Merge pull request #465 from tensorlayer/models-mobilenetv1

tl.models.MobileNetV1

Author: wagamamaz

README.md

@@ -85,7 +85,7 @@ Examples can be found [in this folder](https://github.com/zsdonghao/tensorlayer/
  - VGG 19 (ImageNet). Classification task, see [tutorial_vgg19.py](https://github.com/zsdonghao/tensorlayer/blob/master/example/tutorial_vgg19.py).
  - InceptionV3 (ImageNet). Classification task, see [tutorial\_inceptionV3_tfslim.py](https://github.com/zsdonghao/tensorlayer/blob/master/example/tutorial_inceptionV3_tfslim.py).
  - SqueezeNet (ImageNet). Model compression, see [tl.models.SqueezeNetV1](https://github.com/zsdonghao/tensorlayer/blob/master/example/tutorial_models_squeezenetv1.py) or [tutorial_squeezenet.py](https://github.com/zsdonghao/tensorlayer/blob/master/example/tutorial_squeezenet.py)
- - MobileNet (ImageNet). Model compression, see [tutorial_mobilenet.py](https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_mobilenet.py).
+ - MobileNet (ImageNet). Model compression, see [tl.models.MobileNetV1](https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_models_mobilenetv1.py) or [tutorial_mobilenet.py](https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_mobilenet.py).
  - BinaryNet. Model compression, see [mnist](https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_binarynet_mnist_cnn.py) [cifar10](https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_binarynet_cifar10_tfrecord.py).
  - Ternary Weight Network. Model compression, see [mnist](https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_ternaryweight_mnist_cnn.py) [cifar10](https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_ternaryweight_cifar10_tfrecord.py).
  - DoReFa-Net. Model compression, see [mnist](https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_dorefanet_mnist_cnn.py) [cifar10](https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_dorefanet_cifar10_tfrecord.py).

docs/index.rst

@@ -52,6 +52,7 @@ method, this part of the documentation is for you.
   modules/files
   modules/visualize
   modules/activation
+  modules/models
   modules/distributed
 
 

docs/modules/models.rst

@@ -0,0 +1,26 @@
+API - Models
+================================
+
+TensorLayer provides many pretrained models, you can easily use the whole or a part of the pretrained models via these APIs.
+
+.. automodule:: tensorlayer.models
+
+.. autosummary::
+
+    VGG16
+    SqueezeNetV1
+    MobileNetV1
+
+VGG16
+----------------------
+
+.. autoclass:: VGG16
+
+SqueezeNetV1
+----------------
+.. autoclass:: SqueezeNetV1
+
+MobileNetV1
+----------------
+
+.. autoclass:: MobileNetV1

docs/user/example.rst

@@ -22,7 +22,7 @@ Computer Vision
  - VGG 19 (ImageNet). Classification task, see `tutorial_vgg19.py <https://github.com/zsdonghao/tensorlayer/blob/master/example/tutorial_vgg19.py>`_.
  - InceptionV3 (ImageNet). Classification task, see `tutorial_inceptionV3_tfslim.py <https://github.com/zsdonghao/tensorlayer/blob/master/example/tutorial_inceptionV3_tfslim.py>`_.
  - SqueezeNet (ImageNet). Model compression, see `tl.models.SqueezeNetV1 <https://github.com/zsdonghao/tensorlayer/blob/master/example/tutorial_models_squeezenetv1.py>`__ or `tutorial_squeezenet.py <https://github.com/zsdonghao/tensorlayer/blob/master/example/tutorial_squeezenet.py>`_.
- - MobileNet (ImageNet). Model compression, see `tutorial_mobilenet.py <https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_mobilenet.py>`__.
+ - MobileNet (ImageNet). Model compression, see `tl.models.MobileNetV1 <https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_models_mobilenetv1.py>`__ or `tutorial_mobilenet.py <https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_mobilenet.py>`__.
  - BinaryNet. Model compression, see `mnist <https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_binarynet_mnist_cnn.py>`__ `cifar10 <https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_binarynet_cifar10_tfrecord.py>`__.
  - Ternary Weight Network. Model compression, see `mnist <https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_ternaryweight_mnist_cnn.py>`__ `cifar10 <https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_ternaryweight_cifar10_tfrecord.py>`__.
  - DoReFa-Net. Model compression, see `mnist <https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_dorefanet_mnist_cnn.py>`__ `cifar10 <https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_dorefanet_cifar10_tfrecord.py>`__.

example/tutorial_models_mobilenetv1.py

@@ -0,0 +1,41 @@
+#! /usr/bin/python
+# -*- coding: utf-8 -*-
+"""
+MobileNetV1 for ImageNet using TL models
+
+- mobilenetv2 : https://github.com/tensorflow/models/tree/master/research/slim/nets/mobilenet
+- tf.slim     : https://github.com/tensorflow/models/tree/master/research/slim#pre-trained-models
+"""
+
+import time
+import numpy as np
+import tensorflow as tf
+import tensorlayer as tl
+from tensorlayer.models.imagenet_classes import class_names
+
+x = tf.placeholder(tf.float32, [None, 224, 224, 3])
+
+# get the whole model
+mobilenetv1 = tl.models.MobileNetV1(x)
+
+# restore pre-trained parameters
+sess = tf.InteractiveSession()
+
+mobilenetv1.restore_params(sess)
+
+probs = tf.nn.softmax(mobilenetv1.outputs)
+
+mobilenetv1.print_params(False)
+
+mobilenetv1.print_layers()
+
+img1 = tl.vis.read_image('data/tiger.jpeg')
+img1 = tl.prepro.imresize(img1, (224, 224)) / 255
+
+_ = sess.run(probs, feed_dict={x: [img1]})[0]  # 1st time takes time to compile
+start_time = time.time()
+prob = sess.run(probs, feed_dict={x: [img1]})[0]
+print("  End time : %.5ss" % (time.time() - start_time))
+preds = (np.argsort(prob)[::-1])[0:5]
+for p in preds:
+    print(class_names[p], prob[p])

tensorlayer/layers/shape.py

@@ -37,16 +37,13 @@ class FlattenLayer(Layer):
     def __init__(
             self,
             prev_layer,
-            name='flatten_layer',
+            name='flatten',
     ):
         Layer.__init__(self, prev_layer=prev_layer, name=name)
         self.inputs = prev_layer.outputs
         self.outputs = flatten_reshape(self.inputs, name=name)
         self.n_units = int(self.outputs.get_shape()[-1])
         logging.info("FlattenLayer %s: %d" % (self.name, self.n_units))
-        # self.all_layers = list(layer.all_layers)
-        # self.all_params = list(layer.all_params)
-        # self.all_drop = dict(layer.all_drop)
         self.all_layers.append(self.outputs)
 
 
@@ -76,15 +73,12 @@ def __init__(
             self,
             prev_layer,
             shape,
-            name='reshape_layer',
+            name='reshape',
     ):
         Layer.__init__(self, prev_layer=prev_layer, name=name)
         self.inputs = prev_layer.outputs
         self.outputs = tf.reshape(self.inputs, shape=shape, name=name)
         logging.info("ReshapeLayer %s: %s" % (self.name, self.outputs.get_shape()))
-        # self.all_layers = list(layer.all_layers)
-        # self.all_params = list(layer.all_params)
-        # self.all_drop = dict(layer.all_drop)
         self.all_layers.append(self.outputs)
 
 
@@ -122,10 +116,5 @@ def __init__(
         assert perm is not None
 
         logging.info("TransposeLayer  %s: perm:%s" % (self.name, perm))
-        # with tf.variable_scope(name) as vs:
         self.outputs = tf.transpose(self.inputs, perm=perm, name=name)
-        # self.all_layers = list(layer.all_layers)
-        # self.all_params = list(layer.all_params)
-        # self.all_drop = dict(layer.all_drop)
         self.all_layers.append(self.outputs)
-        # self.all_params.extend( variables )

tensorlayer/models/__init__.py

@@ -2,3 +2,4 @@
 
 from .vgg16 import VGG16
 from .squeezenetv1 import SqueezeNetV1
+from .mobilenetv1 import MobileNetV1

tensorlayer/models/mobilenetv1.py

@@ -0,0 +1,171 @@
+# -*- coding: utf-8 -*-
+"""
+MobileNet for ImageNet.
+"""
+
+import os
+# import numpy as np
+import tensorflow as tf
+from .. import _logging as logging
+from ..layers import (Layer, BatchNormLayer, Conv2d, DepthwiseConv2d, FlattenLayer, GlobalMeanPool2d, InputLayer, ReshapeLayer)
+from ..files import maybe_download_and_extract, assign_params, load_npz
+
+__all__ = [
+    'MobileNetV1',
+]
+
+
+class MobileNetV1(Layer):
+    """Pre-trained MobileNetV1 model.
+
+    Parameters
+    ------------
+    x : placeholder
+        shape [None, 224, 224, 3], value range [0, 1].
+    end_with : str
+        The end point of the model [conv, depth1, depth2 ... depth13, globalmeanpool, out]. Default ``out`` i.e. the whole model.
+    is_train : boolean
+        Whether the model is used for training i.e. enable dropout.
+    reuse : boolean
+        Whether to reuse the model.
+
+    Examples
+    ---------
+    Classify ImageNet classes, see `tutorial_models_mobilenetv1.py <https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_models_mobilenetv1.py>__`
+    >>> x = tf.placeholder(tf.float32, [None, 224, 224, 3])
+    >>> # get the whole model
+    >>> net = tl.models.MobileNetV1(x)
+    >>> # restore pre-trained parameters
+    >>> sess = tf.InteractiveSession()
+    >>> net.restore_params(sess)
+    >>> # use for inferencing
+    >>> probs = tf.nn.softmax(net.outputs)
+
+    Extract features and Train a classifier with 100 classes
+    >>> x = tf.placeholder(tf.float32, [None, 224, 224, 3])
+    >>> # get model without the last layer
+    >>> cnn = tl.models.MobileNetV1(x, end_with='reshape')
+    >>> # add one more layer
+    >>> net = Conv2d(cnn, 100, (1, 1), (1, 1), name='out')
+    >>> net = FlattenLayer(net, name='flatten')
+    >>> # initialize all parameters
+    >>> sess = tf.InteractiveSession()
+    >>> tl.layers.initialize_global_variables(sess)
+    >>> # restore pre-trained parameters
+    >>> cnn.restore_params(sess)
+    >>> # train your own classifier (only update the last layer)
+    >>> train_params = tl.layers.get_variables_with_name('output')
+
+    Reuse model
+    >>> x1 = tf.placeholder(tf.float32, [None, 224, 224, 3])
+    >>> x2 = tf.placeholder(tf.float32, [None, 224, 224, 3])
+    >>> # get VGG without the last layer
+    >>> net1 = tl.models.MobileNetV1(x1, end_with='reshape')
+    >>> # reuse the parameters with different input
+    >>> net2 = tl.models.MobileNetV1(x2, end_with='reshape', reuse=True)
+    >>> # restore pre-trained parameters (as they share parameters, we don’t need to restore net2)
+    >>> sess = tf.InteractiveSession()
+    >>> net1.restore_params(sess)
+
+    """
+
+    def __init__(self, x, end_with='out', is_train=False, reuse=None):
+
+        self.net = self.mobilenetv1(x, end_with, is_train, reuse)
+        self.outputs = self.net.outputs
+        self.all_params = self.net.all_params
+        self.all_layers = self.net.all_layers
+        self.all_drop = self.net.all_drop
+        self.print_layers = self.net.print_layers
+        self.print_params = self.net.print_params
+
+    # @classmethod
+    def mobilenetv1(self, x, end_with='out', is_train=False, reuse=None):
+        with tf.variable_scope("mobilenetv1", reuse=reuse):
+            n = InputLayer(x)
+            n = self.conv_block(n, 32, strides=(2, 2), is_train=is_train, name="conv")
+            if end_with in n.outputs.name:
+                return n
+            n = self.depthwise_conv_block(n, 64, is_train=is_train, name="depth1")
+            if end_with in n.outputs.name:
+                return n
+
+            n = self.depthwise_conv_block(n, 128, strides=(2, 2), is_train=is_train, name="depth2")
+            if end_with in n.outputs.name:
+                return n
+            n = self.depthwise_conv_block(n, 128, is_train=is_train, name="depth3")
+            if end_with in n.outputs.name:
+                return n
+
+            n = self.depthwise_conv_block(n, 256, strides=(2, 2), is_train=is_train, name="depth4")
+            if end_with in n.outputs.name:
+                return n
+            n = self.depthwise_conv_block(n, 256, is_train=is_train, name="depth5")
+            if end_with in n.outputs.name:
+                return n
+
+            n = self.depthwise_conv_block(n, 512, strides=(2, 2), is_train=is_train, name="depth6")
+            if end_with in n.outputs.name:
+                return n
+            n = self.depthwise_conv_block(n, 512, is_train=is_train, name="depth7")
+            if end_with in n.outputs.name:
+                return n
+            n = self.depthwise_conv_block(n, 512, is_train=is_train, name="depth8")
+            if end_with in n.outputs.name:
+                return n
+            n = self.depthwise_conv_block(n, 512, is_train=is_train, name="depth9")
+            if end_with in n.outputs.name:
+                return n
+            n = self.depthwise_conv_block(n, 512, is_train=is_train, name="depth10")
+            if end_with in n.outputs.name:
+                return n
+            n = self.depthwise_conv_block(n, 512, is_train=is_train, name="depth11")
+            if end_with in n.outputs.name:
+                return n
+
+            n = self.depthwise_conv_block(n, 1024, strides=(2, 2), is_train=is_train, name="depth12")
+            if end_with in n.outputs.name:
+                return n
+            n = self.depthwise_conv_block(n, 1024, is_train=is_train, name="depth13")
+            if end_with in n.outputs.name:
+                return n
+
+            n = GlobalMeanPool2d(n, name='globalmeanpool')
+            if end_with in n.outputs.name:
+                return n
+            # n = DropoutLayer(n, 1-1e-3, True, is_train, name='drop')
+            # n = DenseLayer(n, 1000, act=tf.identity, name='output')   # equal
+            n = ReshapeLayer(n, [-1, 1, 1, 1024], name='reshape')
+            if end_with in n.outputs.name:
+                return n
+            n = Conv2d(n, 1000, (1, 1), (1, 1), name='out')
+            n = FlattenLayer(n, name='flatten')
+            if end_with == 'out':
+                return n
+
+            raise Exception("end_with : conv, depth1, depth2 ... depth13, globalmeanpool, out")
+
+    @classmethod
+    def conv_block(cls, n, n_filter, filter_size=(3, 3), strides=(1, 1), is_train=False, name='conv_block'):
+        # ref: https://github.com/keras-team/keras/blob/master/keras/applications/mobilenet.py
+        with tf.variable_scope(name):
+            n = Conv2d(n, n_filter, filter_size, strides, b_init=None, name='conv')
+            n = BatchNormLayer(n, act=tf.nn.relu6, is_train=is_train, name='batchnorm')
+        return n
+
+    @classmethod
+    def depthwise_conv_block(cls, n, n_filter, strides=(1, 1), is_train=False, name="depth_block"):
+        with tf.variable_scope(name):
+            n = DepthwiseConv2d(n, (3, 3), strides, b_init=None, name='depthwise')
+            n = BatchNormLayer(n, act=tf.nn.relu6, is_train=is_train, name='batchnorm1')
+            n = Conv2d(n, n_filter, (1, 1), (1, 1), b_init=None, name='conv')
+            n = BatchNormLayer(n, act=tf.nn.relu6, is_train=is_train, name='batchnorm2')
+        return n
+
+    def restore_params(self, sess, path='models'):
+        logging.info("Restore pre-trained parameters")
+        maybe_download_and_extract(
+            'mobilenet.npz', path, 'https://github.com/tensorlayer/pretrained-models/raw/master/models/', expected_bytes=25600116)  # ls -al
+        params = load_npz(name=os.path.join(path, 'mobilenet.npz'))
+        assign_params(sess, params[:len(self.net.all_params)], self.net)
+        del params

tensorlayer/models/squeezenetv1.py

@@ -31,7 +31,7 @@ class SqueezeNetV1(Layer):
 
     Examples
     ---------
-    Classify ImageNet classes, see `tutorial_models_vgg16.py <https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_models_vgg16.py>__`
+    Classify ImageNet classes, see `tutorial_models_squeezenetv1.py <https://github.com/tensorlayer/tensorlayer/blob/master/example/tutorial_models_squeezenetv1.py>__`
     >>> x = tf.placeholder(tf.float32, [None, 224, 224, 3])
     >>> # get the whole model
     >>> net = tl.models.SqueezeNetV1(x)