can not get 74.8 MOTA in MOT-16 trainging set

[scuShaker]

I download the pre-trained model and test on the MOT-16 training set, only get the result with MOTA 73.1. How can I get 74.8 MOTA described in readme.md? Does the type of GPU matters？or the dataset contains other data？

      IDF1   IDP   IDR  Rcll  Prcn  GT  MT  PT ML   FP    FN  IDs    FM  MOTA  MOTP IDt IDa IDm

MOT16-02 55.6% 60.9% 51.2% 75.3% 89.5% 54 26 23 5 1581 4406 285 480 64.8% 0.197 113 77 11 MOT16-04 72.2% 79.8% 65.9% 78.6% 95.3% 83 45 24 14 1856 10158 158 367 74.4% 0.169 57 62 2 MOT16-05 76.3% 80.2% 72.9% 85.9% 94.5% 125 85 34 6 339 961 172 203 78.4% 0.192 90 58 25 MOT16-09 56.8% 62.2% 52.3% 78.6% 93.4% 25 15 10 0 290 1127 136 143 70.5% 0.175 66 36 7 MOT16-10 66.3% 68.9% 63.8% 82.2% 88.7% 54 30 24 0 1288 2198 295 520 69.3% 0.217 122 72 3 MOT16-11 75.6% 76.5% 74.8% 91.5% 93.6% 69 54 14 1 577 780 82 165 84.3% 0.154 28 37 5 MOT16-13 73.8% 79.3% 69.0% 81.2% 93.3% 107 69 36 2 662 2158 184 467 73.8% 0.212 102 77 35 OVERALL 68.9% 74.4% 64.1% 80.3% 93.1% 517 324 165 28 6593 21788 1312 2345 73.1% 0.183 578 419 88

[ajithvcoder]

@scuShaker when i used pretrained model its giving me infinity and nan in results , could u tell which one of three pretrained model and cgf files u used ?

[ajithvcoder]

please find screen shot, also i used the test set for evaluation

[scuShaker]

please find screen shot, also i used the test set for evaluation

MOT16-01 03 06 07 08 12 14 are belong to MOT-16 testing set，their annotation file not exist actually，you should submit the result file to https://motchallenge.net

[ajithvcoder]

okay got it thanks @scuShaker

[Curry-Christopher]

I download the pre-trained model and test on the MOT-16 training set, only get the result with MOTA 73.1. How can I get 74.8 MOTA described in readme.md? Does the type of GPU matters？or the dataset contains other data？

      IDF1   IDP   IDR  Rcll  Prcn  GT  MT  PT ML   FP    FN  IDs    FM  MOTA  MOTP IDt IDa IDm

MOT16-02 55.6% 60.9% 51.2% 75.3% 89.5% 54 26 23 5 1581 4406 285 480 64.8% 0.197 113 77 11 MOT16-04 72.2% 79.8% 65.9% 78.6% 95.3% 83 45 24 14 1856 10158 158 367 74.4% 0.169 57 62 2 MOT16-05 76.3% 80.2% 72.9% 85.9% 94.5% 125 85 34 6 339 961 172 203 78.4% 0.192 90 58 25 MOT16-09 56.8% 62.2% 52.3% 78.6% 93.4% 25 15 10 0 290 1127 136 143 70.5% 0.175 66 36 7 MOT16-10 66.3% 68.9% 63.8% 82.2% 88.7% 54 30 24 0 1288 2198 295 520 69.3% 0.217 122 72 3 MOT16-11 75.6% 76.5% 74.8% 91.5% 93.6% 69 54 14 1 577 780 82 165 84.3% 0.154 28 37 5 MOT16-13 73.8% 79.3% 69.0% 81.2% 93.3% 107 69 36 2 662 2158 184 467 73.8% 0.212 102 77 35 OVERALL 68.9% 74.4% 64.1% 80.3% 93.1% 517 324 165 28 6593 21788 1312 2345 73.1% 0.183 578 419 88

I got the same result as you.

[SmallGGgg]

I download the pre-trained model and test on the MOT-16 training set, only get the result with MOTA 73.1. How can I get 74.8 MOTA described in readme.md? Does the type of GPU matters？or the dataset contains other data？

      IDF1   IDP   IDR  Rcll  Prcn  GT  MT  PT ML   FP    FN  IDs    FM  MOTA  MOTP IDt IDa IDm

MOT16-02 55.6% 60.9% 51.2% 75.3% 89.5% 54 26 23 5 1581 4406 285 480 64.8% 0.197 113 77 11 MOT16-04 72.2% 79.8% 65.9% 78.6% 95.3% 83 45 24 14 1856 10158 158 367 74.4% 0.169 57 62 2 MOT16-05 76.3% 80.2% 72.9% 85.9% 94.5% 125 85 34 6 339 961 172 203 78.4% 0.192 90 58 25 MOT16-09 56.8% 62.2% 52.3% 78.6% 93.4% 25 15 10 0 290 1127 136 143 70.5% 0.175 66 36 7 MOT16-10 66.3% 68.9% 63.8% 82.2% 88.7% 54 30 24 0 1288 2198 295 520 69.3% 0.217 122 72 3 MOT16-11 75.6% 76.5% 74.8% 91.5% 93.6% 69 54 14 1 577 780 82 165 84.3% 0.154 28 37 5 MOT16-13 73.8% 79.3% 69.0% 81.2% 93.3% 107 69 36 2 662 2158 184 467 73.8% 0.212 102 77 35 OVERALL 68.9% 74.4% 64.1% 80.3% 93.1% 517 324 165 28 6593 21788 1312 2345 73.1% 0.183 578 419 88

Hi bro, could you tell me how to test on mot16 training set, what kind of command line?