Jetson Xavier NX - Numba Problem

After installing numba successfully on jetson Xavier NX
I am adding two arrays with numba and having the following issue:

File “/home/siddhant/c4aarch64_installer/lib/python3.7/site-packages/numba/npyufunc/”, line 120, in wrap
File “/home/siddhant/c4aarch64_installer/lib/python3.7/site-packages/numba/npyufunc/”, line 163, in add
return self._compile_for_argtys(args, return_type)
File “/home/siddhant/c4aarch64_installer/lib/python3.7/site-packages/numba/npyufunc/”, line 209, in _compile_for_argtys
self._dispatcher, self.targetoptions, sig)
File “/home/siddhant/c4aarch64_installer/lib/python3.7/site-packages/numba/npyufunc/”, line 146, in _compile_element_wise_function
cres = nb_func.compile(sig, **targetoptions)
File “/home/siddhant/c4aarch64_installer/lib/python3.7/site-packages/numba/npyufunc/”, line 95, in compile
return self._compile_core(sig, flags, locals)
File “/home/siddhant/c4aarch64_installer/lib/python3.7/site-packages/numba/npyufunc/”, line 130, in _compile_core
flags=flags, locals=locals)
File “/home/siddhant/c4aarch64_installer/lib/python3.7/site-packages/numba/”, line 550, in compile_extra
args, return_type, flags, locals)
File “/home/siddhant/c4aarch64_installer/lib/python3.7/site-packages/numba/”, line 280, in init
File “/home/siddhant/c4aarch64_installer/lib/python3.7/site-packages/numba/typing/”, line 160, in refresh
File “/home/siddhant/c4aarch64_installer/lib/python3.7/site-packages/numba/typing/”, line 400, in _load_builtins
from . import ctypes_utils, bufproto # noqa: F401
File “/home/siddhant/c4aarch64_installer/lib/python3.7/site-packages/numba/typing/”, line 22, in
array.array: types.PyArray,
AttributeError: module ‘array’ has no attribute ‘array’


How do you install numba?

Do you try to use python 3.7 for numpy?
Please noted that we only officially support python 3.6 from Ubuntu 18.04.


I am using python3.6 and I used the following steps to install numba

  • Install conda4aarch64. This will create a minimal conda environment.
  • Add the c4aarch64 and conda-forge channels to your conda configuration:

$ conda config --add channels c4aarch64
$ conda config --add channels conda-forge

  • Then you can install Numba from the numba channel:

$ conda install -c numba numba

When I used Pyhton3.6 it is not able to find numba

Traceback (most recent call last):
File “”, line 3, in
from numba import vectorize, float32
ModuleNotFoundError: No module named ‘numba’


Not sure why the numba is not working with conda installation.

But some of our users can install it with pip directly.
Would you mind to give it a try?