-
Notifications
You must be signed in to change notification settings - Fork 5.5k
/
layers.py
2323 lines (1873 loc) · 83.2 KB
/
layers.py
1
2
3
4
5
6
7
8
9
10
11
12
13
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
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
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
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
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
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
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
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
# Copyright (c) 2018 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.
import collections
import copy
import inspect
import re
import warnings
import weakref
import numpy as np
import paddle
from paddle import nn, profiler
from paddle.fluid import core, framework, unique_name
from paddle.fluid.core import VarDesc
from paddle.fluid.dygraph import no_grad
from paddle.fluid.dygraph.base import (
_convert_into_variable,
in_declarative_mode,
program_desc_tracing_guard,
)
from paddle.fluid.dygraph_utils import _append_activation_in_dygraph
from paddle.fluid.executor import Executor, global_scope
from paddle.fluid.framework import Parameter, Program
from paddle.fluid.framework import _current_expected_place as _get_device
from paddle.fluid.framework import (
_global_flags,
convert_np_dtype_to_dtype_,
default_main_program,
in_dygraph_mode,
)
from paddle.fluid.layer_helper_base import LayerHelperBase
from paddle.fluid.param_attr import ParamAttr
from paddle.profiler.utils import in_profiler_mode
from paddle.utils import deprecated
__all__ = []
_first_cap_re = re.compile('(.)([A-Z][a-z]+)')
_all_cap_re = re.compile('([a-z])([A-Z])')
def record_program_ops_pre_hook(layer, inputs):
"""
A pre-hook to mark op numbers before enter layer.forward.
"""
if not in_dygraph_mode():
if layer._op_recorder.start < 0:
layer._op_recorder.start = len(
default_main_program().current_block().ops
)
layer._op_recorder.is_valid = True
else:
layer._op_recorder.is_valid = False
warnings.warn(
"{} has recorded the op information before. Please check whether you call this layer twice.".format(
layer._full_name
)
)
return None
def set_op_customized_attrs_post_hook(layer, inputs, outputs):
"""
A post-hook to append customized attributes into all operators generated in current layer.
"""
if not in_dygraph_mode() and layer._op_recorder.is_valid:
start = layer._op_recorder.start
end = len(default_main_program().current_block().ops)
assert start >= 0 and end >= start
ops = default_main_program().current_block().ops[start:end]
layer._op_recorder.end = end
layer._op_recorder.ops = ops
for op in ops:
for attr_name, val in layer._customized_attrs.items():
op._set_attr(attr_name, val)
# remove pre-hook and post-hook
for hook_helper in layer._op_recorder.hooks:
hook_helper.remove()
return None
def _scope_dist2single(dist_scope):
mapping = {
"row_parallel_linear": "linear",
"column_parallel_linear": "linear",
"vocab_parallel_embedding": "embedding",
# "parallel_cross_entropy": "cross_entropy", while mp_layer has parallel_cross_entropy,
# but there is no parameters so the mapping of parallel_cross_entropy is not necessary.
}
return mapping.get(dist_scope, dist_scope)
def _convert_camel_to_snake(name):
s1 = _first_cap_re.sub(r'\1_\2', name)
return _all_cap_re.sub(r'\1_\2', s1).lower()
def _addindent(string, indent):
s1 = string.split('\n')
if len(s1) == 1:
return string
s2 = []
for idx, line in enumerate(s1):
if idx > 0:
s2.append(str((indent * ' ') + line))
return s1[0] + '\n' + '\n'.join(s2)
def _layer_trans_dtype(layer, dtype, excluded_layers):
if type(layer) in excluded_layers:
return
layer._to_impl(dtype=dtype, floating_only=True, include_sublayers=False)
class LayerObjectHelper(LayerHelperBase):
def __init__(self, name):
super().__init__(name, layer_type=name)
def append_op(
self,
type=None,
inputs=None,
outputs=None,
attrs=None,
stop_gradient=None,
):
"""append an operator for this layer object.
Args:
type: operator type
inputs: input variable of the operator
dtype: data type of this parameter
is_bias: if this is a bias parameter
default_initializer: set the default initializer for this parameter
Returns created parameter Variable.
"""
return self.main_program.current_block().append_op(
type=type,
inputs=inputs,
outputs=outputs,
attrs=attrs,
stop_gradient=stop_gradient,
)
def _multiple_input(self, inputs_in):
inputs = inputs_in
ret = []
if isinstance(inputs, (list, tuple)):
for inp in inputs:
ret.append(self.to_variable(inp))
else:
ret.append(self.to_variable(inputs))
return ret
# TODO: make it public when we need it
def _input(self, inputs_in):
inputs = self._multiple_input(inputs_in)
if len(inputs) != 1:
raise f"{self.layer_type} layer only takes one input in"
return inputs[0]
def _multiple_param_attr(self, length, param_attr_in=None):
param_attr = param_attr_in
if isinstance(param_attr, ParamAttr):
param_attr = [param_attr]
if len(param_attr) != 1 and len(param_attr) != length:
raise ValueError(f"parameter number mismatch in {self.name}")
elif len(param_attr) == 1 and length != 1:
tmp = [None] * length
for i in range(length):
tmp[i] = copy.deepcopy(param_attr[0])
param_attr = tmp
return param_attr
def iter_inputs_and_params(self, inputs_in, param_attr_in=None):
"""Access all inputs and params one by one
Args:
inputs_in: inputs to be iter
param_attr_in: param_attr to be iter
Returns input, param_attr
"""
param_attr_in = ParamAttr._to_attr(param_attr_in)
if isinstance(param_attr_in, bool):
raise ValueError(f'Param_attr should not be False in {self.name}')
inputs = inputs_in if (inputs_in is not None) else []
inputs = self._multiple_input(inputs)
param_attrs = self._multiple_param_attr(len(inputs), param_attr_in)
yield from zip(inputs, param_attrs)
def input_dtype(self, inputs_in):
"""Get input data type
Args:
inputs_in: inputs wanted know the data type
Returns dtype of the input
"""
inputs_in = inputs_in if (inputs_in is not None) else []
inputs = self._multiple_input(inputs_in)
dtype = None
for each in inputs:
if dtype is None:
dtype = each.dtype
elif dtype != each.dtype:
raise ValueError(
"Data Type mismatch: %d to %d in %s"
% (dtype, each.dtype, self.name)
)
return dtype
def get_parameter(self, name):
"""Get parameter specifically
Args:
name: parameter's name
Returns target parameter
"""
param = self.main_program.global_block().var(name)
if not isinstance(param, Parameter):
raise ValueError(f"no Parameter name {name} found in {self.name}")
return param
# TODO: this should not be called anymore after all activation func move to Layers
def append_activation(self, input_var, act=None, use_cudnn=None):
"""Append activation
Args:
input_var: the input variable. The len(input_var.shape) is
larger or equal than 2.
act: activation type
use_cudnn: if use cudnn
Return the Variable of after append activation
"""
act = act
if act is None:
return input_var
if isinstance(act, str):
act = {'type': act}
else:
raise TypeError(
str(act) + " should be unicode or str in %s ", self.name
)
if (use_cudnn is not None) and use_cudnn:
act['use_cudnn'] = use_cudnn
use_mkldnn = _global_flags()["FLAGS_use_mkldnn"]
if (use_mkldnn is not None) and use_mkldnn:
act['use_mkldnn'] = use_mkldnn
act_type = act.pop('type')
if in_dygraph_mode():
res = _append_activation_in_dygraph(
input_var, act_type, use_cudnn, use_mkldnn
)
return res
else:
tmp = self.create_variable_for_type_inference(dtype=input_var.dtype)
self.append_op(
type=act_type,
inputs={"X": [input_var]},
outputs={"Out": [tmp]},
attrs=act,
)
return tmp
def is_instance(self, param, cls):
"""Check if the input parameter is instance of input class
Args:
param: parameter to be check
cls: class of the parameter
Return result of the check (True or False)
"""
param = param
if not isinstance(param, cls):
raise TypeError(
"The input {0} parameter of method {1} must be {2}, in layer {3}",
param,
self.layer_type,
cls.__name__,
self.name,
)
class LayerOpsRecoder:
"""
Record generated operators information in nn.Layer.
"""
def __init__(self, start=-1, end=-1, ops=None, is_valid=False, hooks=None):
self.start = start
self.end = end
self.ops = ops
self.is_valid = is_valid
self.hooks = hooks
class HookRemoveHelper:
"""A HookRemoveHelper that can be used to remove hook."""
next_hook_id = 0
def __init__(self, hooks):
self._hooks_ref = weakref.ref(hooks)
self._hook_id = HookRemoveHelper.next_hook_id
HookRemoveHelper.next_hook_id += 1
def remove(self):
hooks = self._hooks_ref()
if hooks is not None and self._hook_id in hooks:
del hooks[self._hook_id]
class Layer:
"""
Dynamic graph Layer based on OOD, includes the parameters of the layer, the structure of the forward graph and so on.
Parameters:
name_scope (str, optional): prefix name used by the layer to name parameters.
If prefix is "my_layer", parameter name in MyLayer
can be "my_layer_0.w_n", where "w" is the parameter
base name and "n" is an unique suffix auto-generated.
If None, prefix name will be snake cased class name. Default: None.
dtype(str, optional): data type of this parameter.
If set str, it can be "bool", "float16", "float32", "float64",
"int8", "int16", "int32", "int64", "uint8" or "uint16".
Default: "float32"
Returns:
None
Examples:
.. code-block:: python
import paddle
class MyLayer(paddle.nn.Layer):
def __init__(self):
super().__init__()
self._linear = paddle.nn.Linear(1, 1)
self._dropout = paddle.nn.Dropout(p=0.5)
def forward(self, input):
temp = self._linear(input)
temp = self._dropout(temp)
return temp
x = paddle.randn([10, 1], 'float32')
mylayer = MyLayer()
mylayer.eval() # set mylayer._dropout to eval mode
out = mylayer(x)
mylayer.train() # set mylayer._dropout to train mode
out = mylayer(x)
"""
def __init__(self, name_scope=None, dtype="float32"):
self.training = True
if name_scope is None:
name_scope = _convert_camel_to_snake(self.__class__.__name__)
name_scope = _scope_dist2single(name_scope)
self._full_name = unique_name.generate(name_scope)
self._helper = LayerObjectHelper(self._full_name)
self._built = False
self._dtype = dtype
self._init_in_dynamic_mode = in_dygraph_mode()
self._parameters = collections.OrderedDict()
# Buffers the variable (not parameter) created in layer
self._buffers = collections.OrderedDict()
self._non_persistable_buffer_names_set = set()
self._sub_layers = collections.OrderedDict()
self._loaddict_holder = collections.OrderedDict()
# Record generated op_descs in this layer
self._op_recorder = LayerOpsRecoder(ops=[], hooks=[])
self._customized_attrs = {}
self._forward_pre_hooks = collections.OrderedDict()
self._forward_post_hooks = collections.OrderedDict()
# only used in AMP Training
self._cast_to_low_precison = True
self._state_dict_hooks = collections.OrderedDict()
# Records orignal functions after @to_static to support to rollback
self._original_funcs = collections.OrderedDict()
def train(self):
"""
Sets this Layer and all its sublayers to training mode.
This only effects certain modules like `Dropout` and `BatchNorm`.
Returns:
None
Examples:
.. code-block:: python
import paddle
class MyLayer(paddle.nn.Layer):
def __init__(self):
super().__init__()
self._linear = paddle.nn.Linear(1, 1)
self._dropout = paddle.nn.Dropout(p=0.5)
def forward(self, input):
temp = self._linear(input)
temp = self._dropout(temp)
return temp
x = paddle.randn([10, 1], 'float32')
mylayer = MyLayer()
mylayer.eval() # set mylayer._dropout to eval mode
out = mylayer(x)
mylayer.train() # set mylayer._dropout to train mode
out = mylayer(x)
"""
# global setting in dygraph
# NOTE(chenweihang): nn.Layer also can be used in static mode,
# but _dygraph_tracer() can not be called in static mode
if in_dygraph_mode():
framework._dygraph_tracer().train_mode()
# Layer-level setting
self.training = True
for layer in self.sublayers():
layer.training = True
def eval(self):
"""
Sets this Layer and all its sublayers to evaluation mode.
This only effects certain modules like `Dropout` and `BatchNorm`.
Returns:
None
Example::
.. code-block:: python
import paddle
class MyLayer(paddle.nn.Layer):
def __init__(self):
super().__init__()
self._linear = paddle.nn.Linear(1, 1)
self._dropout = paddle.nn.Dropout(p=0.5)
def forward(self, input):
temp = self._linear(input)
temp = self._dropout(temp)
return temp
x = paddle.randn([10, 1], 'float32')
mylayer = MyLayer()
mylayer.eval() # set mylayer._dropout to eval mode
out = mylayer(x)
print(out)
"""
# global setting in dygraph
# NOTE(chenweihang): nn.Layer also can be used in static mode,
# but _dygraph_tracer() can not be called in static mode
if in_dygraph_mode():
framework._dygraph_tracer().eval_mode()
# Layer-level setting
self.training = False
for layer in self.sublayers():
layer.training = False
def apply(self, fn):
"""
Applies ``fn`` recursively to every sublayer (as returned by ``.sublayers()``)
as well as self. Typical use includes initializing the parameters of a model.
Parameters:
fn (function): a function to be applied to each sublayer
Returns:
Layer, self
Example::
.. code-block:: python
import paddle
import paddle.nn as nn
net = nn.Sequential(nn.Linear(2, 2), nn.Linear(2, 2))
def init_weights(layer):
if type(layer) == nn.Linear:
print('before init weight:', layer.weight.numpy())
new_weight = paddle.full(shape=layer.weight.shape, dtype=layer.weight.dtype, fill_value=0.9)
layer.weight.set_value(new_weight)
print('after init weight:', layer.weight.numpy())
net.apply(init_weights)
print(net.state_dict())
"""
for layer in self.children():
layer.apply(fn)
fn(self)
return self
def full_name(self):
"""
Full name for this layer, composed by name_scope + "/" + MyLayer.__class__.__name__
Returns:
str, full name of this layer.
Example::
.. code-block:: python
import paddle
class LinearNet(paddle.nn.Layer):
def __init__(self):
super().__init__(name_scope = "demo_linear_net")
self._linear = paddle.nn.Linear(1, 1)
def forward(self, x):
return self._linear(x)
linear_net = LinearNet()
print(linear_net.full_name()) # demo_linear_net_0
"""
return self._full_name
def register_forward_post_hook(self, hook):
"""
Register a forward post-hook for Layer. The hook will be called after `forward` function has been computed.
It should have the following form, `input` and `output` of the `hook` is `input` and `output` of the `Layer` respectively.
User can use forward post-hook to change the output of the Layer or perform information statistics tasks on the Layer.
hook(Layer, input, output) -> None or modified output
Parameters:
hook(function): a function registered as a forward post-hook
Returns:
HookRemoveHelper, a HookRemoveHelper object that can be used to remove the added hook by calling `hook_remove_helper.remove()` .
Examples:
.. code-block:: python
import paddle
import numpy as np
# the forward_post_hook change the output of the layer: output = output * 2
def forward_post_hook(layer, input, output):
# user can use layer, input and output for information statistis tasks
# change the output
return output * 2
linear = paddle.nn.Linear(13, 5)
# register the hook
forward_post_hook_handle = linear.register_forward_post_hook(forward_post_hook)
value1 = np.arange(26).reshape(2, 13).astype("float32")
in1 = paddle.to_tensor(value1)
out0 = linear(in1)
# remove the hook
forward_post_hook_handle.remove()
out1 = linear(in1)
# hook change the linear's output to output * 2, so out0 is equal to out1 * 2.
assert (out0.numpy() == (out1.numpy()) * 2).any()
"""
hook_remove_helper = HookRemoveHelper(self._forward_post_hooks)
self._forward_post_hooks[hook_remove_helper._hook_id] = hook
return hook_remove_helper
def register_forward_pre_hook(self, hook):
"""
Register a forward pre-hook for Layer. The hook will be called before `forward` function has been computed.
It should have the following form, `input` of the `hook` is `input` of the `Layer`,
hook can either return a tuple or a single modified value in the hook. We will wrap the value into a tuple if
a single value is returned(unless that value is already a tuple).
User can use forward pre-hook to change the input of the Layer or perform information statistics tasks on the Layer.
hook(Layer, input) -> None or modified input
Parameters:
hook(function): a function registered as a forward pre-hook
Returns:
HookRemoveHelper, a HookRemoveHelper object that can be used to remove the added hook by calling `hook_remove_helper.remove()` .
Examples:
.. code-block:: python
import paddle
import numpy as np
# the forward_pre_hook change the input of the layer: input = input * 2
def forward_pre_hook(layer, input):
# user can use layer and input for information statistis tasks
# change the input
input_return = (input[0] * 2)
return input_return
linear = paddle.nn.Linear(13, 5)
# register the hook
forward_pre_hook_handle = linear.register_forward_pre_hook(forward_pre_hook)
value0 = np.arange(26).reshape(2, 13).astype("float32")
in0 = paddle.to_tensor(value0)
out0 = linear(in0)
# remove the hook
forward_pre_hook_handle.remove()
value1 = value0 * 2
in1 = paddle.to_tensor(value1)
out1 = linear(in1)
# hook change the linear's input to input * 2, so out0 is equal to out1.
assert (out0.numpy() == out1.numpy()).any()
"""
hook_remove_helper = HookRemoveHelper(self._forward_pre_hooks)
self._forward_pre_hooks[hook_remove_helper._hook_id] = hook
return hook_remove_helper
def create_parameter(
self,
shape,
attr=None,
dtype=None,
is_bias=False,
default_initializer=None,
):
"""Create parameters for this layer.
Parameters:
shape(list): Shape of the parameter.
attr(ParamAttr, optional): Parameter attribute of weight. Please refer to :ref:`api_paddle_ParamAttr`. Default: None.
dtype(str, optional): Data type of this parameter.
If set str, it can be "bool", "float16", "float32", "float64",
"int8", "int16", "int32", "int64", "uint8" or "uint16". Default: "float32".
is_bias(bool, optional): if this is a bias parameter. Default: False.
default_initializer(Initializer, optional): the default initializer for this parameter.
If set None, default initializer will be set to paddle.nn.initializer.Xavier and paddle.nn.initializer.Constant
for non-bias and bias parameter, respectively. Default: None.
Returns:
:Tensor, created parameter.
Examples:
.. code-block:: python
import paddle
class MyLayer(paddle.nn.Layer):
def __init__(self):
super().__init__()
self._linear = paddle.nn.Linear(1, 1)
w_tmp = self.create_parameter([1,1])
self.add_parameter("w_tmp", w_tmp)
def forward(self, input):
return self._linear(input)
mylayer = MyLayer()
for name, param in mylayer.named_parameters():
print(name, param) # will print w_tmp,_linear.weight,_linear.bias
"""
temp_attr = copy.deepcopy(attr)
if isinstance(temp_attr, str) and temp_attr == "":
temp_attr = None
return self._helper.create_parameter(
temp_attr, shape, dtype, is_bias, default_initializer
)
@deprecated(
since="2.0.0",
update_to="paddle.nn.Layer.create_tensor",
reason="New api in create_tensor, easier to use.",
)
def create_variable(self, name=None, persistable=None, dtype=None):
"""
Create Tensor for this layer.
Parameters:
name(str, optional): name of the tensor. Please refer to :ref:`api_guide_Name` . Default: None
persistable(bool, optional): if set this tensor persistable. Default: False
dtype(str, optional): data type of this parameter. If set str, it can be "bool", "float16", "float32", "float64","int8", "int16", "int32", "int64", "uint8" or "uint16". If set None, it will be "float32". Default: None
Returns:
Tensor, created Tensor.
Examples:
.. code-block:: python
import paddle
class MyLinear(paddle.nn.Layer):
def __init__(self,
in_features,
out_features):
super().__init__()
self.linear = paddle.nn.Linear( 10, 10)
self.back_var = self.create_variable(name = "linear_tmp_0", dtype=self._dtype)
def forward(self, input):
out = self.linear(input)
paddle.assign( out, self.back_var)
return out
"""
if name is not None:
var_name = ".".join([self._full_name, name])
else:
var_name = unique_name.generate(
".".join([self._full_name, "_generated_var"])
)
return self._helper.main_program.current_block().create_var(
name=var_name,
persistable=persistable,
dtype=dtype,
type=core.VarDesc.VarType.LOD_TENSOR,
)
# TODO: Add more parameter list when we need them
def create_tensor(self, name=None, persistable=None, dtype=None):
"""
Create Tensor for this layer.
Parameters:
name(str, optional): name of the tensor. Please refer to :ref:`api_guide_Name` . Default: None
persistable(bool, optional): if set this tensor persistable. Default: False
dtype(str, optional): data type of this parameter.
If set str, it can be "bool", "float16", "float32", "float64",
"int8", "int16", "int32", "int64", "uint8" or "uint16".
If set None, it will be "float32". Default: None
Returns:
Tensor, created Tensor.
Examples:
.. code-block:: python
import paddle
class MyLinear(paddle.nn.Layer):
def __init__(self,
in_features,
out_features):
super().__init__()
self.linear = paddle.nn.Linear( 10, 10)
self.back_var = self.create_tensor(name = "linear_tmp_0", dtype=self._dtype)
def forward(self, input):
out = self.linear(input)
paddle.assign( out, self.back_var)
return out
"""
if name is not None:
var_name = ".".join([self._full_name, name])
else:
var_name = unique_name.generate(
".".join([self._full_name, "_generated_var"])
)
return self._helper.main_program.current_block().create_var(
name=var_name,
persistable=persistable,
dtype=dtype,
type=core.VarDesc.VarType.LOD_TENSOR,
)
def parameters(self, include_sublayers=True):
"""
Returns a list of all Parameters from current layer and its sub-layers.
Returns:
list of Tensor, a list of Parameters.
Examples:
.. code-block:: python
import paddle
linear = paddle.nn.Linear(1,1)
print(linear.parameters()) # print linear_0.w_0 and linear_0.b_0
"""
ret = [
param
for _, param in self.named_parameters(
include_sublayers=include_sublayers
)
]
return ret
def children(self):
"""
Returns an iterator over immediate children layers.
Yields:
Layer: a child layer
Examples:
.. code-block:: python
import paddle
linear1 = paddle.nn.Linear(10, 3)
linear2 = paddle.nn.Linear(3, 10, bias_attr=False)
model = paddle.nn.Sequential(linear1, linear2)
layer_list = list(model.children())
print(layer_list) # [<paddle.nn.layer.common.Linear object at 0x7f7b8113f830>, <paddle.nn.layer.common.Linear object at 0x7f7b8113f950>]
"""
for _, layer in self.named_children():
yield layer
def named_children(self):
"""Returns an iterator over immediate children layers, yielding both
the name of the layer as well as the layer itself.
Yields:
(string, Layer): Tuple containing a name and child layer
Examples:
.. code-block:: python
import paddle
linear1 = paddle.nn.Linear(10, 3)
linear2 = paddle.nn.Linear(3, 10, bias_attr=False)
model = paddle.nn.Sequential(linear1, linear2)
for prefix, layer in model.named_children():
print(prefix, layer)
# ('0', <paddle.nn.layer.common.Linear object at 0x7fb61ed85830>)
# ('1', <paddle.nn.layer.common.Linear object at 0x7fb61ed85950>)
"""
memo = set()
for name, layer in self._sub_layers.items():
if layer is not None and layer not in memo:
memo.add(layer)
yield name, layer
def sublayers(self, include_self=False):
"""
Returns a list of sub layers.
Parameters:
include_self(bool, optional): Whether return self as sublayers. Default: False
Returns:
list of Layer, a list of sub layers.
Examples:
.. code-block:: python
import paddle
class MyLayer(paddle.nn.Layer):
def __init__(self):
super().__init__()
self._linear = paddle.nn.Linear(1, 1)
self._dropout = paddle.nn.Dropout(p=0.5)
def forward(self, input):
temp = self._linear(input)
temp = self._dropout(temp)
return temp
mylayer = MyLayer()
print(mylayer.sublayers()) # [<paddle.nn.layer.common.Linear object at 0x7f44b58977d0>, <paddle.nn.layer.common.Dropout object at 0x7f44b58978f0>]
"""
ret = [
layer
for _, layer in self.named_sublayers(include_self=include_self)
]
return ret
def named_parameters(self, prefix='', include_sublayers=True):
"""
Returns an iterator over all parameters in the Layer, yielding tuple of name and parameter.
Parameters:
prefix(str, optional): Prefix to prepend to all parameter names. Default: ''.
include_sublayers(bool, optional): Whether include the parameters of sublayers.
If True, also include the named parameters from sublayers. Default: True.
Yields:
(string, Parameter): Tuple of name and Parameter
Examples:
.. code-block:: python
import paddle
fc1 = paddle.nn.Linear(10, 3)
fc2 = paddle.nn.Linear(3, 10, bias_attr=False)
model = paddle.nn.Sequential(fc1, fc2)
for name, param in model.named_parameters():
print(name, param)
"""
params_set = set()
named_sublayers = (
self.named_sublayers(prefix=prefix, include_self=True)
if include_sublayers
else zip([prefix], [self])
)
for layer_prefix, sublayer in named_sublayers:
params = sublayer._parameters.items()
for key, param in params:
if param is None or param in params_set:
continue
params_set.add(param)
name = layer_prefix + ('.' if layer_prefix else '') + key
yield name, param
def named_sublayers(self, prefix='', include_self=False, layers_set=None):
"""
Returns an iterator over all sublayers in the Layer, yielding tuple of name and sublayer.
The duplicate sublayer will only be yielded once.
Parameters:
prefix(str, optional): Prefix to prepend to all parameter names. Default: ''.
include_self(bool, optional): Whether include the Layer itself. Default: False.
layers_set(set, optional): The set to record duplicate sublayers. Default: None.
Yields:
(string, Layer): Tuple of name and Layer
Examples:
.. code-block:: python
import paddle
fc1 = paddle.nn.Linear(10, 3)
fc2 = paddle.nn.Linear(3, 10, bias_attr=False)
model = paddle.nn.Sequential(fc1, fc2)
for prefix, layer in model.named_sublayers():