layer.py 9.9 KB
Newer Older
Q
qiaolongfei 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13
# Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67
"""
Before this new package paddle.v2.layer, users would need to use functions
in paddle.trainer_config_helpers.layers to configure networks.

The Old Way:
=========
This old way requires that the creation of a network be defined in a Python
function, say network_config, and that this Python function being passed to
paddle.trainer_config_helpers.parse_network_config for the creation of
protobuf message description of this network.

```python
def network_config():
  img = paddle.trainer_config_helpers.data_layer(name="pixel", size=784)
  inference = paddle.trainer_config_helpers.fc_layer(
    input=img,
    size=10,
    act=paddle.trainer_config_helpers.SoftmaxActivation())
  cost = paddle.trainer_config_helpers.classification_cost(
    input=inference,
    label=paddle.trainer_config_helpers.data_layer(name="label", size=10))

proto_desc = parse_network_config(network_config)
```

When parse_network_config executes network_config, those layer definition
functions like data_layer and fc_layer would change some Python global variables,
so that after the execution, parse_network_config could collect information from
these global variables and generates the protobuf message.



The New Way:
=========
In this PR, we define a function in paddle.v2.layer which creates a Python
class for each layer creation function in paddle.trainer_config_helpers.layers.
Users can use create a network as follows:

```python
img = paddle.v2.layer.data(name="pixel", size=784)
inference = paddle.v2.layer.fc(input=img, size=10, act=paddle.v2.layer.Softmax())
cost = paddle.v2.layer.classification(
  input=inference,
  label=paddle.v2.layer.data(name="label", size=10))

parameters = paddle.v2.parameters.create(cost)
```

This new way doesn't require those invocations to layer definition functions
to be in a Python function but could be anywhere.

Also, the creation of a protobuf message is hidden in the invocation of
paddle.v2.parameters.create, no longer exposed to users.
"""
Q
qiaolongfei 已提交
68

Q
qiaolongfei 已提交
69 70
import collections

Q
qiaolongfei 已提交
71 72 73 74
import paddle.trainer_config_helpers as conf_helps
from paddle.trainer_config_helpers.config_parser_utils import \
    parse_network_config as __parse__
from paddle.trainer_config_helpers.default_decorators import wrap_name_default
Q
qiaolongfei 已提交
75 76

import data_type
Q
qiaolongfei 已提交
77

Q
qiaolongfei 已提交
78
__all__ = [
L
Luo Tao 已提交
79 80 81 82 83 84 85 86
    'parse_network', 'data', 'fc', 'conv_shift', 'img_conv', 'img_pool', 'spp',
    'maxout', 'img_cmrnorm', 'batch_norm', 'sum_to_one_norm', 'recurrent',
    'lstmemory', 'grumemory', 'pool', 'last_seq', 'first_seq', 'concat',
    'seq_concat', 'block_expand', 'expand', 'repeat', 'seq_reshape', 'addto',
    'linear_comb', 'interpolation', 'bilinear_interp', 'power', 'scaling',
    'slope_intercept', 'tensor', 'cos_sim', 'trans', 'max_id', 'sampling_id',
    'pad', 'classification_cost', 'cross_entropy_cost',
    'cross_entropy_with_selfnorm_cost', 'regression_cost',
L
Luo Tao 已提交
87
    'multi_binary_label_cross_entropy_cost', 'rank_cost', 'lambda_cost',
L
Luo Tao 已提交
88 89
    'sum_cost', 'huber_cost', 'crf', 'crf_decoding', 'ctc', 'warp_ctc', 'nce',
    'hsigmoid', 'eos'
Q
qiaolongfei 已提交
90 91
]

Q
qiaolongfei 已提交
92

Q
qiaolongfei 已提交
93 94 95 96 97 98 99 100 101 102 103 104 105 106 107
def parse_network(*outputs):
    """
    parse all output layers and then generate a model config proto.
    :param outputs:
    :return:
    """

    def __real_func__():
        context = dict()
        real_output = [each.to_proto(context=context) for each in outputs]
        conf_helps.outputs(real_output)

    return __parse__(__real_func__)


Q
qiaolongfei 已提交
108
class Layer(object):
Q
qiaolongfei 已提交
109 110
    def __init__(self, name, parent_layers):
        assert isinstance(parent_layers, dict)
Q
qiaolongfei 已提交
111 112
        assert isinstance(name, basestring)
        self.name = name
Q
qiaolongfei 已提交
113
        self.__parent_layers__ = parent_layers
Q
qiaolongfei 已提交
114 115 116 117 118 119

    def to_proto(self, context):
        """
        function to set proto attribute
        """
        kwargs = dict()
Q
qiaolongfei 已提交
120 121
        for layer_name in self.__parent_layers__:
            if not isinstance(self.__parent_layers__[layer_name],
Q
qiaolongfei 已提交
122
                              collections.Sequence):
Q
qiaolongfei 已提交
123
                v1_layer = self.__parent_layers__[layer_name].to_proto(
Q
qiaolongfei 已提交
124 125
                    context=context)
            else:
Q
qiaolongfei 已提交
126 127 128
                v1_layer = map(lambda x: x.to_proto(context=context),
                               self.__parent_layers__[layer_name])
            kwargs[layer_name] = v1_layer
Q
qiaolongfei 已提交
129 130 131 132 133 134 135 136 137

        if self.name not in context:
            context[self.name] = self.to_proto_impl(**kwargs)
        return context[self.name]

    def to_proto_impl(self, **kwargs):
        raise NotImplementedError()


L
Luo Tao 已提交
138 139
def __convert_to_v2__(method_name, parent_names):
    wrapper = wrap_name_default(name_prefix=method_name)
Q
qiaolongfei 已提交
140

Q
qiaolongfei 已提交
141
    class V2LayerImpl(Layer):
Q
qiaolongfei 已提交
142 143 144 145
        def __init__(self, name=None, **kwargs):
            parent_layers = dict()
            other_kwargs = dict()
            for pname in parent_names:
L
Luo Tao 已提交
146 147
                if kwargs.has_key(pname):
                    parent_layers[pname] = kwargs[pname]
Q
qiaolongfei 已提交
148 149 150 151 152

            for key in kwargs.keys():
                if key not in parent_names:
                    other_kwargs[key] = kwargs[key]

Q
qiaolongfei 已提交
153
            super(V2LayerImpl, self).__init__(name, parent_layers)
Q
qiaolongfei 已提交
154 155 156 157 158 159 160 161 162 163 164 165 166
            self.__other_kwargs__ = other_kwargs

        if wrapper is not None:
            __init__ = wrapper(__init__)

        def to_proto_impl(self, **kwargs):
            args = dict()
            for each in kwargs:
                args[each] = kwargs[each]
            for each in self.__other_kwargs__:
                args[each] = self.__other_kwargs__[each]
            return getattr(conf_helps, method_name)(name=self.name, **args)

Q
qiaolongfei 已提交
167
    return V2LayerImpl
Q
qiaolongfei 已提交
168 169


Q
qiaolongfei 已提交
170 171 172 173 174 175 176
"""
Some layer may need some special config, and can not use __convert_to_v2__ to convert.
So we also need to implement some special LayerV2.
"""


class DataLayerV2(Layer):
Q
qiaolongfei 已提交
177
    def __init__(self, name, type, **kwargs):
178
        assert isinstance(type, data_type.InputType)
Q
qiaolongfei 已提交
179

Q
qiaolongfei 已提交
180
        self.type = type
Q
qiaolongfei 已提交
181 182
        self.__method_name__ = 'data_layer'
        self.__kwargs__ = kwargs
Q
qiaolongfei 已提交
183 184 185 186 187

        super(DataLayerV2, self).__init__(name=name, parent_layers=dict())

    def to_proto_impl(self, **kwargs):
        args = dict()
Q
qiaolongfei 已提交
188
        args['size'] = self.type.dim
Q
qiaolongfei 已提交
189 190
        for each in kwargs:
            args[each] = kwargs[each]
Q
qiaolongfei 已提交
191 192
        for each in self.__kwargs__:
            args[each] = self.__kwargs__[each]
Q
qiaolongfei 已提交
193 194 195 196
        return getattr(conf_helps, self.__method_name__)(name=self.name, **args)


data = DataLayerV2
L
Luo Tao 已提交
197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276
AggregateLevel = conf_helps.layers.AggregateLevel
ExpandLevel = conf_helps.layers.ExpandLevel

layer_list = [
    # [V2LayerImpl, V1_method_name, parent_names]
    # fully connected layers
    ['fc', 'fc_layer', ['input']],
    # conv layers
    ['conv_shift', 'conv_shift_layer', ['a', 'b']],
    ['img_conv', 'img_conv_layer', ['input']],
    # image pooling layers
    ['img_pool', 'img_pool_layer', ['input']],
    ['spp', 'spp_layer', ['input']],
    ['maxout', 'maxout_layer', ['input']],
    # norm layers
    ['img_cmrnorm', 'img_cmrnorm_layer', ['input']],
    ['batch_norm', 'batch_norm_layer', ['input']],
    ['sum_to_one_norm', 'sum_to_one_norm_layer', ['input']],
    # recurrent layers
    ['recurrent', 'recurrent_layer', ['input']],
    ['lstmemory', 'lstmemory', ['input']],
    ['grumemory', 'grumemory', ['input']],
    # aggregate layers
    ['pool', 'pooling_layer', ['input']],
    ['last_seq', 'last_seq', ['input']],
    ['first_seq', 'first_seq', ['input']],
    ['concat', 'concat_layer', ['input']],
    ['seq_concat', 'seq_concat_layer', ['a', 'b']],
    # reshaping layers
    ['block_expand', 'block_expand_layer', ['input']],
    ['expand', 'expand_layer', ['input', 'expand_as']],
    ['repeat', 'repeat_layer', ['input']],
    ['rotate', 'rotate_layer', ['input']],
    ['seq_reshape', 'seq_reshape_layer', ['input']],
    # math layers
    ['addto', 'addto_layer', ['input']],
    ['linear_comb', 'linear_comb_layer', ['weights', 'vectors']],
    ['interpolation', 'interpolation_layer', ['input', 'weight']],
    ['bilinear_interp', 'bilinear_interp_layer', ['input']],
    ['power', 'power_layer', ['input', 'weight']],
    ['scaling', 'scaling_layer', ['input', 'weight']],
    ['slope_intercept', 'slope_intercept_layer', ['input']],
    ['tensor', 'tensor_layer', ['a', 'b']],
    ['cos_sim', 'cos_sim', ['a', 'b']],
    ['trans', 'trans_layer', ['input']],
    # sampling layers
    ['max_id', 'maxid_layer', ['input']],
    ['sampling_id', 'sampling_id_layer', ['input']],
    # slicing and joining layers
    ['pad', 'pad_layer', ['input']],
    # cost layers
    [
        'classification_cost', 'classification_cost',
        ['input', 'label', 'weight']
    ],
    ['regression_cost', 'regression_cost', ['input', 'label', 'weight']],
    ['cross_entropy_cost', 'cross_entropy', ['input', 'label']],
    [
        'cross_entropy_with_selfnorm_cost', 'cross_entropy_with_selfnorm',
        ['input', 'label']
    ],
    [
        'multi_binary_label_cross_entropy_cost',
        'multi_binary_label_cross_entropy', ['input', 'label']
    ],
    ['rank_cost', 'rank_cost', ['left', 'right', 'label', 'weight']],
    ['lambda_cost', 'lambda_cost', ['input', 'score']],
    ['sum_cost', 'sum_cost', ['input']],
    ['huber_cost', 'huber_cost', ['input', 'label']],
    ['crf', 'crf_layer', ['input', 'label']],
    ['crf_decoding', 'crf_decoding_layer', ['input']],
    ['ctc', 'ctc_layer', ['input', 'label']],
    ['warp_ctc', 'warp_ctc_layer', ['input', 'label']],
    ['nce', 'nce_layer', ['input', 'label']],
    ['hsigmoid', 'hsigmoid', ['input', 'label']],
    # check layers
    ['eos', 'eos_layer', ['input']]
]
for l in layer_list:
    globals()[l[0]] = __convert_to_v2__(l[1], l[2])