search

mit-han-lab / spvnas

[ECCV 2020] Searching Efficient 3D Architectures with Sparse Point-Voxel Convolution
http://spvnas.mit.edu/
MIT License
587 stars 109 forks source link

voxel_size question #88

Closed Allianc-rgb closed 2 years ago

Allianc-rgb commented 2 years ago

Thanks for your great work! When i convert voxel_size from 0.05 to [0.05, 0.1, 0.1], Some errors have occurred：

Traceback (most recent call last): File "train.py", line 106, in main() File "train.py", line 101, in main Saver(), File "/home/user1/anaconda3/envs/py-mink1/lib/python3.7/site-packages/torchpack/train/trainer.py", line 39, in train_with_defaults callbacks=callbacks) File "/home/user1/anaconda3/envs/py-mink1/lib/python3.7/site-packages/torchpack/train/trainer.py", line 79, in train output_dict = self.run_step(feed_dict) File "/home/user1/anaconda3/envs/py-mink1/lib/python3.7/site-packages/torchpack/train/trainer.py", line 125, in run_step output_dict = self._run_step(feed_dict) File "/home/user1/Documents/chr/Cylinderspv/code1.0/core/trainers.py", line 39, in _run_step outputs = self.model(inputs) File "/home/user1/anaconda3/envs/py-mink1/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl result = self.forward(*input, kwargs) File "/home/user1/anaconda3/envs/py-mink1/lib/python3.7/site-packages/torch/nn/parallel/distributed.py", line 619, in forward output = self.module(*inputs[0], *kwargs[0]) File "/home/user1/anaconda3/envs/py-mink1/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl result = self.forward(input, kwargs) File "/home/user1/Documents/chr/Cylinderspv/code1.0/core/models/semantic_kitti/spvcnn.py", line 199, in forward x0 = self.stem(x0) File "/home/user1/anaconda3/envs/py-mink1/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl result = self.forward(*input, kwargs) File "/home/user1/anaconda3/envs/py-mink1/lib/python3.7/site-packages/torch/nn/modules/container.py", line 117, in forward input = module(input) File "/home/user1/anaconda3/envs/py-mink1/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl result = self.forward(*input, *kwargs) File "/home/user1/anaconda3/envs/py-mink1/lib/python3.7/site-packages/torchsparse-1.4.0-py3.7-linux-x86_64.egg/torchsparse/nn/modules/conv.py", line 72, in forward transposed=self.transposed) File "/home/user1/anaconda3/envs/py-mink1/lib/python3.7/site-packages/torchsparse-1.4.0-py3.7-linux-x86_64.egg/torchsparse/nn/functional/conv.py", line 182, in conv3d transposed, File "/home/user1/anaconda3/envs/py-mink1/lib/python3.7/site-packages/torch/cuda/amp/autocast_mode.py", line 213, in decorate_fwd return fwd(args, kwargs) File "/home/user1/anaconda3/envs/py-mink1/lib/python3.7/site-packages/torchsparse-1.4.0-py3.7-linux-x86_64.egg/torchsparse/nn/functional/conv.py", line 57, in forward transposed, RuntimeError: expected scalar type Double but found Float 0% 0/2392 [00:01<?, ?it/s]

Primary job terminated normally, but 1 process returned a non-zero exit code.. Per user-direction, the job has been aborted.

mpirun detected that one or more processes exited with non-zero status, thus causing the job to be terminated. The first process to do so was:

Process name: [[41806,1],0] Exit code: 1

It seems that Torchsparse1.4.0 only supports setting voxel size to a single scalar, but Torchsparse1.2.0 can run successfully before that. Can you give me some advice? I look forward to your reply