in range(valid_dataset.num_classes()): false_positives = np.zeros((0,)) true_positives = np.zeros((0,)) scores = np.zeros((0,)) num_annotations = 0.0

for i in range(valid_dataset.num_classes()):
    detections = all_detections[i][label]
    annotations = all_annotations[i][label]
    num_annotations += annotations.shape[0]
    detected_annotations = []

what do you think the code? , and please send me the new result that you tested again by checked code.

[email protected]

From: Qiao Tianwei Date: 2020-01-07 11:23 To: toandaominh1997/EfficientDet.Pytorch CC: y78h11b09; Comment Subject: Re: [toandaominh1997/EfficientDet.Pytorch] Some problems in calculating mAP in eval.py (#61) I have the same problem, I got the hightest mAp just 0.19 after trained 186 epoches you can have a check now, the codes have been updated. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[quangtn266]

Has anyone solutions to this problem?