Tracking using the training files of yolov2

obendidi / Tracking-with-darkflow

Real-time people Multitracker using YOLO v2 and deep_sort with tensorflow

GNU General Public License v3.0

523 stars 174 forks source link

Tracking using the training files of yolov2 #78

Open zanny0713 opened 6 years ago

zanny0713 commented 6 years ago

Hey!! I wanted to ask if anyone had tried or is it possible to track objects in real time using the previously trained weights file for object detection on yolov2? like do we need to train our model again separately for object tracking or the trained annotated files and the weights file which were used in object detection can be reused by tracking algorithm too?

buzdarbalooch commented 5 years ago

@zanny0713 i have a similar question. If you are able to figure out. it would be great if you can share.

fakturk commented 5 years ago

I used yolov2 weights and cfg files and show the location of that files in run.py (don't forget to put yolov2.cfg under cfg folder and yolov2.weights under bin folder) FLAGS.model = "darkflow/cfg/yolov2.cfg" FLAGS.load = "darkflow/bin/yolov2.weights" Also you need to update labels.txt (there are 4 categories right now but I think yolov2 uses 80 categories)