Transformer Layer Calculator
Calculate transformer architecture parameters and requirements.
Architecture Configuration
Total Parameters
108.8M
10,88,05,632 parameters
π―Head Dimension
64
β‘GFLOPs/Forward
77.31
Layer Details
Parameters per Layer7.08M
Attention Parameters2.36M
FFN Parameters4.72M
Embedding Parameters23.83M
Memory Estimates
Activation Memory/Layer1.50 MB
Total Activation Memory18.00 MB
π‘
Help us improve!
How would you rate the Transformer Layer Calculator?