Is it possible to get the output of an intermediate layer?

AI & Data Science Deep Learning (Training & Inference) TensorRT
#1

Hello,

I am using this guide to run YOLO tiny on the Jetson :

But I would like to get the result of the intermediate layers. Wondering if that would be possible.

#2

Hi,

We can set the layer as a network output and bind a buffer to it to collect the output.
We cannot do it while loading. We need to set the layer as a network output and regenerate the engine.

Thank you.

1 Like
#3

Thank you!

But any idea how I can do that? Like a guide or something? I am new to machine learning and trying to learn new stuff

Basically, I want to get outputs of all the layers and save them to a file to see whats inside.

#4

Hi,

Currently, we do not have exact resources to share.
You can check out the following, hope they are helpful.

Thank you.

1 Like
#5

Thanks!

I am using YOLO_to_ONNX and then ONNX_to_TensorRT.

So, can we do this with an existing model or do we need to create our own model?

#8

You can try modifying the ONNX model and re-generate the TensorRT engine.