dsa model training problem #7
Description
Hello,
I am using the newest OpenPCDet, and copy your codes to it.
I met a problem while training DSA model of second/pointpillar/pvrcnn, while not meeting this problem in pointrcnn_dsa, errors is as follows:
Expected to have finished reduction in the prior iteration before starting a new one. This error indicates that your module has parameters that were not used in producing loss. You can enable unused parameter detection by passing the keyword argument `find_unused_parameters=True` to `torch.nn.parallel.DistributedDataParallel`, and by
making sure all `forward` function outputs participate in calculating loss.
If you already have done the above, then the distributed data parallel module wasn't able to locate the output tensors in the return value of your module's `forward` function. Please include the loss function and the structure of the return value of `forward` of your module when reporting this issue (e.g. list, dict, iterable).
Parameter indices which did not receive grad for rank 0: 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87
In addition, you can set the environment variable TORCH_DISTRIBUTED_DEBUG to either INFO or DETAIL to print out information about which particular parameters did not receive gradient on this rank as part of this error
so I follow the instructions and set find_unused_parameters=True in torch.nn.parallel.DistributedDataParallel, it can training normal, did you meet this problem while not setting find_unused_parameters=True in torch.nn.parallel.DistributedDataParallel?
here env is:
pytorch 1.10.1
Name: torch
Version: 1.10.1+cu111
Summary: Tensors and Dynamic neural networks in Python with strong GPU acceleration
Home-page: https://pytorch.org/
Author: PyTorch Team
Author-email: [email protected]
License: BSD-3
Location: /home/deze/.conda/envs/torch110_py38/lib/python3.8/site-packages
Requires: typing-extensions
Required-by: thop, torchvision
cuda 11.1
spconv_cu111
Name: spconv-cu111
Version: 2.1.25
Summary: spatial sparse convolution
Home-page: https://github.com/traveller59/spconv
Author: Yan Yan
Author-email: [email protected]
License: Apache License 2.0
Location: /home/deze/.conda/envs/torch110_py38/lib/python3.8/site-packages
Requires: ccimport, cumm-cu111, fire, numpy, pccm, pybind11
Required-by:
Regards,
Deze