Add optimizer save and load

python/paddle/fluid/dygraph/base.py

@@ -92,3 +92,6 @@ def to_variable(value, block=None, name=None):
         return py_var
     elif isinstance(value, framework.Variable):
         return value
+    else:
+        raise TypeError(
+            "to_variable only accepts 'ndarray' and 'Variable' as value's input")

python/paddle/fluid/dygraph/checkpoint.py

@@ -16,13 +16,18 @@
 
 import os
 import collections
-from .. import core
 from ..framework import Variable, default_main_program
+import pickle
+from . import learning_rate_scheduler
+import warnings
 
 __all__ = ['save_persistables', 'load_persistables']
 
 
-def save_persistables(vardict, dirname, filename=None):
+def save_persistables(model_dict,
+                      optimizer=None,
+                      dirname='save_dir',
+                      filename=None):
     """
     This function filters out all variables in layer.parameters from the
     give `layer` and then trys to load these variables from the folder
@@ -34,12 +39,12 @@ def save_persistables(vardict, dirname, filename=None):
     the file name.
 
     Args:
-        vardict(dict of Parameters): The parameters will
+        model_dict(dict of Parameters): The parameters will
                                     be saved. If it is None, nothing
                                     will be deal.
         dirname(str): The directory path.
         filename(str|None): The file which saved all variables. If variables were
-                            saved in differnet files, set it to None.
+                            saved in different files, set it to None.
                             Default: None
 
     Returns:
@@ -71,11 +76,11 @@ def save_persistables(vardict, dirname, filename=None):
             fluid.dygraph.save_persistables(ptb_model.state_dict(), dirname=param_path,
                                        layer=ptb_model)
     """
-    if isinstance(vardict, collections.OrderedDict):
-        _save_var_to_file(vardict, dirname, filename)
+    if isinstance(model_dict, collections.OrderedDict):
+        _save_var_to_file(model_dict, optimizer, dirname, filename)
 
 
-def load_persistables(dirname):
+def load_persistables(dirname='save_dir'):
     """
     This function trys to load persistable variables from the folder
     `dirname` or the file `filename`.
@@ -86,7 +91,8 @@ def load_persistables(dirname):
     the file name.
 
     Args:
-        dirname(str): The directory path.
+        dirname(str): The directory path. default is save_dir
+        optimizer(Optimizer): Optimizer to be save
 
     Returns:
         dict: The parameter-dict resumed from file
@@ -103,7 +109,7 @@ def load_persistables(dirname):
     return _load_var_from_file(dirname)
 
 
-def _save_var_to_file(stat_dict, file_dir, file_name):
+def _save_var_to_file(stat_dict, optimizers, file_dir, file_name):
     save_block = default_main_program().global_block()
     save_var_map = {}
     for var_key, each_var in stat_dict.items():
@@ -117,6 +123,32 @@ def _save_var_to_file(stat_dict, file_dir, file_name):
                     'file_path': os.path.join(file_dir,
                                               os.path.normpath(each_var.name))
                 })
+    if isinstance(optimizers, (list, tuple)):
+        optimizers = optimizers
+    else:
+        optimizers = [optimizers]
+    if os.path.exists(os.path.join(file_dir, os.path.normpath("optimizers"))):
+        pass
+    else:
+        os.mkdir(os.path.join(file_dir, os.path.normpath("optimizers")))
+    for optimizer in optimizers:
+        if isinstance(optimizer._learning_rate,
+                      learning_rate_scheduler.LearningRateDecay):
+            try:
+                f = open(
+                    os.path.join(file_dir, "optimizers",
+                                 os.path.normpath(str(optimizer._name))), "wb")
+                pickle.dump(optimizer._learning_rate, f, 2)
+                f.close()
+            except ():
+                raise IOError("Can't load %s",
+                              os.path.join(
+                                  file_dir, "optimizers",
+                                  os.path.normpath(str(optimizer._name))))
+        else:
+            warnings.warn(
+                "Optimizer not saved, Only optimizer with 'LearningRateDecay' under DyGraph mode need to be saved"
+            )
 
     if file_name is not None:
         save_var_list = []
@@ -138,6 +170,8 @@ def walk_filename(file_dir):
         var_name_list = []
         if os.path.exists(base_path):
             for dirpath, dirnames, filenames in os.walk(base_path):
+                if "optimizers" in dirpath:
+                    continue
                 pt = dirpath.replace(base_path, "", 1)
                 if pt.startswith("/") or pt.startswith("\\"):
                     pt = pt[1:]
@@ -152,6 +186,7 @@ def walk_filename(file_dir):
 
     load_block = default_main_program().global_block()
     load_var_map = {}
+    load_optimizer_map = {}
     file_var_list = walk_filename(file_dir)
     for var_name in file_var_list:
         new_var = Variable(block=load_block, name=var_name)
@@ -165,8 +200,22 @@ def walk_filename(file_dir):
             })
 
         load_var_map[new_var.name] = new_var
-
-    return load_var_map
+    opt_path = os.path.join(file_dir, "optimizers")
+    for _, _, optimizers in os.walk(opt_path):
+        for optimizer in optimizers:
+            try:
+                f = open(os.path.join(opt_path, optimizer), "rb")
+                load_optimizer_map[optimizer] = pickle.load(f)
+                f.close()
+            except IOError:
+                raise IOError("Can't load %s",
+                              os.path.join(
+                                  file_dir, "optimizers",
+                                  os.path.normpath(str(optimizer._name))))
+    if len(load_optimizer_map) == 0:
+        warnings.warn("No optimizer loaded")
+
+    return load_var_map, load_optimizer_map
 
 
 def _clone_var_in_block_(block, var):

python/paddle/fluid/dygraph/learning_rate_scheduler.py

@@ -63,13 +63,13 @@ def __init__(self, boundaries, values, begin, step=1, dtype='float32'):
 
         self.vars = []
         for value in values:
-            self.vars.append(self.create_lr_var(value))
+            self.vars.append(value)
 
     def step(self):
         for i in range(len(self.boundaries)):
             if self.step_num < self.boundaries[i]:
                 return self.vars[i]
-        return self.vars[len(self.values) - 1]
+        return self.create_lr_var(self.vars[len(self.values) - 1])
 
 
 class NaturalExpDecay(LearningRateDecay):

python/paddle/fluid/optimizer.py

@@ -1,4 +1,4 @@
-# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
+# 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.
@@ -16,25 +16,25 @@
 
 import numpy as np
 from collections import defaultdict
-from functools import reduce
 
-from paddle.fluid import core
 from paddle.fluid.distribute_lookup_table import find_distributed_lookup_table
 from paddle.fluid.framework import Program, Variable, name_scope, default_main_program, default_startup_program
-from paddle.fluid.layers import tensor
 
 from . import framework
 from . import layers
 from . import unique_name
 from .backward import append_backward
 from .clip import append_gradient_clip_ops, error_clip_callback
-from .dygraph import base as imperative_base
-from .dygraph.learning_rate_scheduler import LearningRateDecay
 from .framework import program_guard
 from .initializer import Constant
 from .layer_helper import LayerHelper
 from .layers import ops
 from .regularizer import append_regularization_ops
+from .dygraph import base as imperative_base
+from .dygraph.learning_rate_scheduler import LearningRateDecay
+from paddle.fluid import core
+from paddle.fluid.layers import tensor
+from functools import reduce
 from .wrapped_decorator import signature_safe_contextmanager
 
 __all__ = [
@@ -63,14 +63,18 @@ def __init__(self, learning_rate, regularization=None, name=None):
                 raise TypeError(
                     "learning rate should be float or LearningRateDecay, got %s here"
                     % type(learning_rate))
+            if name is not None:
+                self._name = unique_name.generate(name)
+            else:
+                self._name = unique_name.generate(self.__class__.__name__)
         else:
             if not isinstance(learning_rate, float) and \
                     not isinstance(learning_rate, framework.Variable):
                 raise TypeError(
                     "learning rate should be float or Variable, got %s here" %
                     type(learning_rate))
+            self._name = name
 
-        self._name = name
         self.regularization = regularization
         self._learning_rate = learning_rate
         # the learning rate type should be inferenced from loss
@@ -89,6 +93,19 @@ def __init__(self, learning_rate, regularization=None, name=None):
         self.helper = None
         self._opti_name_list = []
 
+    def load(self, stat_dict):
+        """
+        load optimizer with learning rate decay in dygraph mode
+        :return: None
+
+        Args:
+            stat_dict: the dict load by load_persistable method
+        """
+        if framework.in_dygraph_mode():
+            self._learning_rate = stat_dict[self._name]
+        else:
+            raise TypeError("load_dict can only be used under DyGraph mode")
+
     def get_opti_var_name_list(self):
         return self._opti_name_list
 

python/paddle/fluid/tests/unittests/test_imperative_checkpoint.py

@@ -135,14 +135,14 @@ def test_save_load_persistables(self):
 
                     avg_loss.backward()
                     sgd.minimize(avg_loss)
-                    fluid.dygraph.save_persistables(mnist.state_dict(),
+                    fluid.dygraph.save_persistables(mnist.state_dict(), [sgd],
                                                     "save_dir")
                     mnist.clear_gradients()
 
                     for param in mnist.parameters():
                         dy_param_init_value[param.name] = param.numpy()
 
-                    restore = fluid.dygraph.load_persistables("save_dir")
+                    restore, _ = fluid.dygraph.load_persistables("save_dir")
                     mnist.load_dict(restore)
 
                     self.assertEqual(len(dy_param_init_value), len(restore))