ML Model Parameter Counter

Summary

Add layers (Dense, Conv2D, Embedding, Attention) and instantly see total trainable parameter counts with per-layer breakdowns.

How it works

1. Select a layer type from the dropdown (Dense, Conv2D, Embedding, or Attention).
2. Fill in the layer dimensions — input size, output size, kernel size, etc.
3. Click "Add Layer" to append it to your architecture.
4. Repeat for each layer in your model.
5. Read the total parameter count and per-layer breakdown at the bottom.
6. Click the trash icon to remove any layer, or "Clear All" to reset.

Use cases

• Estimate model memory footprint before training.
• Compare architectures with different hidden sizes.
• Understand how Conv2D kernel dimensions affect parameter count.
• Plan quantization budgets by knowing exact parameter counts.
• Teach students how layer shapes drive model size.
• Validate PyTorch/Keras model.summary() outputs manually.

Frequently Asked Questions