is there something different from pjreddie‘s repo，about saving weights or parsing weights？

AlexeyAB / darknet

YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )

http://pjreddie.com/darknet/

Other

21.8k stars 7.97k forks source link

is there something different from pjreddie‘s repo，about saving weights or parsing weights？ #4107

Open OpenAI2022 opened 5 years ago

OpenAI2022 commented 5 years ago

Hi, @AlexeyAB, I try to convert your yolov3-spp weights to onnx and accelerate it with tensorRT. I accelerated yolov3 model in pjreddie's darknet successfully. But when i imitate yolov3-spp, after speed up by tensorrt, nothing it can detect. I'm quite sure i use all weights in yolov3-spp weights, and build the right network structure. I wonder, is there something different from pjreddie‘s repo，about saving weights or parsing weights？

AlexeyAB commented 5 years ago

Why do you want to convert Darknet to ONNX, while Yolo is natively supported in DeepStream 4.0? https://news.developer.nvidia.com/deepstream-sdk-4-now-available/

Read: https://github.com/AlexeyAB/darknet#yolo-v3-in-other-frameworks

Also, pjreddie's darknet supports yolov3-spp weights too.

luthfianto commented 5 years ago

@15605190982 how do you convert to TensorRT?