Skip to main content

An educational module meant to serve as a prelude to talking about automatic differentiation in deep learning frameworks (such as in the Autograd module of PyTorch)

Project description

Consult the module API page at

https://engineering.purdue.edu/kak/distCGP/ComputationalGraphPrimer-1.0.2.html

for all information related to this module, including information related to the latest changes to the code.

from ComputationalGraphPrimer import *

cgp = ComputationalGraphPrimer(
               expressions = ['xx=xa^2',
                              'xy=ab*xx+ac*xa',
                              'xz=bc*xx+xy',
                              'xw=cd*xx+xz^3'],
               output_vars = ['xw'],
               dataset_size = 10000,
               learning_rate = 1e-6,
               grad_delta    = 1e-4,
               display_vals_how_often = 1000,
      )

cgp.parse_expressions()
cgp.display_network1()
cgp.gen_gt_dataset(vals_for_learnable_params = {'ab':1.0, 'bc':2.0, 'cd':3.0, 'ac':4.0})
cgp.train_on_all_data()
cgp.plot_loss()

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

ComputationalGraphPrimer-1.0.2.tar.gz (32.5 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page