ifzhang / ByteTrack

[ECCV 2022] ByteTrack: Multi-Object Tracking by Associating Every Detection Box
MIT License
4.8k stars 906 forks source link

Relation of image size and model weight #157

Open jackie840129 opened 2 years ago

jackie840129 commented 2 years ago

Hi @ifzhang
In your experiments, you make the input image resized and padded to 1440x800. I wonder if I can use a smaller image size but with the same model weight? ex. With size 1088x608? I know the model can run with input size 1088x608, but I want to know that have you trained multiple models with different input sizes?

Thanks!