Jetson Inference object detection input size

Robotics & Edge Computing Jetson & Embedded Systems Jetson Nano
1

Hi,

As I know ssd mobilenet’s input size is 300x300 but in this tutorial ( jetson-inference/pytorch-ssd.md at master · dusty-nv/jetson-inference · GitHub ) uses mobilenet-v1-ssd-mp-0_675.pth file for training the model.
As I searched this is ssd mobilenet v1 with 0.675 width multiplier. Does this mean input size is 300 * (1.675), 300?

If it does, it is good for me but I had to know correct input size in order to give correct image (it can resize the image but i had to give correct size).

What is the correct size of the ssd mobile net v1 in this tutorial?

thanks, @dusty_nv

Note: No, 0.675 is only performance of the model. input size should be still 300x300.

2

Hi @muhammedsezer12, right - the input size is still 300x300. This constant is defined here in the code:

1 Like