zwqnju
zwqnju
@dllucas I follow the instruction in [Det3D/INSTALLATION.md](https://github.com/poodarchu/Det3D/blob/master/INSTALLATION.md#spconv) to install spconv-v1.0 ``
``` groundtruth_dir (just the label_2 folder in KITTI dataset) |----000000.txt |----000001.txt ``` ``` result_dir |----data |----000000.txt |----000001.txt ``` Run: ``` ./evaluate_object_3d_offline groundtruth_dir result_dir ```
too slow to use even running in local computer
My result was also not as good as the paper. In the paper, the class-level result is: Class Precision Recall IOU Car 66.7 95.4 64.6 Ped 45.2 29.7 21.8 Cyclist...
> i think we need to set the defults parameters of checkpoint in test.py Yes, just add '--checkpoint ./garage/XXX.pth', and test.py can run correctly.
The forward function of SparseConv gives the output of SparseTensor type, which cannot be passed to the batchnorm layer directly. `x = self.bn1(x)` should be replaced by: `x = x.replace_feature(self.bn1(x.features))`...
Can this code be public some day?