Skip to content

README.md

Latest commit

 

History

History
26 lines (19 loc) · 596 Bytes

README.md

File metadata and controls

26 lines (19 loc) · 596 Bytes
Raw

sklearn-hogwild

An implementation of the Hogwild! algorithm for asynchronous SGD that interfaces with sci-kit learn.

Requirements

A requirements.txt file is included and dependencies can be installed via:

pip install -r requirements.txt

Usage

from hogwildsgd import HogWildRegressor
hwsgd = HogWildRegressor(n_epochs = 5,
                         batch_size = 1, 
                         chunk_size = 32,
                         learning_rate = .001,
                         verbose=2)
hwsgd.fit(X,y)

Example

For an example, see the test.py file.