tfm.vision.decoders.FPN

Creates a Feature Pyramid Network (FPN).

tfm.vision.decoders.FPN(
    input_specs: Mapping[str, tf.TensorShape],
    min_level: int = 3,
    max_level: int = 7,
    num_filters: int = 256,
    fusion_type: str = 'sum',
    use_separable_conv: bool = False,
    use_keras_layer: bool = False,
    activation: str = 'relu',
    use_sync_bn: bool = False,
    norm_momentum: float = 0.99,
    norm_epsilon: float = 0.001,
    kernel_initializer: str = 'VarianceScaling',
    kernel_regularizer: Optional[tf_keras.regularizers.Regularizer] = None,
    bias_regularizer: Optional[tf_keras.regularizers.Regularizer] = None,
    **kwargs
)

This implements the paper: Tsung-Yi Lin, Piotr Dollar, Ross Girshick, Kaiming He, Bharath Hariharan, and Serge Belongie. Feature Pyramid Networks for Object Detection. (https://arxiv.org/pdf/1612.03144)

input_specs A dict of input specifications. A dictionary consists of {level: TensorShape} from a backbone.
min_level An int of minimum level in FPN output feature maps.
max_level An int of maximum level in FPN output feature maps.
num_filters An int number of filters in FPN layers.
fusion_type A str of sum or concat. Whether performing sum or concat for feature fusion.
use_separable_conv A bool. If True use separable convolution for convolution in FPN layers.
use_keras_layer A bool. If Ture use keras layers as many as possible.
activation A str name of the activation function.
use_sync_bn A bool. If True, use synchronized batch normalization.
norm_momentum A float of normalization momentum for the moving average.
norm_epsilon A float added to variance to avoid dividing by zero.
kernel_initializer A str name of kernel_initializer for convolutional layers.
kernel_regularizer A tf_keras.regularizers.Regularizer object for Conv2D. Default is None.
bias_regularizer A tf_keras.regularizers.Regularizer object for Conv2D.
**kwargs Additional keyword arguments to be passed.

output_specs A dict of {level: TensorShape} pairs for the model output.

Methods

call

call(
    inputs, training=None, mask=None
)

Calls the model on new inputs and returns the outputs as tensors.

In this case call() just reapplies all ops in the graph to the new inputs (e.g. build a new computational graph from the provided inputs).

Args
inputs Input tensor, or dict/list/tuple of input tensors.
training Boolean or boolean scalar tensor, indicating whether to run the Network in training mode or inference mode.
mask A mask or list of masks. A mask can be either a boolean tensor or None (no mask). For more details, check the guide here.

Returns
A tensor if there is a single output, or a list of tensors if there are more than one outputs.