Datasets and Models
The datasets shipped in this repository are format references. Users should generate their own technology-specific datasets while preserving the same header names and units expected by the scripts.
Reference Datasets
spiral_ind_data.csvtline_data.csvmimcap_dataset.csvsmall_mimcap_dataset.csvcpwd_data.csv
Critical Rule
The scripts depend on exact header names and consistent units. If you change the schema or unit system, model training and primitive optimization can break or become misleading.
Main Schemas
- Inductor:
Radius,Width,Gnd dist,Spacing,Turns,Inductance,Peak Q,Peak Q Freq,SRF - T-line:
Length,Width,Inductance,Peak Q,Peak Q Freq,SRF - MIM cap:
Length,Width,Aspect Ratio,Capacitance,Area - Small MIM cap:
Length,Width,Capacitance - CPWD:
Length,Width,Gap,Char_Imp,beta_l,root_er
Models
The MODELS/ directory is where trained .pkl artifacts are expected to live for faster repeated execution. The actual model files are not committed in the repo.