Neural Architecture Search
Neural Architecture Search (NAS) automates the design of neural network architectures by searching over a defined space of possible configurations.
Search Spaces
| Component | Options |
|---|---|
| Layer types | Dense, Conv1D, Conv2D, LSTM, GRU, Transformer |
| Layer count | 1-10 layers |
| Hidden units | 32, 64, 128, 256, 512, 1024 |
| Activation | ReLU, GELU, SiLU, Tanh |
| Dropout | 0.0 - 0.5 |
| Normalization | BatchNorm, LayerNorm, None |
Source Files
| File | Path |
|---|---|
| NAS Module | data-plane/ml-service/src/automl/nas.py |
| Search Strategies | data-plane/ml-service/src/automl/search.py |