timeserio.keras.utils module¶

timeserio.keras.utils.iterlayers(model)[source]¶

Return iterable over all layers (and sub-layers) of a model.

This works because a keras Model is a sub-class of Layer. Can be used for freezing/un-freezing layers, etc.

timeserio.keras.utils.seed_random()[source]¶

Seed all random number generators to ensure repeatable tests.

Sets python, numpy, and tensorflow random seeds to a repeatable states. This is useful in tests, but should not be used in production.