xuarehere / yolov7_deepsort_tensorrt Goto Github PK

View Code? Open in Web Editor NEW

38.0 2.0 5.0 5.34 MB

MOT using deepsort and yolov7 with c++. It also supports yolov5 as a detector.

CMake 0.86% C++ 96.13% C 1.18% Shell 1.83%

deep-sort deep-sort-tracking deepsort mot mot-tracking tensorrt yolo yolov5 yolov5-deepsort yolov7

yolov7_deepsort_tensorrt's Issues

How to batch input

Hello

Thank you for your good code.

I have a question about batch input....

I change config file BATCH_SIZE: 1 to BATCH_SIZE: 2.
Then, I just change add the same image as like below.

    vec_org_img.push_back(src_img);
    vec_org_img.push_back(src_img);

But ... There is some bug.

What should I do for batch input processing?

:-)

How to get #include <cuda_runtime_api.h>

My cuda,cudnn,yaml,tensorrt is the same as yours
but when i am camke ,it show No CUDA toolset found.

my cmake is 3.20.0.
Can you tell me how to get <cuda_runtime_api.h>?
thank you very much

asert(engine->getNbBindings() == 2) error

I did a custom train at Yolov7
asert(engine->getNbBindings() == 2);
In this code, getNbBindings() has a value of 5
Are there any issues I need to pay attention to in train?

'''
[W] [TRT] CUDA lazy loading is not enabled. Enabling it can significantly reduce device memory usage. See CUDA_MODULE_LOADING in https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#env-vars
alta-yolov7-deepsort: ../alta-yolov7-deepsort/model.cpp:96: void Model::LoadEngine(): Assertion `engine->getNbBindings() == 2' failed.
'''

Recommend Projects