🏗️ Model Architecture
e.g., 224 for ImageNet models
Total depth of network
Model weight count in millions
For inference time estimation
📊 Complexity Metrics
FLOPs (Forward Pass)
—
Memory Usage (Inference)
—
Inference Time (Single Sample)
—
Throughput (Samples/sec)
—
samples/sec
Parameter Memory
—
MB
📈 Architecture Comparison
| Metric | Value | Estimate |
|---|---|---|
| Peak GPU Memory | — | |
| Total Parameters | — | Millions |
| Depth | — | Layers |
| Compute Efficiency | — | FLOPs/Parameter |
| Batch Throughput | — | Samples/sec |