Machine Learning Application Benchmark - MobileNet

This is a re-implementation of MobileNet as described in [1] changed to fulfill hardware constraints of the Vitis AI framework for inference on Xilinx FPGAs.

Pretrained weights are available from the model database as follows:

• ImageNet (with top layers): imagenet_mobilenet.h5
• ImageNet (without top layers): imagenet_mobilenet_notop.h5

Source file: /models/mobilenet_vitis.py

Usage:

models.mobilenet_vitis.mobilenet_vitis(
    input_tensor=None, 
    include_top=True, 
    weight_path=None, 
    return_tensor=False, 
    classes=1000, 
    classifier_activation="softmax",
    alpha=1.0, 
    depth_multiplier=1
)

Arguments:

• input_tensor: optional keras layer, like an input tensor.
• include_top: whether to include the top layers or top.
• weight_path: If not none, these weights will be loaded.
• return_tensor: Whether to return the network as tensor or as tf.keras.model (if true, weights will not be loaded).
• classes: By default the number of classes are 1000 (ImageNet). Only important include_top=True.
• classifier_activation: By default softmax (ImageNet). Only important if include_top=True.
• alpha (float, optional): controls the width of the network. - 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, default number of filters from the paper are used at each layer.
• depth_multiplier (int, optional): The number of depthwise convolution output channels for each input channel. Defaults to 1.

Returns:

The CNN architecture as tf.keras.model if return_tensor=False, otherwise as tf.keras.layers.

References:

[1] A. G. Howard et al., “MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications,” 2017.
[2] O. Russakovsky et al., “ImageNet Large Scale Visual Recognition Challenge,” International Journal of Computer Vision (IJCV), vol. 115, no. 3, pp. 211–252, 2015, doi: 10.1007/s11263-015-0816-y.