basic-tests.class-or-module - created by notespace, Thu Dec 19 01:44:59 IST 2019.

Experimenting with libpython-clj 1.28

(require
  (quote [libpython-clj.require :as req
          :refer [require-python]
          :reload true])
  (quote
    [libpython-clj.python :refer
     [as-python as-jvm ->python ->jvm
      get-attr call-attr call-attr-kw
      get-item att-type-map call call-kw
      initialize! as-numpy as-tensor
      ->numpy run-simple-string
      add-module module-dict
      import-module python-type] :as
     py])
  (quote [basic-tests.utils :refer
          [def+ matplotlib->svg]]))

nil

the dot in python has multiple meanings what are the elements between the dots :

tfds.Split.TRAIN.subsplit

In the following tutorial: https://www.tensorflow.org/tutorials/keras/classification

We find the following code:

import tensorflow_datasets as tfds
train_validation_split = tfds.Split.TRAIN.subsplit ([6, 4])

We can assume that subsplit is a method But what are Split and TRAIN? Are they modules, classes?

import tensorflow_datasets as tfds
type (tfds.Split)
<class 'type'>
type (tfds.Split.TRAIN)
<class 'tensorflow_datasets.core.splits.NamedSplit'>

The best way to implement in clojure seems to be:

(py/import-as tensorflow tf)

(py/import-as tensorflow_datasets tfds)

(def train (py/$.. tfds Split TRAIN))

(py/call-attr train
              "subsplit"
              (py/->py-list [6 4]))

(NamedSplit('train')(tfds.percent[0:60]), NamedSplit('train')(tfds.percent[60:100]))

basic-tests.class-or-module - created by notespace, Thu Dec 19 01:44:59 IST 2019.