EnergyEfficientAI 0.2

A library to calculate power, energy, and training time of machine learning algorithms

EnergyEfficientAI

Overview

The EnergyEfficientAI Library provides a framework for training machine learning models while monitoring CPU and memory utilization. This library is particularly useful for understanding the energy consumption of various machine learning algorithms during training and inference. By tracking system performance metrics, users can make informed decisions about model efficiency and power consumption.

Features

• Monitor CPU and memory utilization during model training.
• Calculate power and energy consumption based on system performance.
• Generate detailed reports including training metrics and classification results.
• Visualize CPU utilization with modern, aesthetically pleasing line graphs.

Installation

To use this library, you need to have Python 3.x installed along with the following dependencies:

• numpy
• psutil
• scikit-learn
• matplotlib
• seaborn

You can install the required packages using pip:

pip install numpy psutil scikit-learn matplotlib seaborn

You can install the library using pip:

pip install EnergyEfficientAI

How to use in Code

import numpy as np
from sklearn.datasets import fetch_openml
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from EnergyEfficientAI import EnergyConsumptionML  # Import the class from the file

mnist = fetch_openml(data_id=554)

# Load MNIST data
# mnist = fetch_openml('mnist_784', as_frame=True)
X, y = mnist.data.astype('float32').to_numpy(), mnist.target.astype('int')

# Flatten the images
X_flatten = np.array([image.flatten() for image in X])

# Split the data into training and testing sets
X_train, X_test, y_train, y_test = train_test_split(X_flatten, y, test_size=0.2, random_state=42)

# Define the model (you can pass any model here)
logreg_model = LogisticRegression()
cpuIdl = 70
cpuFull = 170
# Instantiate the CustomModelTrainer with the model
model_trainer = EnergyConsumptionML(logreg_model, cpuIdl, cpuFull)

# Generate the final report by calling generate_report
model_trainer.generate_report(X_train, y_train, X_test, y_test)
model_trainer.plot_cpu_usage()