sb-ai-lab / eco2ai Goto Github PK

Hey there!

First of all, thank you for the great effort to make AI emissions more transparent, and for drawing more attention to this important topic! I have been using your tool to track the emissions for (small) language models and have a couple of questions regarding the usage of your tool.

Setup:
I have a cluster system with 4 GPUs (Tesla V100) available. I run my models with the CUDA_VISIBLE_DEVICES=gpu_id prefix in the command line to only use 1 specific GPU for my training.

Problem:
In the output csv, it says that 4 GPUs have been tracked, although I enforce only 1 GPU to be used

Questions:

• Does eco2AI count the emissions of the whole rack while runtime (also runs by other people), or just the resources used by the code it is wrapped around? If the second is the case, why does it say 4 GPUs in the output file?
• How would I specify using only 1 GPU in your code, if this is not done automatically?

Thank you very much in advance for your help

Thank you for the helpful package!

A problem I've been having is when I update my environment to use a newer version of pandas, it doesn't allow for higher pandas versions.

Is there any plans to update the package to be compatible with pandas > 2.0?

Hi, i got this error when running the tool (version 0.3.5):

when running top command, it shows me the values with commas:

Desktop

To fix it, i changed line 533 of eco2ai/tools/tools_cpu.py:

from: sum_cpu += float(words[8])
to: sum_cpu += float(words[8].replace(",","."))

Can you add this to the repo? thanks!
Lucía

Hi, we are enjoying using your lib, but we have some problem, that isn't critical.

We have this lib on our linux servers and when user without root trying use this lib, he always has this problem.

The problem is solved by changing the write permissions in the /eco2ai/data/config.txt file to 777
But it would be nice if you changed this in your code, it might be better to create this file in the user's home directory every time.
It's minor problem, but after solving this problem, your lib would be better)

With respect, Mikhail R.

Hello,
I have a couple of questions:

• Can Eco2AI be used in a private cloud in containerized applications orchestrated by K8s?
• Does it track only the code's execution (just the application) or the total energy consumption of the system? and if it can track only the application's, how is it done?
  Thank you :)