How to use the trained model？

MegEngine / FINet

This is the official MegEngine implementation of FINet: Dual Branches Feature Interaction for Partial-to-Partial Point Cloud Registration, AAAI 2022

MIT License

25 stars 1 forks source link

How to use the trained model？ #2

Closed abchears closed 1 year ago

abchears commented 2 years ago

Hi author, thank you very much for your excellent work. I would like to ask, if I now have two tensors of shape n1x3 and n2X3, how should I use the trained model?

hxwork commented 2 years ago

Hi,

Thanks for your interest. We have no plan to provide the inference code, but I think it is easy to evaluate our model. Based on the evaluate.py, you may convert the format of your inputs to that of our data_batch (L39). Then, the transformation (represented by a quaternion and a translation) between the inputs can be obtained from output_batch (L43).

Best regards, Hao