def add_args()

in pytorch_translate/transformer.py [0:0]

• 160 lines of code
• 2 McCabe index (conditional complexity)

    def add_args(parser):
        """Add model-specific arguments to the parser."""
        parser.add_argument(
            "--dropout", type=float, metavar="D", help="dropout probability"
        )
        parser.add_argument(
            "--attention-dropout",
            type=float,
            metavar="D",
            help="dropout probability for attention weights",
        )
        parser.add_argument(
            "--relu-dropout",
            type=float,
            metavar="D",
            help="dropout probability after ReLU in FFN",
        )
        parser.add_argument(
            "--encoder-pretrained-embed",
            type=str,
            metavar="STR",
            help="path to pre-trained encoder embedding",
        )
        parser.add_argument(
            "--encoder-embed-dim",
            type=int,
            metavar="N",
            help="encoder embedding dimension",
        )
        parser.add_argument(
            "--encoder-ffn-embed-dim",
            type=int,
            metavar="N",
            help="encoder embedding dimension for FFN",
        )
        parser.add_argument(
            "--encoder-freeze-embed",
            default=False,
            action="store_true",
            help=(
                "whether to freeze the encoder embedding or allow it to be "
                "updated during training"
            ),
        )
        parser.add_argument(
            "--encoder-layers", type=int, metavar="N", help="num encoder layers"
        )
        parser.add_argument(
            "--encoder-attention-heads",
            type=int,
            metavar="N",
            help="num encoder attention heads",
        )
        parser.add_argument(
            "--encoder-normalize-before",
            default=False,
            action="store_true",
            help="apply layernorm before each encoder block",
        )
        parser.add_argument(
            "--encoder-learned-pos",
            default=False,
            action="store_true",
            help="use learned positional embeddings in the encoder",
        )
        parser.add_argument(
            "--decoder-pretrained-embed",
            type=str,
            metavar="STR",
            help="path to pre-trained decoder embedding",
        )
        parser.add_argument(
            "--decoder-embed-dim",
            type=int,
            metavar="N",
            help="decoder embedding dimension",
        )
        parser.add_argument(
            "--decoder-ffn-embed-dim",
            type=int,
            metavar="N",
            help="decoder embedding dimension for FFN",
        )
        parser.add_argument(
            "--decoder-freeze-embed",
            default=False,
            action="store_true",
            help=(
                "whether to freeze the encoder embedding or allow it to be "
                "updated during training"
            ),
        )
        parser.add_argument(
            "--decoder-layers", type=int, metavar="N", help="num decoder layers"
        )
        parser.add_argument(
            "--decoder-attention-heads",
            type=int,
            metavar="N",
            help="num decoder attention heads",
        )
        parser.add_argument(
            "--decoder-learned-pos",
            default=False,
            action="store_true",
            help="use learned positional embeddings in the decoder",
        )
        parser.add_argument(
            "--decoder-normalize-before",
            default=False,
            action="store_true",
            help="apply layernorm before each decoder block",
        )
        parser.add_argument(
            "--decoder-layerdrop",
            type=float,
            metavar="D",
            default=0,
            help="LayerDrop probability for decoder",
        )
        parser.add_argument(
            "--decoder-layers-to-keep",
            default=None,
            help="which layers to *keep* when pruning as a comma-separated list",
        )
        parser.add_argument(
            "--share-decoder-input-output-embed",
            default=False,
            action="store_true",
            help="share decoder input and output embeddings",
        )
        parser.add_argument(
            "--share-all-embeddings",
            default=False,
            action="store_true",
            help="share encoder, decoder and output embeddings"
            " (requires shared dictionary and embed dim)",
        )
        parser.add_argument(
            "--adaptive-softmax-cutoff",
            default=None,
            metavar="EXPR",
            help="comma separated list of adaptive softmax cutoff points. "
            "Must be used with adaptive_loss criterion",
        )
        parser.add_argument(
            "--decoder-out-embed-dim",
            default=None,
            type=int,
            metavar="N",
            help="decoder output embedding dimension (bottleneck layer before"
            "output layer if specified.)",
        )
        parser.add_argument(
            "--aan",
            default=False,
            action="store_true",
            help="use average attention network (AAN) instead of decoder "
            "self-attention",
        )

        # Args for vocab reduction
        vocab_reduction.add_args(parser)