fairseq/models/bert_seq2seq.py [695:728]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
def base_architecture(args):
    args.encoder_embed_path = getattr(args, 'encoder_embed_path', None)
    args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 512)
    args.encoder_ffn_embed_dim = getattr(args, 'encoder_ffn_embed_dim', args.encoder_embed_dim * 4)
    args.encoder_layers = getattr(args, 'encoder_layers', 6)
    args.encoder_attention_heads = getattr(args, 'encoder_attention_heads', args.encoder_embed_dim // 64)
    args.encoder_normalize_before = getattr(args, 'encoder_normalize_before', False)
    args.encoder_learned_pos = getattr(args, 'encoder_learned_pos', False)
    args.decoder_embed_path = getattr(args, 'decoder_embed_path', None)
    args.decoder_embed_dim = getattr(args, 'decoder_embed_dim', args.encoder_embed_dim)
    args.decoder_ffn_embed_dim = getattr(args, 'decoder_ffn_embed_dim', args.encoder_ffn_embed_dim)
    args.decoder_layers = getattr(args, 'decoder_layers', 6)
    args.decoder_attention_heads = getattr(args, 'decoder_attention_heads', args.encoder_attention_heads)
    args.decoder_normalize_before = getattr(args, 'decoder_normalize_before', False)
    args.decoder_learned_pos = getattr(args, 'decoder_learned_pos', False)
    args.attention_dropout = getattr(args, 'attention_dropout', 0.)
    args.relu_dropout = getattr(args, 'relu_dropout', 0.)
    args.dropout = getattr(args, 'dropout', 0.1)
    args.adaptive_softmax_cutoff = getattr(args, 'adaptive_softmax_cutoff', None)
    args.adaptive_softmax_dropout = getattr(args, 'adaptive_softmax_dropout', 0)
    args.share_decoder_input_output_embed = getattr(args, 'share_decoder_input_output_embed', False)
    args.share_all_embeddings = getattr(args, 'share_all_embeddings', False)
    args.no_enc_token_positional_embeddings = getattr(args, 'no_enc_token_positional_embeddings', False)
    args.no_dec_token_positional_embeddings = getattr(args, 'no_dec_token_positional_embeddings', False)
    args.embedding_only = getattr(args, 'embedding_only', False)

    args.decoder_output_dim = getattr(args, 'decoder_output_dim', args.decoder_embed_dim)
    args.decoder_input_dim = getattr(args, 'decoder_input_dim', args.decoder_embed_dim)

    args.decoder_embed_scale = getattr(args, 'decoder_embed_scale', None)
    args.encoder_embed_scale = getattr(args, 'encoder_embed_scale', None)

    args.bilm_mask_last_state = getattr(args, 'bilm_mask_last_state', False)
    args.bilm_add_bos = getattr(args, 'bilm_add_bos', False)
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



fairseq/models/disco_transformer.py [342:375]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
def base_architecture(args):
    args.encoder_embed_path = getattr(args, 'encoder_embed_path', None)
    args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 512)
    args.encoder_ffn_embed_dim = getattr(args, 'encoder_ffn_embed_dim', args.encoder_embed_dim * 4)
    args.encoder_layers = getattr(args, 'encoder_layers', 6)
    args.encoder_attention_heads = getattr(args, 'encoder_attention_heads', args.encoder_embed_dim // 64)
    args.encoder_normalize_before = getattr(args, 'encoder_normalize_before', False)
    args.encoder_learned_pos = getattr(args, 'encoder_learned_pos', False)
    args.decoder_embed_path = getattr(args, 'decoder_embed_path', None)
    args.decoder_embed_dim = getattr(args, 'decoder_embed_dim', args.encoder_embed_dim)
    args.decoder_ffn_embed_dim = getattr(args, 'decoder_ffn_embed_dim', args.encoder_ffn_embed_dim)
    args.decoder_layers = getattr(args, 'decoder_layers', 6)
    args.decoder_attention_heads = getattr(args, 'decoder_attention_heads', args.encoder_attention_heads)
    args.decoder_normalize_before = getattr(args, 'decoder_normalize_before', False)
    args.decoder_learned_pos = getattr(args, 'decoder_learned_pos', False)
    args.attention_dropout = getattr(args, 'attention_dropout', 0.)
    args.relu_dropout = getattr(args, 'relu_dropout', 0.)
    args.dropout = getattr(args, 'dropout', 0.1)
    args.adaptive_softmax_cutoff = getattr(args, 'adaptive_softmax_cutoff', None)
    args.adaptive_softmax_dropout = getattr(args, 'adaptive_softmax_dropout', 0)
    args.share_decoder_input_output_embed = getattr(args, 'share_decoder_input_output_embed', False)
    args.share_all_embeddings = getattr(args, 'share_all_embeddings', False)
    args.no_enc_token_positional_embeddings = getattr(args, 'no_enc_token_positional_embeddings', False)
    args.no_dec_token_positional_embeddings = getattr(args, 'no_dec_token_positional_embeddings', False)
    args.embedding_only = getattr(args, 'embedding_only', False)

    args.decoder_output_dim = getattr(args, 'decoder_output_dim', args.decoder_embed_dim)
    args.decoder_input_dim = getattr(args, 'decoder_input_dim', args.decoder_embed_dim)

    args.decoder_embed_scale = getattr(args, 'decoder_embed_scale', None)
    args.encoder_embed_scale = getattr(args, 'encoder_embed_scale', None)

    args.bilm_mask_last_state = getattr(args, 'bilm_mask_last_state', False)
    args.bilm_add_bos = getattr(args, 'bilm_add_bos', False)
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -