search

agoryuno / deepsolo-onnx

An ONNX exporter fot the DeepSolo scene text recognition model
3 stars 0 forks source link

question about dynamic resolution #1

Open Zalways opened 12 months ago

Zalways commented 12 months ago

in the detectron2 tutorials, it said that export model into torchscript by traceing can infer on dynamic resolution images: image

but when i export the model, the torchscript model only can inferrence on the images that have same height and width with the dummy image i used for exporting the model. i'd like to know whether your exported model can inference on dynamic resolution images.

agoryuno commented 12 months ago

Unfortunately, no. The image dimensions are fixed in the exported ONNX model. Can you link to the tutorial that mentions dynamic resolution?

YuMJie commented 11 months ago

in the detectron2 tutorials, it said that export model into torchscript by traceing can infer on dynamic resolution images: image

but when i export the model, the torchscript model only can inferrence on the images that have same height and width with the dummy image i used for exporting the model. i'd like to know whether your exported model can inference on dynamic resolution images.

Hi, have you fix it? : )

YuMJie commented 11 months ago

Unfortunately, no. The image dimensions are fixed in the exported ONNX model. Can you link to the tutorial that mentions dynamic resolution?

The image dimensions are fixed in the exported ONNX model, however , the image dimensions are unfixed on test and inference with pth file , which may cause the different results between them. Can you provice the dynamic resolution code for me? Thank you !