Closed tavareshugo closed 1 year ago
We will increase the space to avoid this problem again.
Another thing that should do is cache the packages needed for numpy
and matplotlib
to make installation quicker.
This should work:
conda install -y -c conda-forge numpy=1.23.1 matplotlib=3.5.2
conda remove -y --force numpy matplotlib
This should remove those two packages, without removing the dependencies. So the installation process later on is quicker.
Also, update the materials to install those specific versions of the packages, otherwise newer versions (with different dependencies) might end up being downloaded.
Add TODOs to this comment:
/home
spaceconda environment now lives on /scratch
In the course iteration 2022-07-14, when we all created the
scipy
environment (here), we started getting errors because/home
ran out of space.Some solutions:
Our quick-and-dirty solution was: