e8ee5ce8baba0330f5b64e36c34b12f52a5ad29e,pyannote/audio/embedding/models.py,ClopiNet,__call__,#ClopiNet#Any#,289

Before Change

        // stack (bidirectional) LSTM layers
        for i, output_dim in enumerate(self.lstm):

            if i:
                lstm = LSTM(output_dim,
                            name="lstm_{i:d}".format(i=i),
                            return_sequences=True,
                            activation="tanh")
            else:
                // we need to provide input_shape to first LSTM
                lstm = LSTM(output_dim,
                            input_shape=input_shape,
                            name="lstm_{i:d}".format(i=i),
                            return_sequences=True,
                            activation="tanh")

            // bi-directional LSTM
            if self.bidirectional:
                lstm = Bidirectional(lstm, merge_mode=self.bidirectional)

            // (actually) stack LSTM

After Change

        // stack (bidirectional) LSTM layers
        for i, output_dim in enumerate(self.lstm):

            params = {
                "name": "lstm_{i:d}".format(i=i),
                "return_sequences": True,
                // "go_backwards": False,
                // "stateful": False,
                // "unroll": False,
                // "implementation": 0,
                "activation": "tanh",
                // "recurrent_activation": "hard_sigmoid",
                // "use_bias": True,
                // "kernel_initializer": "glorot_uniform",
                // "recurrent_initializer": "orthogonal",
                // "bias_initializer": "zeros",
                // "unit_forget_bias": True,
                // "kernel_regularizer": None,
                // "recurrent_regularizer": None,
                // "bias_regularizer": None,
                // "activity_regularizer": None,
                // "kernel_constraint": None,
                // "recurrent_constraint": None,
                // "bias_constraint": None,
                // "dropout": 0.0,
                // "recurrent_dropout": 0.0,
            }

            // first LSTM needs to be given the input shape
            if i == 0:
                params["input_shape"] = input_shape

            lstm = LSTM(output_dim, **params)

            // bi-directional LSTM
            if self.bidirectional:
                lstm = Bidirectional(lstm, merge_mode=self.bidirectional)

Instances