ontolearn.nces_modules

NCES modules.

Classes

MAB	MAB module.
SAB	SAB module.
ISAB	ISAB module.
PMA	PMA module.
ConEx	Convolutional Complex Knowledge Graph Embeddings

Module Contents

class ontolearn.nces_modules.MAB(dim_Q, dim_K, dim_V, num_heads, ln=False)

Bases: torch.nn.Module

MAB module.

dim_V

num_heads

fc_q

fc_k

fc_v

fc_o

forward(Q, K)

class ontolearn.nces_modules.SAB(dim_in, dim_out, num_heads, ln=False)

Bases: torch.nn.Module

SAB module.

mab

forward(X)

class ontolearn.nces_modules.ISAB(dim_in, dim_out, num_heads, m, ln=False)

Bases: torch.nn.Module

ISAB module.

I

mab0

mab1

forward(X)

class ontolearn.nces_modules.PMA(dim, num_heads, num_seeds, ln=False)

Bases: torch.nn.Module

PMA module.

S

mab

forward(X)

class ontolearn.nces_modules.ConEx(embedding_dim, num_entities, num_relations, input_dropout, feature_map_dropout, kernel_size, num_of_output_channels)

Bases: torch.nn.Module

Convolutional Complex Knowledge Graph Embeddings

name = 'ConEx'

loss

embedding_dim

num_entities

num_relations

input_dropout

feature_map_dropout

kernel_size

num_of_output_channels

emb_ent_real

emb_ent_i

emb_rel_real

emb_rel_i

input_dp_ent_real

input_dp_ent_i

input_dp_rel_real

input_dp_rel_i

bn_ent_real

bn_ent_i

bn_rel_real

bn_rel_i

conv1

fc_num_input

fc

bn_conv1

bn_conv2

feature_dropout

residual_convolution(C_1, C_2)

forward_head_batch(*, e1_idx, rel_idx)

Given a head entity and a relation (h,r), we compute scores for all entities. [score(h,r,x)|x in Entities] => [0.0,0.1,…,0.8], shape=> (1, |Entities|) Given a batch of head entities and relations => shape (size of batch,| Entities|)

forward_head_and_loss(e1_idx, rel_idx, targets)

init()

get_embeddings()

forward_triples(*, e1_idx, rel_idx, e2_idx)

forward_triples_and_loss(e1_idx, rel_idx, e2_idx, targets)