We deployed GATorch! This is a python library seamlessly integrated with PyTorch that enables ML developers to generate an energy consumption report. By attaching your model, the tool automatically tracks the energy consumption of your model’s training and generates graphs and plots to gain in-depth insights into the energy consumption of your model.
Here is an example of the plots you can generate:
We integrated tensorboard functionalities for an in depth analysis of your models energy consumption. Through the tensorboard dashboard it is possible to visualize how much energy your model consumes per forward pass or how much energy is needed to decrease your loss by a constant amount.
We hope this tool can be another step towards developing energy aware models!
You can also consult the Official GATorch Repository and its Manual .