As we discussed in #5, since we refactored the code in our open source project, we need to double-check the checkpoints on each dataset to ensure the correctness. However, recently, MOTChallenge server seems to encounter some problems, so we can not get our evaluation results by submitting tracker files. Therefore, although we upload the checkpoint on MOT17, we lack the required double-check yet.
I will close this issue till I finish this double-check. If you successfully obtain the MOT17 evaluation result, you can also let me know through this issue. I will be grateful.
As we discussed in #5, since we refactored the code in our open source project, we need to double-check the checkpoints on each dataset to ensure the correctness. However, recently, MOTChallenge server seems to encounter some problems, so we can not get our evaluation results by submitting tracker files. Therefore, although we upload the checkpoint on MOT17, we lack the required double-check yet.
I will close this issue till I finish this double-check. If you successfully obtain the MOT17 evaluation result, you can also let me know through this issue. I will be grateful.