PyTorch sequential regression model example

An example of a sequential network that solves a regression task used as an OpenML flow.

import torch.nn
import torch.optim

import openml
import openml.extensions.pytorch

import logging

Enable logging in order to observe the progress while running the example.

openml.config.logger.setLevel(logging.DEBUG)
openml.extensions.pytorch.config.logger.setLevel(logging.DEBUG)

Define a sequential network with 1 input layer, 3 hidden layers and 1 output layer, using the LeakyReLU activation function and a dropout rate of 0.5.

model = torch.nn.Sequential(
    torch.nn.Linear(in_features=13, out_features=256),
    torch.nn.LeakyReLU(),
    torch.nn.Dropout(),
    torch.nn.Linear(in_features=256, out_features=256),
    torch.nn.LeakyReLU(),
    torch.nn.Dropout(),
    torch.nn.Linear(in_features=256, out_features=256),
    torch.nn.LeakyReLU(),
    torch.nn.Dropout(),
    torch.nn.Linear(in_features=256, out_features=256),
    torch.nn.LeakyReLU(),
    torch.nn.Dropout(),
    torch.nn.Linear(in_features=256, out_features=1)
)

Download the OpenML task for the cholesterol dataset.

task = openml.tasks.get_task(2295)

Run the model on the task (requires an API key).

run = openml.runs.run_model_on_task(model, task, avoid_duplicate_runs=False)
# Publish the experiment on OpenML (optional, requires an API key).
run.publish()

print('URL for run: %s/run/%d' % (openml.config.server, run.run_id))