You can deploy this model and start earning money today!

mayankawrr/
tds-ga8-green-ai-carbon-audit

Name: tds-ga8-green-ai-carbon-audit
Availability: InStock

co2_eq_emissions

Views

No views yet

TDS GA8 — Green AI Carbon Accounting Audit

Carbon accounting for an assigned GPU training run, reported with the standard Hugging Face co2_eq_emissions model-card metadata block.

Training run log

Field	Value
GPU type	NVIDIA H100
GPU count	8
GPU hours	30.9
Region	europe-west4
Power Usage Effectiveness (PUE)	1.13
Training type	fine-tuning

Reference values

NVIDIA H100 Thermal Design Power: 700 W
europe-west4 grid carbon intensity: 200 gCO2eq/kWh

Calculation

text

energy_kWh = (TDP_watts * num_gpus * gpu_hours * PUE) / 1000
           = (700 * 8 * 30.9 * 1.13) / 1000
           = 195.53519999999997 kWh

co2_kg     = (energy_kWh * grid_gCO2_per_kWh) / 1000
           = (195.53519999999997 * 200) / 1000
           = 39.10703999999999 kg CO2eq
           = 39.107 kg CO2eq (rounded to 3 dp)

Total energy: 195.53519999999997 kWh
Total emissions: 39.107 kg CO2eq

Notes

Running this job in europe-north1 (120 gCO2eq/kWh) instead of europe-west4 would cut emissions to 23.464 kg CO2eq — the same compute, a 40% lower footprint. Region choice dominates the carbon result once hardware and runtime are fixed.

mayankawrr/tds-ga8-green-ai-carbon-audit