selflein / GraphNN-Multi-Object-Tracking

Unofficial PyTorch implementation of "Learning a Neural Solver for Multiple Object Tracking"
122 stars 23 forks source link

test sequences (for test additional work necessary) #14

Closed Rajkumarsaswamy closed 2 years ago

Rajkumarsaswamy commented 2 years ago

Hi, thank you for sharing nice work. I really appreciate taking your time to answer some of my questions.

May I ask how is the implementation for test sequences.

Sofar I have got results for training sequence in Mot16 dataset.

But to use test dataset I did following:

1.python src/data_utils/preprocessing.py --dataset_path. /data/test_preprocessing --mode test

After this I should execute below code to obtain detections? I have doubt why we are obtaining detections. How it is useful in when executing inference.py in third step?

  1. python src/data_utils/run_obj_detect.py

next

  1. python inference.py --preprocessed_sequence /data/test_preprocessing

After execution this command, I checked the output txt file for example MOT16-01.txt but there was no tracking details. So there is still some work need to be done to implement test sequences, isn't?

selflein commented 2 years ago

Reply is here: https://github.com/selflein/GraphNN-Multi-Object-Tracking/issues/9#issuecomment-915120902