test on the MOT17 test set, results NaN.

[yuzhiyiliu]

$ python test.py tracking --exp_id mot17_fulltrain_public --dataset mot --dataset_version 17test --pre_hm --ltrb_amodal --track_thresh 0.4 --pre_thresh 0.5 --load_model ../models/mot17_fulltrain_sc.pth --public_det --load_results ../data/mot17/results/test_det.json

|nemot17_half_public |################################| [2651/2652]|Tot: 0:18:20 |ETA: 0:00:01 |tot 0.411s (0.415s) |load 0.000s (0.000s) |pre 0.004s (0.005s) |net 0.405s (0.405s) |dec 0.001s (0.003s) |post 0.000s (0.001s) |merge 0.000s (0.000s) |track 0.000s (0.000s)
gt_type _val_half
gt_files ['../data/mot17/train/MOT17-02-FRCNN/gt/gt_val_half.txt', '../data/mot17/train/MOT17-05-FRCNN/gt/gt_val_half.txt', '../data/mot17/train/MOT17-04-FRCNN/gt/gt_val_half.txt', '../data/mot17/train/MOT17-13-FRCNN/gt/gt_val_half.txt', '../data/mot17/train/MOT17-11-FRCNN/gt/gt_val_half.txt', '../data/mot17/train/MOT17-10-FRCNN/gt/gt_val_half.txt', '../data/mot17/train/MOT17-09-FRCNN/gt/gt_val_half.txt']
02:06:40 INFO - Found 7 groundtruths and 7 test files.
02:06:40 INFO - Available LAP solvers ['scipy']
02:06:40 INFO - Default LAP solver 'scipy'
02:06:40 INFO - Loading files.
02:06:41 INFO - Comparing MOT17-02-FRCNN...
02:06:42 INFO - Comparing MOT17-05-FRCNN...
02:06:42 INFO - Comparing MOT17-11-FRCNN...
02:06:42 INFO - Comparing MOT17-09-FRCNN...
02:06:42 INFO - Comparing MOT17-04-FRCNN...
02:06:43 INFO - Comparing MOT17-13-FRCNN...
02:06:43 INFO - Comparing MOT17-10-FRCNN...
02:06:43 INFO - Running metrics
02:06:43 INFO - partials: 0.503 seconds.
02:06:44 INFO - mergeOverall: 0.513 seconds.
Rcll Prcn GT MT PT ML FP FN IDs FM MOTA MOTP num_objects
MOT17-02-FRCNN 0.0% nan% 53 0.0% 0.0% 100.0% 0.0% 100.0% 0.0% 0.0% 0.0% nan 9880
MOT17-05-FRCNN 0.0% nan% 71 0.0% 0.0% 100.0% 0.0% 100.0% 0.0% 0.0% 0.0% nan 3357
MOT17-11-FRCNN 0.0% nan% 44 0.0% 0.0% 100.0% 0.0% 100.0% 0.0% 0.0% 0.0% nan 4517
MOT17-09-FRCNN 0.0% nan% 22 0.0% 0.0% 100.0% 0.0% 100.0% 0.0% 0.0% 0.0% nan 2879
MOT17-04-FRCNN 0.0% nan% 69 0.0% 0.0% 100.0% 0.0% 100.0% 0.0% 0.0% 0.0% nan 24178
MOT17-13-FRCNN 0.0% nan% 44 0.0% 0.0% 100.0% 0.0% 100.0% 0.0% 0.0% 0.0% nan 3156
MOT17-10-FRCNN 0.0% nan% 36 0.0% 0.0% 100.0% 0.0% 100.0% 0.0% 0.0% 0.0% nan 5923
OVERALL 0.0% nan% 339 0.0% 0.0% 100.0% 0.0% 100.0% 0.0% 0.0% 0.0% nan 53890
02:06:44 INFO - partials: 0.811 seconds.
02:06:44 INFO - mergeOverall: 0.824 seconds.
IDF1 IDP IDR Rcll Prcn GT MT PT ML FP FN IDs FM MOTA MOTP IDt IDa IDm num_objects
MOT17-02-FRCNN 0.0% nan% 0.0% 0.0% nan% 53 0 0 53 0 9880 0 0 0.0% nan 0 0 0 9880
MOT17-05-FRCNN 0.0% nan% 0.0% 0.0% nan% 71 0 0 71 0 3357 0 0 0.0% nan 0 0 0 3357
MOT17-11-FRCNN 0.0% nan% 0.0% 0.0% nan% 44 0 0 44 0 4517 0 0 0.0% nan 0 0 0 4517
MOT17-09-FRCNN 0.0% nan% 0.0% 0.0% nan% 22 0 0 22 0 2879 0 0 0.0% nan 0 0 0 2879
MOT17-04-FRCNN 0.0% nan% 0.0% 0.0% nan% 69 0 0 69 0 24178 0 0 0.0% nan 0 0 0 24178
MOT17-13-FRCNN 0.0% nan% 0.0% 0.0% nan% 44 0 0 44 0 3156 0 0 0.0% nan 0 0 0 3156
MOT17-10-FRCNN 0.0% nan% 0.0% 0.0% nan% 36 0 0 36 0 5923 0 0 0.0% nan 0 0 0 5923
OVERALL 0.0% nan% 0.0% 0.0% nan% 339 0 0 339 0 53890 0 0 0.0% nan 0 0 0 53890
02:06:44 INFO - Completed

I'm sure the env set up correctly, because test the tracking performance on MOT17 with our pretrained model is normal, got 66.1%.

[Hidehidden]

Excuse me.Have you found the solution?Thank you in advance.

[amoskalev]

@xingyizhou

Hello, I have the same problem while using -public_det

[LesterLian]

Check the JSON file in mot17/results and see if it contains any bbox. If not, maybe my changes to the result converting script in #232 can help.

[Li1zz]

I have the same problem with you. Have you solved the problem yet? Thanks

[LesterLian]

@Li1zz
Have you tried my fix? If it doesn't help, you are mostly on your own. There're incomplete code here and there, yet I don't think the author actively maintains this project.

[Li1zz]

@LesterLian
Hello, I saw the solution you provide , but i dont know how to implement it at that time. I guess it is to replace convert_mot_det_to_results.py file with the fix you provided, and then run datasetup again, is it right?

[Li1zz]

@LesterLian
Running is successful after I tried your solution to this. Thank you so much !