How can I print output tensor values at different stages in the model during runtime?
Using get_output() on the GraphModule after run() gives the final output of the overall model. I was wondering if it was possible to dump the output of a specific operator (such as a specific convolution layer) in the graph during runtime?