Difference in False Detections when compared with yolov4 and yolov4+Tensor RT

[anusha657]

I have used this repositry for integrating yolov4 with Tensor RT .
GeForce GTX 1070 Mobile
Ubunto - 16.4
TensorRT 5.0

cfg file - As suggested I modified accordingly .[NOTE - I used SAME yolov4,.cfg and yolov4.weights for BOTH TensorRT+yolov4 and yolov4 (darkent)]

Case 1 - when I run the following command for TensorRT+yolov4 --- " ./runDet -i model/yolov4.engine -p output_Side.mp4 "
We are getting false detections in the video file.

Case 2 when I run the following command for yolov4 --- "./darknet detector demo cfg/coco.data cfg/yolov4.cfg yolov4.weights -dont_show /home/test/Anusha/output_Side.mp4 -i 0 -avgframes 1 -out_filename outyolov4.avi"
We are getting False Detetctions in the video file .

However the false Detections in case 1 and case 2 are not matching , which Ideally should match when I am using same config and weights file .

[akashAD98]

@anusha657 @CaoWGG im also facing the same issue. have you solved the problem??
im using this repo.
jkjung-avt/tensorrt_demos#473
jkjung-avt/tensorrt_demos#315