crested.tl.zoo.basenji#
- crested.tl.zoo.basenji(seq_len, num_classes, first_activation='gelu', activation='gelu', output_activation='softplus', first_filters=256, filters=2048, first_kernel_size=15, kernel_size=3)#
Construct a Basenji model.
- Parameters:
seq_len (
int
) – Width of the input region.num_classes (
int
) – Number of classes to predict.first_activation (
str
(default:'gelu'
)) – Activation function for the first convolutional block.activation (
str
(default:'gelu'
)) – Activation function for subsequent convolutional blocks.output_activation (
str
(default:'softplus'
)) – Activation function for the output layer.first_filters (
int
(default:256
)) – Number of filters in the first convolutional block.filters (
int
(default:2048
)) – Number of filters in subsequent convolutional blocks.first_kernel_size (
int
(default:15
)) – Size of the kernel in the first convolutional block.kernel_size (
int
(default:3
)) – Size of the kernel in subsequent convolutional blocks.
- Return type:
Model
- Returns:
A Keras model.