Hi,
Thanks for this wonderful repo!
I have been able to convert the custom yolov8s model to onnx and engine file and I am able to save the outputs as well as per your script.
I am trying to validate the metrics on the validation set for the custom dataset I have through model.val() functionality of Ultralytics using following command
yolo val task=segment model=/workspace/tensorrt/bin/custom-model-v5-ultraonnx-enginetrtexec.engine data=custom.yaml imgsz=960 conf=0.5 iou=0.7
But I am getting the results 0 as in this image
I posted the same issue on Ultralytics page as well and the author said that meta-data is very much essential in converting the onnx to engine. I tried shutting down meta-data in their repo and tried inferencing the engine file. I was able to obtain the results.
Posted here: ultralytics/ultralytics#5791 (comment)
Can you please suggest what thing is going wrong here?
I have two classes in my dataset; top_card and a bottom_card
Hi,
Thanks for this wonderful repo!
I have been able to convert the custom yolov8s model to onnx and engine file and I am able to save the outputs as well as per your script.
I am trying to validate the metrics on the validation set for the custom dataset I have through model.val() functionality of Ultralytics using following command
yolo val task=segment model=/workspace/tensorrt/bin/custom-model-v5-ultraonnx-enginetrtexec.engine data=custom.yaml imgsz=960 conf=0.5 iou=0.7
But I am getting the results 0 as in this image
I posted the same issue on Ultralytics page as well and the author said that meta-data is very much essential in converting the onnx to engine. I tried shutting down meta-data in their repo and tried inferencing the engine file. I was able to obtain the results.
Posted here: ultralytics/ultralytics#5791 (comment)
Can you please suggest what thing is going wrong here?
I have two classes in my dataset; top_card and a bottom_card