tf-deformable-conv-layer:可变形卷积层的TensorFlow实现

# TensorFlow implementation of Deformable Convolution This is a TensorFlow implementation of the following paper: > Dai, Jifeng, Haozhi Qi, Yuwen Xiong, Yi Li, Guodong Zhang, Han Hu, and Yichen Wei. 2017. > Deformable Convolutional Networks. arXiv [cs.CV]. arXiv. http://arxiv.org/abs/1703.06211 The code can only run in the [Eager Execution](https://www.tensorflow.org/guide/eager). Rotated training image <img src="images/rotated_image.png" width="300px"> Sampling locations <img src="images/sampling_locations.png" width="330px"> ## Basic Usage `DeformableConvLayer` is a custom Keras layer, so you can use it like any other standard layer, such as `Dense`, `Conv2D`. This is a simple example: ```python inputs = tf.zeros([16, 28, 28, 3]) model = tf.keras.Sequential() model.add(DeformableConvLayer(filters=6, kernel_size=3, strides=1, padding='valid', dilation_rate=1, num_deformable_group=1)) model.add(Conv2D(6, 3)) outputs = model(inputs) ```