when I convert my YOLOv3 model to tensorRT, some objects that can be detected in YOLOv3 are not detected in TensorRT with FP32 mode. The project is based on this repository: https://github.com/lewes6369/TensorRT-Yolov3/
Objects detected in Darknet YOLOv3
Objects detected in TensorRT YOLOv3
Should I modify the training data or modify some settings in TensorRT ?
what version of TRT are you using? and just to confirm, https://github.com/lewes6369/TensorRT-Yolov3/ is the repro that demonstrates the issue are you are seeing?
HI NVES, the version of TensorRT I’m using is TensorRT 188.8.131.52. The example YOLOv3 model in the repository works fine, but when I train the architecture with my own dataset, the problem appears.
to help us debug, please share a repro package that contains the inference code, model, and dataset used that demonstrate the issues you are seeing. you can dm me a google drive, dropbox link, or github link.
HI NVES, it seems my preprocess function cutoff some area from the original image. Now this issue is solved, thank you.