Distributed Grid Descent: an algorithm for hyperparameter tuning guided by Bayesian inference, designed to run on multiple processes and potentially many machines with no central point of control

Last update: Jan 01, 2022

Related tags

Algorithms DistributedGridDescent

Overview

Distributed Grid Descent

An implementation of Distributed Grid Descent: an algorithm for hyperparameter tuning guided by Bayesian inference, designed to run on multiple processes and potentially many machines with no central point of control as described in Appendix B of Working Memory Graphs [Loynd et al., 2019].

Note: This project is a work in progress. Please contact me if you like to contribute and help to develop a fully fledged python library out of it.

Usage

import numpy as np
from dgd import DistributedGridDescent

model = ... # model wrapper
data = {
    "train_data": ...
}

param_grid = {
    "learning_rate":[3e-3, 1e-3, 3e-4, 1e-4, 3e-5, 1e-5],
    "optimizer":["adam", "rmsprop"],
    "lr_annealing":[False, 0.95, 0.99],
    "batch_size":[32, 64, 128, 256, 1024],
    "num_linear_layers":[1, 2, 4, 8, 16],
    "num_neurons":[512, 256, 128, 64, 32, 16],
    "dropout":[0.0, 0.1, 0.3, 0.5],
    "l2":[0.0, 0.01, 0.1]
}

dgd = DistributedGridDescent(model, param_grid, metric=np.mean, n_jobs=-1)
dgd.run(data)

print(dgd.best_params_)
df = pd.DataFrame(dgd.results_).set_index("ID").sort_values(by=["metric"],ascending=False)

Examples and Tutorials

See sklearn_example.py, pytorch_example.py, rosenbrock_example.py and tensorflow_example.py in the examples folder for examples of basic usage of dgd.
See rosenbrock_server_example.py for an example of distributed usage.

Strong and weak scaling analysis

Algorithm

Input: Set of hyperparameters H, each having a discrete, ordered set of possible values.  
Input: Maximum number of training steps N per run.  
repeat  
    Download any new results.  
    if no results so far then
        Choose a random configuration C from the grid defined by H.
    else
        Identify the run set S with the highest metric.
        Initialize neighborhood B to contain only S.
        Expand B by adding all possible sets whose configurations differ from that of S by one step in exactly one hyperparameter setting.
        Calculate a ceiling M = Count(B) + 1.
        Weight each run set x in B M - Count(x).
        Sample a random run set S' from B according to run set weights.
        Choose configuration C from S'.
    end if
    Perform one training run of N steps using C.
    Calculate the runs Metric.
    Log the result on shared storage.
until terminated by user.

See Appendix B of Loynd et al., 2019 for details.

Distributed Grid Descent: an algorithm for hyperparameter tuning guided by Bayesian inference, designed to run on multiple processes and potentially many machines with no central point of control

Related tags

Overview

Distributed Grid Descent

Usage

Examples and Tutorials

Strong and weak scaling analysis

Algorithm

Owner

Martin

My dynamic programming algorithms for exercise and fun

CLI Eight Puzzle mini-game featuring BFS, DFS, Greedy and A* searches as solver algorithms.

Implementation of core NuPIC algorithms in C++

FLIght SCheduling OPTimization - a simple optimization library for flight scheduling and related problems in the discrete domain

Implements (high-dimenstional) clustering algorithm

Python-Strongest-Encrypter - Transform your text into encrypted symbols using their dictionary

Primedice like provably fair algorithm

marching Squares algorithm in python with clean code.

Implementation of Apriori algorithms via Python

Infomap is a network clustering algorithm based on the Map equation.

This python algorithm creates a simple house floor plan based on a user-provided CSV file.

sudoku solver using CSP forward-tracking algorithms.

Python algorithm to determine the optimal elevation threshold of a GNSS receiver, by using a statistical test known as the Brown-Forsynthe test.

A lightweight, pure-Python mobile robot simulator designed for experiments in Artificial Intelligence (AI) and Machine Learning, especially for Jupyter Notebooks

A simple python implementation of A* and bfs algorithm solving Eight-Puzzle

FPE - Format Preserving Encryption with FF3 in Python

Genetic Algorithm for Robby Robot based on Complexity a Guided Tour by Melanie Mitchell

Implemented page rank program

Exam Schedule Generator using Genetic Algorithm

Algorithms and data structures for educational, demonstrational and experimental purposes.