from recwizard.configuration_utils import BaseConfig
[docs]class KGSFGenConfig(BaseConfig):
def __init__(self,
dictionary = None,
subkg = None,
mask4movie = None,
mask4key = None,
embedding_data = None,
edge_set = None,
batch_size: int = 32,
max_r_length: int = 30,
embedding_size: int = 300,
n_concept: int = 29308,
dim: int = 128,
n_entity: int = 64368,
num_bases: int = 8,
n_positions: int = None,
truncate: int = 0,
text_truncate: int = 0,
label_truncate: int = 0,
padding_idx: int = 0,
start_idx: int = 1,
end_idx: int = 2,
longest_label: int = 1,
pretrain: bool = False,
n_heads: int = 2,
n_layers: int = 2,
ffn_size: int = 300,
dropout: float = 0.1,
attention_dropout: float = 0.0,
relu_dropout: float = 0.1,
learn_positional_embeddings: bool = False,
embeddings_scale: bool = True,
**kwargs):
super().__init__(**kwargs)
self.batch_size = batch_size
self.max_r_length = max_r_length
self.embedding_size = embedding_size
self.n_concept = n_concept
self.dim = dim
self.n_entity = n_entity
if n_positions == None: # should default to 1024 if truncate,text_truncate,label_truncate are 0
if max(truncate,label_truncate,label_truncate) != 0:
self.n_positions = max(truncate,label_truncate,label_truncate)
else:
self.n_positions = 1024
else:
self.n_positions = n_positions
self.num_bases = num_bases
self.truncate = truncate
self.text_truncate = text_truncate
self.label_truncate = label_truncate
self.padding_idx = padding_idx
self.start_idx = start_idx
self.end_idx = end_idx
self.longest_label = longest_label
self.pretrain = pretrain
# for building encoder
self.n_heads = n_heads
self.n_layers = n_layers
self.ffn_size = ffn_size
self.dropout = dropout
self.attention_dropout = attention_dropout
self.relu_dropout = relu_dropout
self.learn_positional_embeddings = learn_positional_embeddings
self.embeddings_scale = embeddings_scale
self.dictionary = dictionary
self.subkg = subkg
self.mask4key = mask4key
self.mask4movie = mask4movie
self.embedding_data = embedding_data
self.edge_set = edge_set