Commit c1371d2d authored by Rudolf Chrispens's avatar Rudolf Chrispens
Browse files

fixed vocab creaton bug

parent f06cc002

caption-lib/lstm/Readers/coco_reader.py

+7 −2
Original line number Diff line number Diff line
@@ -192,10 +192,15 @@ class coco_reader: 
        return self.create_vocab_from_data(data, filename)

    """



    def create_vocab_from_data(self, data, save_filename_prefix):

    def create_vocab_from_data(self, data_train, data_valid, data_test, save_filename_prefix):

        counter: int = 0

        vocab_id_word = {}

        for i in data['captions']:

        print("TODO add testset!")

        tr = data_train['captions']

        val = data_valid['captions']

        data_combined = np.concatenate((tr, val), axis=0)



        for i in data_combined:

            for c in i:

                if c not in vocab_id_word.values():

                    vocab_id_word[counter] = c

caption-lib/lstm/lstm_main.py

+2 −1
Original line number Diff line number Diff line
@@ -30,7 +30,8 @@ def main(self, parameter_list): 
    # load train dataset

    current_reader = _reader.coco_reader()

    train_data_decoded = current_reader.load_data_decoded()

    word_to_idx = current_reader.create_vocab_from_data(data=train_data_decoded, save_filename_prefix=_config.path.input_train)

    valid_data_decoded = current_reader.load_data_decoded(filename=_config.path.input_validate)

    word_to_idx = current_reader.create_vocab_from_data(data_train=train_data_decoded, data_valid=valid_data_decoded, data_test=None, save_filename_prefix=_config.path.input_train)

    train_data = current_reader.load_data_encoded()



    #test1 = current_reader.get_stored_vocab_id_word(filename_prefix=_config.path.read_train)