Preset: GPT-style Model
Preset: Edge Deployment
GreenAISim: AI Sustainability Simulator
1. Model Parameters
Model Type:
Transformer (e.g., BERT)
CNN
MLP
Model Size (parameters in billions):
Training Epochs:
Precision:
FP32
FP16
INT8
2. Hardware Configuration
Processor Type:
GPU
TPU
Hardware Model:
NVIDIA A100
NVIDIA V100
NVIDIA T4
Google TPU v4
Google TPU v5e
Google TPU v5p
Compute Hours:
3. Deployment Location
Region:
US West (high carbon)
EU Central (medium)
Nordics (low carbon)
Power Usage Effectiveness (PUE):
1.3
4. Inference Profile
Model reused or fine-tuned (shared cost)
Inference Frequency (per day):
Deployment Duration (months):
Simulate Environmental Impact