Help me! Some problems about submitting the results of multi-target tracking test set to Kitti's server
lipeng-gu opened this issue · comments
Hi，
I can see from Kitti's website that your multi-target tracking indicators are very good，but I meet some problems about submitting the results of multi-target tracking test set to Kitti's server. I have sent an email to your mailbox. I wonder if you have seen it
That porblem is that the data in your zip archive submitted to Kitti's server has the false format.
I have submitted the results of multi-target tracking test sets to the server twice, according to the document requirements, but the server still says that the data in your zip archive has the false format. I can't find the problems. Because I am writing a paper, I am very anxious to solve this problem. I have attached the submitted compressed file. If you are free, can you help me to see if there is any problem with my file format? Or can you provide a document in the correct format that you submitted？
thank you very much!
Hi @axiaoxiaobai ,
I did not find attachments in your email.
You can check the label of the tracking data, they should have the same format.
The results *.txt should directly inside the zip file.
@ZwwWayne Sorry, I forgot to add the attachment just now. Do you have time to have a look now.
Maybe you should change those -1s to 0.
I am not sure whether you modified this codebase. If you did not, then the function write_kitti_result should be called and the results should be OK.
Yes. And I suggest you also check the tracking devkit and the instructions in the benchmark.
@ZwwWayne
hi，I submitted the result again, but it was still in the wrong format. Can you send a txt file of a sequence of txt files to my mailbox(2181330@mail.dhu.edu.cn)？
You should read and check the readme.txt carefully, and compare the formats between your results and the tracking labels.
@ZwwWayne I read and check the readme.txt carefully，and use writeLabels.m for writing your results. I submitted the results to the server, but the server replied that the file I submitted was in the wrong format.You can send the result of a sequence to my email（2181330@mail.dhu.edu.cn）, and I would like to thank you very much
This is one of the sequence in the submission files.
0001.txt
@ZwwWayne
Thank you very much!
I've compared your results with mine. I'd like to ask you a few more questions.
1)Are the unused values are set to 0.0000?
2)All your values retain four decimal, such as 1021.7770. Why does the value in the fifth column only retain one decimal, such as 0.0, or it doesn't matter?
3)Some of the values in the 17th column are set to 0.000, but some are set to -0.0001 and -0.0000. What's the difference?
I'm confused now. Thank you very much!
@ZwwWayne
There's one problem. If there are several frames without targets, what do you do? Do you want to skip these frames in the txt file without recording any data?
@ZwwWayne
There's one problem. If there are several frames without targets, what do you do? Do you want to skip these frames in the txt file without recording any data?
Yes, just skip them.
@ZwwWayne
Thank you very much! If you are free, can you take a look at the following questions for me?
I've compared your results with mine. I'd like to ask you a few more questions.
1)Are the unused values are set to 0.0000?
2)All your values retain four decimal, such as 1021.7770. Why does the value in the fifth column only retain one decimal, such as 0.0, or it doesn't matter?
3)Some of the values in the 17th column are set to 0.000, but some are set to -0.0001 and -0.0000. What's the difference?
I'm confused now. Thank you very much!
- Yes.
- It should do not matter. There has been a while since the code is written, so I cannot remember the reason. But it should works because I use them to submit the results.
- Same reply as 2.
A good way to debug your results is to use thedevkitto evaluate your predictions on the validation set or the total training set and see what happens. If your results could be evaluated by the devkit successfully, you should consider other problems or email the benchmark manager for help.