How to transform dict X = { 'layer': layer_array, # shape (#layers, #layer-features) 'levels': levels_array, # shape (#levels, #level-features) 'globals': globals_array (#global-features,) } to the form th
(self, X: Dict[str, np.ndarray])
| 32 | return self._out_dim |
| 33 | |
| 34 | def transform(self, X: Dict[str, np.ndarray]) -> Any: |
| 35 | """ |
| 36 | How to transform dict |
| 37 | X = { |
| 38 | 'layer': layer_array, # shape (#layers, #layer-features) |
| 39 | 'levels': levels_array, # shape (#levels, #level-features) |
| 40 | 'globals': globals_array (#global-features,) |
| 41 | } |
| 42 | to the form the model will use/receive it in forward. |
| 43 | Implementation will be applied (with multi-processing) in the _get_item(.) method of the dataset |
| 44 | --> IMPORTANT: the arrays in X will *not* have the batch dimension! |
| 45 | """ |
| 46 | raise NotImplementedError |
| 47 | |
| 48 | def batched_transform(self, X: Dict[str, np.ndarray]) -> Any: |
| 49 | """ |
nothing calls this directly
no outgoing calls
no test coverage detected