diff --git a/mambular/configs/mambular_config.py b/mambular/configs/mambular_config.py index c6fcd89..72b0b60 100644 --- a/mambular/configs/mambular_config.py +++ b/mambular/configs/mambular_config.py @@ -75,9 +75,9 @@ class DefaultMambularConfig: Whether to shuffle the embeddings before being passed to the Mamba layers. layer_norm_eps : float, default=1e-05 Epsilon value for layer normalization. - AD_weight_decay : bool, default=False + AD_weight_decay : bool, default=True whether weight decay is also applied to A-D matrices - BC_layer_norm: bool, default=True + BC_layer_norm: bool, default=False whether to apply layer normalization to B-C matrices """ @@ -114,5 +114,5 @@ class DefaultMambularConfig: use_cls: bool = False shuffle_embeddings: bool = False layer_norm_eps: float = 1e-05 - AD_weight_decay: bool = False - BC_layer_norm: bool = True + AD_weight_decay: bool = True + BC_layer_norm: bool = False