$ python main.py --num_hidden_layers 1 --num_units 2 --batch_size 10

$ python main.py -h

usage: main.py [-h] [--batch_size BATCH_SIZE]
               [--max_train_data_size MAX_TRAIN_DATA_SIZE]
               [--num_units NUM_UNITS] [--num_hidden_layers NUM_HIDDEN_LAYERS]
               [--learning_rate LEARNING_RATE]
               [--learning_rate_decay_factor LEARNING_RATE_DECAY_FACTOR]
               [--max_gradient_norm MAX_GRADIENT_NORM]
               [--num_samples NUM_SAMPLES]
               [--en_vocabulary_size EN_VOCABULARY_SIZE]
               [--fr_vocabulary_size FR_VOCABULARY_SIZE]
               [--target_vocab TARGET_VOCAB]
               [--checkpoint_step CHECKPOINT_STEP] [--train [TRAIN]]
               [--notrain] [--dataset DATASET] [--model_name MODEL_NAME]
               [--dataset_dir DATASET_DIR] [--checkpoint_dir CHECKPOINT_DIR]

optional arguments:
  -h, --help            show this help message and exit
  --batch_size BATCH_SIZE
                        Size of training batch
  --max_train_data_size MAX_TRAIN_DATA_SIZE
                        Limit on the size of training data (0: no limit)
  --num_units NUM_UNITS
                        Number of units in LSTM layer
  --num_hidden_layers NUM_HIDDEN_LAYERS
                        Number of hidden LSTM layers
  --learning_rate LEARNING_RATE
                        Initial learning rate
  --learning_rate_decay_factor LEARNING_RATE_DECAY_FACTOR
                        Learning rate decays by this much
  --max_gradient_norm MAX_GRADIENT_NORM
                        Clip gradients to this norm
  --num_samples NUM_SAMPLES
                        Number of samples for sampled softmax
  --en_vocabulary_size EN_VOCABULARY_SIZE
                        English vocabulary size
  --fr_vocabulary_size FR_VOCABULARY_SIZE
                        French vocabulary size
  --target_vocab TARGET_VOCAB
                        Target vocabulary (en/fr)
  --checkpoint_step CHECKPOINT_STEP
                        Number of training steps per checkpoint
  --train [TRAIN]       True for training, False for validating
  --notrain
  --dataset DATASET     Name of the dataset file
  --model_name MODEL_NAME
                        Name of the model
  --dataset_dir DATASET_DIR
                        Directory name for the dataset
  --checkpoint_dir CHECKPOINT_DIR
                        Directory name to save the checkpoint