def get_input_fn(mode, data_files, meta_data_file):
return plx.processing.create_input_data_fn(
mode=mode,
pipeline_config=TFRecordImagePipelineConfig(
shuffle=plx.Modes.is_train(mode),
dynamic_pad=False,
batch_size=64 if plx.Modes.is_train(mode) else 32,
data_files=data_files,
meta_data_file=meta_data_file,
feature_processors=FeatureProcessorsConfig.from_dict(
{'image': {
'input_layers': [['image', 0, 0]],
'output_layers': [['std', 0, 0]],
'layers': [
{'Cast': {
'name': 'cast',
'dtype': 'float32',
'inbound_nodes': [['image', 0, 0]]
}},
{'Standardization': {
'name': 'std',
'inbound_nodes': [['cast', 0, 0]]
}},
]
}})
)
)
def get_pipeline_config(mode):
return TFRecordImagePipelineConfig(
dynamic_pad=False,
data_files=train_data_file
if Modes.is_train(mode) else eval_data_file,
meta_data_file=meta_data_filename,
feature_processors=FeatureProcessorsConfig({
'image':
GraphConfig(input_layers=[['image', 0, 0]],
output_layers=[['image_out', 0, 0]],
layers=[
CastConfig(dtype='float32',
name='image_out',
inbound_nodes=[['image', 0, 0]])
])
}))
def test_feature_processors(self):
config_dict = {
'image1': {
'input_layers': ['image'],
'output_layers': ['reshap_0'],
'layers': [{
'Resize': {
'height': 28,
'width': 28
}
}, {
'Reshape': {
'target_shape': [784]
}
}]
},
'image2': {
'input_layers': ['image'],
'output_layers': ['reshap_0'],
'layers': [{
'Standardization': {}
}, {
'Resize': {
'height': 28,
'width': 28
}
}, {
'Reshape': {
'target_shape': [784]
}
}]
}
}
config = FeatureProcessorsConfig.from_dict(config_dict)
config_to_dict = config.to_dict()
assert_equal_feature_processors(config_to_dict, config_dict)