search

mit-han-lab / torchsparse

[MICRO'23, MLSys'22] TorchSparse: Efficient Training and Inference Framework for Sparse Convolution on GPUs.
https://torchsparse.mit.edu
MIT License
1.16k stars 132 forks source link

Update backbone example for v2.1.0 #224

Closed ys-2020 closed 11 months ago

ys-2020 commented 1 year ago

Update examples/backbone.py for TorchSparse v2.1.0. We moved batch index to the first dim of coordinates.