Move trainer to contrib

01.fit_a_line/README.cn.md

@@ -164,7 +164,7 @@ place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 训练器会读入一个训练程序和一些必要的其他参数：
 
 ```python
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=train_program,
     place=place,
     optimizer_func=optimizer_program)
@@ -194,7 +194,7 @@ step = 0
 # event_handler prints training and testing info
 def event_handler_plot(event):
     global step
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if step % 10 == 0:   # record a train cost every 10 batches
             plot_cost.append(train_title, step, event.metrics[0])
 
@@ -210,7 +210,7 @@ def event_handler_plot(event):
                 trainer.stop()
         step += 1
 
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         if event.epoch % 10 == 0:
             # We can save the trained parameters for the inferences later
             if params_dirname is not None:
@@ -254,7 +254,7 @@ def inference_program():
 预测器会从`params_dirname`中读取已经训练好的模型，来对从未遇见过的数据进行预测。
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
     infer_func=inference_program, param_path=params_dirname, place=place)
 
 batch_size = 10

01.fit_a_line/README.md

@@ -180,7 +180,7 @@ place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 The trainer will take the `train_program` as input.
 
 ```python
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=train_program,
     place=place,
     optimizer_func=optimizer_program)
@@ -213,7 +213,7 @@ step = 0
 # event_handler prints training and testing info
 def event_handler_plot(event):
     global step
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if step % 10 == 0:   # record a train cost every 10 batches
             plot_cost.append(train_title, step, event.metrics[0])
 
@@ -229,7 +229,7 @@ def event_handler_plot(event):
                 trainer.stop()
         step += 1
 
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         if event.epoch % 10 == 0:
             # We can save the trained parameters for the inferences later
             if params_dirname is not None:
@@ -276,7 +276,7 @@ def inference_program():
 Inferencer will load the trained model from `params_dirname` and use it to infer the unseen data.
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
     infer_func=inference_program, param_path=params_dirname, place=place)
 
 batch_size = 10

01.fit_a_line/index.cn.html

@@ -206,7 +206,7 @@ place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 训练器会读入一个训练程序和一些必要的其他参数：
 
 ```python
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=train_program,
     place=place,
     optimizer_func=optimizer_program)
@@ -236,7 +236,7 @@ step = 0
 # event_handler prints training and testing info
 def event_handler_plot(event):
     global step
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if step % 10 == 0:   # record a train cost every 10 batches
             plot_cost.append(train_title, step, event.metrics[0])
 
@@ -252,7 +252,7 @@ def event_handler_plot(event):
                 trainer.stop()
         step += 1
 
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         if event.epoch % 10 == 0:
             # We can save the trained parameters for the inferences later
             if params_dirname is not None:
@@ -296,7 +296,7 @@ def inference_program():
 预测器会从`params_dirname`中读取已经训练好的模型，来对从未遇见过的数据进行预测。
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
     infer_func=inference_program, param_path=params_dirname, place=place)
 
 batch_size = 10

01.fit_a_line/index.html

@@ -222,7 +222,7 @@ place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 The trainer will take the `train_program` as input.
 
 ```python
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=train_program,
     place=place,
     optimizer_func=optimizer_program)
@@ -255,7 +255,7 @@ step = 0
 # event_handler prints training and testing info
 def event_handler_plot(event):
     global step
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if step % 10 == 0:   # record a train cost every 10 batches
             plot_cost.append(train_title, step, event.metrics[0])
 
@@ -271,7 +271,7 @@ def event_handler_plot(event):
                 trainer.stop()
         step += 1
 
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         if event.epoch % 10 == 0:
             # We can save the trained parameters for the inferences later
             if params_dirname is not None:
@@ -318,7 +318,7 @@ def inference_program():
 Inferencer will load the trained model from `params_dirname` and use it to infer the unseen data.
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
     infer_func=inference_program, param_path=params_dirname, place=place)
 
 batch_size = 10

01.fit_a_line/train.py

@@ -15,6 +15,18 @@
 from __future__ import print_function
 import paddle
 import paddle.fluid as fluid
+import sys
+
+try:
+    from paddle.fluid.contrib.trainer import *
+    from paddle.fluid.contrib.inferencer import *
+except ImportError:
+    print(
+        "In the fluid 1.0, the trainer and inferencer are moving to paddle.fluid.contrib",
+        file=sys.stderr)
+    from paddle.fluid.trainer import *
+    from paddle.fluid.inferencer import *
+
 import numpy
 
 BATCH_SIZE = 20
@@ -49,7 +61,7 @@ def optimizer_program():
 use_cuda = False
 place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 
-trainer = fluid.Trainer(
+trainer = Trainer(
     train_func=train_program, place=place, optimizer_func=optimizer_program)
 
 feed_order = ['x', 'y']
@@ -59,6 +71,7 @@ params_dirname = "fit_a_line.inference.model"
 
 # Plot data
 from paddle.v2.plot import Ploter
+
 train_title = "Train cost"
 test_title = "Test cost"
 plot_cost = Ploter(train_title, test_title)
@@ -69,7 +82,7 @@ step = 0
 # event_handler prints training and testing info
 def event_handler_plot(event):
     global step
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, EndStepEvent):
         if step % 10 == 0:  # record a train cost every 10 batches
             plot_cost.append(train_title, step, event.metrics[0])
 
@@ -85,7 +98,7 @@ def event_handler_plot(event):
                 trainer.stop()
         step += 1
 
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, EndEpochEvent):
         if event.epoch % 10 == 0:
             # We can save the trained parameters for the inferences later
             if params_dirname is not None:
@@ -106,7 +119,7 @@ def inference_program():
     return y_predict
 
 
-inferencer = fluid.Inferencer(
+inferencer = Inferencer(
     infer_func=inference_program, param_path=params_dirname, place=place)
 
 batch_size = 10

02.recognize_digits/README.cn.md

@@ -274,7 +274,7 @@ test_reader = paddle.batch(
 use_cuda = False # set to True if training with GPU
 place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=train_program, place=place, optimizer_func=optimizer_program)
 ```
 
@@ -290,14 +290,14 @@ Fluid API 在训练期间为回调函数提供了一个钩子。用户能够通
 params_dirname = "recognize_digits_network.inference.model"
 lists = []
 def event_handler(event):
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if event.step % 100 == 0:
             # event.metrics maps with train program return arguments.
             # event.metrics[0] will yeild avg_cost and event.metrics[1] will yeild acc in this example.
             print("Pass %d, Batch %d, Cost %f" % (
                 event.step, event.epoch, event.metrics[0]))
 
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         avg_cost, acc = trainer.test(
             reader=test_reader, feed_order=['img', 'label'])
 
@@ -328,14 +328,14 @@ lists = []
 # event_handler to plot a figure
 def event_handler_plot(event):
     global step
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if step % 100 == 0:
             # event.metrics maps with train program return arguments.
             # event.metrics[0] will yeild avg_cost and event.metrics[1] will yeild acc in this example.
             cost_ploter.append(train_title, step, event.metrics[0])
             cost_ploter.plot()
         step += 1
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         # save parameters
         trainer.save_params(params_dirname)
 
@@ -380,7 +380,7 @@ Test with Epoch 0, avg_cost: 0.053097883707459624, acc: 0.9822850318471338
 
 ## 应用模型
 
-可以使用训练好的模型对手写体数字图片进行分类，下面程序展示了如何使用 `fluid.Inferencer` 接口进行推断。
+可以使用训练好的模型对手写体数字图片进行分类，下面程序展示了如何使用 `fluid.contrib.inferencer.Inferencer` 接口进行推断。
 
 ### Inference 配置
 
@@ -388,7 +388,7 @@ Test with Epoch 0, avg_cost: 0.053097883707459624, acc: 0.9822850318471338
 我们可以简单地插入在此之前定义的分类器。
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
     # infer_func=softmax_regression, # uncomment for softmax regression
     # infer_func=multilayer_perceptron, # uncomment for MLP
     infer_func=convolutional_neural_network,  # uncomment for LeNet5

02.recognize_digits/README.md

@@ -281,7 +281,7 @@ Now, we need to setup the trainer. The trainer need to take in `train_program`,
 use_cuda = False # set to True if training with GPU
 place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=train_program, place=place, optimizer_func=optimizer_program)
  ```
 
@@ -297,14 +297,14 @@ We will demonstrate two event handlers here. Please feel free to modify on the J
 params_dirname = "recognize_digits_network.inference.model"
 lists = []
 def event_handler(event):
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if event.step % 100 == 0:
             # event.metrics maps with train program return arguments.
             # event.metrics[0] will yeild avg_cost and event.metrics[1] will yeild acc in this example.
             print("Pass %d, Batch %d, Cost %f" % (
                 event.step, event.epoch, event.metrics[0]))
 
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         avg_cost, acc = trainer.test(
             reader=test_reader, feed_order=['img', 'label'])
 
@@ -331,14 +331,14 @@ lists = []
 # event_handler to plot a figure
 def event_handler_plot(event):
     global step
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if step % 100 == 0:
             # event.metrics maps with train program return arguments.
             # event.metrics[0] will yeild avg_cost and event.metrics[1] will yeild acc in this example.
             cost_ploter.append(train_title, step, event.metrics[0])
             cost_ploter.plot()
         step += 1
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         # save parameters
         trainer.save_params(params_dirname)
 
@@ -390,7 +390,7 @@ Usually, with MNIST data, the softmax regression model achieves an accuracy arou
 
 ## Application
 
-After training, users can use the trained model to classify images. The following code shows how to inference MNIST images through `fluid.Inferencer`.
+After training, users can use the trained model to classify images. The following code shows how to inference MNIST images through `fluid.contrib.inferencer.Inferencer`.
 
 ### Create Inferencer
 
@@ -398,7 +398,7 @@ The `Inferencer` takes an `infer_func` and `param_path` to setup the network and
 We can simply plug-in the classifier defined earlier here.
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
     # infer_func=softmax_regression, # uncomment for softmax regression
     # infer_func=multilayer_perceptron, # uncomment for MLP
     infer_func=convolutional_neural_network,  # uncomment for LeNet5

02.recognize_digits/index.cn.html

@@ -316,7 +316,7 @@ test_reader = paddle.batch(
 use_cuda = False # set to True if training with GPU
 place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=train_program, place=place, optimizer_func=optimizer_program)
 ```
 
@@ -332,14 +332,14 @@ Fluid API 在训练期间为回调函数提供了一个钩子。用户能够通
 params_dirname = "recognize_digits_network.inference.model"
 lists = []
 def event_handler(event):
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if event.step % 100 == 0:
             # event.metrics maps with train program return arguments.
             # event.metrics[0] will yeild avg_cost and event.metrics[1] will yeild acc in this example.
             print("Pass %d, Batch %d, Cost %f" % (
                 event.step, event.epoch, event.metrics[0]))
 
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         avg_cost, acc = trainer.test(
             reader=test_reader, feed_order=['img', 'label'])
 
@@ -370,14 +370,14 @@ lists = []
 # event_handler to plot a figure
 def event_handler_plot(event):
     global step
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if step % 100 == 0:
             # event.metrics maps with train program return arguments.
             # event.metrics[0] will yeild avg_cost and event.metrics[1] will yeild acc in this example.
             cost_ploter.append(train_title, step, event.metrics[0])
             cost_ploter.plot()
         step += 1
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         # save parameters
         trainer.save_params(params_dirname)
 
@@ -422,7 +422,7 @@ Test with Epoch 0, avg_cost: 0.053097883707459624, acc: 0.9822850318471338
 
 ## 应用模型
 
-可以使用训练好的模型对手写体数字图片进行分类，下面程序展示了如何使用 `fluid.Inferencer` 接口进行推断。
+可以使用训练好的模型对手写体数字图片进行分类，下面程序展示了如何使用 `fluid.contrib.inferencer.Inferencer` 接口进行推断。
 
 ### Inference 配置
 
@@ -430,7 +430,7 @@ Test with Epoch 0, avg_cost: 0.053097883707459624, acc: 0.9822850318471338
 我们可以简单地插入在此之前定义的分类器。
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
     # infer_func=softmax_regression, # uncomment for softmax regression
     # infer_func=multilayer_perceptron, # uncomment for MLP
     infer_func=convolutional_neural_network,  # uncomment for LeNet5

02.recognize_digits/index.html

@@ -323,7 +323,7 @@ Now, we need to setup the trainer. The trainer need to take in `train_program`,
 use_cuda = False # set to True if training with GPU
 place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=train_program, place=place, optimizer_func=optimizer_program)
  ```
 
@@ -339,14 +339,14 @@ We will demonstrate two event handlers here. Please feel free to modify on the J
 params_dirname = "recognize_digits_network.inference.model"
 lists = []
 def event_handler(event):
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if event.step % 100 == 0:
             # event.metrics maps with train program return arguments.
             # event.metrics[0] will yeild avg_cost and event.metrics[1] will yeild acc in this example.
             print("Pass %d, Batch %d, Cost %f" % (
                 event.step, event.epoch, event.metrics[0]))
 
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         avg_cost, acc = trainer.test(
             reader=test_reader, feed_order=['img', 'label'])
 
@@ -373,14 +373,14 @@ lists = []
 # event_handler to plot a figure
 def event_handler_plot(event):
     global step
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if step % 100 == 0:
             # event.metrics maps with train program return arguments.
             # event.metrics[0] will yeild avg_cost and event.metrics[1] will yeild acc in this example.
             cost_ploter.append(train_title, step, event.metrics[0])
             cost_ploter.plot()
         step += 1
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         # save parameters
         trainer.save_params(params_dirname)
 
@@ -432,7 +432,7 @@ Usually, with MNIST data, the softmax regression model achieves an accuracy arou
 
 ## Application
 
-After training, users can use the trained model to classify images. The following code shows how to inference MNIST images through `fluid.Inferencer`.
+After training, users can use the trained model to classify images. The following code shows how to inference MNIST images through `fluid.contrib.inferencer.Inferencer`.
 
 ### Create Inferencer
 
@@ -440,7 +440,7 @@ The `Inferencer` takes an `infer_func` and `param_path` to setup the network and
 We can simply plug-in the classifier defined earlier here.
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
     # infer_func=softmax_regression, # uncomment for softmax regression
     # infer_func=multilayer_perceptron, # uncomment for MLP
     infer_func=convolutional_neural_network,  # uncomment for LeNet5

02.recognize_digits/train.py

@@ -5,6 +5,16 @@ import numpy as np
 import paddle
 import paddle.fluid as fluid
 
+try:
+    from paddle.fluid.contrib.trainer import *
+    from paddle.fluid.contrib.inferencer import *
+except ImportError:
+    print(
+        "In the fluid 1.0, the trainer and inferencer are moving to paddle.fluid.contrib",
+        file=sys.stderr)
+    from paddle.fluid.trainer import *
+    from paddle.fluid.inferencer import *
+
 
 def softmax_regression():
     img = fluid.layers.data(name='img', shape=[1, 28, 28], dtype='float32')
@@ -77,7 +87,7 @@ def main():
     use_cuda = False  # set to True if training with GPU
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 
-    trainer = fluid.Trainer(
+    trainer = Trainer(
         train_func=train_program, place=place, optimizer_func=optimizer_program)
 
     # Save the parameter into a directory. The Inferencer can load the parameters from it to do infer
@@ -86,14 +96,14 @@ def main():
     lists = []
 
     def event_handler(event):
-        if isinstance(event, fluid.EndStepEvent):
+        if isinstance(event, EndStepEvent):
             if event.step % 100 == 0:
                 # event.metrics maps with train program return arguments.
                 # event.metrics[0] will yeild avg_cost and event.metrics[1] will yeild acc in this example.
                 print("Pass %d, Batch %d, Cost %f" % (event.step, event.epoch,
                                                       event.metrics[0]))
 
-        if isinstance(event, fluid.EndEpochEvent):
+        if isinstance(event, EndEpochEvent):
             avg_cost, acc = trainer.test(
                 reader=test_reader, feed_order=['img', 'label'])
 
@@ -125,7 +135,7 @@ def main():
 
     cur_dir = os.path.dirname(os.path.realpath(__file__))
     img = load_image(cur_dir + '/image/infer_3.png')
-    inferencer = fluid.Inferencer(
+    inferencer = Inferencer(
         # infer_func=softmax_regression, # uncomment for softmax regression
         # infer_func=multilayer_perceptron, # uncomment for MLP
         infer_func=convolutional_neural_network,  # uncomment for LeNet5

03.image_classification/README.cn.md

@@ -346,7 +346,7 @@ def optimizer_program():
 ```python
 use_cuda = False
 place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=train_program,
     optimizer_func=optimizer_program,
     place=place)
@@ -394,12 +394,12 @@ cost_ploter = Ploter(train_title, test_title)
 step = 0
 def event_handler_plot(event):
     global step
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if step % 1 == 0:
             cost_ploter.append(train_title, step, event.metrics[0])
             cost_ploter.plot()
         step += 1
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         avg_cost, accuracy = trainer.test(
             reader=test_reader,
             feed_order=['pixel', 'label'])
@@ -417,7 +417,7 @@ params_dirname = "image_classification_resnet.inference.model"
 
 # event handler to track training and testing process
 def event_handler(event):
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if event.step % 100 == 0:
             print("\nPass %d, Batch %d, Cost %f, Acc %f" %
                   (event.step, event.epoch, event.metrics[0],
@@ -426,7 +426,7 @@ def event_handler(event):
             sys.stdout.write('.')
             sys.stdout.flush()
 
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         # Test against with the test dataset to get accuracy.
         avg_cost, accuracy = trainer.test(
             reader=test_reader, feed_order=['pixel', 'label'])
@@ -475,7 +475,7 @@ Test with Pass 0, Loss 1.1, Acc 0.6
 
 ## 应用模型
 
-可以使用训练好的模型对图片进行分类，下面程序展示了如何使用 `fluid.Inferencer` 接口进行推断，可以打开注释，更改加载的模型。
+可以使用训练好的模型对图片进行分类，下面程序展示了如何使用 `fluid.contrib.inferencer.Inferencer` 接口进行推断，可以打开注释，更改加载的模型。
 
 ### 生成预测输入数据
 
@@ -513,7 +513,7 @@ img = load_image(cur_dir + '/image/dog.png')
 现在我们准备做预测。
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
     infer_func=inference_program, param_path=params_dirname, place=place)
 label_list = ["airplane", "automobile", "bird", "cat", "deer", "dog", "frog", "horse", "ship", "truck"]
 # inference

03.image_classification/README.md

@@ -348,7 +348,7 @@ Here we specify `Adam` optimization algorithm via `fluid.optimizer`.
 ```python
 use_cuda = False
 place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=train_program,
     optimizer_func=optimizer_program,
     place=place)
@@ -392,12 +392,12 @@ cost_ploter = Ploter(train_title, test_title)
 step = 0
 def event_handler_plot(event):
     global step
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if step % 1 == 0:
             cost_ploter.append(train_title, step, event.metrics[0])
             cost_ploter.plot()
         step += 1
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         avg_cost, accuracy = trainer.test(
             reader=test_reader,
             feed_order=['pixel', 'label'])
@@ -415,7 +415,7 @@ params_dirname = "image_classification_resnet.inference.model"
 
 # event handler to track training and testing process
 def event_handler(event):
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if event.step % 100 == 0:
             print("\nPass %d, Batch %d, Cost %f, Acc %f" %
                   (event.step, event.epoch, event.metrics[0],
@@ -424,7 +424,7 @@ def event_handler(event):
             sys.stdout.write('.')
             sys.stdout.flush()
 
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         # Test against with the test dataset to get accuracy.
         avg_cost, accuracy = trainer.test(
             reader=test_reader, feed_order=['pixel', 'label'])
@@ -474,7 +474,7 @@ Figure 12. The error rate of VGG model on CIFAR10
 
 ## Application
 
-After training is completed, users can use the trained model to classify images. The following code shows how to infer through `fluid.Inferencer` interface. You can uncomment some lines from below to change the model name.
+After training is completed, users can use the trained model to classify images. The following code shows how to infer through `fluid.contrib.inferencer.Inferencer` interface. You can uncomment some lines from below to change the model name.
 
 ### Generate input data for inferring
 
@@ -512,7 +512,7 @@ We can simply plug-in the inference_program defined earlier here.
 Now we are ready to do inference.
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
     infer_func=inference_program, param_path=params_dirname, place=place)
 
 label_list = ["airplane", "automobile", "bird", "cat", "deer", "dog", "frog", "horse", "ship", "truck"]

03.image_classification/index.cn.html

@@ -388,7 +388,7 @@ def optimizer_program():
 ```python
 use_cuda = False
 place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=train_program,
     optimizer_func=optimizer_program,
     place=place)
@@ -436,12 +436,12 @@ cost_ploter = Ploter(train_title, test_title)
 step = 0
 def event_handler_plot(event):
     global step
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if step % 1 == 0:
             cost_ploter.append(train_title, step, event.metrics[0])
             cost_ploter.plot()
         step += 1
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         avg_cost, accuracy = trainer.test(
             reader=test_reader,
             feed_order=['pixel', 'label'])
@@ -459,7 +459,7 @@ params_dirname = "image_classification_resnet.inference.model"
 
 # event handler to track training and testing process
 def event_handler(event):
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if event.step % 100 == 0:
             print("\nPass %d, Batch %d, Cost %f, Acc %f" %
                   (event.step, event.epoch, event.metrics[0],
@@ -468,7 +468,7 @@ def event_handler(event):
             sys.stdout.write('.')
             sys.stdout.flush()
 
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         # Test against with the test dataset to get accuracy.
         avg_cost, accuracy = trainer.test(
             reader=test_reader, feed_order=['pixel', 'label'])
@@ -517,7 +517,7 @@ Test with Pass 0, Loss 1.1, Acc 0.6
 
 ## 应用模型
 
-可以使用训练好的模型对图片进行分类，下面程序展示了如何使用 `fluid.Inferencer` 接口进行推断，可以打开注释，更改加载的模型。
+可以使用训练好的模型对图片进行分类，下面程序展示了如何使用 `fluid.contrib.inferencer.Inferencer` 接口进行推断，可以打开注释，更改加载的模型。
 
 ### 生成预测输入数据
 
@@ -555,7 +555,7 @@ img = load_image(cur_dir + '/image/dog.png')
 现在我们准备做预测。
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
     infer_func=inference_program, param_path=params_dirname, place=place)
 label_list = ["airplane", "automobile", "bird", "cat", "deer", "dog", "frog", "horse", "ship", "truck"]
 # inference

03.image_classification/index.html

@@ -390,7 +390,7 @@ Here we specify `Adam` optimization algorithm via `fluid.optimizer`.
 ```python
 use_cuda = False
 place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=train_program,
     optimizer_func=optimizer_program,
     place=place)
@@ -434,12 +434,12 @@ cost_ploter = Ploter(train_title, test_title)
 step = 0
 def event_handler_plot(event):
     global step
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if step % 1 == 0:
             cost_ploter.append(train_title, step, event.metrics[0])
             cost_ploter.plot()
         step += 1
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         avg_cost, accuracy = trainer.test(
             reader=test_reader,
             feed_order=['pixel', 'label'])
@@ -457,7 +457,7 @@ params_dirname = "image_classification_resnet.inference.model"
 
 # event handler to track training and testing process
 def event_handler(event):
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         if event.step % 100 == 0:
             print("\nPass %d, Batch %d, Cost %f, Acc %f" %
                   (event.step, event.epoch, event.metrics[0],
@@ -466,7 +466,7 @@ def event_handler(event):
             sys.stdout.write('.')
             sys.stdout.flush()
 
-    if isinstance(event, fluid.EndEpochEvent):
+    if isinstance(event, fluid.contrib.trainer.EndEpochEvent):
         # Test against with the test dataset to get accuracy.
         avg_cost, accuracy = trainer.test(
             reader=test_reader, feed_order=['pixel', 'label'])
@@ -516,7 +516,7 @@ Figure 12. The error rate of VGG model on CIFAR10
 
 ## Application
 
-After training is completed, users can use the trained model to classify images. The following code shows how to infer through `fluid.Inferencer` interface. You can uncomment some lines from below to change the model name.
+After training is completed, users can use the trained model to classify images. The following code shows how to infer through `fluid.contrib.inferencer.Inferencer` interface. You can uncomment some lines from below to change the model name.
 
 ### Generate input data for inferring
 
@@ -554,7 +554,7 @@ We can simply plug-in the inference_program defined earlier here.
 Now we are ready to do inference.
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
     infer_func=inference_program, param_path=params_dirname, place=place)
 
 label_list = ["airplane", "automobile", "bird", "cat", "deer", "dog", "frog", "horse", "ship", "truck"]

03.image_classification/train.py

@@ -58,7 +58,7 @@ def train(use_cuda, train_program, params_dirname):
         paddle.dataset.cifar.test10(), batch_size=BATCH_SIZE)
 
     def event_handler(event):
-        if isinstance(event, fluid.EndStepEvent):
+        if isinstance(event, EndStepEvent):
             if event.step % 100 == 0:
                 print("\nPass %d, Batch %d, Cost %f, Acc %f" %
                       (event.step, event.epoch, event.metrics[0],
@@ -67,7 +67,7 @@ def train(use_cuda, train_program, params_dirname):
                 sys.stdout.write('.')
                 sys.stdout.flush()
 
-        if isinstance(event, fluid.EndEpochEvent):
+        if isinstance(event, EndEpochEvent):
             avg_cost, accuracy = trainer.test(
                 reader=test_reader, feed_order=['pixel', 'label'])
 
@@ -77,7 +77,7 @@ def train(use_cuda, train_program, params_dirname):
                 trainer.save_params(params_dirname)
 
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
-    trainer = fluid.Trainer(
+    trainer = Trainer(
         train_func=train_program, optimizer_func=optimizer_program, place=place)
 
     trainer.train(
@@ -89,7 +89,7 @@ def train(use_cuda, train_program, params_dirname):
 
 def infer(use_cuda, inference_program, params_dirname=None):
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
-    inferencer = fluid.Inferencer(
+    inferencer = Inferencer(
         infer_func=inference_program, param_path=params_dirname, place=place)
 
     # Prepare testing data. 

04.word2vec/README.cn.md

@@ -309,7 +309,7 @@ def train(use_cuda, train_program, params_dirname):
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 
     def event_handler(event):
-        if isinstance(event, fluid.EndStepEvent):
+        if isinstance(event, fluid.contrib.trainer.EndStepEvent):
             # We output cost every 10 steps.
             if event.step % 10 == 0:
                 outs = trainer.test(
@@ -329,7 +329,7 @@ def train(use_cuda, train_program, params_dirname):
                 if math.isnan(avg_cost):
                     sys.exit("got NaN loss, training failed.")
 
-    trainer = fluid.Trainer(
+    trainer = fluid.contrib.trainer.Trainer(
         train_func=train_program,
         optimizer_func=optimizer_func,
         place=place)
@@ -360,7 +360,7 @@ Step 20: Average Cost 5.766995
 ```python
 def infer(use_cuda, inference_program, params_dirname=None):
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
-    inferencer = fluid.Inferencer(
+    inferencer = fluid.contrib.inferencer.Inferencer(
         infer_func=inference_program, param_path=params_dirname, place=place)
 
     # Setup inputs by creating 4 LoDTensors representing 4 words. Here each word

04.word2vec/README.md

@@ -321,7 +321,7 @@ def train(use_cuda, train_program, params_dirname):
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 
     def event_handler(event):
-        if isinstance(event, fluid.EndStepEvent):
+        if isinstance(event, fluid.contrib.trainer.EndStepEvent):
             outs = trainer.test(
                 reader=test_reader,
                 feed_order=['firstw', 'secondw', 'thirdw', 'fourthw', 'nextw'])
@@ -341,7 +341,7 @@ def train(use_cuda, train_program, params_dirname):
             if math.isnan(avg_cost):
                 sys.exit("got NaN loss, training failed.")
 
-    trainer = fluid.Trainer(
+    trainer = fluid.contrib.trainer.Trainer(
         train_func=train_program,
         # Note here we need to choose more sophisticated optimizer
         # such as AdaGrad with a decay rate. The normal SGD converges
@@ -378,7 +378,7 @@ We can use our trained model to predict the next word given its previous N-gram.
 ```python
 def infer(use_cuda, inference_program, params_dirname=None):
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
-    inferencer = fluid.Inferencer(
+    inferencer = fluid.contrib.inferencer.Inferencer(
         infer_func=inference_program, param_path=params_dirname, place=place)
 
     # Setup inputs by creating 4 LoDTensors representing 4 words. Here each word

04.word2vec/index.cn.html

@@ -351,7 +351,7 @@ def train(use_cuda, train_program, params_dirname):
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 
     def event_handler(event):
-        if isinstance(event, fluid.EndStepEvent):
+        if isinstance(event, fluid.contrib.trainer.EndStepEvent):
             # We output cost every 10 steps.
             if event.step % 10 == 0:
                 outs = trainer.test(
@@ -371,7 +371,7 @@ def train(use_cuda, train_program, params_dirname):
                 if math.isnan(avg_cost):
                     sys.exit("got NaN loss, training failed.")
 
-    trainer = fluid.Trainer(
+    trainer = fluid.contrib.trainer.Trainer(
         train_func=train_program,
         optimizer_func=optimizer_func,
         place=place)
@@ -402,7 +402,7 @@ Step 20: Average Cost 5.766995
 ```python
 def infer(use_cuda, inference_program, params_dirname=None):
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
-    inferencer = fluid.Inferencer(
+    inferencer = fluid.contrib.inferencer.Inferencer(
         infer_func=inference_program, param_path=params_dirname, place=place)
 
     # Setup inputs by creating 4 LoDTensors representing 4 words. Here each word

04.word2vec/index.html

@@ -363,7 +363,7 @@ def train(use_cuda, train_program, params_dirname):
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 
     def event_handler(event):
-        if isinstance(event, fluid.EndStepEvent):
+        if isinstance(event, fluid.contrib.trainer.EndStepEvent):
             outs = trainer.test(
                 reader=test_reader,
                 feed_order=['firstw', 'secondw', 'thirdw', 'fourthw', 'nextw'])
@@ -383,7 +383,7 @@ def train(use_cuda, train_program, params_dirname):
             if math.isnan(avg_cost):
                 sys.exit("got NaN loss, training failed.")
 
-    trainer = fluid.Trainer(
+    trainer = fluid.contrib.trainer.Trainer(
         train_func=train_program,
         # Note here we need to choose more sophisticated optimizer
         # such as AdaGrad with a decay rate. The normal SGD converges
@@ -420,7 +420,7 @@ We can use our trained model to predict the next word given its previous N-gram.
 ```python
 def infer(use_cuda, inference_program, params_dirname=None):
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
-    inferencer = fluid.Inferencer(
+    inferencer = fluid.contrib.inferencer.Inferencer(
         infer_func=inference_program, param_path=params_dirname, place=place)
 
     # Setup inputs by creating 4 LoDTensors representing 4 words. Here each word

04.word2vec/train.py

@@ -97,7 +97,7 @@ def train(use_cuda, train_program, params_dirname):
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 
     def event_handler(event):
-        if isinstance(event, fluid.EndStepEvent):
+        if isinstance(event, EndStepEvent):
             outs = trainer.test(
                 reader=test_reader,
                 feed_order=['firstw', 'secondw', 'thirdw', 'fourthw', 'nextw'])
@@ -116,7 +116,7 @@ def train(use_cuda, train_program, params_dirname):
             if math.isnan(avg_cost):
                 sys.exit("got NaN loss, training failed.")
 
-    trainer = fluid.Trainer(
+    trainer = Trainer(
         train_func=train_program,
         # optimizer=fluid.optimizer.SGD(learning_rate=0.001),
         optimizer_func=optimizer_func,
@@ -131,7 +131,7 @@ def train(use_cuda, train_program, params_dirname):
 
 def infer(use_cuda, inference_program, params_dirname=None):
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
-    inferencer = fluid.Inferencer(
+    inferencer = Inferencer(
         infer_func=inference_program, param_path=params_dirname, place=place)
 
     # Setup inputs by creating 4 LoDTensors representing 4 words. Here each word

05.recommender_system/README.cn.md

@@ -409,7 +409,7 @@ test_reader = paddle.batch(
 训练器需要一个训练程序和一个训练优化函数。
 
 ```python
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=train_program, place=place, optimizer_func=optimizer_func)
 ```
 
@@ -437,7 +437,7 @@ plot_cost = Ploter(test_title)
 
 
 def event_handler(event):
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         avg_cost_set = trainer.test(
             reader=test_reader, feed_order=feed_order)
 
@@ -473,7 +473,7 @@ trainer.train(
 传入`inference_program`和`params_dirname`来初始化一个预测器, `params_dirname`用来存放训练过程中的各个参数。
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
         inference_program, param_path=params_dirname, place=place)
 ```
 

05.recommender_system/README.md

@@ -382,7 +382,7 @@ test_reader = paddle.batch(
 Create a trainer that takes `train_program` as input and specify optimizer function.
 
 ```python
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=train_program, place=place, optimizer_func=optimizer_func)
 ```
 
@@ -412,7 +412,7 @@ plot_cost = Ploter(test_title)
 
 
 def event_handler(event):
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         avg_cost_set = trainer.test(
             reader=test_reader, feed_order=feed_order)
 
@@ -451,7 +451,7 @@ trainer.train(
 Initialize Inferencer with `inference_program` and `params_dirname` which is where we save params from training.
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
         inference_program, param_path=params_dirname, place=place)  
 ```
 

05.recommender_system/index.cn.html

@@ -451,7 +451,7 @@ test_reader = paddle.batch(
 训练器需要一个训练程序和一个训练优化函数。
 
 ```python
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=train_program, place=place, optimizer_func=optimizer_func)
 ```
 
@@ -479,7 +479,7 @@ plot_cost = Ploter(test_title)
 
 
 def event_handler(event):
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         avg_cost_set = trainer.test(
             reader=test_reader, feed_order=feed_order)
 
@@ -515,7 +515,7 @@ trainer.train(
 传入`inference_program`和`params_dirname`来初始化一个预测器, `params_dirname`用来存放训练过程中的各个参数。
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
         inference_program, param_path=params_dirname, place=place)
 ```
 

05.recommender_system/index.html

@@ -424,7 +424,7 @@ test_reader = paddle.batch(
 Create a trainer that takes `train_program` as input and specify optimizer function.
 
 ```python
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=train_program, place=place, optimizer_func=optimizer_func)
 ```
 
@@ -454,7 +454,7 @@ plot_cost = Ploter(test_title)
 
 
 def event_handler(event):
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         avg_cost_set = trainer.test(
             reader=test_reader, feed_order=feed_order)
 
@@ -493,7 +493,7 @@ trainer.train(
 Initialize Inferencer with `inference_program` and `params_dirname` which is where we save params from training.
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
         inference_program, param_path=params_dirname, place=place)  
 ```
 

05.recommender_system/train.py

@@ -160,7 +160,7 @@ def optimizer_func():
 def train(use_cuda, train_program, params_dirname):
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
 
-    trainer = fluid.Trainer(
+    trainer = Trainer(
         train_func=train_program, place=place, optimizer_func=optimizer_func)
 
     feed_order = [
@@ -169,7 +169,7 @@ def train(use_cuda, train_program, params_dirname):
     ]
 
     def event_handler(event):
-        if isinstance(event, fluid.EndStepEvent):
+        if isinstance(event, EndStepEvent):
             test_reader = paddle.batch(
                 paddle.dataset.movielens.test(), batch_size=BATCH_SIZE)
             avg_cost_set = trainer.test(
@@ -202,7 +202,7 @@ def train(use_cuda, train_program, params_dirname):
 
 def infer(use_cuda, inference_program, params_dirname):
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
-    inferencer = fluid.Inferencer(
+    inferencer = Inferencer(
         inference_program, param_path=params_dirname, place=place)
 
     # Use the first data from paddle.dataset.movielens.test() as input.

06.understand_sentiment/README.cn.md

@@ -249,7 +249,7 @@ train_reader = paddle.batch(
 训练器需要一个训练程序和一个训练优化函数。
 
 ```python
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=partial(train_program, word_dict),
     place=place,
     optimizer_func=optimizer_func)
@@ -272,7 +272,7 @@ feed_order = ['words', 'label']
 params_dirname = "understand_sentiment_conv.inference.model"
 
 def event_handler(event):
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         print("Step {0}, Epoch {1} Metrics {2}".format(
                 event.step, event.epoch, map(np.array, event.metrics)))
 
@@ -300,7 +300,7 @@ trainer.train(
 传入`inference_program`和`params_dirname`来初始化一个预测器, `params_dirname`用来存放训练过程中的各个参数。
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
         infer_func=partial(inference_program, word_dict), param_path=params_dirname, place=place)
 ```
 

06.understand_sentiment/README.md

@@ -255,7 +255,7 @@ train_reader = paddle.batch(
 Create a trainer that takes `train_program` as input and specify optimizer function.
 
 ```python
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=partial(train_program, word_dict),
     place=place,
     optimizer_func=optimizer_func)
@@ -279,7 +279,7 @@ For example, we can check the cost by `trainer.test` when `EndStepEvent` occurs
 params_dirname = "understand_sentiment_conv.inference.model"
 
 def event_handler(event):
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         print("Step {0}, Epoch {1} Metrics {2}".format(
                 event.step, event.epoch, map(np.array, event.metrics)))
 
@@ -307,7 +307,7 @@ trainer.train(
 Initialize Inferencer with `inference_program` and `params_dirname` which is where we save params from training.
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
         infer_func=partial(inference_program, word_dict),
         param_path=params_dirname,
         place=place)

06.understand_sentiment/index.cn.html

@@ -291,7 +291,7 @@ train_reader = paddle.batch(
 训练器需要一个训练程序和一个训练优化函数。
 
 ```python
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=partial(train_program, word_dict),
     place=place,
     optimizer_func=optimizer_func)
@@ -314,7 +314,7 @@ feed_order = ['words', 'label']
 params_dirname = "understand_sentiment_conv.inference.model"
 
 def event_handler(event):
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         print("Step {0}, Epoch {1} Metrics {2}".format(
                 event.step, event.epoch, map(np.array, event.metrics)))
 
@@ -342,7 +342,7 @@ trainer.train(
 传入`inference_program`和`params_dirname`来初始化一个预测器, `params_dirname`用来存放训练过程中的各个参数。
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
         infer_func=partial(inference_program, word_dict), param_path=params_dirname, place=place)
 ```
 

06.understand_sentiment/index.html

@@ -297,7 +297,7 @@ train_reader = paddle.batch(
 Create a trainer that takes `train_program` as input and specify optimizer function.
 
 ```python
-trainer = fluid.Trainer(
+trainer = fluid.contrib.trainer.Trainer(
     train_func=partial(train_program, word_dict),
     place=place,
     optimizer_func=optimizer_func)
@@ -321,7 +321,7 @@ For example, we can check the cost by `trainer.test` when `EndStepEvent` occurs
 params_dirname = "understand_sentiment_conv.inference.model"
 
 def event_handler(event):
-    if isinstance(event, fluid.EndStepEvent):
+    if isinstance(event, fluid.contrib.trainer.EndStepEvent):
         print("Step {0}, Epoch {1} Metrics {2}".format(
                 event.step, event.epoch, map(np.array, event.metrics)))
 
@@ -349,7 +349,7 @@ trainer.train(
 Initialize Inferencer with `inference_program` and `params_dirname` which is where we save params from training.
 
 ```python
-inferencer = fluid.Inferencer(
+inferencer = fluid.contrib.inferencer.Inferencer(
         infer_func=partial(inference_program, word_dict),
         param_path=params_dirname,
         place=place)

06.understand_sentiment/train_conv.py

@@ -83,7 +83,7 @@ def train(use_cuda, train_program, params_dirname):
     test_reader = paddle.batch(
         paddle.dataset.imdb.test(word_dict), batch_size=BATCH_SIZE)
 
-    trainer = fluid.Trainer(
+    trainer = Trainer(
         train_func=partial(train_program, word_dict),
         place=place,
         optimizer_func=optimizer_func)
@@ -91,7 +91,7 @@ def train(use_cuda, train_program, params_dirname):
     feed_order = ['words', 'label']
 
     def event_handler(event):
-        if isinstance(event, fluid.EndStepEvent):
+        if isinstance(event, EndStepEvent):
             if event.step % 10 == 0:
                 avg_cost, acc = trainer.test(
                     reader=test_reader, feed_order=feed_order)
@@ -102,7 +102,7 @@ def train(use_cuda, train_program, params_dirname):
                 print("Step {0}, Epoch {1} Metrics {2}".format(
                     event.step, event.epoch, map(np.array, event.metrics)))
 
-        elif isinstance(event, fluid.EndEpochEvent):
+        elif isinstance(event, EndEpochEvent):
             trainer.save_params(params_dirname)
 
     trainer.train(
@@ -116,7 +116,7 @@ def infer(use_cuda, inference_program, params_dirname=None):
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
     word_dict = paddle.dataset.imdb.word_dict()
 
-    inferencer = fluid.Inferencer(
+    inferencer = Inferencer(
         infer_func=partial(inference_program, word_dict),
         param_path=params_dirname,
         place=place)

06.understand_sentiment/train_dyn_rnn.py

@@ -100,7 +100,7 @@ def train(use_cuda, train_program, params_dirname):
     test_reader = paddle.batch(
         paddle.dataset.imdb.test(word_dict), batch_size=BATCH_SIZE)
 
-    trainer = fluid.Trainer(
+    trainer = Trainer(
         train_func=partial(train_program, word_dict),
         place=place,
         optimizer_func=optimizer_func)
@@ -108,7 +108,7 @@ def train(use_cuda, train_program, params_dirname):
     feed_order = ['words', 'label']
 
     def event_handler(event):
-        if isinstance(event, fluid.EndStepEvent):
+        if isinstance(event, EndStepEvent):
             if event.step % 10 == 0:
                 avg_cost, acc = trainer.test(
                     reader=test_reader, feed_order=feed_order)
@@ -119,7 +119,7 @@ def train(use_cuda, train_program, params_dirname):
                 print("Step {0}, Epoch {1} Metrics {2}".format(
                     event.step, event.epoch, map(np.array, event.metrics)))
 
-        elif isinstance(event, fluid.EndEpochEvent):
+        elif isinstance(event, EndEpochEvent):
             trainer.save_params(params_dirname)
 
     trainer.train(
@@ -133,7 +133,7 @@ def infer(use_cuda, inference_program, params_dirname=None):
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
     word_dict = paddle.dataset.imdb.word_dict()
 
-    inferencer = fluid.Inferencer(
+    inferencer = Inferencer(
         infer_func=partial(inference_program, word_dict),
         param_path=params_dirname,
         place=place)

06.understand_sentiment/train_stacked_lstm.py

@@ -91,7 +91,7 @@ def train(use_cuda, train_program, params_dirname):
     test_reader = paddle.batch(
         paddle.dataset.imdb.test(word_dict), batch_size=BATCH_SIZE)
 
-    trainer = fluid.Trainer(
+    trainer = Trainer(
         train_func=partial(train_program, word_dict),
         place=place,
         optimizer_func=optimizer_func)
@@ -99,7 +99,7 @@ def train(use_cuda, train_program, params_dirname):
     feed_order = ['words', 'label']
 
     def event_handler(event):
-        if isinstance(event, fluid.EndStepEvent):
+        if isinstance(event, EndStepEvent):
             if event.step % 10 == 0:
                 avg_cost, acc = trainer.test(
                     reader=test_reader, feed_order=feed_order)
@@ -110,7 +110,7 @@ def train(use_cuda, train_program, params_dirname):
                 print("Step {0}, Epoch {1} Metrics {2}".format(
                     event.step, event.epoch, map(np.array, event.metrics)))
 
-        elif isinstance(event, fluid.EndEpochEvent):
+        elif isinstance(event, EndEpochEvent):
             trainer.save_params(params_dirname)
 
     trainer.train(
@@ -124,7 +124,7 @@ def infer(use_cuda, inference_program, params_dirname=None):
     place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
     word_dict = paddle.dataset.imdb.word_dict()
 
-    inferencer = fluid.Inferencer(
+    inferencer = Inferencer(
         infer_func=partial(inference_program, word_dict),
         param_path=params_dirname,
         place=place)