Computing & Software
Dense Neural Network Parameter & Weight Memory Calculator
Count fully connected layer weights and biases from an architecture’s layer widths.
How the calculation works
Trainable dense weights and biases = dense_parameters(layers) parameters. Weight payload at selected format = dense_parameters(layers)*bytes_per_parameter bytes.
Worked example
Example inputs — Input and subsequent layer widths, including output: 10, 5, 2; Stored bytes per parameter: 4. Expected results — Trainable dense weights and biases: 67 parameters; Weight payload at selected format: 268 bytes.
Assumptions and limits
Analytical software/hardware scenario, not a benchmark or a live provider quote. Use the actual architecture, workload and billing definitions; overheads not named in the model are excluded. Bias on every non-input layer; fully connected adjacent layers only. Activations, gradients, optimizer states, padding, embeddings and shared weights are excluded.
Reference and further reading
Read the underlying subject reference. Calculith is an independent implementation; the reference does not endorse or certify this tool.
Related calculators
- Password Entropy & Brute-Force Time Estimate
- CPU Instruction Count & CPI Runtime Calculator
- Amdahl & Gustafson Parallel Speedup Calculator
- Roofline Arithmetic Intensity Performance Calculator
- Memory Bus Theoretical Bandwidth Calculator
Example values are editable. Results are estimates where assumptions apply. Calculations run in your browser. Updated 2026-11-16.