refactor: replace .hdf5 with .keras format

aucmedi/automl/block_pred.py

@@ -120,7 +120,7 @@ def block_predict(config):
                                  standardize_mode=model.meta_standardize,
                                  **paras_datagen)
         # Load model
-        path_model = os.path.join(config["path_modeldir"], "model.last.hdf5")
+        path_model = os.path.join(config["path_modeldir"], "model.last.keras")
         model.load(path_model)
         # Start model inference
         preds = model.predict(prediction_generator=pred_gen)
@@ -139,7 +139,7 @@ def block_predict(config):
                                  **paras_datagen)
         # Load model
         path_model = os.path.join(config["path_modeldir"],
-                                  "model.best_loss.hdf5")
+                                  "model.best_loss.keras")
         model.load(path_model)
         # Start model inference via Augmenting
         preds = predict_augmenting(model, pred_gen)

aucmedi/automl/block_train.py

@@ -105,7 +105,7 @@ def block_train(config):
     callbacks = []
     if config["analysis"] == "standard":
         cb_loss = ModelCheckpoint(os.path.join(config["path_modeldir"],
-                                               "model.best_loss.hdf5"),
+                                               "model.best_loss.keras"),
                                   monitor="val_loss", verbose=1,
                                   save_best_only=True)
         callbacks.append(cb_loss)
@@ -214,7 +214,7 @@ def block_train(config):
         # Start model training
         hist = model.train(training_generator=train_gen, **paras_train)
         # Store model
-        path_model = os.path.join(config["path_modeldir"], "model.last.hdf5")
+        path_model = os.path.join(config["path_modeldir"], "model.last.keras")
         model.dump(path_model)
     elif config["analysis"] == "standard":
         # Setup neural network
@@ -247,7 +247,7 @@ def block_train(config):
                            validation_generator=val_gen,
                            **paras_train)
         # Store model
-        path_model = os.path.join(config["path_modeldir"], "model.last.hdf5")
+        path_model = os.path.join(config["path_modeldir"], "model.last.keras")
         model.dump(path_model)
     else:
         # Sanity check of architecutre config

aucmedi/ensemble/bagging.py

@@ -164,7 +164,7 @@ def train(self, training_generator, epochs=20, iterations=None,
             # Extend Callback list
             cb_mc = ModelCheckpoint(os.path.join(self.cache_dir.name,
                                                  "cv_" + str(i) + \
-                                                 ".model.hdf5"),
+                                                 ".model.keras"),
                                     monitor="val_loss", verbose=1,
                                     save_best_only=True, mode="min")
             cb_cl = CSVLogger(os.path.join(self.cache_dir.name,
@@ -306,7 +306,7 @@ def predict(self, prediction_generator, aggregate="mean",
         for i in range(self.k_fold):
             # Identify path to fitted model
             path_model = os.path.join(path_model_dir,
-                                      "cv_" + str(i) + ".model.hdf5")
+                                      "cv_" + str(i) + ".model.keras")
 
             # Gather NeuralNetwork parameters
             model_paras = {
@@ -385,7 +385,7 @@ def load(self, directory_path):
         # Check model existence
         for i in range(self.k_fold):
             path_model = os.path.join(directory_path,
-                                      "cv_" + str(i) + ".model.hdf5")
+                                      "cv_" + str(i) + ".model.keras")
             if not os.path.exists(path_model):
                 raise FileNotFoundError("Bagging model for fold " + str(i) + \
                                         " does not exist!", path_model)

aucmedi/ensemble/composite.py

@@ -220,7 +220,7 @@ def train(self, training_generator, epochs=20, iterations=None,
             callbacks_model = callbacks.copy()
             # Extend Callback list
             path_model = os.path.join(self.cache_dir.name,
-                                      "cv_" + str(i) + ".model.hdf5")
+                                      "cv_" + str(i) + ".model.keras")
             cb_mc = ModelCheckpoint(path_model,
                                     monitor="val_loss", verbose=1,
                                     save_best_only=True, mode="min")
@@ -336,7 +336,7 @@ def train_metalearner(self, training_generator):
         for i in range(len(self.model_list)):
             # Load current model
             path_model = os.path.join(path_model_dir,
-                                      "cv_" + str(i) + ".model.hdf5")
+                                      "cv_" + str(i) + ".model.keras")
 
             # Gather NeuralNetwork parameters
             model_paras = {
@@ -447,7 +447,7 @@ def predict(self, prediction_generator, return_ensemble=False):
         # Sequentially iterate over model list
         for i in range(len(self.model_list)):
             path_model = os.path.join(path_model_dir,
-                                      "cv_" + str(i) + ".model.hdf5")
+                                      "cv_" + str(i) + ".model.keras")
 
             # Gather NeuralNetwork parameters
             model_paras = {
@@ -554,7 +554,7 @@ def load(self, directory_path):
         # Check model existence
         for i in range(len(self.model_list)):
             path_model = os.path.join(directory_path,
-                                      "cv_" + str(i) + ".model.hdf5")
+                                      "cv_" + str(i) + ".model.keras")
             if not os.path.exists(path_model):
                 raise FileNotFoundError("Composite model " + str(i) + \
                                         " does not exist!", path_model)

aucmedi/ensemble/stacking.py

@@ -208,7 +208,7 @@ def train(self, training_generator, epochs=20, iterations=None,
             callbacks_model = callbacks.copy()
             # Extend Callback list
             path_model = os.path.join(self.cache_dir.name,
-                                      "nn_" + str(i) + ".model.hdf5")
+                                      "nn_" + str(i) + ".model.keras")
             cb_mc = ModelCheckpoint(path_model,
                                     monitor="val_loss", verbose=1,
                                     save_best_only=True, mode="min")
@@ -324,7 +324,7 @@ def train_metalearner(self, training_generator):
         for i in range(len(self.model_list)):
             #  Load current model
             path_model = os.path.join(path_model_dir,
-                                      "nn_" + str(i) + ".model.hdf5")
+                                      "nn_" + str(i) + ".model.keras")
 
             # Gather NeuralNetwork parameters
             model_paras = {
@@ -434,7 +434,7 @@ def predict(self, prediction_generator, return_ensemble=False):
         # Sequentially iterate over model list
         for i in range(len(self.model_list)):
             path_model = os.path.join(path_model_dir,
-                                      "nn_" + str(i) + ".model.hdf5")
+                                      "nn_" + str(i) + ".model.keras")
 
             # Gather NeuralNetwork parameters
             model_paras = {
@@ -541,7 +541,7 @@ def load(self, directory_path):
         # Check model existence
         for i in range(len(self.model_list)):
             path_model = os.path.join(directory_path,
-                                      "nn_" + str(i) + ".model.hdf5")
+                                      "nn_" + str(i) + ".model.keras")
             if not os.path.exists(path_model):
                 raise FileNotFoundError("Stacking model " + str(i) + \
                                         " does not exist!", path_model)

aucmedi/neural_network/model.py

@@ -372,7 +372,7 @@ def reset_weights(self):
     def dump(self, file_path):
         """ Store model to disk.
 
-        Recommended to utilize the file format ".hdf5".
+        Recommended to utilize the file format ".keras".
 
         Args:
             file_path (str):    Path to store the model on disk.
@@ -385,7 +385,7 @@ def load(self, file_path, custom_objects={}):
 
         After loading, the model will be compiled.
 
-        If loading a model in ".hdf5" format, it is not necessary to define any custom_objects.
+        If loading a model in ".keras" format, it is not necessary to define any custom_objects.
 
         Args:
             file_path (str):            Input path, from which the model will be loaded.

aucmedi/xai/__init__.py

@@ -35,7 +35,7 @@
     # Get a model
     model = NeuralNetwork(n_labels=3, channels=3, architecture="Xception",
                            input_shape=(299,299))
-    model.load("model.xray.hdf5")
+    model.load("model.xray.keras")
 
     # Make some predictions
     preds = model.predict(datagen)

aucmedi/xai/decoder.py

@@ -55,7 +55,7 @@ def xai_decoder(data_gen, model, preds=None, method="gradcam", layerName=None,
         # Get a model
         model = NeuralNetwork(n_labels=3, channels=3, architecture="Xception",
                                input_shape=(299,299))
-        model.load("model.xray.hdf5")
+        model.load("model.xray.keras")
 
         # Make some predictions
         preds = model.predict(datagen)

docs/automl/cli/usage.md

@@ -44,11 +44,11 @@ To enable them in other operations, rebuild TensorFlow with the appropriate comp
 Epoch 1/10
 2022-07-18 12:57:32.516662: I tensorflow/stream_executor/cuda/cuda_dnn.cc:368] Loaded cuDNN version 8100
 177/177 [==============================] - ETA: 0s - loss: 0.5103 - auc: 0.9563 - f1_score: 0.7556   
-Epoch 1: val_loss improved from inf to 0.20766, saving model to model/model.best_loss.hdf5
+Epoch 1: val_loss improved from inf to 0.20766, saving model to model/model.best_loss.keras
 177/177 [==============================] - 21s 92ms/step - loss: 0.5103 - auc: 0.9563 - f1_score: 0.7556 - val_loss: 0.2077 - val_auc: 0.9864 - val_f1_score: 0.8958 - lr: 1.0000e-04
 Epoch 2/10
 177/177 [==============================] - ETA: 0s - loss: 0.1932 - auc: 0.9893 - f1_score: 0.8842
-Epoch 2: val_loss improved from 0.20766 to 0.18348, saving model to model/model.best_loss.hdf5
+Epoch 2: val_loss improved from 0.20766 to 0.18348, saving model to model/model.best_loss.keras
 177/177 [==============================] - 15s 84ms/step - loss: 0.1932 - auc: 0.9893 - f1_score: 0.8842 - val_loss: 0.1835 - val_auc: 0.9891 - val_f1_score: 0.9010 - lr: 1.0000e-04
 ...
 Epoch 23/25
@@ -79,8 +79,8 @@ cwd/
 └── model/
     ├── logs.training.csv
     ├── meta.training.json
-    ├── model.best_loss.hdf5
-    ├── model.last.hdf5
+    ├── model.best_loss.keras
+    ├── model.last.keras
     └── plot.fitting_course.png
 ```
 

docs/automl/docker/usage.md

@@ -95,11 +95,11 @@ To enable them in other operations, rebuild TensorFlow with the appropriate comp
 Epoch 1/10
 2022-07-18 12:57:32.516662: I tensorflow/stream_executor/cuda/cuda_dnn.cc:368] Loaded cuDNN version 8100
 177/177 [==============================] - ETA: 0s - loss: 0.5103 - auc: 0.9563 - f1_score: 0.7556   
-Epoch 1: val_loss improved from inf to 0.20766, saving model to model/model.best_loss.hdf5
+Epoch 1: val_loss improved from inf to 0.20766, saving model to model/model.best_loss.keras
 177/177 [==============================] - 21s 92ms/step - loss: 0.5103 - auc: 0.9563 - f1_score: 0.7556 - val_loss: 0.2077 - val_auc: 0.9864 - val_f1_score: 0.8958 - lr: 1.0000e-04
 Epoch 2/10
 177/177 [==============================] - ETA: 0s - loss: 0.1932 - auc: 0.9893 - f1_score: 0.8842
-Epoch 2: val_loss improved from 0.20766 to 0.18348, saving model to model/model.best_loss.hdf5
+Epoch 2: val_loss improved from 0.20766 to 0.18348, saving model to model/model.best_loss.keras
 177/177 [==============================] - 15s 84ms/step - loss: 0.1932 - auc: 0.9893 - f1_score: 0.8842 - val_loss: 0.1835 - val_auc: 0.9891 - val_f1_score: 0.9010 - lr: 1.0000e-04
 ...
 Epoch 23/25
@@ -130,8 +130,8 @@ aucmedi.data/
 └── model/
     ├── logs.training.csv
     ├── meta.training.json
-    ├── model.best_loss.hdf5
-    ├── model.last.hdf5
+    ├── model.best_loss.keras
+    ├── model.last.keras
     └── plot.fitting_course.png
 ```
 

tests/test_automl_block_train.py

@@ -116,7 +116,7 @@ def test_minimal(self):
         # Run AutoML training block
         block_train(config)
 
-        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.last.hdf5")))
+        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.last.keras")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "logs.training.csv")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "meta.training.json")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "plot.fitting_course.png")))
@@ -144,7 +144,7 @@ def test_minimal_multilabel(self):
         # Run AutoML training block
         block_train(config)
 
-        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.last.hdf5")))
+        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.last.keras")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "logs.training.csv")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "meta.training.json")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "plot.fitting_course.png")))
@@ -172,7 +172,7 @@ def test_minimal_3D(self):
         # Run AutoML training block
         block_train(config)
 
-        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.last.hdf5")))
+        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.last.keras")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "logs.training.csv")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "meta.training.json")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "plot.fitting_course.png")))
@@ -202,8 +202,8 @@ def test_standard(self):
         # Run AutoML training block
         block_train(config)
 
-        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.last.hdf5")))
-        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.best_loss.hdf5")))
+        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.last.keras")))
+        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.best_loss.keras")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "logs.training.csv")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "meta.training.json")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "plot.fitting_course.png")))
@@ -230,8 +230,8 @@ def test_standard_multilabel(self):
         # Run AutoML training block
         block_train(config)
 
-        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.last.hdf5")))
-        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.best_loss.hdf5")))
+        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.last.keras")))
+        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.best_loss.keras")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "logs.training.csv")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "meta.training.json")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "plot.fitting_course.png")))
@@ -259,8 +259,8 @@ def test_standard_3D(self):
         # Run AutoML training block
         block_train(config)
 
-        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.last.hdf5")))
-        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.best_loss.hdf5")))
+        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.last.keras")))
+        self.assertTrue(os.path.exists(os.path.join(output_dir.name, "model.best_loss.keras")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "logs.training.csv")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "meta.training.json")))
         self.assertTrue(os.path.exists(os.path.join(output_dir.name, "plot.fitting_course.png")))