config.py

class Config(object):
    """ Wrapper class for various (hyper)parameters. """
    def __init__(self):
        """Sets the default model hyperparameters."""
        # File pattern of sharded TFRecord file containing SequenceExample protos.
        # Must be provided in training and evaluation modes.
        self.input_file_pattern = None

        # Image format ("jpeg" or "png").
        self.image_format = "jpeg"
        # Dimensions of VGG-16 input images.
        self.image_height = 224
        self.image_width = 224
    
        # Approximate number of values per input shard. Used to ensure sufficient
        # mixing between shards in training.
        self.values_per_input_shard = 100
        # Minimum number of shards to keep in the input queue.
        self.input_queue_capacity_factor = 2
        # Number of threads for prefetching SequenceExample protos.
        self.num_input_reader_threads = 1
        # Number of threads for image preprocessing. Should be a multiple of 2.
        self.num_preprocess_threads = 2

        # about the model architecture
        self.cnn = 'rpn'               # 'rpn', 'vgg16' or 'resnet50'

        self.max_caption_length = 20
        self.dim_embedding = 512
        self.num_lstm_units = 512
        self.num_initalize_layers = 2    # 1 or 2
        self.dim_initalize_layer = 512
        self.num_decode_layers = 2       # 1 or 2
        self.dim_decode_layer = 1024
        
        # attention mechanism
        self.attention_mechanism = "fc2"       # "fc1", "fc2", "rnn" or "bias"
        self.dim_attend_layer = 512     # for "fc1" and "fc2" only
        self.dim_rnn_att_state = 256    # for rnn only

        # about the weight initialization and regularization
        self.fc_kernel_initializer_scale = 0.08
        self.fc_kernel_regularizer_scale = 1e-4
        self.fc_activity_regularizer_scale = 0.0
        self.conv_kernel_regularizer_scale = 1e-4
        self.conv_activity_regularizer_scale = 0.0
        self.fc_drop_rate = 0.5
        self.lstm_drop_rate = 0.3
        self.attention_loss_factor = 0.01

        # about the optimization
        self.max_checkpoints_to_keep = 5
        self.log_every_n_steps = 10

        self.num_epochs = 100
        self.batch_size = 4
        self.optimizer = 'Adam'    # 'Adam', 'RMSProp', 'Momentum' or 'SGD'
        self.initial_learning_rate = 0.0001
        self.learning_rate_decay_factor = 1.0
        self.num_steps_per_decay = 100000
        self.clip_gradients = 5.0
        self.momentum = 0.0
        self.use_nesterov = True
        self.decay = 0.9
        self.centered = True
        self.beta1 = 0.9
        self.beta2 = 0.999
        self.epsilon = 1e-6

        # about the saver
        self.save_period = 1000
        self.save_dir = '/home/hillyess/ai/project-image-caption/output/model'
        self.eval_dir = '/home/hillyess/ai/project-image-caption/output/eval'
        # about the vocabulary
        self.vocabulary_file = './vocabulary.csv'
        # Number of unique words in the vocab (plus 1, for <UNK>).
        # The default value is larger than the expected actual vocab size to allow
        # for differences between tokenizer versions used in preprocessing. There is
        # no harm in using a value greater than the actual vocab size, but using a
        # value less than the actual vocab size will result in an error.
        self.vocabulary_size = 12000

        # # about the training
        # self.train_image_dir = './train/images/'
        # self.train_caption_file = './train/captions_train2014.json'
        # self.temp_annotation_file = './train/anns.csv'
        # self.temp_data_file = './train/data.npy'

        # about the evaluation
        # self.eval_image_dir = './val/images/'
        # self.eval_caption_file = './val/captions_val2014.json'
        # self.eval_result_dir = '../val/results/'
        # self.eval_result_file = '../val/results.json'
        # self.save_eval_result_as_image = False

        # about the testing
        # self.test_image_dir = './test/images/'
        # self.test_result_dir = './test/results/'
        # self.test_result_file = './test/results.csv'