1cda66554188c4504aa94d1bcffd4cef9ee8a997,autokeras/net_transformer.py,,transform,#Any#,79
Before Change
def transform(model):
models = []
layers = model.layers
for index in range(len(layers) - 1):
if isinstance(layers[index], tuple(WEIGHTED_LAYER_FUNC_LIST)):
models.append(to_deeper_model(model, index))
models.append(to_wider_model(model, index))
return models
After Change
// search dense variation
if isinstance(layer, Dense):
models.append(Sequential(copy_conv_model(conv_model), to_deeper_dense_model(dense_model, index)))
models.append(Sequential(copy_conv_model(conv_model), to_wider_dense_model(dense_model, index)))
for layer in conv_model.layers:
// search conv variation
In pattern: SUPERPATTERN
Frequency: 4
Non-data size: 3
Instances
Project Name: jhfjhfj1/autokeras
Commit Name: 1cda66554188c4504aa94d1bcffd4cef9ee8a997
Time: 2017-12-28
Author: jhfjhfj1@gmail.com
File Name: autokeras/net_transformer.py
Class Name:
Method Name: transform
Project Name: keras-team/autokeras
Commit Name: 074d16ac217b9cddc04d093d019c61e393522eb5
Time: 2017-12-30
Author: jhfjhfj1@gmail.com
File Name: autokeras/net_transformer.py
Class Name:
Method Name: transform