Zero-values when generating carbon impact statement

[wolf019]

Hi everyone.

I am writing my thesis on deriving an approximated cost model for NLP deployment, and this tracker would help me a lot.

When executing the example, everything runs smooth, but when generating the carbon impact statement, I only get zero values. I have tried to make the example work harder and, together with my deployment code with only zero values as a result. I can see in the JSON file that a lot of information is missing to calculate a non-zero value.

I have tested the example on a Linux machine and a macOS (Big Sur) machine with the same result. (Python 3.7, Pandas 1.3.5)

I have tested parts of the code where I get information about the system. I could continue to go through the code, but the time is short, and I am not using Intel Power Gadget as a backup solution.

Terminal printout during execution:
experiment_impact_tracker.compute_tracker.ImpactTracker - WARNING - Starting process to monitor power
experiment_impact_tracker.compute_tracker.ImpactTracker - WARNING - Datapoint timestamp took 0.0006999969482421875 seconds
experiment_impact_tracker.compute_tracker.ImpactTracker - WARNING - Datapoint cpu_count_adjusted_average_load took 7.891654968261719e-05 seconds
experiment_impact_tracker.compute_tracker.ImpactTracker - WARNING - Datapoint disk_write_speed took 0.5056366920471191 seconds
experiment_impact_tracker.compute_tracker.ImpactTracker - WARNING - Datapoint timestamp took 0.0004968643188476562 seconds
experiment_impact_tracker.compute_tracker.ImpactTracker - WARNING - Datapoint cpu_count_adjusted_average_load took 2.288818359375e-05 seconds

Generate carbon impact statement:
$ generate-carbon-impact-statement my_directories that_contain all_my_experiments "USA"

loading region bounding boxes for computing carbon emissions region, this may take a moment...
454/454... rate=707.43 Hz, eta=0:00:00, total=0:00:000
Done!
/Users/xxxxx/opt/anaconda3/envs/nlp_energy_project/bin/generate-carbon-impact-statement:37: FutureWarning: Passing a negative integer is deprecated in version 1.0 and will not be supported in future version. Instead, use None to not limit the column width.
pd.set_option('display.max_colwidth', -1)
This work contributed 0.000 kg of $\text{CO}_{2eq}$ to the atmosphere and used 0.000 kWh of electricity, having a USA-specific social cost of carbon of $0.00 ($0.00, $0.00). Carbon accounting information can be found here: ....
.
.
.

Has anybody had the same issue and found a solution?

Thanks all