CS443-2021-ecoML project update

EPFLiGHT / cumulator

A tool to quantify and report the carbon footprint of machine learning computations and communication

MIT License

21 stars 5 forks source link

CS443-2021-ecoML project update #17

Closed giovannimonea closed 2 years ago

giovannimonea commented 2 years ago

Changes

Updated README.rst with changes from ecoML project
Implemented a small web app for predicting consumption and accuracy

Grim-bot commented 2 years ago

Thank you for updating the README file and implementig a web app! How can one access the web app ? I didn't find any link to it from the readme file with your proposed changes.

Grim-bot commented 2 years ago

Hi! In an effort to try out the web app, I cloned your fork and navigated to the CS433-2021-ecoML-update branch. I followed the instructions from the readme file and got a ModuleNotFoundError. What should I do?

C:\Users\felix\Documents\cumulator>python .\src\cumulator\web_app\app.py
Traceback (most recent call last):
  File ".\src\cumulator\web_app\app.py", line 4, in <module>
    from cumulator.prediction_feature.prediction_helper import compute_features, get_predictions
ModuleNotFoundError: No module named 'cumulator.prediction_feature'

Grim-bot commented 2 years ago

Do you think some of you will continue developing ecoML ? In this case, I have the following ideas:

Reduce the number of digits displayed (e.g., I had roughly 991 gCO2eq with more than 10 digits after the decimal point)
Write unit tests for the web app
Host the web app for anyone to use. Careful, this is not as simple as finding a hosting service: You would at least also need to think about input sanitization and about GDPR

Grim-bot commented 2 years ago

Thank you for the fixes! I tried the web app, it's really cool! I used a dataset that isn't very well suited, so I got terrible results, but that's a good sign !