reader.py 13.6 KB
Newer Older
S
sneaxiy 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14
# Copyright (c) 2019 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.

S
sneaxiy 已提交
15
from . import core
S
sneaxiy 已提交
16 17
import six
import threading
S
sneaxiy 已提交
18 19
from .framework import Program, Variable, program_guard, default_main_program, default_startup_program
from .executor import global_scope
S
sneaxiy 已提交
20
from .data_feeder import DataFeeder, BatchedTensorProvider
S
sneaxiy 已提交
21
from .layers.io import monkey_patch_reader_methods, _copy_reader_var_, double_buffer
S
sneaxiy 已提交
22
from .unique_name import UniqueNameGenerator
S
sneaxiy 已提交
23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

__all__ = ['PyReader']


def _convert_places(places):
    if not isinstance(places, (list, tuple)):
        places = [places]

    ret = []
    for p in places:
        if not isinstance(p, core.Place):
            tmp = core.Place()
            tmp.set_place(p)
            p = tmp

        ret.append(p)
    return ret


S
sneaxiy 已提交
42
class PyReader(object):
S
sneaxiy 已提交
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 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
    """
    Create a reader object for data feeding in Python. 
    Data would be prefetched using Python thread and be pushed
    into a queue asynchronously. Data in the queue would be extracted 
    automatically when `Executor.run(...)` is called.

    Args:  
        feed_list (list(Variable)|tuple(Variable)): feed variable list.
            The variables should be created by :code:`fluid.layers.data()`. 
        capacity (int): capacity of the queue maintained in PyReader object. 
        use_double_buffer (bool): whether to use double_buffer_reader to 
            speed up data feeding. 
        iterable (bool): whether the created reader object is iterable.   

    Returns:
        reader (Reader): the created reader object.

    Examples:
        1. If iterable = False, the created PyReader object is almost the
           same as :code:`fluid.layers.py_reader()`. Operators would be 
           inserted into the program. User should call :code:`start()` 
           before each epoch and catch :code:`fluid.core.EOFException`
           thrown by :code:`Executor.run()` when epoch ends. Once the 
           exception is caught, user should call :code:`reset()` to reset 
           the reader manually.

        .. code-block:: python
            
            image = fluid.layers.data(
                        name='image', shape=[784], dtype='float32')
            label = fluid.layers.data(
                        name='label', shape=[1], dtype='int64')
            
            reader = fluid.io.PyReader(feed_list=[image, label], 
                        capacity=4, iterable=False)
            reader.decorate_sample_list_generator(user_defined_reader)
            ... # definition of network is omitted
            executor.run(fluid.default_main_program())
            for _ in range(EPOCH_NUM):
                reader.start()
                while True:
                    try:
                        executor.run(feed=None, ...)
                    except fluid.core.EOFException:
                        reader.reset()
                        break
                    
        2. If iterable=True, the created PyReader object is decoupled with
           the program. No operator would be inserted into the program. 
           In this case, the created reader is a Python generator, which 
           is iterable. User should feed the data yielded from PyReader 
           object into :code:`Executor.run(feed=...)`.  

        .. code-block:: python

            image = fluid.layers.data(
                        name='image', shape=[784], dtype='float32')
            label = fluid.layers.data(
                        name='label', shape=[1], dtype='int64')

            reader = fluid.io.PyReader(feed_list=[image, label], 
                        capacity=4, iterable=True)
            reader.decorate_sample_list_generator(user_defined_reader, 
                        places=fluid.cuda_places())
            ... # definition of network is omitted
            executor.run(fluid.default_main_program())
            for _ in range(EPOCH_NUM):
                for data in reader():
                    executor.run(feed=data, ...)
    """

S
sneaxiy 已提交
114
    unique_name_generator = UniqueNameGenerator()
S
sneaxiy 已提交
115 116 117 118 119

    def __init__(self,
                 feed_list,
                 capacity,
                 use_double_buffer=True,
S
sneaxiy 已提交
120
                 iterable=False):
S
sneaxiy 已提交
121 122
        self._tensor_reader = None
        self._thread = None
S
sneaxiy 已提交
123 124 125
        self._iterable = iterable
        self._use_double_buffer = use_double_buffer
        self._capacity = capacity
S
sneaxiy 已提交
126
        self._feed_list = feed_list
S
sneaxiy 已提交
127 128
        if not self._iterable:
            self._init_non_iterable()
S
sneaxiy 已提交
129

S
sneaxiy 已提交
130 131
    def _init_iterable(self, places):
        self._var_names = [v.name for v in self._feed_list]
S
sneaxiy 已提交
132
        self._places = _convert_places(places)
S
sneaxiy 已提交
133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155
        self._queue = core.init_lod_tensor_blocking_queue(core.Variable(),
                                                          self._capacity)
        self._reader = core.create_py_reader(
            self.queue, self._var_names, self._places, self._use_double_buffer)

    def _init_non_iterable(self):
        lod_levels = []
        dtypes = []
        shape_concat = []
        ranks = []
        shapes = []

        for feed_data in self._feed_list:
            dtypes.append(feed_data.dtype)
            shape_concat.extend(feed_data.shape)
            ranks.append(len(feed_data.shape))
            shapes.append(feed_data.shape)
            lod_levels.append(feed_data.lod_level)

        queue_name = PyReader.unique_name_generator('lod_tensor_blocking_queue')
        reader_name = PyReader.unique_name_generator('create_py_reader')
        double_buffer_name = PyReader.unique_name_generator('double_buffer')

S
sneaxiy 已提交
156
        var = global_scope().var(queue_name)
S
sneaxiy 已提交
157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203
        self._queue = core.init_lod_tensor_blocking_queue(var, self._capacity)

        startup_blk = default_startup_program().current_block()
        startup_var = startup_blk.create_var(name=reader_name)

        startup_blk.append_op(
            type='create_py_reader',
            inputs={'blocking_queue': [queue_name]},
            outputs={'Out': [startup_var]},
            attrs={
                'shape_concat': shape_concat,
                'lod_levels': lod_levels,
                'ranks': ranks
            })

        startup_var.desc.set_dtypes(dtypes)
        startup_var.persistable = True

        main_prog_var = _copy_reader_var_(
            default_main_program().current_block(), startup_var)

        main_prog_var.stop_gradient = True
        main_prog_var.persistable = True

        reader = monkey_patch_reader_methods(main_prog_var)
        if self._use_double_buffer:
            double_buffer_reader = double_buffer(
                reader, name=double_buffer_name)
            # we return a double buffer reader. However, the reset method comes from
            # py_reader.
            double_buffer_reader.reset = reader.reset
            reader = double_buffer_reader

        self._reader = reader

        default_main_program().current_block().append_op(
            type='read',
            inputs={'Reader': [self._reader]},
            outputs={'Out': self._feed_list})

    @property
    def queue(self):
        return self._queue

    @property
    def iterable(self):
        return self._iterable
S
sneaxiy 已提交
204 205

    def __call__(self):
S
sneaxiy 已提交
206
        assert self.iterable, "PyReader is not iterable"
S
sneaxiy 已提交
207 208 209 210 211
        assert self._tensor_reader is not None, \
            "Data source of PyReader has not set yet"

        class Iterator(object):
            def __init__(self, reader):
S
sneaxiy 已提交
212 213
                self._reader = reader._reader
                self._reset = reader._reset
S
sneaxiy 已提交
214 215 216 217

            def __iter__(self):
                return self

S
sneaxiy 已提交
218 219 220
            def __next__(self):
                return self.next()

S
sneaxiy 已提交
221
            def next(self):
S
sneaxiy 已提交
222
                ret = self._reader.read_next()
S
sneaxiy 已提交
223
                if ret:
S
sneaxiy 已提交
224 225
                    return ret
                else:
S
sneaxiy 已提交
226
                    self._reset()
S
sneaxiy 已提交
227 228
                    raise StopIteration

S
sneaxiy 已提交
229
        self._start()
S
sneaxiy 已提交
230 231
        return Iterator(self)

S
sneaxiy 已提交
232
    def _reset(self):
S
sneaxiy 已提交
233 234 235 236
        self._reader.reset()
        self._thread.join()

    def start(self):
S
add doc  
sneaxiy 已提交
237 238 239 240
        '''
        Start the data feeding thread. 
        Can only call when the reader object is not iterable.  
        '''
S
sneaxiy 已提交
241 242
        assert not self._iterable, "start() cannot be called when PyReader is iterable"
        self._start()
S
sneaxiy 已提交
243

S
sneaxiy 已提交
244
    def reset(self):
S
add doc  
sneaxiy 已提交
245 246 247 248
        '''
        Reset the reader object when :code:`fluid.core.EOFException` raises. 
        Can only call when the reader object is not iterable.
        '''
S
sneaxiy 已提交
249 250 251 252
        assert not self._iterable, "reset() cannot be called when PyReader is iterable"
        self._reset()

    def _start(self):
S
sneaxiy 已提交
253
        def __thread_main__():
S
sneaxiy 已提交
254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271
            try:
                for tensors in self._tensor_reader():
                    array = core.LoDTensorArray()
                    for item in tensors:
                        if not isinstance(item, core.LoDTensor):
                            tmp = core.LoDTensor()
                            tmp.set(item, core.CPUPlace())
                            item = tmp

                        array.append(item)

                    if not self._queue.push(array):
                        break

                self._queue.close()
            except Exception as ex:
                self._queue.close()
                raise ex
S
sneaxiy 已提交
272 273 274 275 276

        self._thread = threading.Thread(target=__thread_main__)
        self._thread.daemon = True
        self._thread.start()

S
sneaxiy 已提交
277 278 279 280 281 282 283 284 285 286 287 288 289 290
    def decorate_sample_generator(self,
                                  sample_generator,
                                  batch_size,
                                  drop_last=True,
                                  places=None):
        '''
        Set the data source of the PyReader object.
        
        The provided :code:`sample_generator` should be a Python generator,
        which yields numpy.ndarray typed data of each sample.

        :code:`places` must be set when the PyReader object is iterable.

        If all inputs have no lods, this method is faster than 
S
sneaxiy 已提交
291
        :code:`decorate_sample_list_generator(paddle.batch(sample_generator, ...))` .
S
sneaxiy 已提交
292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309

        Args:
            sample_generator (generator): Python generator that yields
                numpy.ndarray-typed sample data.
            batch_size (int): batch size. Must be larger than 0.
            drop_last (bool): Whether to drop the last batch when sample number
                is less than batch_size. 
            places (None|list(CUDAPlace)|list(CPUPlace)): place list. Must
                be provided when PyReader is iterable.
        '''
        assert batch_size > 0, "batch_size must be larger than 0"
        has_lod = False
        for f in self._feed_list:
            if f.lod_level != 0:
                has_lod = True
                break

        if has_lod:
S
sneaxiy 已提交
310
            self.decorate_sample_list_generator(
S
sneaxiy 已提交
311 312 313 314 315 316 317 318 319 320 321 322
                paddle.batch(
                    sample_generator,
                    batch_size=batch_size,
                    drop_last=drop_last),
                places=places)
        else:
            reader = BatchedTensorProvider(
                feed_list=self._feed_list,
                place=core.CPUPlace(),
                batch_size=batch_size,
                generator=sample_generator,
                drop_last=drop_last)
S
sneaxiy 已提交
323
            self.decorate_batch_generator(reader, places=places)
S
sneaxiy 已提交
324

S
sneaxiy 已提交
325
    def decorate_sample_list_generator(self, reader, places=None):
S
add doc  
sneaxiy 已提交
326 327 328 329
        '''
        Set the data source of the PyReader object. 

        The provided :code:`reader` should be a Python generator,
S
sneaxiy 已提交
330
        which yields list(numpy.ndarray) typed batched data. 
S
add doc  
sneaxiy 已提交
331 332 333 334
        
        :code:`places` must be set when the PyReader object is iterable.

        Args:
S
sneaxiy 已提交
335 336 337 338
            reader (generator): Python generator that yields 
                list(numpy.ndarray)-typed batched data. 
            places (None|list(CUDAPlace)|list(CPUPlace)): place list. Must
                be provided when PyReader is iterable.
S
add doc  
sneaxiy 已提交
339
        '''
S
sneaxiy 已提交
340 341 342 343 344 345 346 347 348 349 350
        assert self._tensor_reader is None, \
            "Cannot reset the data source of PyReader"
        with program_guard(Program(), Program()):
            feeder = DataFeeder(
                feed_list=self._feed_list, place=core.CPUPlace())
            paddle_reader = feeder.decorate_reader(reader, multi_devices=False)

        def __tensor_reader_impl__():
            for slots in paddle_reader():
                yield [slots[var.name] for var in self._feed_list]

S
sneaxiy 已提交
351
        self.decorate_batch_generator(__tensor_reader_impl__, places)
S
sneaxiy 已提交
352

S
sneaxiy 已提交
353
    def decorate_batch_generator(self, reader, places=None):
S
add doc  
sneaxiy 已提交
354 355 356 357
        '''
        Set the data source of the PyReader object.

        The provided :code:`reader` should be a Python generator,
S
sneaxiy 已提交
358
        which yields numpy.ndarray-typed or LoDTensor-typed batched data.
S
add doc  
sneaxiy 已提交
359 360 361 362 363 364

        :code:`places` must be set when the PyReader object is iterable.

        Args:
            reader (generator): Python generator that yields LoDTensor-typed
                batched data.
S
sneaxiy 已提交
365
            places (None|list(CUDAPlace)|list(CPUPlace)): place list. Must
S
sneaxiy 已提交
366
                be provided when PyReader is iterable.
S
add doc  
sneaxiy 已提交
367
        '''
S
sneaxiy 已提交
368 369 370
        assert self._tensor_reader is None, \
            "Cannot reset the data source of PyReader"
        self._tensor_reader = reader
S
sneaxiy 已提交
371 372 373
        if self._iterable:
            assert places is not None, "Places cannot be None when py_reader is iterable"
            self._init_iterable(places)