How to map train_edge.py files to .step source data？

[zctlovesy0313]

How do I map the result labels predicted by the model to the STEP model for visualization?

The shape of logits is 704, what does 704 represent?
The events I run are:https://www.sciencedirect.com/science/article/pii/S0010448522000240#appB

[AndrewColligan]

Hi, in the Hierarchical CADNet repo there is a script called test_and_save.py that can be used to predict the machining features for individual CAD models and save the results to a CAD model (with a label per B-Rep face) using a trained NN.

The logit shape is the total number of B-Rep faces predicted per minibatch. The minibatch is made up of a number of the graphs of different CAD models, each of these CAD models will have a different number of B-Rep faces associated with them. Therefore, logit shape will be different each minibatch as the number of B-Rep faces of each CAD model will differ.

Logit shape = total number of B-Rep faces in each CAD model in the minibatch

Hope this helps 😄

[zctlovesy0313]

Thank you! Another question: how to get string data representing different polygon labels, e.g. [1,24,2,12,6,......]. Each number represents a different face.And can I get, for example, the relative position of a hole relative to the entire drawing, so that I can generate the machining path of the CNC machine? Do you have any good suggestions, please?