How to control input image dimensions when executing tao export

Hi, I have a questions about TAO Toolkit.

At the tao-converter section in the below document says
-d : Comma-separated list of input dimensions that should match the dimensions used for tao dssd export
But I don’t get how to specify input dimensions when using tao export command.

I want to train a model with 300x300 image size, then export etlt and convert it to TensorRT model with different image size than training image size, like 600x600.

• Network Type : dssd (Object Detection)

• TLT Version

tao info

Configuration of the TAO Toolkit Instance
dockers: [‘nvidia/tao/tao-toolkit-tf’, ‘nvidia/tao/tao-toolkit-pyt’, ‘nvidia/tao/tao-toolkit-lm’]
format_version: 2.0
toolkit_version: 3.21.11
published_date: 11/08/2021


The input size of a model is usually fixed. After training, you have gotten a 300x300 model.
It cannot be changed.

Thanks for the reply.
As you said, it looks like DSSD can not change input image size after training. But DetectNetV2 can! I can change input image size freely when using DetectNetV2.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.