shape inference of TRT::EfficientNMS_TRT type is missing

WongKinYiu / yolov7

Implementation of paper - YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors

GNU General Public License v3.0

13.43k stars 4.23k forks source link

shape inference of TRT::EfficientNMS_TRT type is missing #2011

Open Kuchchi12 opened 7 months ago

Kuchchi12 commented 7 months ago

I trained yolov7 tiny for 4 classes for recognition of face landmarks face, eye, nose and mouth further i am trying to convert .pt --> .onnx and facing this warning Screenshot 2024-04-04 160704 which ig is causing problem in .onnx -->.trt conversion because of not getting desired shape Screenshot 2024-04-04 160825 also it would be great help if anyone can teach me how to decrease model size while conversion

python : Python 3.8.10 torch : 2.2.0+cu118

dsbyprateekg commented 7 months ago

Try this notebook for the conversion- https://colab.research.google.com/gist/AlexeyAB/fcb47ae544cf284eb24d8ad8e880d45c/yolov7trtlinaom.ipynb