Hwallace/patches/fix model build

src/MaCh3PythonUtils/machine_learning/file_ml_interface.py

@@ -213,8 +213,6 @@ def test_model(self):
 
         print("Training Results!")
         train_prediction = self.model_predict(self._training_data)
-        print(self._training_data)
-        print(train_prediction)
         train_as_numpy = self.scale_labels(self._training_labels).T[0]
         self.evaluate_model(train_prediction, train_as_numpy, "train_qq_plot.pdf")
 

src/MaCh3PythonUtils/machine_learning/tensorflow/tf_interface.py

@@ -5,7 +5,7 @@
 import numpy as np
 import tensorflow.keras as tfk
 import tensorflow_probability as tfp
-
+from typing import Iterable
 
 
 class TfInterface(FileMLInterface):
@@ -36,7 +36,7 @@ def add_layer(self, layer_id: str, layer_args: dict):
 
         self._layers.append(self.__TF_LAYER_IMPLEMENTATIONS[layer_id.lower()](**layer_args))
 
-    def build_model(self, _: dict):
+    def build_model(self, **kwargs):
         return None
 
 
@@ -93,11 +93,14 @@ def model_predict(self, test_data: pd.DataFrame):
         if self._model is None:
             return np.zeros(len(test_data))
 
-        pred = self._model.predict(scaled_data, verbose=False)
-        print(f"PREDICTION: {pred}")
-        # return self._model.predict(scaled_data, verbose=False).T[0]
-        return pred
+        return self._model.predict(scaled_data, verbose=False).T[0]
 
     def model_predict_no_scale(self, test_data):
         # Same as above but specifically for TF, optimised to avoid if statement...
         return self._model(test_data, training=False)
+
+    def evaluate_model(self, predicted_values: Iterable, true_values: Iterable, outfile: str = ""):
+
+        # CODE TO DO TF SPECIFIC PLOTS GOES HERE
+
+        return super().evaluate_model(predicted_values, true_values, outfile)

src/MaCh3PythonUtils/machine_learning/tensorflow/tf_normalizing_flow_model.py

@@ -54,10 +54,10 @@ def __call__(self):
 
 
 class TfNormalizingFlowModel(TfManualInterface):
-    def build_model(self, model_args):
+    def build_model(self, **kwargs):
         input_dim = self.chain.ndim-1
-        self._model = NormalizingFlow(model_args.get("hidden_units", [100]), model_args.get("n_bijectors", 1), input_dim)()
-        self._optimizer = tfk.optimizers.Adam(model_args.get("learning_rate", 1e-3))
+        self._model = NormalizingFlow(kwargs.get("hidden_units", [100]), kwargs.get("n_bijectors", 1), input_dim)()
+        self._optimizer = tfk.optimizers.Adam(kwargs.get("learning_rate", 1e-3))
 
     def nll_loss(self, features):
         return -tf.reduce_mean(self._model.log_prob(features))

src/MaCh3PythonUtils/machine_learning/tensorflow/tf_residual_model.py

@@ -3,7 +3,7 @@
 
 
 class TfResidualModel(TfManualLayeredInterface):
-    def build_model(self, model_args: dict):
+    def build_model(self, **kwargs):
         input_shape = self.training_data.shape[1:]  # Assuming shape is (batch_size, features)
         network_input = tfk.layers.Input(shape=input_shape)
 
@@ -23,9 +23,9 @@ def build_model(self, model_args: dict):
 
         # Define and compile the model
         self._model = tfk.Model(inputs=network_input, outputs=x)
-        optimizer = tfk.optimizers.AdamW(learning_rate=model_args.get("learning_rate", 1e-5),
+        optimizer = tfk.optimizers.AdamW(learning_rate=kwargs.get("learning_rate", 1e-5),
                         weight_decay=1e-4, clipnorm=1.0)
 
-        _ = model_args.pop("learning_rate", None)
+        _ = kwargs.pop("learning_rate", None)
 
-        self._model.compile(optimizer=optimizer, **model_args)
+        self._model.compile(optimizer=optimizer, **kwargs)

src/MaCh3PythonUtils/machine_learning/tensorflow/tf_sequential_model.py

@@ -4,11 +4,11 @@
 
 class TfSequentialModel(TfManualLayeredInterface):
 
-    def build_model(self, model_args: dict):
+    def build_model(self, **kwargs: dict):
         """Build and compile TF model
 
-        :param model_args: Model arguments as dictionary
-        :type model_args: dict
+        :param kwargs: Model arguments as dictionary
+        :type kwargs: dict
         :raises ValueError: Model not set up yet
         """
         self._model = tfk.Sequential()
@@ -20,10 +20,10 @@ def build_model(self, model_args: dict):
             self._model.add(layer)
 
         self._model.build()
-        optimizer = tfk.optimizers.AdamW(learning_rate=model_args.get("learning_rate", 1e-5),
+        optimizer = tfk.optimizers.AdamW(learning_rate=kwargs.get("learning_rate", 1e-5),
                         weight_decay=1e-4, clipnorm=1.0)
 
-        model_args.pop("learning_rate", None)
+        kwargs.pop("learning_rate", None)
 
 
-        self._model.compile(**model_args, optimizer=optimizer)
+        self._model.compile(**kwargs, optimizer=optimizer)