bug_fix

autokeras/constant.py

@@ -37,7 +37,7 @@ class Constant:
     MAX_NO_IMPROVEMENT_NUM = 5
     MAX_BATCH_SIZE = 128
     LIMIT_MEMORY = False
-    SEARCH_MAX_ITER = 200
+    SEARCH_MAX_ITER = 1
 
     # text preprocessor
 

autokeras/net_transformer.py

@@ -45,9 +45,18 @@ def to_skip_connection_graph(graph):
     return graph
 
 
-def create_new_layer(input_shape, n_dim):
+def create_new_layer(layer, n_dim):
+    input_shape = layer.output.shape
     dense_deeper_classes = [StubDense, get_dropout_class(n_dim), StubReLU]
     conv_deeper_classes = [get_conv_class(n_dim), get_batch_norm_class(n_dim), StubReLU]
+    if is_layer(layer, 'ReLU'):
+        conv_deeper_classes = [get_conv_class(n_dim), get_batch_norm_class(n_dim)]
+        dense_deeper_classes = [StubDense, get_dropout_class(n_dim)]
+    elif is_layer(layer, 'Dropout'):
+        dense_deeper_classes = [StubDense, StubReLU]
+    elif is_layer(layer, 'BatchNormalization'):
+        conv_deeper_classes = [get_conv_class(n_dim), StubReLU]
+
     if len(input_shape) == 1:
         # It is in the dense layer part.
         layer_class = sample(dense_deeper_classes, 1)[0]
@@ -85,7 +94,7 @@ def to_deeper_graph(graph):
 
     for layer_id in deeper_layer_ids:
         layer = graph.layer_list[layer_id]
-        new_layer = create_new_layer(layer.output.shape, graph.n_dim)
+        new_layer = create_new_layer(layer, graph.n_dim)
         graph.to_deeper_model(layer_id, new_layer)
     return graph
 

autokeras/nn/graph.py

@@ -2,7 +2,6 @@
 from copy import deepcopy, copy
 from queue import Queue
 
-import keras
 import numpy as np
 import torch
 
@@ -505,6 +504,8 @@ def _insert_pooling_layer_chain(self, start_node_id, end_node_id):
             if is_layer(new_layer, 'Conv'):
                 filters = self.node_list[start_node_id].shape[-1]
                 new_layer = get_conv_class(self.n_dim)(filters, filters, 1, layer.stride)
+                if self.weighted:
+                    init_conv_weight(new_layer)
             else:
                 new_layer = deepcopy(layer)
             skip_output_id = self.add_layer(new_layer, skip_output_id)
@@ -685,6 +686,7 @@ def set_weight_to_graph(self):
 
 class KerasModel:
     def __init__(self, graph):
+        import keras
         self.graph = graph
         self.layers = []
         for layer in graph.layer_list:

autokeras/nn/layers.py

@@ -2,7 +2,6 @@
 
 import torch
 from torch import nn
-from keras import layers
 from torch.nn import functional
 
 from autokeras.constant import Constant
@@ -417,6 +416,7 @@ def forward(self, input_tensor):
 
 
 def keras_dropout(layer, rate):
+    from keras import layers
     input_dim = len(layer.input.shape)
     if input_dim == 2:
         return layers.SpatialDropout1D(rate)
@@ -429,6 +429,7 @@ def keras_dropout(layer, rate):
 
 
 def to_real_keras_layer(layer):
+    from keras import layers
     if is_layer(layer, 'Dense'):
         return layers.Dense(layer.units, input_shape=(layer.input_units,))
     if is_layer(layer, 'Conv'):

autokeras/search.py

@@ -113,6 +113,7 @@ def add_model(self, metric_value, loss, graph, model_id):
         if self.verbose:
             print('\nSaving model.')
 
+        graph.clear_operation_history()
         pickle_to_file(graph, os.path.join(self.path, str(model_id) + '.graph'))
 
         ret = {'model_id': model_id, 'loss': loss, 'metric_value': metric_value}

autokeras/text/text_preprocessor.py

@@ -3,12 +3,6 @@
 
 import GPUtil
 import numpy as np
-import tensorflow as tf
-from keras import Input, Model
-from keras import backend
-from keras.layers import Embedding
-from keras_preprocessing.sequence import pad_sequences
-from keras_preprocessing.text import Tokenizer
 
 from autokeras.constant import Constant
 from autokeras.utils import download_file_with_extract, temp_path_generator, ensure_dir
@@ -61,6 +55,8 @@ def tokenlize_text(max_num_words, max_seq_length, x_train):
         x_train: Tokenlized input data.
         word_index: Dictionary contains word with tokenlized index.
     """
+    from keras_preprocessing.sequence import pad_sequences
+    from keras_preprocessing.text import Tokenizer
     print("tokenlizing texts...")
     tokenizer = Tokenizer(num_words=max_num_words)
     tokenizer.fit_on_texts(x_train)
@@ -138,6 +134,7 @@ def processing(path, word_index, input_length, x_train):
     Returns:
         x_train: Numpy array as processed x_train.
     """
+    import tensorflow as tf
 
     embedding_matrix = load_pretrain(path=path, word_index=word_index)
 
@@ -149,6 +146,9 @@ def processing(path, word_index, input_length, x_train):
     os.environ["CUDA_VISIBLE_DEVICES"] = str(device_id)
     device = '/gpu:0'
     with tf.device(device):
+        from keras import Input, Model
+        from keras import backend
+        from keras.layers import Embedding
         config = tf.ConfigProto(allow_soft_placement=True)
         config.gpu_options.allow_growth = True
         sess = tf.Session(config=config)

tests/nn/test_graph.py

@@ -163,6 +163,7 @@ def test_node_consistency():
 
 
 def test_produce_keras_model():
+    import keras
     for graph in [get_conv_dense_model(),
                   get_add_skip_model(),
                   get_pooling_model(),