PyTorch benchmark harness comparing full / fine-tuned / quantised NLP models on accuracy, latency, memory, and energy per 1,000 predictions. Produces accuracy-vs-emissions trade-off curves for stakeholder consumption.
python benchmarking inference-optimization green-ai model-quantization ai-sustainability trade-off-analysis
-
Updated
May 11, 2026 - Python