How can I get intermediate output from a torch._C.Graph

jit
1

In one .py file, I save a torch.jit model:

torch.jit.save(torch.jit.trace(model_quantized, example_inputs=input), “model.pt”)

in another .py file, I load and run the model:

model_quantized = torch.jit.load(“model.pt”)
model_quantized(input)
graph = model_quantized.graph

how can I modify the graph to get intermediate output?

2

Is this what you are looking for?