Eco2AI

Eco2AI - is a python library which calculates power consumption and CO2 emission during running code. Emission information is written to file. Every single run of tracker creates line in this file with following items:

• project_name
• experiment_description
• start_time
• duration(s)
• power_consumption(kWTh)
• CO2_emissions(kg)
• CPU_name
• GPU_name
• OS
• country

Installation

To install eco2ai library run next command:

pip install eco2ai

Use examples

eco2ai's interface is quite simple. Here is a the most straightforward usage example:

import eco2ai

tracker = eco2ai.Tracker(project_name="YourProjectName", experiment_description="training the <your model> model")

tracker.start()

<your gpu &(or) cpu calculations>

tracker.stop()

eco2ai also supports decorators. Once decorated function executed, emissions info will be written to the file. See example below:

from eco2ai import track

@track
def train_func(model, dataset, optimizer, epochs):
    ...

train_func(your_model, your_dataset, your_optimizer, your_epochs)

For your convenience every time you initilize a Tracker object with your custom parameters, this settings will be saved until library is uninstalled, and then every new tracker will be created with your custom settings(if you will create new tracker with new parameters, then they will be saved instead of old ones). For example:

import eco2ai

tracker = eco2ai.Tracker(
    project_name="YourProjectName", 
    experiment_description="training <your model> model",
    file_name="emission.csv"
    )

tracker.start()
<your gpu &(or) cpu calculations>
tracker.stop()

...

# now, we want to create a new tracker for new calculations
tracker = eco2ai.Tracker()
# now, it's equivelent to:
# tracker = eco2ai.Tracker(
#     project_name="YourProjectName", 
#     experiment_description="training the <your model> model",
#     file_name="emission.csv"
# )
tracker.start()
<your gpu &(or) cpu calculations>
tracker.stop()

You can also set parameters using set_params() function, like in the example below:

from eco2ai import set_params, Tracker

set_params(
    project_name="My_default_project_name",
    experiment_description="We trained...",
    file_name="my_emission_file.csv"
)

tracker = Tracker()
# now, it's equivelent to:
# tracker = Tracker(
#     project_name="My_default_project_name",
#     experiment_description="We trained...",
#     file_name="my_emission_file.csv"
# )
tracker.start()
<your code>
tracker.stop()

Important note

According to climate transparency report for each kilowatt hour of electricity generated in Russia in 2020, an average of 310 g of CO2 was emitted. This constant is used in CO2 estimation by default.

In order to calculate gpu & cpu power consumption correctly you should create the 'Tracker' before any gpu or cpu usage

For every new calculation create a new “Tracker.”

Feedback

If you had some problems while working with our tracker, please, give us a feedback comments in document