Search code examples
pandasnumpynumbacudf

from numba import cuda, numpy_support and ImportError: cannot import name 'numpy_support' from 'numba'

I am changing pandas into cudf to make faster aggregating and reduce the processing speed. I figure out one library which works on GPU with pandas.

"CUDF LINK" https://github.com/rapidsai/cudf

When I entered the below to install in my project it gives an error and I also tried many version of numba.

conda install -c rapidsai -c nvidia -c numba -c conda-forge \
    cudf=0.13 python=3.7 cudatoolkit=10.2

Traceback

Traceback (most recent call last):
  File "/home/khawar/deepface/tests/Ensemble-Face-Recognition.py", line 5, in <module>
    import cudf
  File "/home/khawar/anaconda3/envs/deepface/lib/python3.7/site-packages/cudf/__init__.py", line 7, in <module>
    from cudf import core, datasets
  File "/home/khawar/anaconda3/envs/deepface/lib/python3.7/site-packages/cudf/core/__init__.py", line 3, in <module>
    from cudf.core import buffer, column
  File "/home/khawar/anaconda3/envs/deepface/lib/python3.7/site-packages/cudf/core/column/__init__.py", line 1, in <module>
    from cudf.core.column.categorical import CategoricalColumn  # noqa: F401
  File "/home/khawar/anaconda3/envs/deepface/lib/python3.7/site-packages/cudf/core/column/categorical.py", line 11, in <module>
    import cudf._libxx as libcudfxx
  File "/home/khawar/anaconda3/envs/deepface/lib/python3.7/site-packages/cudf/_libxx/__init__.py", line 5, in <module>
    from . import (
  File "cudf/_libxx/aggregation.pxd", line 9, in init cudf._libxx.reduce
  File "cudf/_libxx/aggregation.pyx", line 11, in init cudf._libxx.aggregation
  File "/home/khawar/anaconda3/envs/deepface/lib/python3.7/site-packages/cudf/utils/cudautils.py", line 7, in <module>
    from numba import cuda, numpy_support
ImportError: cannot import name 'numpy_support' from 'numba' (/home/khawar/anaconda3/envs/deepface/lib/python3.7/site-packages/numba/__init__.py)
Solution

  • When trying to install cuDF 0.13, conda is apparently finding a numba version that is incompatible with that cuDF 0.13.

    cuDF 0.13 is out of date. The current stable release is 0.17 and the nightly is 0.18. We'll update the README, as it should provide installation instructions for the current version.

    We recommend creating a fresh conda environment. Please try the following conda install command, found here:

    conda create -n rapids-0.17 -c rapidsai -c nvidia -c conda-forge \
    -c defaults rapids-blazing=0.17 python=3.7 cudatoolkit=10.2