Install blas & mkl (Kaggle#1060)

* Install blas & mkl

These are necessary for the TPU pytorch XLA.
They also provide a significant perforance boost.

They used to be automatically included when we were installing torch
from conda but they are not now that we install torch from pip.

I added a test to prevent regression.

http://b/195414519

* separate conda commands to speed up resolution

Install blas & mkl (Kaggle#1060)

* Install blas & mkl

These are necessary for the TPU pytorch XLA.
They also provide a significant perforance boost.

They used to be automatically included when we were installing torch
from conda but they are not now that we install torch from pip.

I added a test to prevent regression.

http://b/195414519

* separate conda commands to speed up resolution

Upgrade JAX to 0.2.16 (Kaggle#1039)

Upgrade JAX to 0.2.16 (Kaggle#1039)

Unpin wand (Kaggle#1003)

Latest version of `wand` is compatible with latest version of tesseract

Unpin wand (Kaggle#1003)

Latest version of `wand` is compatible with latest version of tesseract

Pin pystan (Kaggle#978)

pystan 3.x released on March 25th, 2021 is not compatible with
fbprophet and causing a build failure.

http://b/183944405

Pin pystan (Kaggle#978)

pystan 3.x released on March 25th, 2021 is not compatible with
fbprophet and causing a build failure.

http://b/183944405

Install openssh-client (Kaggle#971)

In Kaggle#967, I upgraded the base image from `m61` to `m64`.

The new  `m64` base image doesn't include a SSH client anymore.

This was causing issues with a learn course relying on `stable-baselines`. This package isn't included in our base image but is installed at runtime.

Install openssh-client (Kaggle#971)

In Kaggle#967, I upgraded the base image from `m61` to `m64`.

The new  `m64` base image doesn't include a SSH client anymore.

This was causing issues with a learn course relying on `stable-baselines`. This package isn't included in our base image but is installed at runtime.