Python library for the Pythia platform.
from pythia import Experiment
# init experiment
# implies that an instance of Pythia is listening on localhost
exp = Experiment('http://localhost:5667/api/v1/')
exp.name = "some catchy name"
exp.description = "describe the experiment here"
# create a model
model = exp.create_model()
model.name = "2 layers foo bar dropout softmax"
for i in range(NUM_EPOCHS):
for j in range(NUM_EXAMPLES):
# the following two lines are not really valid
train_accuracy = accuracy(classifier)
# save the measured training accuracy
* [MNIST ConvNet in Tensorflow](https://github.com/ChristianSch/PyPythia/blob/master/examples/Tensorflow%20MNIST.ipynb)
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.