Carbon Accounting Model Card

This repository documents the carbon footprint of a fine-tuning run, computed from the assigned GPU training run log (carbon_run_log_22f3000233.json).

Run details

GPU: NVIDIA RTX 4090 (TDP 450 W)
Number of GPUs: 5
GPU hours: 200.1
Power Usage Effectiveness (PUE): 1.4
Region: europe-west4 (grid carbon intensity 200 gCO2eq/kWh)
Training type: fine-tuning

Calculations

Total energy (kWh):

energy_kWh = (TDP_watts * num_gpus * gpu_hours * PUE) / 1000
           = (450 * 5 * 200.1 * 1.4) / 1000
           = 630.315 kWh

Carbon emissions (kg CO2eq):

co2_kg = (energy_kWh * region_carbon_intensity_gCO2_per_kWh) / 1000
       = (630.315 * 200) / 1000
       = 126.063 kg CO2eq

Results

Total energy: 630.315 kWh
Total emissions: 126.063 kg CO2eq

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support