unknown device type error

[lonlonago]

I use k2200 GPU , does it not supported ?

[traveller59]

I think k2200 is supported, but you may need to use cuda 9.0+.

[lonlonago]

I use cuda 10, win10, compile is suscess. I run simple_inference code. seems much changes need to fix this code, here are some places need to fix:
1 change_detection_range() model_config proto changed;
2 voxels, coords, num_points = voxel_generator.generate(points, max_voxels=90000) too many value to unpack ;
3 pred = net(example)[0] assert unknown device type error;

[MartinHahner]

I get the same error using PCDet.

Traceback (most recent call last):
  File "train.py", line 215, in <module>
    main()
  File "train.py", line 210, in main
    max_ckpt_save_num=arguments.max_ckpt_save_num)
  File "/scratch_net/hox/mhahner/repositories/PCDet/tools/train_utils/train_utils.py", line 80, in train_model
    leave_pbar=(cur_epoch + 1 == total_epochs)
  File "/scratch_net/hox/mhahner/repositories/PCDet/tools/train_utils/train_utils.py", line 36, in train_one_epoch
    loss, tb_dict, disp_dict = model_func(model, batch)
  File "/scratch_net/hox/mhahner/repositories/PCDet/pcdet/models/__init__.py", line 25, in model_func
    ret_dict, tb_dict, disp_dict = model(input_dict)
  File "/home/mhahner/scratch/apps/anaconda3/envs/apex/lib/python3.7/site-packages/torch/nn/modules/module.py", line 532, in __call__
    result = self.forward(*input, **kwargs)
  File "/scratch_net/hox/mhahner/repositories/PCDet/pcdet/models/detectors/PartA2_net.py", line 106, in forward
    rpn_ret_dict = self.forward_rpn(**input_dict)
  File "/scratch_net/hox/mhahner/repositories/PCDet/pcdet/models/detectors/PartA2_net.py", line 33, in forward_rpn
    **kwargs
  File "/home/mhahner/scratch/apps/anaconda3/envs/apex/lib/python3.7/site-packages/torch/nn/modules/module.py", line 532, in __call__
    result = self.forward(*input, **kwargs)
  File "/scratch_net/hox/mhahner/repositories/PCDet/pcdet/models/rpn/rpn_unet.py", line 471, in forward
    x = self.conv_input(input_sp_tensor)
  File "/home/mhahner/scratch/apps/anaconda3/envs/apex/lib/python3.7/site-packages/torch/nn/modules/module.py", line 532, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/mhahner/scratch/apps/anaconda3/envs/apex/lib/python3.7/site-packages/spconv/modules.py", line 130, in forward
    input = module(input)
  File "/home/mhahner/scratch/apps/anaconda3/envs/apex/lib/python3.7/site-packages/torch/nn/modules/module.py", line 532, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/mhahner/scratch/apps/anaconda3/envs/apex/lib/python3.7/site-packages/spconv/conv.py", line 177, in forward
    use_hash=self.use_hash)
  File "/home/mhahner/scratch/apps/anaconda3/envs/apex/lib/python3.7/site-packages/spconv/ops.py", line 93, in get_indice_pairs
    stride, padding, dilation, out_padding, int(subm), int(transpose), int(use_hash))
ValueError: /scratch_net/hox/mhahner/repositories/spconv/include/spconv/spconv_ops.h 132
false assert faild. unknown device type

# packages in environment at /home/mhahner/scratch/apps/anaconda3/envs/apex:
#
# Name                    Version                   Build  Channel
_libgcc_mutex             0.1                 conda_forge    conda-forge
_openmp_mutex             4.5                      1_llvm    conda-forge
bzip2                     1.0.8                h516909a_2    conda-forge
ca-certificates           2020.1.1                      0
certifi                   2020.4.5.1               py37_0
cmake                     3.17.0               h28c56e5_0    conda-forge
coloredlogs               14.0                     pypi_0    pypi
cudatoolkit               10.1.243             h6bb024c_0
cudatoolkit-dev           10.1.243             h516909a_3    conda-forge
cudnn                     7.6.5                cuda10.1_0
cycler                    0.10.0                   pypi_0    pypi
decorator                 4.4.2                    pypi_0    pypi
easydict                  1.9                      pypi_0    pypi
expat                     2.2.9                he1b5a44_2    conda-forge
humanfriendly             8.2                      pypi_0    pypi
imagecodecs               2020.2.18                pypi_0    pypi
imageio                   2.8.0                    pypi_0    pypi
kiwisolver                1.2.0                    pypi_0    pypi
krb5                      1.17.1               h2fd8d38_0    conda-forge
libblas                   3.8.0               16_openblas    conda-forge
libcblas                  3.8.0               16_openblas    conda-forge
libcurl                   7.69.1               hf7181ac_0    conda-forge
libedit                   3.1.20181209         hc058e9b_0
libffi                    3.2.1             he1b5a44_1007    conda-forge
libgcc-ng                 9.2.0                h24d8f2e_2    conda-forge
libgfortran-ng            7.5.0                hdf63c60_6    conda-forge
liblapack                 3.8.0               16_openblas    conda-forge
libopenblas               0.3.9                h5ec1e0e_0    conda-forge
libssh2                   1.8.2                h22169c7_2    conda-forge
libstdcxx-ng              9.2.0                hdf63c60_2    conda-forge
libuv                     1.34.0               h516909a_0    conda-forge
llvm-openmp               10.0.0               hc9558a2_0    conda-forge
llvmlite                  0.32.1                   pypi_0    pypi
matplotlib                3.2.1                    pypi_0    pypi
mkl                       2020.1                      217    conda-forge
ncurses                   6.1               hf484d3e_1002    conda-forge
networkx                  2.4                      pypi_0    pypi
ninja                     1.10.0               hc9558a2_0    conda-forge
numba                     0.49.1                   pypi_0    pypi
numpy                     1.18.4           py37h8960a57_0    conda-forge
openssl                   1.1.1g               h7b6447c_0
pcdet                     0.1.0+2244be4             dev_0    <develop>
pillow                    7.1.2                    pypi_0    pypi
pip                       20.1               pyh9f0ad1d_0    conda-forge
protobuf                  3.11.3                   pypi_0    pypi
pyparsing                 2.4.7                    pypi_0    pypi
python                    3.7.5                h0371630_0
python-dateutil           2.8.1                    pypi_0    pypi
python_abi                3.7                     1_cp37m    conda-forge
pytorch                   1.4.0           py3.7_cuda10.1.243_cudnn7.6.3_0    pytorch
pywavelets                1.1.1                    pypi_0    pypi
pyyaml                    5.3.1                    pypi_0    pypi
readline                  7.0               hf8c457e_1001    conda-forge
rhash                     1.3.6             h14c3975_1001    conda-forge
scikit-image              0.17.2                   pypi_0    pypi
scipy                     1.4.1                    pypi_0    pypi
setuptools                46.3.0           py37hc8dfbb8_0    conda-forge
six                       1.14.0                   pypi_0    pypi
spconv                    1.1                      pypi_0    pypi
sqlite                    3.31.1               h7b6447c_0
tensorboardx              2.0                      pypi_0    pypi
tifffile                  2020.5.11                pypi_0    pypi
tk                        8.6.10               hed695b0_0    conda-forge
tqdm                      4.46.0                   pypi_0    pypi
wheel                     0.34.2                     py_1    conda-forge
xz                        5.2.5                h516909a_0    conda-forge
zlib                      1.2.11            h516909a_1006    conda-forge

[MartinHahner]

In case anyone wants to build spconv v.1.0 (commit 8da6f96)
with a fairly up-to-date conda environment, here is how we in our lab were able to do it:

• download and unzip our hotfixes

• conda create --name spconv python=3.7 pytorch=1.4 cudatoolkit=10.1 cudatoolkit-dev=10.1 cmake --channel pytorch

• conda activate spconv

• conda install cudnn

• git clone https://github.com/traveller59/spconv --recursive

• cd spconv

• git checkout 8da6f967fb9a054d8870c3515b1b44eca2103634

• git am <path_to_hotfixes>/0001-fix-problem-with-torch-1.4.patch

• git am <path_to_hotfixes>/0001-Allow-to-specifiy-CUDA_ROOT-directory-and-pick-corre.patch

• CUDA_ROOT=<path_to_your_conda_installation>/envs/spconv python setup.py bdist_wheel

• cd dist/

• pip install *

Thank me later.

[MartinHahner]

Here is an even better solution:

conda create --name PCDet python=3.6 pytorch=1.1 cudatoolkit=9.2 cudatoolkit-dev=9.2 \
cmake --channel pytorch --channel=conda-forge
conda activate PCDet
conda install cudnn
conda install boost
git clone https://github.com/traveller59/spconv spconv_8da6f96 --recursive
cd spconv_8da6f96
git checkout 8da6f967fb9a054d8870c3515b1b44eca2103634

If necessary:
Download and extract
0001-Allow-to-specifiy-CUDA_ROOT-directory-and-pick-corre.patch.zip
and patch spconv via:
git am <PATH_TO_EXTRACTED_FILE>/0001-Allow-to-specifiy-CUDA_ROOT-directory-and-pick-corre.patch

CUDA_ROOT=<PATH_TO_YOUR_CONDA_INSTALLATION>/conda_envs/PCDet python setup.py bdist_wheel
cd dist/
pip install *

Test spconv via:
python -c 'import spconv'
(should just return and not raise any errors)

Done!

I hope these instructions help someone else who struggles to build spconv as well.

[SternSu]

Here is an even better solution:

conda create --name PCDet python=3.6 pytorch=1.1 cudatoolkit=9.2 cudatoolkit-dev=9.2 \
cmake --channel pytorch --channel=conda-forge
conda activate PCDet
conda install cudnn
conda install boost
git clone https://github.com/traveller59/spconv spconv_8da6f96 --recursive
cd spconv_8da6f96
git checkout 8da6f967fb9a054d8870c3515b1b44eca2103634

If necessary:
Download and extract
0001-Allow-to-specifiy-CUDA_ROOT-directory-and-pick-corre.patch.zip
and patch spconv via:
git am <PATH_TO_EXTRACTED_FILE>/0001-Allow-to-specifiy-CUDA_ROOT-directory-and-pick-corre.patch

CUDA_ROOT=<PATH_TO_YOUR_CONDA_INSTALLATION>/conda_envs/PCDet python setup.py bdist_wheel
cd dist/
pip install *

Test spconv via:
python -c 'import spconv'
(should just return and not raise any errors)

Done!

I hope these instructions help someone else who struggles to build spconv as well.

Thank you very much for your help! This problem has puzzled me for a long time.

[lipengfeizju]

Hi, I use a 2080ti GPU and encounter similar problem (unknown device type). Is this model not supported ? I am in spconv v1.2

[MartinHahner]

Hi, I use a 2080ti GPU and encounter similar problem (unknown device type). Is this model not supported ? I am in spconv v1.2

If you don't need features in spconv v1.2 specifically,
try spconv v1.0 (commit 8da6f96) and the introductions mentioned above.

[lipengfeizju]

Hi, I use a 2080ti GPU and encounter similar problem (unknown device type). Is this model not supported ? I am in spconv v1.2

If you don't need features in spconv v1.2 specifically,
try spconv v1.0 (commit 8da6f96) and the introductions mentioned above.

Thanks! It works now

[gxj98]

Hi, I use a 2080ti GPU and encounter similar problem (unknown device type). Is this model not supported ? I am in spconv v1.2

If you don't need features in spconv v1.2 specifically,
try spconv v1.0 (commit 8da6f96) and the introductions mentioned above.

Hi, martin, I install spconv and run it correctly in a docker image, I release this image and use Microsoft azurel cloud server to pull this image, the program doesn't work with an unknown device type error. Do you know the reason, thanks.

[MartinHahner]

Hi, I use a 2080ti GPU and encounter similar problem (unknown device type). Is this model not supported ? I am in spconv v1.2

If you don't need features in spconv v1.2 specifically,
try spconv v1.0 (commit 8da6f96) and the introductions mentioned above.

Hi, martin, I install spconv and run it correctly in a docker image, I release this image and use Microsoft azurel cloud server to pull this image, the program doesn't work with an unknown device type error. Do you know the reason, thanks.

Hi Xiangjun, you probably need to install spconv directly on the server, otherwise the device type (GPU) does not match.
I assume the server uses different GPUs.

[gxj98]

Hi, I use a 2080ti GPU and encounter similar problem (unknown device type). Is this model not supported ? I am in spconv v1.2

If you don't need features in spconv v1.2 specifically,
try spconv v1.0 (commit 8da6f96) and the introductions mentioned above.

Hi, martin, I install spconv and run it correctly in a docker image, I release this image and use Microsoft azurel cloud server to pull this image, the program doesn't work with an unknown device type error. Do you know the reason, thanks.

Hi Xiangjun, you probably need to install spconv directly on the server, otherwise the device type (GPU) does not match.
I assume the server uses different GPUs.

Yeah, I build it on Tesla M40, but run it on V100, anyway, I will try to build it on V100.