search

indra4837 / yolov4_trt_ros

YOLOv4 object detector using TensorRT engine
MIT License
104 stars 31 forks source link

can this repo used in pc x86 platform? #3

Closed yantaixu0120 closed 3 years ago

yantaixu0120 commented 3 years ago

hello, can this repo used in pc x86 platform?

indra4837 commented 3 years ago

Yes you can use it on x86 platform. You need to download and install TensorRT 7+ on your pc.

You can only use the inference node and not the video capture node as it is written with the jetson device in mind.

Also, remove the lines that indicate jetson in the CMakelists.txt and the video_capture node accordingly.

Let me know if you need any help, thanks!

yantaixu0120 commented 3 years ago

Yes you can use it on x86 platform. You need to download and install TensorRT 7+ on your pc.

You can only use the inference node and not the video capture node as it is written with the jetson device in mind.

Also, remove the lines that indicate jetson in the CMakelists.txt and the video_capture node accordingly.

Let me know if you need any help, thanks!

thanks very much!!!,I have placed a CATKIN_IGNORE in the ros_deep_learning package,but i get the follow errros when i run the catkin_make -j12 command:

"CMake Error at yolov4_trt_ros/CMakeLists.txt:42 (add_executable): add_executable cannot create target "video_source" because another target with the same name already exists."

Should I write someting in the CATKIN_IGNORE file in the ros_deep_learning?

indra4837 commented 3 years ago

you can remove the executable line (line 42) in the CMakeLists.txt as that executable requires a jetson device. Let me know if that works!

yantaixu0120 commented 3 years ago

you can remove the executable line (line 42) in the CMakeLists.txt as that executable requires a jetson device. Let me know if that works!

Hello,This approach is works,thanks very much! but there seem to be tow mistakes in the yolo_to_onnx.py file,there have no the verify_classes() function and get_input_wh() function in the line 879 and 884 Respectively...

yantaixu0120 commented 3 years ago

you can remove the executable line (line 42) in the CMakeLists.txt as that executable requires a jetson device. Let me know if that works!

Hello,This approach is works,thanks very much! but there seem to be tow mistakes in the yolo_to_onnx.py file,there have no the verify_classes() function and get_input_wh() function in the line 879 and 884 Respectively...

O,I have found the two functions ,thanks!!!