T-GANs

code for Training Generative Adversarial Networks Via Turing Test
(There may be a delay of arxiv version, please get lastest version here)

Require: Keras 2.2.0+

the GANs with Turing test can be found here.

before using the code, you must change the source code of keras !!
(you can run it without changing keras, but the effect will be worse.)

keras/engine/base_layer.py, replace

    def add_weight(self,
                   name,
                   shape,
                   dtype=None,
                   initializer=None,
                   regularizer=None,
                   trainable=True,
                   constraint=None):
        """Adds a weight variable to the layer.
        # Arguments
            name: String, the name for the weight variable.
            shape: The shape tuple of the weight.
            dtype: The dtype of the weight.
            initializer: An Initializer instance (callable).
            regularizer: An optional Regularizer instance.
            trainable: A boolean, whether the weight should
                be trained via backprop or not (assuming
                that the layer itself is also trainable).
            constraint: An optional Constraint instance.
        # Returns
            The created weight variable.
        """
        initializer = initializers.get(initializer)
        if dtype is None:
            dtype = K.floatx()
        weight = K.variable(initializer(shape),
                            dtype=dtype,
                            name=name,
                            constraint=constraint)
        if regularizer is not None:
            with K.name_scope('weight_regularizer'):
                self.add_loss(regularizer(weight))
        if trainable:
            self._trainable_weights.append(weight)
        else:
            self._non_trainable_weights.append(weight)
        return weight

with

    def add_weight(self,
                   name,
                   shape,
                   dtype=None,
                   initializer=None,
                   regularizer=None,
                   trainable=True,
                   constraint=None):
        """Adds a weight variable to the layer.
        # Arguments
            name: String, the name for the weight variable.
            shape: The shape tuple of the weight.
            dtype: The dtype of the weight.
            initializer: An Initializer instance (callable).
            regularizer: An optional Regularizer instance.
            trainable: A boolean, whether the weight should
                be trained via backprop or not (assuming
                that the layer itself is also trainable).
            constraint: An optional Constraint instance.
        # Returns
            The created weight variable.
        """
        initializer = initializers.get(initializer)
        if dtype is None:
            dtype = K.floatx()
        weight = K.variable(initializer(shape),
                            dtype=dtype,
                            name=name,
                            constraint=None)
        if regularizer is not None:
            with K.name_scope('weight_regularizer'):
                self.add_loss(regularizer(weight))
        if trainable:
            self._trainable_weights.append(weight)
        else:
            self._non_trainable_weights.append(weight)
        if constraint is not None:
            return constraint(weight)
        return weight

other platforms

• tensorflow version: https://github.com/TwistedW/tf-T-GANs
• pytorch version: https://github.com/rahulbhalley/turing-gan.pytorch