Python load_inputs_and_targets示例

示例#1

    def transform(self, item):
        batch = load_inputs_and_targets(item, True, self.use_speaker_embedding)

        # added eos into input sequence
        eos = int(item[0][1]['output'][0]['shape'][1]) - 1
        xs = [np.append(x, eos) for x in batch[1]]
        if self.use_speaker_embedding:
            return batch[0], xs, batch[2]
        else:
            return batch[0], xs

示例#2

 def transform(self, item):
     return load_inputs_and_targets(item)