# Copyright 2018 The TensorFlow 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. # ============================================================================== """MobileNet v2 models for TF-Keras. MobileNetV2 is a general architecture and can be used for multiple use cases. Depending on the use case, it can use different input layer size and different width factors. This allows different width models to reduce the number of multiply-adds and thereby reduce inference cost on mobile devices. MobileNetV2 is very similar to the original MobileNet, except that it uses inverted residual blocks with bottlenecking features. It has a drastically lower parameter count than the original MobileNet. MobileNets support any input size greater than 32 x 32, with larger image sizes offering better performance. The number of parameters and number of multiply-adds can be modified by using the `alpha` parameter, which increases/decreases the number of filters in each layer. By altering the image size and `alpha` parameter, all 22 models from the paper can be built, with ImageNet weights provided. The paper demonstrates the performance of MobileNets using `alpha` values of 1.0 (also called 100 % MobileNet), 0.35, 0.5, 0.75, 1.0, 1.3, and 1.4 For each of these `alpha` values, weights for 5 different input image sizes are provided (224, 192, 160, 128, and 96). The following table describes the performance of MobileNet on various input sizes: ------------------------------------------------------------------------ MACs stands for Multiply Adds Classification Checkpoint|MACs (M)|Parameters (M)|Top 1 Accuracy|Top 5 Accuracy --------------------------|------------|---------------|---------|------------ | [mobilenet_v2_1.4_224] | 582 | 6.06 | 75.0 | 92.5 | | [mobilenet_v2_1.3_224] | 509 | 5.34 | 74.4 | 92.1 | | [mobilenet_v2_1.0_224] | 300 | 3.47 | 71.8 | 91.0 | | [mobilenet_v2_1.0_192] | 221 | 3.47 | 70.7 | 90.1 | | [mobilenet_v2_1.0_160] | 154 | 3.47 | 68.8 | 89.0 | | [mobilenet_v2_1.0_128] | 99 | 3.47 | 65.3 | 86.9 | | [mobilenet_v2_1.0_96] | 56 | 3.47 | 60.3 | 83.2 | | [mobilenet_v2_0.75_224] | 209 | 2.61 | 69.8 | 89.6 | | [mobilenet_v2_0.75_192] | 153 | 2.61 | 68.7 | 88.9 | | [mobilenet_v2_0.75_160] | 107 | 2.61 | 66.4 | 87.3 | | [mobilenet_v2_0.75_128] | 69 | 2.61 | 63.2 | 85.3 | | [mobilenet_v2_0.75_96] | 39 | 2.61 | 58.8 | 81.6 | | [mobilenet_v2_0.5_224] | 97 | 1.95 | 65.4 | 86.4 | | [mobilenet_v2_0.5_192] | 71 | 1.95 | 63.9 | 85.4 | | [mobilenet_v2_0.5_160] | 50 | 1.95 | 61.0 | 83.2 | | [mobilenet_v2_0.5_128] | 32 | 1.95 | 57.7 | 80.8 | | [mobilenet_v2_0.5_96] | 18 | 1.95 | 51.2 | 75.8 | | [mobilenet_v2_0.35_224] | 59 | 1.66 | 60.3 | 82.9 | | [mobilenet_v2_0.35_192] | 43 | 1.66 | 58.2 | 81.2 | | [mobilenet_v2_0.35_160] | 30 | 1.66 | 55.7 | 79.1 | | [mobilenet_v2_0.35_128] | 20 | 1.66 | 50.8 | 75.0 | | [mobilenet_v2_0.35_96] | 11 | 1.66 | 45.5 | 70.4 | Reference: - [MobileNetV2: Inverted Residuals and Linear Bottlenecks]( https://arxiv.org/abs/1801.04381) (CVPR 2018) """ import tensorflow.compat.v2 as tf from tf_keras.src import backend from tf_keras.src.applications import imagenet_utils from tf_keras.src.engine import training from tf_keras.src.layers import VersionAwareLayers from tf_keras.src.utils import data_utils from tf_keras.src.utils import layer_utils # isort: off from tensorflow.python.platform import tf_logging as logging from tensorflow.python.util.tf_export import keras_export BASE_WEIGHT_PATH = ( "https://storage.googleapis.com/tensorflow/keras-applications/mobilenet_v2/" ) layers = None @keras_export( "keras.applications.mobilenet_v2.MobileNetV2", "keras.applications.MobileNetV2", ) def MobileNetV2( input_shape=None, alpha=1.0, include_top=True, weights="imagenet", input_tensor=None, pooling=None, classes=1000, classifier_activation="softmax", **kwargs, ): """Instantiates the MobileNetV2 architecture. MobileNetV2 is very similar to the original MobileNet, except that it uses inverted residual blocks with bottlenecking features. It has a drastically lower parameter count than the original MobileNet. MobileNets support any input size greater than 32 x 32, with larger image sizes offering better performance. Reference: - [MobileNetV2: Inverted Residuals and Linear Bottlenecks]( https://arxiv.org/abs/1801.04381) (CVPR 2018) This function returns a TF-Keras image classification model, optionally loaded with weights pre-trained on ImageNet. For image classification use cases, see [this page for detailed examples]( https://keras.io/api/applications/#usage-examples-for-image-classification-models). For transfer learning use cases, make sure to read the [guide to transfer learning & fine-tuning]( https://keras.io/guides/transfer_learning/). Note: each TF-Keras Application expects a specific kind of input preprocessing. For MobileNetV2, call `tf.keras.applications.mobilenet_v2.preprocess_input` on your inputs before passing them to the model. `mobilenet_v2.preprocess_input` will scale input pixels between -1 and 1. Args: input_shape: Optional shape tuple, to be specified if you would like to use a model with an input image resolution that is not (224, 224, 3). It should have exactly 3 inputs channels (224, 224, 3). You can also omit this option if you would like to infer input_shape from an input_tensor. If you choose to include both input_tensor and input_shape then input_shape will be used if they match, if the shapes do not match then we will throw an error. E.g. `(160, 160, 3)` would be one valid value. alpha: Float, larger than zero, controls the width of the network. This is known as the width multiplier in the MobileNetV2 paper, but the name is kept for consistency with `applications.MobileNetV1` model in TF-Keras. - If `alpha` < 1.0, proportionally decreases the number of filters in each layer. - If `alpha` > 1.0, proportionally increases the number of filters in each layer. - If `alpha` = 1.0, default number of filters from the paper are used at each layer. include_top: Boolean, whether to include the fully-connected layer at the top of the network. Defaults to `True`. weights: String, one of `None` (random initialization), 'imagenet' (pre-training on ImageNet), or the path to the weights file to be loaded. input_tensor: Optional TF-Keras tensor (i.e. output of `layers.Input()`) to use as image input for the model. pooling: String, optional pooling mode for feature extraction when `include_top` is `False`. - `None` means that the output of the model will be the 4D tensor output of the last convolutional block. - `avg` means that global average pooling will be applied to the output of the last convolutional block, and thus the output of the model will be a 2D tensor. - `max` means that global max pooling will be applied. classes: Optional integer number of classes to classify images into, only to be specified if `include_top` is True, and if no `weights` argument is specified. classifier_activation: A `str` or callable. The activation function to use on the "top" layer. Ignored unless `include_top=True`. Set `classifier_activation=None` to return the logits of the "top" layer. When loading pretrained weights, `classifier_activation` can only be `None` or `"softmax"`. **kwargs: For backwards compatibility only. Returns: A `keras.Model` instance. """ global layers if "layers" in kwargs: layers = kwargs.pop("layers") else: layers = VersionAwareLayers() if kwargs: raise ValueError(f"Unknown argument(s): {kwargs}") if not (weights in {"imagenet", None} or tf.io.gfile.exists(weights)): raise ValueError( "The `weights` argument should be either " "`None` (random initialization), `imagenet` " "(pre-training on ImageNet), " "or the path to the weights file to be loaded. " f"Received `weights={weights}`" ) if weights == "imagenet" and include_top and classes != 1000: raise ValueError( 'If using `weights` as `"imagenet"` with `include_top` ' f"as true, `classes` should be 1000. Received `classes={classes}`" ) # Determine proper input shape and default size. # If both input_shape and input_tensor are used, they should match if input_shape is not None and input_tensor is not None: try: is_input_t_tensor = backend.is_keras_tensor(input_tensor) except ValueError: try: is_input_t_tensor = backend.is_keras_tensor( layer_utils.get_source_inputs(input_tensor) ) except ValueError: raise ValueError( f"input_tensor: {input_tensor}" "is not type input_tensor. " f"Received `type(input_tensor)={type(input_tensor)}`" ) if is_input_t_tensor: if backend.image_data_format() == "channels_first": if backend.int_shape(input_tensor)[1] != input_shape[1]: raise ValueError( "input_shape[1] must equal shape(input_tensor)[1] " "when `image_data_format` is `channels_first`; " "Received `input_tensor.shape=" f"{input_tensor.shape}`" f", `input_shape={input_shape}`" ) else: if backend.int_shape(input_tensor)[2] != input_shape[1]: raise ValueError( "input_tensor.shape[2] must equal input_shape[1]; " "Received `input_tensor.shape=" f"{input_tensor.shape}`, " f"`input_shape={input_shape}`" ) else: raise ValueError( "input_tensor is not a TF-Keras tensor; " f"Received `input_tensor={input_tensor}`" ) # If input_shape is None, infer shape from input_tensor. if input_shape is None and input_tensor is not None: try: backend.is_keras_tensor(input_tensor) except ValueError: raise ValueError( "input_tensor must be a valid TF-Keras tensor type; " f"Received {input_tensor} of type {type(input_tensor)}" ) if input_shape is None and not backend.is_keras_tensor(input_tensor): default_size = 224 elif input_shape is None and backend.is_keras_tensor(input_tensor): if backend.image_data_format() == "channels_first": rows = backend.int_shape(input_tensor)[2] cols = backend.int_shape(input_tensor)[3] else: rows = backend.int_shape(input_tensor)[1] cols = backend.int_shape(input_tensor)[2] if rows == cols and rows in [96, 128, 160, 192, 224]: default_size = rows else: default_size = 224 # If input_shape is None and no input_tensor elif input_shape is None: default_size = 224 # If input_shape is not None, assume default size. else: if backend.image_data_format() == "channels_first": rows = input_shape[1] cols = input_shape[2] else: rows = input_shape[0] cols = input_shape[1] if rows == cols and rows in [96, 128, 160, 192, 224]: default_size = rows else: default_size = 224 input_shape = imagenet_utils.obtain_input_shape( input_shape, default_size=default_size, min_size=32, data_format=backend.image_data_format(), require_flatten=include_top, weights=weights, ) if backend.image_data_format() == "channels_last": row_axis, col_axis = (0, 1) else: row_axis, col_axis = (1, 2) rows = input_shape[row_axis] cols = input_shape[col_axis] if weights == "imagenet": if alpha not in [0.35, 0.50, 0.75, 1.0, 1.3, 1.4]: raise ValueError( "If imagenet weights are being loaded, " "alpha must be one of `0.35`, `0.50`, `0.75`, " "`1.0`, `1.3` or `1.4` only;" f" Received `alpha={alpha}`" ) if rows != cols or rows not in [96, 128, 160, 192, 224]: rows = 224 logging.warning( "`input_shape` is undefined or non-square, " "or `rows` is not in [96, 128, 160, 192, 224]. " "Weights for input shape (224, 224) will be " "loaded as the default." ) if input_tensor is None: img_input = layers.Input(shape=input_shape) else: if not backend.is_keras_tensor(input_tensor): img_input = layers.Input(tensor=input_tensor, shape=input_shape) else: img_input = input_tensor channel_axis = 1 if backend.image_data_format() == "channels_first" else -1 first_block_filters = _make_divisible(32 * alpha, 8) x = layers.Conv2D( first_block_filters, kernel_size=3, strides=(2, 2), padding="same", use_bias=False, name="Conv1", )(img_input) x = layers.BatchNormalization( axis=channel_axis, epsilon=1e-3, momentum=0.999, name="bn_Conv1" )(x) x = layers.ReLU(6.0, name="Conv1_relu")(x) x = _inverted_res_block( x, filters=16, alpha=alpha, stride=1, expansion=1, block_id=0 ) x = _inverted_res_block( x, filters=24, alpha=alpha, stride=2, expansion=6, block_id=1 ) x = _inverted_res_block( x, filters=24, alpha=alpha, stride=1, expansion=6, block_id=2 ) x = _inverted_res_block( x, filters=32, alpha=alpha, stride=2, expansion=6, block_id=3 ) x = _inverted_res_block( x, filters=32, alpha=alpha, stride=1, expansion=6, block_id=4 ) x = _inverted_res_block( x, filters=32, alpha=alpha, stride=1, expansion=6, block_id=5 ) x = _inverted_res_block( x, filters=64, alpha=alpha, stride=2, expansion=6, block_id=6 ) x = _inverted_res_block( x, filters=64, alpha=alpha, stride=1, expansion=6, block_id=7 ) x = _inverted_res_block( x, filters=64, alpha=alpha, stride=1, expansion=6, block_id=8 ) x = _inverted_res_block( x, filters=64, alpha=alpha, stride=1, expansion=6, block_id=9 ) x = _inverted_res_block( x, filters=96, alpha=alpha, stride=1, expansion=6, block_id=10 ) x = _inverted_res_block( x, filters=96, alpha=alpha, stride=1, expansion=6, block_id=11 ) x = _inverted_res_block( x, filters=96, alpha=alpha, stride=1, expansion=6, block_id=12 ) x = _inverted_res_block( x, filters=160, alpha=alpha, stride=2, expansion=6, block_id=13 ) x = _inverted_res_block( x, filters=160, alpha=alpha, stride=1, expansion=6, block_id=14 ) x = _inverted_res_block( x, filters=160, alpha=alpha, stride=1, expansion=6, block_id=15 ) x = _inverted_res_block( x, filters=320, alpha=alpha, stride=1, expansion=6, block_id=16 ) # no alpha applied to last conv as stated in the paper: # if the width multiplier is greater than 1 we increase the number of output # channels. if alpha > 1.0: last_block_filters = _make_divisible(1280 * alpha, 8) else: last_block_filters = 1280 x = layers.Conv2D( last_block_filters, kernel_size=1, use_bias=False, name="Conv_1" )(x) x = layers.BatchNormalization( axis=channel_axis, epsilon=1e-3, momentum=0.999, name="Conv_1_bn" )(x) x = layers.ReLU(6.0, name="out_relu")(x) if include_top: x = layers.GlobalAveragePooling2D()(x) imagenet_utils.validate_activation(classifier_activation, weights) x = layers.Dense( classes, activation=classifier_activation, name="predictions" )(x) else: if pooling == "avg": x = layers.GlobalAveragePooling2D()(x) elif pooling == "max": x = layers.GlobalMaxPooling2D()(x) # Ensure that the model takes into account any potential predecessors of # `input_tensor`. if input_tensor is not None: inputs = layer_utils.get_source_inputs(input_tensor) else: inputs = img_input # Create model. model = training.Model(inputs, x, name=f"mobilenetv2_{alpha:0.2f}_{rows}") # Load weights. if weights == "imagenet": if include_top: model_name = ( "mobilenet_v2_weights_tf_dim_ordering_tf_kernels_" + str(float(alpha)) + "_" + str(rows) + ".h5" ) weight_path = BASE_WEIGHT_PATH + model_name weights_path = data_utils.get_file( model_name, weight_path, cache_subdir="models" ) else: model_name = ( "mobilenet_v2_weights_tf_dim_ordering_tf_kernels_" + str(float(alpha)) + "_" + str(rows) + "_no_top" + ".h5" ) weight_path = BASE_WEIGHT_PATH + model_name weights_path = data_utils.get_file( model_name, weight_path, cache_subdir="models" ) model.load_weights(weights_path) elif weights is not None: model.load_weights(weights) return model def _inverted_res_block(inputs, expansion, stride, alpha, filters, block_id): """Inverted ResNet block.""" channel_axis = 1 if backend.image_data_format() == "channels_first" else -1 in_channels = backend.int_shape(inputs)[channel_axis] pointwise_conv_filters = int(filters * alpha) # Ensure the number of filters on the last 1x1 convolution is divisible by # 8. pointwise_filters = _make_divisible(pointwise_conv_filters, 8) x = inputs prefix = f"block_{block_id}_" if block_id: # Expand with a pointwise 1x1 convolution. x = layers.Conv2D( expansion * in_channels, kernel_size=1, padding="same", use_bias=False, activation=None, name=prefix + "expand", )(x) x = layers.BatchNormalization( axis=channel_axis, epsilon=1e-3, momentum=0.999, name=prefix + "expand_BN", )(x) x = layers.ReLU(6.0, name=prefix + "expand_relu")(x) else: prefix = "expanded_conv_" # Depthwise 3x3 convolution. if stride == 2: x = layers.ZeroPadding2D( padding=imagenet_utils.correct_pad(x, 3), name=prefix + "pad" )(x) x = layers.DepthwiseConv2D( kernel_size=3, strides=stride, activation=None, use_bias=False, padding="same" if stride == 1 else "valid", name=prefix + "depthwise", )(x) x = layers.BatchNormalization( axis=channel_axis, epsilon=1e-3, momentum=0.999, name=prefix + "depthwise_BN", )(x) x = layers.ReLU(6.0, name=prefix + "depthwise_relu")(x) # Project with a pointwise 1x1 convolution. x = layers.Conv2D( pointwise_filters, kernel_size=1, padding="same", use_bias=False, activation=None, name=prefix + "project", )(x) x = layers.BatchNormalization( axis=channel_axis, epsilon=1e-3, momentum=0.999, name=prefix + "project_BN", )(x) if in_channels == pointwise_filters and stride == 1: return layers.Add(name=prefix + "add")([inputs, x]) return x def _make_divisible(v, divisor, min_value=None): if min_value is None: min_value = divisor new_v = max(min_value, int(v + divisor / 2) // divisor * divisor) # Make sure that round down does not go down by more than 10%. if new_v < 0.9 * v: new_v += divisor return new_v @keras_export("keras.applications.mobilenet_v2.preprocess_input") def preprocess_input(x, data_format=None): return imagenet_utils.preprocess_input( x, data_format=data_format, mode="tf" ) @keras_export("keras.applications.mobilenet_v2.decode_predictions") def decode_predictions(preds, top=5): return imagenet_utils.decode_predictions(preds, top=top) preprocess_input.__doc__ = imagenet_utils.PREPROCESS_INPUT_DOC.format( mode="", ret=imagenet_utils.PREPROCESS_INPUT_RET_DOC_TF, error=imagenet_utils.PREPROCESS_INPUT_ERROR_DOC, ) decode_predictions.__doc__ = imagenet_utils.decode_predictions.__doc__