Polyaxon allows to schedule Fastai experiments, and supports tracking metrics, outputs, and models.

Overview

Polyaxon provides a tracking API to track experiment and report metrics, artifacts, logs, and results to the Polyaxon dashboard.

Tracking Fastai experiments

Polyaxon provides a Fastai callback. You can use this callback with your experiment to report metrics automatically:

As a fit callback

from polyaxon_client.tracking import Experiment
from polyaxon_client.tracking.contrib.fastai import PolyaxonFastai

...
experiment = Experiment()
...
learn.fit_one_cycle(1, 0.02, callbacks=[PolyaxonFastai(learn=learn, experiment=experiment, monitor='accuracy')])

As a partial function

from functools import partial

from polyaxon_client.tracking import Experiment
from polyaxon_client.tracking.contrib.fastai import PolyaxonFastai

...
experiment = Experiment()
...
Learner(..., callback_fns=partial(PolyaxonFastai, experiment=experiment, ...), ...)