optimizer.py 19.9 KB
Newer Older
1
from collections import defaultdict
Q
Qiao Longfei 已提交
2

3 4 5 6 7 8
import framework
from backward import append_backward_ops
from framework import unique_name
from initializer import Constant
from layer_helper import LayerHelper
from regularizer import append_regularization_ops
9

10
__all__ = ['SGD', 'Momentum', 'Adagrad', 'Adam', 'Adamax', 'DecayedAdagrad']
Q
Qiao Longfei 已提交
11 12 13 14 15 16


class Optimizer(object):
    """Optimizer Base class.

    Define the common interface of an optimizer.
17 18
    User should not use this class directly,
    but need to use one of it's implementation.
Q
Qiao Longfei 已提交
19 20
    """

21 22
    def __init__(self, global_step=None):
        self._global_step = global_step
23 24 25 26 27
        # Dictionary of accumulators. Some optimizer subclasses need to
        # allocate and manage extra variables associated with the parameters
        # to train. These variables are called accumulators.
        # {accum_name : { paramter_name : accumulator_for_parameter, ...}, ...}
        self._accumulators = defaultdict(lambda: dict())
Q
Qiao Longfei 已提交
28
        self.helper = None
Q
Qiao Longfei 已提交
29 30 31 32 33 34

    def _append_optimize_op(self, block, param_and_grad):
        """ append optimize operator to block and return all the added optimize_op
        """
        raise NotImplementedError()

35 36 37 38 39 40 41 42 43 44 45 46 47
    def _create_param_lr(self, param_and_grad):
        # create learning rate variable for every parameter
        param = param_and_grad[0]
        param_lr = param.optimize_attr['learning_rate']
        param_lr_shape = [1]
        param_lr_var = self.helper.create_global_variable(
            name=unique_name("learning_rate"),
            dtype='float32',
            shape=param_lr_shape,
            lod_level=1,
            persistable=True)
        param_lr = param_lr * self._learning_rate
        self.helper.set_variable_initializer(
48
            var=param_lr_var, initializer=Constant(param_lr))
49
        return param_lr_var
50 51 52 53 54 55 56

    def _create_accumulators(self, block, parameters):
        """Create all accumulators needed by the parameters

        Args:
            block: the block in which the loss variable is present
            parameters: list of parameter variables for the optimizer
Q
Qiao Longfei 已提交
57
        """
58 59
        pass

60 61 62 63 64 65 66 67 68 69 70 71 72
    def _finish_update(self, block):
        """Finish any custom updates needed
           before completing an optimization step

        Args:
            block: the block in which the loss variable is present
            parameters: list of parameter variables for the optimizer

        Returns:
            list of finish ops or None
        """
        pass

Q
Qiao Longfei 已提交
73
    def _add_accumulator(self, name, param, dtype=None, fill_value=0.0):
74 75 76 77 78 79 80 81 82 83 84
        """Utility function to add an accumulator for a parameter

        Args:
            block: the block in which the loss variable is present
            name: name of the accumulator
            param: parameter variable for which accumulator is to be added
            dtype: data type of the accumulator variable
            fill_value: value to initialize the accumulator variable
        """
        if (name in self._accumulators and
                param.name in self._accumulators[name]):
85
            raise Exception("Accumulator {} already exists for parameter {}".
86
                            format(name, param.name))
Q
Qiao Longfei 已提交
87 88 89 90 91

        assert isinstance(self.helper, LayerHelper)
        var = self.helper.create_global_variable(
            name=unique_name(name),
            persistable=True,
F
fengjiayi 已提交
92
            dtype=dtype or param.dtype,
Q
Qiao Longfei 已提交
93 94 95
            type=param.type,
            shape=param.shape)
        self.helper.set_variable_initializer(
96
            var, initializer=Constant(value=float(fill_value)))
Q
Qiao Longfei 已提交
97
        self._accumulators[name][param.name] = var
98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114

    def _get_accumulator(self, name, param):
        """Utility function to fetch an accumulator for a parameter

        Args:
            name: name of the accumulator
            param: parameter variable for which accumulator is to be fetched

        Returns:
            accumulator variable for the parameter
        """
        if (name not in self._accumulators or
                param.name not in self._accumulators[name]):
            raise Exception("Accumulator {} does not exist for parameter {}".
                            format(name, param.name))
        return self._accumulators[name][param.name]

115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134
    def _increment_global_step(self, block):
        """Increment the global step by 1 after every iteration

        Args:
            block: the block in which the loss variable is present

        Returns:
            list with global_step increment op as its only element
        """
        assert isinstance(block, framework.Block)
        assert self._global_step is not None
        # create the increment op
        increment_op = block.append_op(
            type="increment",
            inputs={"X": self._global_step},
            outputs={"Out": self._global_step},
            attrs={"step": 1.0})

        return increment_op

Q
Qiao Longfei 已提交
135 136 137
    def create_optimization_pass(self,
                                 parameters_and_grads,
                                 loss,
138
                                 startup_program=None):
Q
Qiao Longfei 已提交
139 140 141 142 143 144 145
        """Add optimization operators to update gradients to variables.

        Args:
          loss: the target that this optimization is for.
          parameters_and_grads: a list of (variable, gradient) pair to update.

        Returns:
146 147 148 149
          return_op_list: a list of operators that will complete one step of
          optimization. This will include parameter update ops, global step
          update ops and any other custom ops required by subclasses to manage
          their internal state.
150
          :param startup_program: 
Q
Qiao Longfei 已提交
151
        """
152 153 154 155 156
        # This is a default implementation of create_optimization_pass that
        # can be shared by most optimizers. This implementation assumes that
        # the subclass will implement the _append_optimize_op method and the
        #  _initialize_tensors method. The subclass can extend the
        # _create_accumulators method if it needs to create accumulators
157
        # for parameters and extend _finish_update method to add custom ops.
158 159

        # Create any accumulators
Q
Qiao Longfei 已提交
160 161
        program = loss.block.program
        self.helper = LayerHelper(
162 163 164
            self.__class__.__name__,
            main_program=program,
            startup_program=startup_program)
165 166 167
        self._create_accumulators(loss.block,
                                  [p[0] for p in parameters_and_grads])

Q
Qiao Longfei 已提交
168 169
        optimize_ops = []
        for param_and_grad in parameters_and_grads:
Q
Qiao Longfei 已提交
170 171
            if param_and_grad[0].trainable is True and param_and_grad[
                    1] is not None:
Q
Qiao Longfei 已提交
172 173 174
                optimize_op = self._append_optimize_op(loss.block,
                                                       param_and_grad)
                optimize_ops.append(optimize_op)
175

176 177 178 179 180 181 182 183 184 185
        # Returned list of ops can include more ops in addition
        # to optimization ops
        return_ops = optimize_ops

        # Get custom finish ops for subclasses
        # FIXME: Need to fix this once we figure out how to handle dependencies
        finish_ops = self._finish_update(loss.block)
        if finish_ops is not None:
            return_ops += finish_ops

186 187
        if self._global_step is not None:
            return_ops.append(self._increment_global_step(loss.block))
188
        return return_ops
Q
Qiao Longfei 已提交
189

Q
Qiao Longfei 已提交
190 191
    def minimize(self,
                 loss,
192
                 startup_program=None,
Q
Qiao Longfei 已提交
193 194
                 parameter_list=None,
                 no_grad_set=None):
Q
Qiao Longfei 已提交
195 196
        """Add operations to minimize `loss` by updating `parameter_list`.

197
        This method combines interface `append_backward_ops()` and
Q
Qiao Longfei 已提交
198 199
        `create_optimization_pass()` into one.
        """
200 201
        params_grads = append_backward_ops(loss, parameter_list, no_grad_set or
                                           set())
F
fengjiayi 已提交
202
        # Add regularization if any
203
        params_grads = append_regularization_ops(params_grads)
Q
Qiao Longfei 已提交
204
        optimize_ops = self.create_optimization_pass(params_grads, loss,
205
                                                     startup_program)
Q
Qiao Longfei 已提交
206 207 208 209 210 211 212
        return optimize_ops


class SGDOptimizer(Optimizer):
    """ Simple SGD optimizer without any state.
    """

213
    def __init__(self, learning_rate, global_step=None):
Q
Qiao Longfei 已提交
214
        assert learning_rate is not None
215
        super(SGDOptimizer, self).__init__(global_step)
Q
Qiao Longfei 已提交
216 217 218
        self.type = "sgd"
        self._learning_rate = learning_rate

219 220
    def _append_optimize_op(self, block, param_and_grad):
        assert isinstance(block, framework.Block)
221

Q
Qiao Longfei 已提交
222 223 224 225 226 227
        # create the optimize op
        sgd_op = block.append_op(
            type=self.type,
            inputs={
                "Param": param_and_grad[0],
                "Grad": param_and_grad[1],
228
                "LearningRate": self._create_param_lr(param_and_grad)
Q
Qiao Longfei 已提交
229
            },
230
            outputs={"ParamOut": param_and_grad[0]})
Q
Qiao Longfei 已提交
231 232

        return sgd_op
233 234 235 236 237 238 239


class MomentumOptimizer(Optimizer):
    """Simple Momentum optimizer with velocity state
    """
    _velocity_acc_str = "velocity"

240 241 242 243 244
    def __init__(self,
                 learning_rate,
                 momentum,
                 use_nesterov=False,
                 global_step=None):
245 246
        assert learning_rate is not None
        assert momentum is not None
247
        super(MomentumOptimizer, self).__init__(global_step)
248 249 250
        self.type = "momentum"
        self._learning_rate = learning_rate
        self._momentum = momentum
251
        self._use_nesterov = bool(use_nesterov)
252 253 254 255 256

    def _create_accumulators(self, block, parameters):
        assert isinstance(block, framework.Block)

        for p in parameters:
Q
Qiao Longfei 已提交
257
            self._add_accumulator(self._velocity_acc_str, p)
258 259 260 261 262 263 264 265 266 267 268 269 270

    def _append_optimize_op(self, block, param_and_grad):
        assert isinstance(block, framework.Block)

        velocity_acc = self._get_accumulator(self._velocity_acc_str,
                                             param_and_grad[0])
        # create the momentum optimize op
        momentum_op = block.append_op(
            type=self.type,
            inputs={
                "Param": param_and_grad[0],
                "Grad": param_and_grad[1],
                "Velocity": velocity_acc,
271
                "LearningRate": self._create_param_lr(param_and_grad)
272 273 274 275 276
            },
            outputs={
                "ParamOut": param_and_grad[0],
                "VelocityOut": velocity_acc
            },
277
            attrs={"mu": self._momentum,
278
                   "use_nesterov": self._use_nesterov})
279 280

        return momentum_op
281 282 283 284 285 286 287


class AdagradOptimizer(Optimizer):
    """Simple Adagrad optimizer with moment state
    """
    _moment_acc_str = "moment"

288
    def __init__(self, learning_rate, epsilon=1.0e-6, global_step=None):
289 290
        assert learning_rate is not None
        assert epsilon is not None
291
        super(AdagradOptimizer, self).__init__(global_step)
292 293 294 295 296 297 298 299
        self.type = "adagrad"
        self._learning_rate = learning_rate
        self._epsilon = epsilon

    def _create_accumulators(self, block, parameters):
        assert isinstance(block, framework.Block)

        for p in parameters:
Q
Qiao Longfei 已提交
300
            self._add_accumulator(self._moment_acc_str, p)
301 302 303 304 305 306 307

    def _append_optimize_op(self, block, param_and_grad):
        assert isinstance(block, framework.Block)

        moment_acc = self._get_accumulator(self._moment_acc_str,
                                           param_and_grad[0])

308
        # Create the adagrad optimizer op
309 310 311 312 313 314
        adagrad_op = block.append_op(
            type=self.type,
            inputs={
                "Param": param_and_grad[0],
                "Grad": param_and_grad[1],
                "Moment": moment_acc,
315
                "LearningRate": self._create_param_lr(param_and_grad)
316 317 318 319 320 321
            },
            outputs={"ParamOut": param_and_grad[0],
                     "MomentOut": moment_acc},
            attrs={"epsilon": self._epsilon})

        return adagrad_op
322 323 324 325 326 327 328 329 330 331 332 333


class AdamOptimizer(Optimizer):
    """Implements the Adam Optimizer
    """
    _moment1_acc_str = "moment1"
    _moment2_acc_str = "moment2"

    def __init__(self,
                 learning_rate=0.001,
                 beta1=0.9,
                 beta2=0.999,
334 335
                 epsilon=1e-8,
                 global_step=None):
336 337 338 339
        assert learning_rate is not None
        assert beta1 is not None
        assert beta2 is not None
        assert epsilon is not None
340
        super(AdamOptimizer, self).__init__(global_step)
341 342 343 344 345 346 347 348 349
        self.type = "adam"
        self._learning_rate = learning_rate
        self._beta1 = beta1
        self._beta2 = beta2
        self._epsilon = epsilon

    def _create_accumulators(self, block, parameters):
        assert isinstance(block, framework.Block)

Q
Qiao Longfei 已提交
350
        main_block = block.program.global_block()
351 352
        # Create beta1 and beta2 power tensors
        beta_shape = [1]
Q
Qiao Longfei 已提交
353 354 355 356 357 358 359
        self._beta1_pow_acc = self.helper.create_global_variable(
            name=unique_name('beta1_pow_acc'),
            dtype='float32',
            shape=beta_shape,
            lod_level=0,
            persistable=True)
        self.helper.set_variable_initializer(
360
            self._beta1_pow_acc, initializer=Constant(self._beta1))
Q
Qiao Longfei 已提交
361 362 363 364 365 366 367 368 369

        self._beta2_pow_acc = self.helper.create_global_variable(
            name=unique_name('beta2_pow_acc'),
            dtype='float32',
            shape=beta_shape,
            lod_level=0,
            persistable=True)

        self.helper.set_variable_initializer(
370
            self._beta2_pow_acc, initializer=Constant(self._beta2))
371 372 373

        # Create accumulator tensors for first and second moments
        for p in parameters:
Q
Qiao Longfei 已提交
374 375
            self._add_accumulator(self._moment1_acc_str, p)
            self._add_accumulator(self._moment2_acc_str, p)
376 377 378 379 380 381 382 383

    def _append_optimize_op(self, block, param_and_grad):
        assert isinstance(block, framework.Block)

        moment1 = self._get_accumulator(self._moment1_acc_str,
                                        param_and_grad[0])
        moment2 = self._get_accumulator(self._moment2_acc_str,
                                        param_and_grad[0])
384
        # create the adam optimize op
385 386 387 388 389
        adam_op = block.append_op(
            type=self.type,
            inputs={
                "Param": param_and_grad[0],
                "Grad": param_and_grad[1],
390
                "LearningRate": self._create_param_lr(param_and_grad),
391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412
                "Moment1": moment1,
                "Moment2": moment2,
                "Beta1Pow": self._beta1_pow_acc,
                "Beta2Pow": self._beta2_pow_acc
            },
            outputs={
                "ParamOut": param_and_grad[0],
                "Moment1Out": moment1,
                "Moment2Out": moment2
            },
            attrs={
                "beta1": self._beta1,
                "beta2": self._beta2,
                "epsilon": self._epsilon
            })

        return adam_op

    def _finish_update(self, block):
        """Update Beta1 and Beta2 Power accumulators
        """
        assert isinstance(block, framework.Block)
Q
Qiao Longfei 已提交
413 414
        main_block = block.program.global_block()
        scale_beta1 = main_block.append_op(
415 416 417 418 419
            type="scale",
            inputs={"X": self._beta1_pow_acc},
            outputs={"Out": self._beta1_pow_acc},
            attrs={"scale": self._beta1})

Q
Qiao Longfei 已提交
420
        scale_beta2 = main_block.append_op(
421 422 423 424 425 426
            type="scale",
            inputs={"X": self._beta2_pow_acc},
            outputs={"Out": self._beta2_pow_acc},
            attrs={"scale": self._beta2})

        return [scale_beta1, scale_beta2]
427 428 429 430 431 432 433 434 435 436 437 438


class AdamaxOptimizer(Optimizer):
    """Implements the Adamax Optimizer
    """
    _moment_acc_str = "moment"
    _inf_norm_acc_str = "inf_norm"

    def __init__(self,
                 learning_rate=0.001,
                 beta1=0.9,
                 beta2=0.999,
439 440
                 epsilon=1e-8,
                 global_step=None):
441 442 443 444 445 446 447 448 449 450 451 452 453 454
        assert learning_rate is not None
        assert beta1 is not None
        assert beta2 is not None
        assert epsilon is not None
        super(AdamaxOptimizer, self).__init__()
        self.type = "adamax"
        self._learning_rate = learning_rate
        self._beta1 = beta1
        self._beta2 = beta2
        self._epsilon = epsilon

    def _create_accumulators(self, block, parameters):
        # Create beta1 power accumulator tensor
        beta_shape = [1]
Q
Qiao Longfei 已提交
455 456 457 458 459 460 461
        self._beta1_pow_acc = self.helper.create_global_variable(
            name=unique_name('beta1_pow_acc'),
            dtype='float32',
            shape=beta_shape,
            lod_level=0,
            persistable=True)
        self.helper.set_variable_initializer(
462
            self._beta1_pow_acc, initializer=Constant(self._beta1))
463 464 465

        # Create accumulator tensors for first moment and infinity norm
        for p in parameters:
Q
Qiao Longfei 已提交
466 467
            self._add_accumulator(self._moment_acc_str, p)
            self._add_accumulator(self._inf_norm_acc_str, p)
468 469 470 471 472 473 474 475 476 477 478 479 480

    def _append_optimize_op(self, block, param_and_grad):
        assert isinstance(block, framework.Block)

        moment = self._get_accumulator(self._moment_acc_str, param_and_grad[0])
        inf_norm = self._get_accumulator(self._inf_norm_acc_str,
                                         param_and_grad[0])
        # create the adamax optimize op
        adamax_op = block.append_op(
            type=self.type,
            inputs={
                "Param": param_and_grad[0],
                "Grad": param_and_grad[1],
481
                "LearningRate": self._create_param_lr(param_and_grad),
482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502
                "Moment": moment,
                "InfNorm": inf_norm,
                "Beta1Pow": self._beta1_pow_acc
            },
            outputs={
                "ParamOut": param_and_grad[0],
                "MomentOut": moment,
                "InfNormOut": inf_norm
            },
            attrs={
                "beta1": self._beta1,
                "beta2": self._beta2,
                "epsilon": self._epsilon
            })

        return adamax_op

    def _finish_update(self, block):
        """Update Beta1 Power accumulator
        """
        assert isinstance(block, framework.Block)
Q
Qiao Longfei 已提交
503 504
        main_block = block.program.global_block()
        scale_beta1 = main_block.append_op(
505 506 507 508 509 510
            type="scale",
            inputs={"X": self._beta1_pow_acc},
            outputs={"Out": self._beta1_pow_acc},
            attrs={"scale": self._beta1})

        return [scale_beta1]
511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558


class DecayedAdagradOptimizer(Optimizer):
    """Simple Decayed Adagrad optimizer with moment state
    """
    _moment_acc_str = "moment"

    def __init__(self,
                 learning_rate,
                 decay=0.95,
                 epsilon=1.0e-6,
                 global_step=None):
        assert learning_rate is not None
        assert decay is not None
        assert epsilon is not None

        super(DecayedAdagradOptimizer, self).__init__(global_step)
        self.type = "decayed_adagrad"
        self._learning_rate = learning_rate
        self._decay = decay
        self._epsilon = epsilon

    def _create_accumulators(self, block, parameters):
        assert isinstance(block, framework.Block)

        for p in parameters:
            self._add_accumulator(self._moment_acc_str, p)

    def _append_optimize_op(self, block, param_and_grad):
        assert isinstance(block, framework.Block)

        moment_acc = self._get_accumulator(self._moment_acc_str,
                                           param_and_grad[0])

        # Create the decayed adagrad optimizer op
        decayed_adagrad_op = block.append_op(
            type=self.type,
            inputs={
                "Param": param_and_grad[0],
                "Grad": param_and_grad[1],
                "Moment": moment_acc,
                "LearningRate": self._create_param_lr(param_and_grad)
            },
            outputs={"ParamOut": param_and_grad[0],
                     "MomentOut": moment_acc},
            attrs={"epsilon": self._epsilon})

        return decayed_adagrad_op
559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574


# We short the class name, since users will use the optimizer with the package
# name. The sample code:
#
# import paddle.fluid as fluid
#
# sgd = fluid.optimizer.SGD(...)
#
# It is no need to add an `Optimizer` as the class suffix
SGD = SGDOptimizer
Momentum = MomentumOptimizer
Adagrad = AdagradOptimizer
Adam = AdamOptimizer
Adamax = AdamaxOptimizer
DecayedAdagrad = DecayedAdagradOptimizer