diff --git a/models/pretraining.py b/models/pretraining.py index fa25741..e824e17 100644 --- a/models/pretraining.py +++ b/models/pretraining.py @@ -52,6 +52,7 @@ def __init__(self, args): num_layers=args.graph_num_layers, num_heads=args.graph_num_heads, dropout=args.dropout, + attention_dropout=getattr(args, "attention_dropout", args.dropout), dim_feedforward=args.dim_feedforward, add_positional_encoding=args.add_positional_encoding, use_edge_bias=args.use_edge_bias, diff --git a/pretrain.py b/pretrain.py index 372459d..e5d4767 100644 --- a/pretrain.py +++ b/pretrain.py @@ -26,6 +26,7 @@ def build_parser(): parser.add_argument("--dim_feedforward", type=int, default=512) parser.add_argument("--dropout", type=float, default=0.25) parser.add_argument("--mlp_dropout", type=float, default=0.1) + parser.add_argument("--attention_dropout", type=float, default=0.1) parser.add_argument("--act", type=str, default="gelu") parser.add_argument("--curve_emb_dim", type=int, default=64) parser.add_argument("--surface_emb_dim", type=int, default=64)