DEFT

DEFT copied to clipboard

Hi, The length of the trajectory samples in the [training set](https://github.com/MedChaabane/DEFT/blob/71abba2c69a9d3a4d07154a1ed871f8846ba8648/src/lib/dataset/trajectory_dataset.py#L150) does not seem to be 10 as stated in the paper, but rather 9,10,11. Is this the reason why...

xiaocc612

AttributeError: 'ModleWithLoss' object has no attribute 's_det'

3

I am very interested in this work, I would like to reproduce it, but can you help me, I can run test.py file, but when I want to train yes,...

yan1617262965

What is the exact inference speed you are getting on the MOT 17 dataset?

JaiDoshi

Hi, Thanks for your work. I am wondering if there is a full metrics on nuScenes validation split (Tracking). I read your paper, and checked some of the issues, didn't...

a1600012888

Some problems when working on reproducing the results

1

Hi, thank you every much for releasing the code ! I’ve been working on reproducing the results on the nuScences dataset and met some problems. 1. Hyper-parameters differs between code...

junjunxia

Training is stuck at the beginning on custom dataset

4

Hi and thanks for your work! I'm trying to train DEFT on a dataset which I have used to train CenterTrack successfully before. When running train.py the output seems promising:...

rvrsprdx

How can I use the model to predict my own video

Linkcy97

What's the pretrained model of kitii_tracking ?

2

Hi, I have some questions about the experiment on kitii_tracking. I notice that the paper said "All object detectors were pre-trained on the COCO dataset". But the setting in your...

lianghao12138

ssbilakeri

balancing weights for loss

1

Hi, thanks for your amazing work. In the paper, you mentioned using lambda_1 and lambda_2, which balancing weights to the two tasks. However, I found self.s_det and self.s_id are not...

Chenzhaowei13