Examples

There are several ways to use this package depending on how you develop. You just have to import the PowerMeter object, initialise it and call the function you want to monitor. Please provide a description of the running function, the dataset, the model, or any info you deem useful.

In most cases, you only have to declare a PowerMeter once per file/project.

Easy information filling

To avoid filling information regarding the project at each instanciation, it is possible to use a configuration file to do it only once:

{
    "project_name": "Project X",
    "program_name": "Program X",
    "client_name": "Client X",
    "cpu_power_log_path": "",
    "get_country": true,
    "is_online": false,
    "user_name": "customUsername",
    "filepath": "",
    "api_endpoint": "...",
    "location":"FR"
}

The JSON-formatted file is then handled by the method PowerMeter.from_config where an absolute path must be given.

from carbonai import PowerMeter
PROJECT_PATH = ".../path-to-the-config.json"
power_meter = PowerMeter.from_config(PROJECT_PATH)

Function decorator

To monitor the power consumption of a function, follow this example:

from carbonai import PowerMeter
power_meter = PowerMeter(project_name="MNIST classifier", is_online=False, location="FR")

@power_meter.measure_power(
    package="sklearn",
    algorithm="RandomForestClassifier",
    data_type="tabular",
    data_shape=<your_data>.shape,
    algorithm_params="n_estimators=300, max_depth=15",
    comments="Classifier trained on the MNIST dataset, 3rd test"
)
def my_func(arg1, arg2, ...):
    # Do something

Jupyter notebook magic function

To monitor the power consumption of a jupyter notebook cell, follow this example:

%load_ext carbonai.MagicPowerMeter
from carbonai import PowerMeter
power_meter = PowerMeter(project_name="MNIST classifier", is_online=False, location="FR")

%%measure_power power_meter "package_name_used" "algorithm" --data_type "tabular" --data_shape "your_data_shape" --algorithm_params "n_estimators=300, max_depth=15" --comments "Classifier trained on the MNIST dataset, 3rd test"
# Do something

with statement

This is the prefered inline method as it will stop the process even if you get an error

from carbonai import PowerMeter
power_meter = PowerMeter(project_name="MNIST classifier", is_online=False, location="FR")

with power_meter(
    package="sklearn",
    algorithm="RandomForestClassifier",
    data_type="tabular",
    data_shape=<your_data>.shape,
    algorithm_params="n_estimators=300, max_depth=15",
    comments="Classifier trained on the MNIST dataset, 3rd test"
):
    # Do something

Start and stop method

This method won’t stop the monitoring process unless told to do so, and therefore is not recommended

from carbonai import PowerMeter
power_meter = PowerMeter(project_name="MNIST classifier", is_online=False, location="FR")

power_meter.start_measure(
    package="sklearn",
    algorithm="RandomForestClassifier",
    data_type="tabular",
    data_shape=<your_data>.shape,
    algorithm_params="n_estimators=300, max_depth=15",
    comments="Classifier trained on the MNIST dataset, 3rd test"
)
# Do something
power_meter.stop_measure()

Installation Data information