I have followed the instructions to install some dependencies. Everything seems OK. But:
1) When I try to run:
python tools/dist_test.py configs/nusc/voxelnet/nusc_centerpoint_voxelnet_0075voxel_fix_bn_z.py --work_dir work_dirs/nusc_centerpoint_voxelnet_0075voxel_fix_bn_z --checkpoint work_dirs/nusc_centerpoint_voxelnet_0075voxel_fix_bn_z/epoch_20.pth --speed_test
And I found that it is actually caused by the function load_state_dict() in det3d/torchie/trainer/checkpoint.py in the line own_state[name].copy_(param) when name is backbone.conv3.0.weight.
2) When I try to run:
python tools/train.py configs/nusc/voxelnet/nusc_centerpoint_voxelnet_0075voxel_fix_bn_z.py
Hi @tianweiy, thanks for the great works!
I have followed the instructions to install some dependencies. Everything seems OK. But: 1) When I try to run:
python tools/dist_test.py configs/nusc/voxelnet/nusc_centerpoint_voxelnet_0075voxel_fix_bn_z.py --work_dir work_dirs/nusc_centerpoint_voxelnet_0075voxel_fix_bn_z --checkpoint work_dirs/nusc_centerpoint_voxelnet_0075voxel_fix_bn_z/epoch_20.pth --speed_test
It throws an error:
And I found that it is actually caused by the function
load_state_dict()
indet3d/torchie/trainer/checkpoint.py
in the lineown_state[name].copy_(param)
whenname
isbackbone.conv3.0.weight
.2) When I try to run:
python tools/train.py configs/nusc/voxelnet/nusc_centerpoint_voxelnet_0075voxel_fix_bn_z.py
The same error occurs:
But this time it is caused by the function
train_detector()
indet3d/torchie/apis/train.py
with the linemodel = model.cuda()
.Can you please help me to solve this error? Thanks a lot!