| --- |
| library_name: transformers |
| tags: |
| - dia |
| - carbon-footprint |
| - energy-efficiency |
| - sustainability |
| dia_version: '0.1' |
| dia_report: |
| scope: incremental |
| lineage: |
| - model: distilbert-base-uncased |
| relation: finetune |
| compute: |
| hardware: |
| gpu: cpu-80core |
| count: 1 |
| duration_gpu_hours: 0.5381 |
| footprint: |
| energy_kwh: |
| value: 0.026 |
| quality: measured |
| carbon_kgco2eq: |
| value: 0.0017 |
| quality: measured |
| water_liters: |
| value: |
| - 0.047 |
| - 0.104 |
| quality: estimated-from-default-wue |
| context: |
| region: ca-on |
| carbon_intensity: 0.03 |
| wue_l_per_kwh: |
| - 1.8 |
| - 4.0 |
| tool: codecarbon |
| license: apache-2.0 |
| pipeline_tag: text-classification |
| base_model: distilbert-base-uncased |
| --- |
| |
| # DistilBERT — SST-2 sentiment (CPU (80-core)) |
|
|
| A demo model from the **Data & Impact Accounting (DIA)** lab. It performs |
| binary sentiment classification (SST-2) via **full fine-tune**, with the base model `distilbert-base-uncased`, trained on |
| **CPU (80-core)**. |
|
|
| The point of this repo is not the model itself but its **`dia_report`** — a |
| standardized record of the energy, carbon, and water used to train it, embedded |
| in this card's metadata. |
| |
| This footprint feeds the DIA dashboard, which rolls up a base model and all its derivatives to show the **cumulative** carbon, water, and energy cost of a model family. |
| |
| ## Training footprint |
| |
| | Metric | Value | |
| |---|---| |
| | Hardware | 1× cpu-80core | |
| | Compute | 0.5381 GPU-hours | |
| | Energy | 0.026 (measured) kWh | |
| | Carbon | 0.0017 (measured) kgCO₂eq | |
| | Water | 0.047–0.104 (estimated-from-default-wue) L | |
| | Grid region | ca-on | |
| |
| *Energy and carbon are measured with [CodeCarbon](https://github.com/mlco2/codecarbon); |
| water is estimated from a default water-usage-effectiveness range. Carbon uses the |
| local grid's intensity (Ontario, ~0.03 kgCO₂eq/kWh).* |
| |
| ## Reproduce |
| |
| ```bash |
| REPO=DIA-MVP/my-bert-sentiment-cpu python scripts/train_bert_demo.py |
| ``` |
| |
| ## Links |
| |
| - **Footprint table (dataset):** [DIA-MVP/dia-state-lab-2026](https://huggingface.co/datasets/DIA-MVP/dia-state-lab-2026) |
| - **Project / paper:** [ai-impact-accounting](https://github.com/VectorInstitute/ai-impact-accounting) |
| - **Lab workflow:** see `LAB.md` in the repo |
| |
