honnet_brain_architecture.yaml

class_name: Model
config:
  input_layers:
  - [hero_input, 0, 0]
  layers:
  - class_name: InputLayer
    config:
      batch_input_shape: !!python/tuple [null, 520]
      input_dtype: float32
      name: hero_input
      sparse: false
    inbound_nodes: []
    name: hero_input
  - class_name: Reshape
    config:
      batch_input_shape: !!python/tuple [null, 520]
      input_dtype: float32
      name: reshape_1
      target_shape: !!python/tuple [1, 2, 260]
      trainable: true
    inbound_nodes:
    - - [hero_input, 0, 0]
    name: reshape_1
  - class_name: Permute
    config:
      dims: !!python/tuple [2, 3, 1]
      name: permute_1
      trainable: true
    inbound_nodes:
    - - [reshape_1, 0, 0]
    name: permute_1
  - class_name: Convolution2D
    config:
      W_constraint: null
      W_regularizer: null
      activation: linear
      activity_regularizer: null
      b_constraint: null
      b_regularizer: null
      bias: true
      border_mode: valid
      dim_ordering: tf
      init: glorot_uniform
      name: convolution2d_1
      nb_col: 260
      nb_filter: 135
      nb_row: 1
      subsample: !!python/tuple [1, 1]
      trainable: true
    inbound_nodes:
    - - [permute_1, 0, 0]
    name: convolution2d_1
  - class_name: BatchNormalization
    config: {axis: -1, beta_regularizer: null, epsilon: 0.001, gamma_regularizer: null,
      mode: 0, momentum: 0.99, name: batchnormalization_1, trainable: true}
    inbound_nodes:
    - - [convolution2d_1, 0, 0]
    name: batchnormalization_1
  - class_name: Activation
    config: {activation: relu, name: activation_1, trainable: true}
    inbound_nodes:
    - - [batchnormalization_1, 0, 0]
    name: activation_1
  - class_name: Convolution2D
    config:
      W_constraint: null
      W_regularizer: null
      activation: linear
      activity_regularizer: null
      b_constraint: null
      b_regularizer: null
      bias: true
      border_mode: valid
      dim_ordering: tf
      init: glorot_uniform
      name: convolution2d_2
      nb_col: 1
      nb_filter: 64
      nb_row: 1
      subsample: !!python/tuple [1, 1]
      trainable: true
    inbound_nodes:
    - - [activation_1, 0, 0]
    name: convolution2d_2
  - class_name: BatchNormalization
    config: {axis: -1, beta_regularizer: null, epsilon: 0.001, gamma_regularizer: null,
      mode: 0, momentum: 0.99, name: batchnormalization_2, trainable: true}
    inbound_nodes:
    - - [convolution2d_2, 0, 0]
    name: batchnormalization_2
  - class_name: Activation
    config: {activation: relu, name: activation_2, trainable: true}
    inbound_nodes:
    - - [batchnormalization_2, 0, 0]
    name: activation_2
  - class_name: Flatten
    config: {name: flatten_1, trainable: true}
    inbound_nodes:
    - - [activation_2, 0, 0]
    name: flatten_1
  - class_name: Dense
    config:
      W_constraint: null
      W_regularizer: null
      activation: linear
      activity_regularizer: null
      b_constraint: null
      b_regularizer: null
      batch_input_shape: !!python/tuple [null, 128]
      bias: true
      init: glorot_uniform
      input_dim: !!python/object/apply:numpy.core.multiarray.scalar
      - !!python/object/apply:numpy.dtype
        args: [i8, 0, 1]
        state: !!python/tuple [3, <, null, null, null, -1, -1, 0]
      - !!binary |
        gAAAAAAAAAA=
      input_dtype: float32
      name: dense_1
      output_dim: 128
      trainable: true
    inbound_nodes:
    - - [flatten_1, 0, 0]
    name: dense_1
  - class_name: BatchNormalization
    config: {axis: -1, beta_regularizer: null, epsilon: 0.001, gamma_regularizer: null,
      mode: 0, momentum: 0.99, name: batchnormalization_3, trainable: true}
    inbound_nodes:
    - - [dense_1, 0, 0]
    name: batchnormalization_3
  - class_name: Activation
    config: {activation: relu, name: activation_3, trainable: true}
    inbound_nodes:
    - - [batchnormalization_3, 0, 0]
    name: activation_3
  - class_name: Dropout
    config: {name: dropout_1, p: 0.5, trainable: true}
    inbound_nodes:
    - - [activation_3, 0, 0]
    name: dropout_1
  - class_name: Dense
    config:
      W_constraint: null
      W_regularizer: null
      activation: linear
      activity_regularizer: null
      b_constraint: null
      b_regularizer: null
      batch_input_shape: !!python/tuple [null, 128]
      bias: true
      init: glorot_uniform
      input_dim: 128
      input_dtype: float32
      name: dense_2
      output_dim: 128
      trainable: true
    inbound_nodes:
    - - [dropout_1, 0, 0]
    name: dense_2
  - class_name: BatchNormalization
    config: {axis: -1, beta_regularizer: null, epsilon: 0.001, gamma_regularizer: null,
      mode: 0, momentum: 0.99, name: batchnormalization_4, trainable: true}
    inbound_nodes:
    - - [dense_2, 0, 0]
    name: batchnormalization_4
  - class_name: Activation
    config: {activation: relu, name: activation_4, trainable: true}
    inbound_nodes:
    - - [batchnormalization_4, 0, 0]
    name: activation_4
  - class_name: Dropout
    config: {name: dropout_2, p: 0.5, trainable: true}
    inbound_nodes:
    - - [activation_4, 0, 0]
    name: dropout_2
  - class_name: Dense
    config:
      W_constraint: null
      W_regularizer: null
      activation: linear
      activity_regularizer: null
      b_constraint: null
      b_regularizer: null
      batch_input_shape: !!python/tuple [null, 128]
      bias: true
      init: glorot_uniform
      input_dim: 128
      input_dtype: float32
      name: dense_3
      output_dim: 128
      trainable: true
    inbound_nodes:
    - - [dropout_2, 0, 0]
    name: dense_3
  - class_name: BatchNormalization
    config: {axis: -1, beta_regularizer: null, epsilon: 0.001, gamma_regularizer: null,
      mode: 0, momentum: 0.99, name: batchnormalization_5, trainable: true}
    inbound_nodes:
    - - [dense_3, 0, 0]
    name: batchnormalization_5
  - class_name: Activation
    config: {activation: relu, name: activation_5, trainable: true}
    inbound_nodes:
    - - [batchnormalization_5, 0, 0]
    name: activation_5
  - class_name: Dropout
    config: {name: dropout_3, p: 0.5, trainable: true}
    inbound_nodes:
    - - [activation_5, 0, 0]
    name: dropout_3
  - class_name: Dense
    config:
      W_constraint: null
      W_regularizer: null
      activation: softmax
      activity_regularizer: null
      b_constraint: null
      b_regularizer: null
      batch_input_shape: !!python/tuple [null, 128]
      bias: true
      init: glorot_uniform
      input_dim: 128
      input_dtype: float32
      name: out_winner
      output_dim: 2
      trainable: true
    inbound_nodes:
    - - [dropout_3, 0, 0]
    name: out_winner
  - class_name: Dense
    config:
      W_constraint: null
      W_regularizer: null
      activation: sigmoid
      activity_regularizer: null
      b_constraint: null
      b_regularizer: null
      batch_input_shape: !!python/tuple [null, 128]
      bias: true
      init: glorot_uniform
      input_dim: 128
      input_dtype: float32
      name: out_concede
      output_dim: 1
      trainable: true
    inbound_nodes:
    - - [dropout_3, 0, 0]
    name: out_concede
  - class_name: Dense
    config:
      W_constraint: null
      W_regularizer: null
      activation: linear
      activity_regularizer: null
      b_constraint: null
      b_regularizer: null
      batch_input_shape: !!python/tuple [null, 128]
      bias: true
      init: glorot_uniform
      input_dim: 128
      input_dtype: float32
      name: out_secs
      output_dim: 1
      trainable: true
    inbound_nodes:
    - - [dropout_3, 0, 0]
    name: out_secs
  name: model_1
  output_layers:
  - [out_winner, 0, 0]
  - [out_concede, 0, 0]
  - [out_secs, 0, 0]
keras_version: 1.2.1