another way

to train one epoch

runner blocks

Enabling partial epochs

This reverts commit 128e28b, reversing
changes made to 26b4337.

…epochs""

This reverts commit 11fd6e7.

* changing the number of val batches and training batches to apply locally
to 50 and 250 respectively. Timeouts will still apply to stop early if
it is taking too long. Amount of training completed will be computed off
of these new maxes, i.e. training_completed of 1.0 will mean all of the
250 batches were trained. With this change we will want model updates to
be counted according to local data size, therefore it is important that
the training config set the train_completion_dampener at the value: 0.0.

* setting the train_completion_dampener to 0.0

* Evan had the following order specified in his 'order' file given to us related to his new initial model:
order = t1 t2 flair t1c

* making train/val split random operations depend on a seed, default seed provided

* non default argument had been following default argument

* testing a training config

* putting capability to set seed at top level

* docstring change

* docstring

---------

Co-authored-by: hasan7n <[email protected]>