Add more Talos testing

jacobbieker · Feb 9, 2019 · 1bb85ff · 1bb85ff
1 parent 1c35a0e
commit 1bb85ff
Show file tree

Hide file tree

Showing 5 changed files with 269 additions and 6 deletions.
diff --git a/examples/talos_classification.py b/examples/talos_classification.py
@@ -7,6 +7,8 @@
 from talos.model.early_stopper import early_stopper
 from talos.model.layers import hidden_layers
 from talos.model.normalizers import lr_normalizer
+from talos.metrics.keras_metrics import root_mean_squared_error, fmeasure_acc, matthews_correlation_acc, precision_acc, recall_acc
+
 
 from factnn.utils.cross_validate import get_chunk_of_data
 
@@ -77,7 +79,11 @@ def input_model(x_train, y_train, x_val, y_val, params):
 
     model.compile(loss=params['losses'],
                   optimizer=params['optimizer'](lr_normalizer(params['lr'], params['optimizer'])),
-                  metrics=['accuracy'])
+                  metrics=['acc',
+                           fmeasure_acc,
+                           matthews_correlation_acc,
+                           precision_acc,
+                           recall_acc])
 
     out = model.fit(x_train, y_train,
                     batch_size=params['batch_size'],
@@ -89,12 +95,12 @@ def input_model(x_train, y_train, x_val, y_val, params):
     return out, model
 
 
-directory = "/media/jacob/WDRed8TB1/Insync/iact_events/"
-gamma_dir = [directory + "gamma/no_clean/"]
-proton_dir = [directory + "proton/no_clean/"]
+directory = "/home/jacob/Documents/iact_events/"
+gamma_dir = [directory + "gammaFeature/no_clean/"]
+proton_dir = [directory + "protonFeature/no_clean/"]
 
 x, y = get_chunk_of_data(directory=gamma_dir, proton_directory=proton_dir, indicies=(30, 129, 5), rebin=75,
-                         chunk_size=10000)
+                         chunk_size=1000)
 print("Got data")
 print("X Shape", x.shape)
 print("Y Shape", y.shape)
@@ -104,4 +110,4 @@ def input_model(x_train, y_train, x_val, y_val, params):
                   experiment_no='1',
                   model=input_model,
                   search_method='random',
-                  grid_downsample=0.001)
+                  grid_downsample=0.00001)
diff --git a/examples/talos_energy.py b/examples/talos_energy.py
@@ -0,0 +1,119 @@
+import talos as ta
+from keras.activations import relu, elu, softmax, hard_sigmoid, tanh
+from keras.layers import Flatten, ConvLSTM2D, Dense, Conv2D, MaxPooling2D, Dropout
+from keras.losses import mean_squared_error, mean_absolute_error
+from keras.models import Sequential
+from keras.optimizers import adam, nadam, rmsprop
+from talos.model.early_stopper import early_stopper
+from talos.model.layers import hidden_layers
+from talos.model.normalizers import lr_normalizer
+from talos.metrics.keras_metrics import root_mean_squared_error
+
+from factnn.utils.cross_validate import get_chunk_of_data
+
+# Parameter Dictionary for talos
+
+params = {'lr': (1, 10, 50),
+          'first_neuron': [4, 16, 32],
+          'last_neuron': [4, 8, 16],
+          'hidden_layers': [2, 3, 4],
+          'batch_size': [2, 8, 32],
+          'epochs': [500],
+          'dropout': (0, 0.40, 5),
+          'weight_regulizer': [None],
+          'emb_output_dims': [None],
+          'optimizer': [adam, nadam, rmsprop],
+          'losses': [mean_squared_error, mean_absolute_error],
+          'activation': [relu, elu],
+          'last_activation': [softmax],
+
+          'neuron_1': [8, 16, 32, 64],
+          'kernel_1': [1, 3, 5],
+          'stride_1': [1, 2, 3],
+          'layer_drop': [0.0, 0.4, 5],
+          'layers': [2,3,4],
+          'pool': [0,1]
+
+          }
+'''
+'rec_dropout': [0.0, 0.4, 5],
+'rec_act': [hard_sigmoid, tanh],
+'pool': [0, 1],
+'neuron_2': [4, 8, 16, 32, 64],
+'kernel_2': [1, 2, 3],
+'stride_2': [1, 2],
+'three': [0, 1],
+'neuron_3': [4, 8, 16, 32, 64],
+'kernel_3': [1, 2, 3],
+'stride_3': [1, 2],
+'''
+
+def input_model(x_train, y_train, x_val, y_val, params):
+    model = Sequential()
+
+    model.add(Conv2D(params['neuron_1'], kernel_size=params['kernel_1'], strides=params['stride_1'],
+                         padding='same',
+                         input_shape=(100, 100, 3),
+                         activation=params['activation']))
+    if params['pool']:
+        model.add(MaxPooling2D())
+    if params['dropout'] > 0.001:
+        model.add(Dropout(params['layer_drop']))
+    model.add(Conv2D(params['neuron_1'], kernel_size=params['kernel_1'], strides=params['stride_1'],
+                     padding='same', activation=params['activation']))
+    #if params['pool']:
+    #    model.add(MaxPooling2D())
+    if params['dropout'] > 0.001:
+        model.add(Dropout(params['layer_drop']))
+
+    if params['layers'] >= 3:
+        model.add(Conv2D(params['neuron_1'], kernel_size=params['kernel_1'], strides=params['stride_1'],
+                         padding='same', activation=params['activation']))
+        if params['pool']:
+            model.add(MaxPooling2D())
+        if params['dropout'] > 0.001:
+            model.add(Dropout(params['layer_drop']))
+
+    if params['layers'] >= 4:
+        model.add(Conv2D(params['neuron_1'], kernel_size=params['kernel_1'], strides=params['stride_1'],
+                         padding='same', activation=params['activation']))
+     #   if params['pool']:
+     #       model.add(MaxPooling2D())
+        if params['dropout'] > 0.001:
+            model.add(Dropout(params['layer_drop']))
+    model.add(Flatten())
+
+    hidden_layers(model, params, params['last_neuron'])
+
+    model.add(Dense(1, activation='linear'))
+
+    model.compile(loss=params['losses'],
+                  optimizer=params['optimizer'](lr_normalizer(params['lr'], params['optimizer'])),
+                  metrics=['mse'])
+
+    out = model.fit(x_train, y_train,
+                    batch_size=params['batch_size'],
+                    epochs=params['epochs'], verbose=0,
+                    validation_data=[x_val, y_val],
+                    callbacks=[early_stopper(params['epochs'],
+                                             mode='moderate')])
+
+    return out, model
+
+
+directory = "/home/jacob/Documents/iact_events/"
+gamma_dir = [directory + "gammaFeature/no_clean/"]
+proton_dir = [directory + "protonFeature/no_clean/"]
+
+x, y = get_chunk_of_data(directory=gamma_dir, indicies=(30, 129, 3), rebin=100,
+                         chunk_size=1000, as_channels=True, model_type='Energy')
+print("Got data")
+print("X Shape", x.shape)
+print("Y Shape", y.shape)
+history = ta.Scan(x, y,
+                  params=params,
+                  dataset_name='energy_test',
+                  experiment_no='1',
+                  model=input_model,
+                  search_method='random',
+                  grid_downsample=0.00001)
diff --git a/examples/talos_flat_classification.py b/examples/talos_flat_classification.py
@@ -0,0 +1,136 @@
+import talos as ta
+from keras.activations import relu, elu, softmax, hard_sigmoid, tanh
+from keras.layers import Flatten, ConvLSTM2D, Dense, Conv2D, MaxPooling2D, Dropout
+from keras.losses import mean_squared_error, mean_absolute_error
+from keras.models import Sequential
+from keras.optimizers import adam, nadam, rmsprop
+from talos.model.early_stopper import early_stopper
+from talos.model.layers import hidden_layers
+from talos.model.normalizers import lr_normalizer
+from talos.metrics.keras_metrics import root_mean_squared_error, fmeasure_acc, matthews_correlation_acc, precision_acc, recall_acc
+
+from factnn.utils.cross_validate import get_chunk_of_data
+
+# Parameter Dictionary for talos
+
+params = {'lr': (1, 10, 20),
+          'first_neuron': [4, 16, 32, 128],
+          'last_neuron': [4, 8, 16, 64],
+          'hidden_layers': [2, 3, 4, 6],
+          'batch_size': [2, 8, 32, 64],
+          'epochs': [500],
+          'dropout': (0, 0.40, 4),
+          'weight_regulizer': [None],
+          'emb_output_dims': [None],
+          'optimizer': [adam, nadam, rmsprop],
+          'losses': [mean_squared_error, mean_absolute_error],
+          'activation': [relu, elu, hard_sigmoid, tanh],
+          'last_activation': [softmax],
+
+          'neuron_1': [8, 16, 32, 64, 128],
+          'kernel_1': [1, 3, 5],
+          'stride_1': [1, 2, 3],
+          'layer_drop': [0.0, 0.4, 4],
+          'layers': [2,3,4],
+          'second_conv': [0,1],
+          'pool': [0,1]
+
+          }
+'''
+'rec_dropout': [0.0, 0.4, 5],
+'rec_act': [hard_sigmoid, tanh],
+'pool': [0, 1],
+'neuron_2': [4, 8, 16, 32, 64],
+'kernel_2': [1, 2, 3],
+'stride_2': [1, 2],
+'three': [0, 1],
+'neuron_3': [4, 8, 16, 32, 64],
+'kernel_3': [1, 2, 3],
+'stride_3': [1, 2],
+'''
+
+def input_model(x_train, y_train, x_val, y_val, params):
+    model = Sequential()
+
+    model.add(Conv2D(params['neuron_1'], kernel_size=params['kernel_1'], strides=params['stride_1'],
+                     padding='same',
+                     input_shape=(100, 100, 3),
+                     activation=params['activation']))
+    if params['second_conv']:
+        model.add(Conv2D(params['neuron_1']/2, kernel_size=params['kernel_1'], strides=1,
+                         padding='same', activation=params['activation']))
+    if params['pool']:
+        model.add(MaxPooling2D())
+    if params['dropout'] > 0.001:
+        model.add(Dropout(params['layer_drop']))
+    model.add(Conv2D(params['neuron_1'], kernel_size=params['kernel_1'], strides=params['stride_1'],
+                     padding='same', activation=params['activation']))
+    if params['second_conv']:
+        model.add(Conv2D(params['neuron_1']/2, kernel_size=params['kernel_1'], strides=1,
+                         padding='same', activation=params['activation']))
+    #if params['pool']:
+    #    model.add(MaxPooling2D())
+    if params['dropout'] > 0.001:
+        model.add(Dropout(params['layer_drop']))
+
+    if params['layers'] >= 3:
+        model.add(Conv2D(params['neuron_1'], kernel_size=params['kernel_1'], strides=params['stride_1'],
+                         padding='same', activation=params['activation']))
+        if params['second_conv']:
+            model.add(Conv2D(params['neuron_1']/2, kernel_size=params['kernel_1'], strides=1,
+                             padding='same', activation=params['activation']))
+        if params['pool']:
+            model.add(MaxPooling2D())
+        if params['dropout'] > 0.001:
+            model.add(Dropout(params['layer_drop']))
+
+    if params['layers'] >= 4:
+        model.add(Conv2D(params['neuron_1'], kernel_size=params['kernel_1'], strides=params['stride_1'],
+                         padding='same', activation=params['activation']))
+        if params['second_conv']:
+            model.add(Conv2D(params['neuron_1']/2, kernel_size=params['kernel_1'], strides=1,
+                             padding='same', activation=params['activation']))
+        #   if params['pool']:
+        #       model.add(MaxPooling2D())
+        if params['dropout'] > 0.001:
+            model.add(Dropout(params['layer_drop']))
+    model.add(Flatten())
+
+    hidden_layers(model, params, params['last_neuron'])
+
+    model.add(Dense(2, activation='softmax'))
+
+    model.compile(loss=params['losses'],
+                  optimizer=params['optimizer'](lr_normalizer(params['lr'], params['optimizer'])),
+                  metrics=['acc',
+                           fmeasure_acc,
+                           matthews_correlation_acc,
+                           precision_acc,
+                           recall_acc])
+
+    out = model.fit(x_train, y_train,
+                    batch_size=params['batch_size'],
+                    epochs=params['epochs'], verbose=0,
+                    validation_data=[x_val, y_val],
+                    callbacks=[early_stopper(params['epochs'],
+                                             mode='moderate')])
+
+    return out, model
+
+
+directory = "/home/jacob/Documents/iact_events/"
+gamma_dir = [directory + "gammaFeature/no_clean/"]
+proton_dir = [directory + "protonFeature/no_clean/"]
+
+x, y = get_chunk_of_data(directory=gamma_dir, proton_directory=proton_dir, indicies=(30, 129, 3), rebin=100,
+                         chunk_size=1000, as_channels=True)
+print("Got data")
+print("X Shape", x.shape)
+print("Y Shape", y.shape)
+history = ta.Scan(x, y,
+                  params=params,
+                  dataset_name='flat_separation_test',
+                  experiment_no='1',
+                  model=input_model,
+                  search_method='random',
+                  grid_downsample=0.00001)
diff --git a/factnn/__init__.py b/factnn/__init__.py
@@ -4,3 +4,4 @@
 from factnn.data.preprocess.simulation_preprocessors import GammaDiffusePreprocessor, GammaPreprocessor, ProtonPreprocessor
 from factnn.generator.generator.source_generators import DispGenerator, SignGenerator
 from factnn.utils import plotting
+from factnn.data.preprocess.eventfile_preprocessor import EventFilePreprocessor
diff --git a/factnn/utils/augment.py b/factnn/utils/augment.py
@@ -542,6 +542,7 @@ def augment_image_batch(images, proton_images=None, type_training=None, augment=
     elif type_training == "Energy":
         labels = [item[data_format["Energy"]] for item in training_data]
         labels = np.array(labels)
+        training_data = [item[data_format["Image"]] for item in training_data]
     elif type_training == "Disp":
         labels = [euclidean_distance(item[data_format['Source_X']], item[data_format['Source_Y']],
                                      item[data_format['COG_X']], item[data_format['COG_Y']]) for item in training_data]