ByteTrack

ByteTrack copied to clipboard

显存溢出

1

想问下按照您的方式进行训练，一张3090ti使用yolox-m的模型在batchsize为1时为什么显存在6000-19000来回横跳，使用bs=2时，显存就会溢出

AllesOderNicht

Post processing sequence of lost list and remove list

3

https://github.com/ifzhang/ByteTrack/blob/72ca8b45d36caf5a39e949c6aa815d9abffd1ab5/yolox/tracker/byte_tracker.py#L283 For example, id = 7, state = lost, if the current frame does not match, it will be placed in removed_ stacks, the state is changed to state =...

YuNaruto

Hi. I was curious how we can get the ID of the person who has just left the field of view of the camera while being tracked in the camera...

jamshaidsohail5

How to detect an object only once and track it in other frames?

aminromi11

Frame interval for tracker

1

Thanks for the good work! I am looking to change the interval of detection. For instance, run detection every other frame, but run tracker on every frame. This is important...

mayur-ag

Use YoloV7 detector with Byte Track

2

Firstly, thank you so much for this awesome repo ! I want to use YoloV7 Object Detector in place of YoloX detector with ByteTrack, as YoloV7 tends to perform better...

vineetsharma14

Has anybody attempted using ONNX for YOLOv5? I was able to change the script to generate yolov5n6.onnx but I am getting shape errors through demo_postprocess in onnx_inference.py. @iamrajee @ifzhang

parthmalpathak

TensorRT C++ inference results do not have bbox with custom dataset

2

I can use author's 'model_trt.engine' deploy 'palace.mp4' normally.  However, I generate 'bytetrack_s.onnx' , with my dataset,and make 'model_trt.engine', there have no bbox when inference. Who can tell me why?...

zzh805780186

YOLOv5 version 6 support for ByteTrack

1

Added support for YOLOv5 tiny and nano models to Bytetrack. Many thanks to @ifzhang and @gmt710 for their work on ByteTrack and YOLOv5 implementation.

parthmalpathak

跟踪其他物体？

9

我想直接使用yolox在coco上训练的模型以跟踪其他物体，但是貌似没有能够检测到任何物体，请问是什么问题？

Naughty-Galileo