CODEWITHSUNDEEP
pythonnvidianumbanvidia-jetson
Juliet Teoh
Juliet Teoh

Reputation: 11

How to install numba library on Jetson tx2?

I want to install numba on my jetson tx2 so that I can use tf-pose-estimation. All my libraries are not on the anaconda python3 but on the base python instead so I would like to download the numba library without using anaconda. Can you walk me through how to install the numba library on the jetson (etc. what additional packages I have to install before install the numba library)

I have tried pip3 install numba but the install fails when building llvmlite. My python version is 3.5.

Upvotes: 1

Views: 1148

Answers (2)

vee ven
vee ven

Reputation: 1

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

Upvotes: 0

jefflgaol
jefflgaol

Reputation: 73

Juliet Teoh. You need to install LLVM 7.0.1 first and then Numba. For LLVM, you need to run this:

$ wget http://releases.llvm.org/7.0.1/llvm-7.0.1.src.tar.xz
$ tar -xvf llvm-7.0.1.src.tar.xz
$ cd llvm-7.0.1.src.tar.xz
$ mkdir llvm_build_dir
$ cd llvm_build_dir/
$ cmake ../ -DCMAKE_BUILD_TYPE=Release -DLLVM_TARGETS_TO_BUILD="ARM;X86;AArch64"
$ make -j4
$ sudo make install
$ cd bin/
$ echo "export LLVM_CONFIG=\""`pwd`"/llvm-config\"" >> ~/.bashrc
$ echo "alias llvm='"`pwd`"/llvm-lit'" >> ~/.bashrc
$ source ~/.bashrc
$ sudo pip3 install llvmlite

For Numba, you can run this:

$ sudo pip3 install numba

I put this instruction in https://github.com/jefflgaol/Install-Packages-Jetson-ARM-Family. You may find another installation tutorial there too.

Upvotes: 1

Related Questions