diff --git a/_modules/besskge/embedding.html b/_modules/besskge/embedding.html index 9c3af28..e7e8b24 100644 --- a/_modules/besskge/embedding.html +++ b/_modules/besskge/embedding.html @@ -100,6 +100,24 @@

Source code for besskge.embedding

     return torch.nn.functional.normalize(torch.nn.init.uniform(embedding_table), dim=-1)
+
[docs]def init_xavier_norm(embedding_table: torch.Tensor, gain: float = 1.0) -> torch.Tensor: + """ + Initialize embeddings according to Xavier normal scheme, with + `fan_in = 0`, `fan_out=row_size`. + + :param embedding_table: + Tensor of embedding parameters to initialize. + :param gain: + Scaling factor for standard deviation. Default: 1.0. + + :return: + Initialized tensor. + """ + return torch.nn.init.normal_( + embedding_table, std=gain * np.sqrt(2.0 / embedding_table.shape[-1]) + )
+ +
[docs]def init_KGE_uniform( embedding_table: torch.Tensor, b: float = 1.0, divide_by_embedding_size: bool = True ) -> torch.Tensor: diff --git a/_modules/besskge/scoring.html b/_modules/besskge/scoring.html index bfa52ed..9da5b9d 100644 --- a/_modules/besskge/scoring.html +++ b/_modules/besskge/scoring.html @@ -89,6 +89,7 @@

Source code for besskge.scoring

     init_KGE_normal,
     init_KGE_uniform,
     init_uniform_norm,
+    init_xavier_norm,
     initialize_entity_embedding,
     initialize_relation_embedding,
     refactor_embedding_sharding,
@@ -365,7 +366,7 @@ 
Source code for besskge.scoring
         :param relation_initializer:
             Initialization function or table for relation embeddings.
         :param inverse_relations:
-            If True, learn embeddings for inverse relations. Default: False
+            If True, learn embeddings for inverse relations. Default: False.
         """
         super(TransE, self).__init__(
             negative_sample_sharing=negative_sample_sharing, scoring_norm=scoring_norm
@@ -465,7 +466,7 @@ 
Source code for besskge.scoring
         :param relation_initializer:
             Initialization function or table for relation embeddings.
         :param inverse_relations:
-            If True, learn embeddings for inverse relations. Default: False
+            If True, learn embeddings for inverse relations. Default: False.
         """
         super(RotatE, self).__init__(
             negative_sample_sharing=negative_sample_sharing, scoring_norm=scoring_norm
@@ -576,7 +577,7 @@ 
Source code for besskge.scoring
             If True, L2-normalize head and tail entity embeddings before projecting,
             as in :cite:p:`PairRE`. Default: True.
         :param inverse_relations:
-            If True, learn embeddings for inverse relations. Default: False
+            If True, learn embeddings for inverse relations. Default: False.
         """
         super(PairRE, self).__init__(
             negative_sample_sharing=negative_sample_sharing, scoring_norm=scoring_norm
@@ -711,7 +712,7 @@ 
Source code for besskge.scoring
             Offset factor for head/tail relation projections, as in TripleREv2.
             Default: 0.0 (no offset).
         :param inverse_relations:
-            If True, learn embeddings for inverse relations. Default: False
+            If True, learn embeddings for inverse relations. Default: False.
         """
         super(TripleRE, self).__init__(
             negative_sample_sharing=negative_sample_sharing, scoring_norm=scoring_norm
@@ -850,7 +851,7 @@ 
Source code for besskge.scoring
         :param relation_initializer:
             Initialization function or table for relation embeddings.
         :param inverse_relations:
-            If True, learn embeddings for inverse relations. Default: False
+            If True, learn embeddings for inverse relations. Default: False.
         """
         super(DistMult, self).__init__(negative_sample_sharing=negative_sample_sharing)
 
@@ -944,7 +945,7 @@ 
Source code for besskge.scoring
         :param relation_initializer:
             Initialization function or table for relation embeddings.
         :param inverse_relations:
-            If True, learn embeddings for inverse relations. Default: False
+            If True, learn embeddings for inverse relations. Default: False.
         """
         super(ComplEx, self).__init__(negative_sample_sharing=negative_sample_sharing)
 
@@ -1018,6 +1019,206 @@ 
Source code for besskge.scoring
         )

 
 
+
[docs]class ConvE(MatrixDecompositionScoreFunction):
+    """
+    ConvE scoring function :cite:p:`ConvE`.
+
+    Note that, differently from :cite:p:`ConvE`, the scores returned by this class
+    have not been passed through a final sigmoid layer, as we assume that this is
+    included in the loss function.
+
+    By design, this scoring function should be used in combination with a
+    negative/candidate sampler that only corrupts tails (possibly after
+    including all inverse triples in the dataset, see the `add_inverse_triples`
+    argument in :func:`besskge.sharding.PartitionedTripleSet.create_from_dataset`).
+    """
+
+    def __init__(
+        self,
+        negative_sample_sharing: bool,
+        sharding: Sharding,
+        n_relation_type: int,
+        embedding_size: int,
+        embedding_height: int,
+        embedding_width: int,
+        entity_initializer: Union[torch.Tensor, List[Callable[..., torch.Tensor]]] = [
+            init_xavier_norm,
+            torch.nn.init.zeros_,
+        ],
+        relation_initializer: Union[torch.Tensor, List[Callable[..., torch.Tensor]]] = [
+            init_xavier_norm,
+        ],
+        inverse_relations: bool = True,
+        input_channels: int = 1,
+        output_channels: int = 32,
+        kernel_height: int = 3,
+        kernel_width: int = 3,
+        input_dropout: float = 0.2,
+        feature_map_dropout: float = 0.2,
+        hidden_dropout: float = 0.3,
+        batch_normalization: bool = True,
+    ) -> None:
+        """
+        Initialize ConvE model.
+
+        :param negative_sample_sharing:
+            see :meth:`DistanceBasedScoreFunction.__init__`
+        :param sharding:
+            Entity sharding.
+        :param n_relation_type:
+            Number of relation types in the knowledge graph.
+        :param embedding_size:
+            Size of entity and relation embeddings.
+        :param embedding_height:
+            Height of the 2D-reshaping of the concatenation of
+            head and relation embeddings.
+        :param embedding_width:
+            Width of the 2D-reshaping of the concatenation of
+            head and relation embeddings.
+        :param entity_initializer:
+            Initialization functions or table for entity embeddings.
+            If not passing a table, two functions are needed: the initializer
+            for entity embeddings and initializer for (scalar) tail biases.
+        :param relation_initializer:
+            Initialization function or table for relation embeddings.
+        :param inverse_relations:
+            If True, learn embeddings for inverse relations. Default: True.
+        :param input_channels:
+            Number of input channels of the Conv2D operator. Default: 1.
+        :param output_channels:
+            Number of output channels of the Conv2D operator. Default: 32.
+        :param kernel_height:
+            Height of the Conv2D kernel. Default: 3.
+        :param kernel_width:
+            Width of the Conv2D kernel. Default: 3.
+        :param input_dropout:
+            Rate of Dropout applied before the convolution. Default: 0.2.
+        :param feature_map_dropout:
+            Rate of Dropout applied after the convolution. Default: 0.2.
+        :param hidden_dropout:
+            Rate of Dropout applied after the Linear layer. Default: 0.3.
+        :param batch_normalization:
+            If True, apply batch normalization before and after the
+            convolution and after the Linear layer. Default: True.
+        """
+        super(ConvE, self).__init__(negative_sample_sharing=negative_sample_sharing)
+
+        self.sharding = sharding
+
+        if input_channels * embedding_width * embedding_height != embedding_size:
+            raise ValueError(
+                "`embedding_size` needs to be equal to"
+                " `input_channels * embedding_width * embedding_height`"
+            )
+
+        # self.entity_embedding[..., :embedding_size] entity_embeddings
+        # self.entity_embedding[..., -1] tail biases
+        self.entity_embedding = initialize_entity_embedding(
+            self.sharding, entity_initializer, [embedding_size, 1]
+        )
+        self.relation_embedding = initialize_relation_embedding(
+            n_relation_type, inverse_relations, relation_initializer, [embedding_size]
+        )
+        assert (
+            self.entity_embedding.shape[-1] - 1
+            == self.relation_embedding.shape[-1]
+            == embedding_size
+        ), (
+            "ConvE requires `embedding_size + 1` embedding parameters for each entity"
+            " and `embedding_size` embedding parameters for each relation"
+        )
+        self.embedding_size = embedding_size
+
+        self.inp_channels = input_channels
+        self.emb_h = embedding_height
+        self.emb_w = embedding_width
+        conv_layers = [
+            torch.nn.Dropout(input_dropout),
+            torch.nn.Conv2d(
+                in_channels=self.inp_channels,
+                out_channels=output_channels,
+                kernel_size=(kernel_height, kernel_width),
+            ),
+            torch.nn.ReLU(),
+            torch.nn.Dropout2d(feature_map_dropout),
+        ]
+        fc_layers = [
+            torch.nn.Linear(
+                output_channels
+                * (2 * self.emb_h - kernel_height + 1)
+                * (self.emb_w - kernel_width + 1),
+                embedding_size,
+            ),
+            torch.nn.Dropout(hidden_dropout),
+            torch.nn.ReLU(),
+        ]
+        if batch_normalization:
+            conv_layers.insert(0, torch.nn.BatchNorm2d(input_channels))
+            conv_layers.insert(3, torch.nn.BatchNorm2d(output_channels))
+            fc_layers.insert(2, torch.nn.BatchNorm1d(embedding_size))
+        self.conv_layers = torch.nn.Sequential(*conv_layers)
+        self.fc_layers = torch.nn.Sequential(*fc_layers)
+
+    # docstr-coverage: inherited
+
[docs]    def score_triple(
+        self,
+        head_emb: torch.Tensor,
+        relation_id: torch.Tensor,
+        tail_emb: torch.Tensor,
+    ) -> torch.Tensor:
+        relation_emb = torch.index_select(
+            self.relation_embedding, index=relation_id, dim=0
+        )
+        # Discard bias for heads
+        head_emb = head_emb[..., :-1]
+        tail_emb, tail_bias = torch.split(tail_emb, self.embedding_size, dim=-1)
+        hr_cat = torch.cat(
+            [
+                head_emb.view(-1, self.inp_channels, self.emb_h, self.emb_w),
+                relation_emb.view(-1, self.inp_channels, self.emb_h, self.emb_w),
+            ],
+            dim=-2,
+        )
+        hr_conv = self.fc_layers(self.conv_layers(hr_cat).flatten(start_dim=1))
+        return self.reduce_embedding(hr_conv * tail_emb) + tail_bias.squeeze(-1)

+
+    # docstr-coverage: inherited
+
[docs]    def score_heads(
+        self,
+        head_emb: torch.Tensor,
+        relation_id: torch.Tensor,
+        tail_emb: torch.Tensor,
+    ) -> torch.Tensor:
+        raise NotImplementedError("ConvE should not be used with head corruption")

+
+    # docstr-coverage: inherited
+
[docs]    def score_tails(
+        self,
+        head_emb: torch.Tensor,
+        relation_id: torch.Tensor,
+        tail_emb: torch.Tensor,
+    ) -> torch.Tensor:
+        relation_emb = torch.index_select(
+            self.relation_embedding, index=relation_id, dim=0
+        )
+        # Discard bias for heads
+        head_emb = head_emb[..., :-1]
+        tail_emb, tail_bias = torch.split(tail_emb, self.embedding_size, dim=-1)
+        if self.negative_sample_sharing:
+            tail_bias = tail_bias.view(1, -1)
+        else:
+            tail_bias = tail_bias.squeeze(-1)
+        hr_cat = torch.cat(
+            [
+                head_emb.view(-1, self.inp_channels, self.emb_h, self.emb_w),
+                relation_emb.view(-1, self.inp_channels, self.emb_h, self.emb_w),
+            ],
+            dim=-2,
+        )
+        hr_conv = self.fc_layers(self.conv_layers(hr_cat).flatten(start_dim=1))
+        return self.broadcasted_dot_product(hr_conv, tail_emb) + tail_bias

+
+
 
[docs]class BoxE(DistanceBasedScoreFunction):
     """
     BoxE scoring function :cite:p:`BoxE`.
@@ -1074,7 +1275,7 @@ 
Source code for besskge.scoring
             Softening parameter for geometric normalization of box widths.
             Default: 1e-6.
         :param inverse_relations:
-            If True, learn embeddings for inverse relations. Default: False
+            If True, learn embeddings for inverse relations. Default: False.
         """
         super(BoxE, self).__init__(
             negative_sample_sharing=negative_sample_sharing, scoring_norm=scoring_norm
@@ -1332,7 +1533,7 @@ 
Source code for besskge.scoring
         :param offset:
             Offset applied to auxiliary entity embeddings. Default: 1.0.
         :param inverse_relations:
-            If True, learn embeddings for inverse relations. Default: False
+            If True, learn embeddings for inverse relations. Default: False.
         """
         super(InterHT, self).__init__(
             negative_sample_sharing=negative_sample_sharing, scoring_norm=scoring_norm
@@ -1489,7 +1690,7 @@ 
Source code for besskge.scoring
         :param offset:
             Offset applied to tilde entity embeddings. Default: 1.0.
         :param inverse_relations:
-            If True, learn embeddings for inverse relations. Default: False
+            If True, learn embeddings for inverse relations. Default: False.
         """
         super(TranS, self).__init__(
             negative_sample_sharing=negative_sample_sharing, scoring_norm=scoring_norm
diff --git a/_sources/generated/besskge.embedding.init_xavier_norm.rst.txt b/_sources/generated/besskge.embedding.init_xavier_norm.rst.txt
new file mode 100644
index 0000000..ced220a
--- /dev/null
+++ b/_sources/generated/besskge.embedding.init_xavier_norm.rst.txt
@@ -0,0 +1,6 @@
+besskge.embedding.init\_xavier\_norm
+====================================
+
+.. currentmodule:: besskge.embedding
+
+.. autofunction:: init_xavier_norm
\ No newline at end of file
diff --git a/_sources/generated/besskge.embedding.rst.txt b/_sources/generated/besskge.embedding.rst.txt
index 0f9b9d7..901b5be 100644
--- a/_sources/generated/besskge.embedding.rst.txt
+++ b/_sources/generated/besskge.embedding.rst.txt
@@ -21,6 +21,7 @@ besskge.embedding
       init_KGE_normal
       init_KGE_uniform
       init_uniform_norm
+      init_xavier_norm
       initialize_entity_embedding
       initialize_relation_embedding
       refactor_embedding_sharding
diff --git a/_sources/generated/besskge.scoring.ConvE.rst.txt b/_sources/generated/besskge.scoring.ConvE.rst.txt
new file mode 100644
index 0000000..8240814
--- /dev/null
+++ b/_sources/generated/besskge.scoring.ConvE.rst.txt
@@ -0,0 +1,12 @@
+..
+   # Copyright (c) 2023 Graphcore Ltd. All rights reserved.
+   # Copyright (c) 2007-2023 by the Sphinx team. All rights reserved.
+
+besskge.scoring.ConvE
+=====================
+
+.. currentmodule:: besskge.scoring
+
+.. autoclass:: ConvE
+   :members:
+   :inherited-members: Module
\ No newline at end of file
diff --git a/_sources/generated/besskge.scoring.rst.txt b/_sources/generated/besskge.scoring.rst.txt
index dd39740..0fbbdc5 100644
--- a/_sources/generated/besskge.scoring.rst.txt
+++ b/_sources/generated/besskge.scoring.rst.txt
@@ -26,6 +26,7 @@ besskge.scoring
       BaseScoreFunction
       BoxE
       ComplEx
+      ConvE
       DistMult
       DistanceBasedScoreFunction
       InterHT
diff --git a/generated/besskge.dataset.KGDataset.html b/generated/besskge.dataset.KGDataset.html
index 49baa16..7d5eba9 100644
--- a/generated/besskge.dataset.KGDataset.html
+++ b/generated/besskge.dataset.KGDataset.html
@@ -222,12 +222,12 @@ 
besskge.dataset.KGDataset
 
Parameters:

 

 
 
  • besskge.embedding.init_uniform_norm
    • 
+
  • besskge.embedding.init_xavier_norm
    • 
 
  • besskge.embedding.initialize_entity_embedding
    • 
 
  • besskge.embedding.initialize_relation_embedding
    • 
 
  • besskge.embedding.refactor_embedding_sharding
    • 
diff --git a/generated/besskge.embedding.init_uniform_norm.html b/generated/besskge.embedding.init_uniform_norm.html
index 423faae..036d1b5 100644
--- a/generated/besskge.embedding.init_uniform_norm.html
+++ b/generated/besskge.embedding.init_uniform_norm.html
@@ -18,7 +18,7 @@
     
     
     
-    
+    
      
 
 
@@ -62,6 +62,7 @@
 
  • init_uniform_norm()
    • 
 
 
+
  • besskge.embedding.init_xavier_norm
    • 
 
  • besskge.embedding.initialize_entity_embedding
    • 
 
  • besskge.embedding.initialize_relation_embedding
    • 
 
  • besskge.embedding.refactor_embedding_sharding
    • 
@@ -127,7 +128,7 @@ 
    besskge.embedding.init_uniform_norm
         
-        
+        
     
    
 
   
    
diff --git a/generated/besskge.embedding.init_xavier_norm.html b/generated/besskge.embedding.init_xavier_norm.html
new file mode 100644
index 0000000..4d22e0f
--- /dev/null
+++ b/generated/besskge.embedding.init_xavier_norm.html
@@ -0,0 +1,160 @@
+
+
+
+  
+
+  
+  besskge.embedding.init_xavier_norm — BESS-KGE  documentation
+      
+      
+      
+  
+  
+        
+        
+        
+    
+    
+    
+    
+     
+
+
+ 
+  
    
+    
+
+    
    
+
+      
    
+        
    
+          
    
+  
+  
    
+
    
+          
    
+           
    
+             
+  
    
+
    besskge.embedding.init_xavier_norm
    
+
    
+
    
+besskge.embedding.init_xavier_norm(embedding_table, gain=1.0)[source]
    
+
    Initialize embeddings according to Xavier normal scheme, with
+fan_in = 0, fan_out=row_size.
    
+
    
+
    Parameters:
    
+
      
+
    • embedding_table (Tensor) – Tensor of embedding parameters to initialize.
      • 
+
    • gain (float) – Scaling factor for standard deviation. Default: 1.0.
      • 
+
    
+
    
+
    Return type:
    
+
    Tensor
    
+
    
+
    Returns:
    
+
    Initialized tensor.
    
+
    
+
    
+
    
+
+
    
+
+
+           
    
+          
    
+          
+        
    
+      
    
+    
    
+  
    
+   
+
+
+
\ No newline at end of file
diff --git a/generated/besskge.embedding.initialize_entity_embedding.html b/generated/besskge.embedding.initialize_entity_embedding.html
index 85624e2..a97967b 100644
--- a/generated/besskge.embedding.initialize_entity_embedding.html
+++ b/generated/besskge.embedding.initialize_entity_embedding.html
@@ -19,7 +19,7 @@
     
     
     
-     
+     
 
 
  
@@ -59,6 +59,7 @@
 
  • besskge.embedding.init_KGE_normal
    • 
 
  • besskge.embedding.init_KGE_uniform
    • 
 
  • besskge.embedding.init_uniform_norm
    • 
+
  • besskge.embedding.init_xavier_norm
    • 
 
  • besskge.embedding.initialize_entity_embedding
@@ -139,7 +140,7 @@ 
    besskge.embedding.initialize_entity_embedding
-        
+        
         
     
    • 
 
diff --git a/generated/besskge.embedding.initialize_relation_embedding.html b/generated/besskge.embedding.initialize_relation_embedding.html
index 0fbc006..e137d13 100644
--- a/generated/besskge.embedding.initialize_relation_embedding.html
+++ b/generated/besskge.embedding.initialize_relation_embedding.html
@@ -59,6 +59,7 @@
 
  • besskge.embedding.init_KGE_normal
    • 
 
  • besskge.embedding.init_KGE_uniform
    • 
 
  • besskge.embedding.init_uniform_norm
    • 
+
  • besskge.embedding.init_xavier_norm
    • 
 
  • besskge.embedding.initialize_entity_embedding
    • 
 
  • besskge.embedding.initialize_relation_embedding
 
 
diff --git a/generated/besskge.scoring.ComplEx.html b/generated/besskge.scoring.ComplEx.html
index 2354795..eceab6d 100644
--- a/generated/besskge.scoring.ComplEx.html
+++ b/generated/besskge.scoring.ComplEx.html
@@ -18,7 +18,7 @@
     
     
     
-    
+    
      
 
 
@@ -59,6 +59,7 @@
 
  • ComplEx
    • 
 
 
+
  • besskge.scoring.ConvE
    • 
 
  • besskge.scoring.DistMult
    • 
 
  • besskge.scoring.DistanceBasedScoreFunction
    • 
 
  • besskge.scoring.InterHT
    • 
@@ -122,7 +123,7 @@ 
    besskge.scoring.ComplEx
    embedding_size (int) – Complex size of entity and relation embeddings.
    
 
  • entity_initializer (Union[Tensor, List[Callable[..., Tensor]]]) – Initialization function or table for entity embeddings.
    • 
 
  • relation_initializer (Union[Tensor, List[Callable[..., Tensor]]]) – Initialization function or table for relation embeddings.
    • 
-
  • inverse_relations (bool) – If True, learn embeddings for inverse relations. Default: False
    • 
+
  • inverse_relations (bool) – If True, learn embeddings for inverse relations. Default: False.
    • 
 
 
 
@@ -315,7 +316,7 @@ 
    besskge.scoring.ComplEx
           
    
 
   
    
diff --git a/generated/besskge.scoring.ConvE.html b/generated/besskge.scoring.ConvE.html
new file mode 100644
index 0000000..523e5f4
--- /dev/null
+++ b/generated/besskge.scoring.ConvE.html
@@ -0,0 +1,367 @@
+
+
+
+  
+
+  
+  besskge.scoring.ConvE — BESS-KGE  documentation
+      
+      
+      
+  
+  
+        
+        
+        
+    
+    
+    
+    
+     
+
+
+ 
+  
    
+    
+
+    
    
+
+      
    
+        
    
+          
    
+  
+  
    
+
    
+          
    
+           
    
+             
+  
    
+
    besskge.scoring.ConvE
    
+
    
+
    
+class besskge.scoring.ConvE(negative_sample_sharing, sharding, n_relation_type, embedding_size, embedding_height, embedding_width, entity_initializer=[<function init_xavier_norm>, <function zeros_>], relation_initializer=[<function init_xavier_norm>], inverse_relations=True, input_channels=1, output_channels=32, kernel_height=3, kernel_width=3, input_dropout=0.2, feature_map_dropout=0.2, hidden_dropout=0.3, batch_normalization=True)[source]
    
+
    ConvE scoring function [DPPR18].
    
+
    Note that, differently from [DPPR18], the scores returned by this class
+have not been passed through a final sigmoid layer, as we assume that this is
+included in the loss function.
    
+
    By design, this scoring function should be used in combination with a
+negative/candidate sampler that only corrupts tails (possibly after
+including all inverse triples in the dataset, see the add_inverse_triples
+argument in besskge.sharding.PartitionedTripleSet.create_from_dataset()).
    
+
    Initialize ConvE model.
    
+
    
+
    Parameters:
    
+
      
+
    • negative_sample_sharing (bool) – see DistanceBasedScoreFunction.__init__()
      • 
+
    • sharding (Sharding) – Entity sharding.
      • 
+
    • n_relation_type (int) – Number of relation types in the knowledge graph.
      • 
+
    • embedding_size (int) – Size of entity and relation embeddings.
      • 
+
    • embedding_height (int) – Height of the 2D-reshaping of the concatenation of
+head and relation embeddings.
      • 
+
    • embedding_width (int) – Width of the 2D-reshaping of the concatenation of
+head and relation embeddings.
      • 
+
    • entity_initializer (Union[Tensor, List[Callable[..., Tensor]]]) – Initialization functions or table for entity embeddings.
+If not passing a table, two functions are needed: the initializer
+for entity embeddings and initializer for (scalar) tail biases.
      • 
+
    • relation_initializer (Union[Tensor, List[Callable[..., Tensor]]]) – Initialization function or table for relation embeddings.
      • 
+
    • inverse_relations (bool) – If True, learn embeddings for inverse relations. Default: True.
      • 
+
    • input_channels (int) – Number of input channels of the Conv2D operator. Default: 1.
      • 
+
    • output_channels (int) – Number of output channels of the Conv2D operator. Default: 32.
      • 
+
    • kernel_height (int) – Height of the Conv2D kernel. Default: 3.
      • 
+
    • kernel_width (int) – Width of the Conv2D kernel. Default: 3.
      • 
+
    • input_dropout (float) – Rate of Dropout applied before the convolution. Default: 0.2.
      • 
+
    • feature_map_dropout (float) – Rate of Dropout applied after the convolution. Default: 0.2.
      • 
+
    • hidden_dropout (float) – Rate of Dropout applied after the Linear layer. Default: 0.3.
      • 
+
    • batch_normalization (bool) – If True, apply batch normalization before and after the
+convolution and after the Linear layer. Default: True.
      • 
+
    
+
    
+
    
+
    
+
    
+broadcasted_dot_product(v1, v2)
    
+
    Broadcasted dot product of queries against sets of entities.
    
+
    For each query and candidate, computes the dot product of
+the embeddings.
    
+
    
+
    Parameters:
    
+
      
+
    • v1 (Tensor) – shape: (batch_size, embedding_size)
+Batch queries.
      • 
+
    • v2 (Tensor) – shape: (B, n_neg, embedding_size) with B = 1, batch_size
+Negative embeddings to score against queries.
      • 
+
    
+
    
+
    Return type:
    
+
    Tensor
    
+
    
+
    Returns:
    
+
    shape: (batch_size, B * n_neg) if
+BaseScoreFunction.negative_sample_sharing
+else (batch_size, n_neg)
    
+
    
+
    
+
    
+
+
    
+
    
+entity_embedding: Parameter
    
+
    Entity embedding table
    
+
    
+
+
    
+
    
+forward(head_emb, relation_id, tail_emb)
    
+
    see BaseScoreFunction.score_triple()
    
+
    
+
    Return type:
    
+
    Tensor
    
+
    
+
    Parameters:
    
+
    
+
    
+
    
+
    
+
+
    
+
    
+negative_sample_sharing: bool
    
+
    Share negative entities to construct negative samples
    
+
    
+
+
    
+
    
+reduce_embedding(v)
    
+
    Sum reduction along the embedding dimension.
    
+
    
+
    Parameters:
    
+
    v (Tensor) – shape: (*, embedding_size)
+The tensor to reduce.
    
+
    
+
    Return type:
    
+
    Tensor
    
+
    
+
    Returns:
    
+
    shape: (*,)
+Sum reduction.
    
+
    
+
    
+
    
+
+
    
+
    
+relation_embedding: Parameter
    
+
    Relation embedding table
    
+
    
+
+
    
+
    
+score_heads(head_emb, relation_id, tail_emb)[source]
    
+
    Score sets of head entities against fixed (r,t) queries.
    
+
    
+
    Parameters:
    
+
      
+
    • head_emb (Tensor) – shape: (B, n_heads, embedding_size) with B = 1, batch_size
+Embeddings of head entities.
      • 
+
    • relation_id (Tensor) – shape: (batch_size,)
+IDs of relation types in batch.
      • 
+
    • tail_emb (Tensor) – shape: (batch_size, embedding_size)
+Embedding of tail entities in batch.
      • 
+
    
+
    
+
    Return type:
    
+
    Tensor
    
+
    
+
    Returns:
    
+
    shape: (batch_size, B * n_heads)
+if BaseScoreFunction.negative_sample_sharing
+else (batch_size, n_heads).
+Scores of broadcasted triples.
    
+
    
+
    
+
    
+
+
    
+
    
+score_tails(head_emb, relation_id, tail_emb)[source]
    
+
    Score sets of tail entities against fixed (h,r) queries.
    
+
    
+
    Parameters:
    
+
      
+
    • head_emb (Tensor) – shape: (batch_size, embedding_size)
+Embeddings of head entities in batch.
      • 
+
    • relation_id (Tensor) – shape: (batch_size,)
+IDs of relation types in batch.
      • 
+
    • tail_emb (Tensor) – shape: (B, n_tails, embedding_size) with B = 1, batch_size
+Embedding of tail entities.
      • 
+
    
+
    
+
    Return type:
    
+
    Tensor
    
+
    
+
    Returns:
    
+
    shape: (batch_size, B * n_tails)
+if BaseScoreFunction.negative_sample_sharing
+else (batch_size, n_tails)
+Scores of broadcasted triples.
    
+
    
+
    
+
    
+
+
    
+
    
+score_triple(head_emb, relation_id, tail_emb)[source]
    
+
    Score a batch of (h,r,t) triples.
    
+
    
+
    Parameters:
    
+
      
+
    • head_emb (Tensor) – shape: (batch_size, embedding_size)
+Embeddings of head entities in batch.
      • 
+
    • relation_id (Tensor) – shape: (batch_size,)
+IDs of relation types in batch.
      • 
+
    • tail_emb (Tensor) – shape: (batch_size, embedding_size)
+Embeddings of tail entities in batch.
      • 
+
    
+
    
+
    Return type:
    
+
    Tensor
    
+
    
+
    Returns:
    
+
    shape: (batch_size,)
+Scores of a batch of triples.
    
+
    
+
    
+
    
+
+
    
+
    
+sharding: Sharding
    
+
    Sharding of entities
    
+
    
+
+
    
+
    
+update_sharding(new_sharding)
    
+
    Change the sharding of the entity embedding table.
    
+
    
+
    Parameters:
    
+
    new_sharding (Sharding) – The new entity sharding.
    
+
    
+
    Return type:
    
+
    None
    
+
    
+
    
+
    
+
+
    
+
+
    
+
+
+           
    
+          
    
+          
+        
    
+      
    
+    
    
+  
    
+   
+
+
+
\ No newline at end of file
diff --git a/generated/besskge.scoring.DistMult.html b/generated/besskge.scoring.DistMult.html
index 5e107f8..a133b02 100644
--- a/generated/besskge.scoring.DistMult.html
+++ b/generated/besskge.scoring.DistMult.html
@@ -19,7 +19,7 @@
     
     
     
-     
+     
 
 
  
@@ -56,6 +56,7 @@
 
  • besskge.scoring.BaseScoreFunction
    • 
 
  • besskge.scoring.BoxE
    • 
 
  • besskge.scoring.ComplEx
    • 
+
  • besskge.scoring.ConvE
    • 
 
  • besskge.scoring.DistMult
@@ -122,7 +123,7 @@ 
    besskge.scoring.DistMultint) – Size of entity and relation embeddings.
    • 
 
  • entity_initializer (Union[Tensor, List[Callable[..., Tensor]]]) – Initialization function or table for entity embeddings.
    • 
 
  • relation_initializer (Union[Tensor, List[Callable[..., Tensor]]]) – Initialization function or table for relation embeddings.
    • 
-
  • inverse_relations (bool) – If True, learn embeddings for inverse relations. Default: False
    • 
+
  • inverse_relations (bool) – If True, learn embeddings for inverse relations. Default: False.
    • 
 
 
 
@@ -314,7 +315,7 @@ 
    besskge.scoring.DistMult
-        
+        
         
     
    
 
diff --git a/generated/besskge.scoring.DistanceBasedScoreFunction.html b/generated/besskge.scoring.DistanceBasedScoreFunction.html
index 1a795bd..a32fdf9 100644
--- a/generated/besskge.scoring.DistanceBasedScoreFunction.html
+++ b/generated/besskge.scoring.DistanceBasedScoreFunction.html
@@ -56,6 +56,7 @@
 
  • besskge.scoring.BaseScoreFunction
    • 
 
  • besskge.scoring.BoxE
    • 
 
  • besskge.scoring.ComplEx
    • 
+
  • besskge.scoring.ConvE
    • 
 
  • besskge.scoring.DistMult
    • 
 
  • besskge.scoring.DistanceBasedScoreFunction
 
 
diff --git a/generated/besskge.scoring.RotatE.html b/generated/besskge.scoring.RotatE.html
index 9606fbe..4d010d4 100644
--- a/generated/besskge.scoring.RotatE.html
+++ b/generated/besskge.scoring.RotatE.html
@@ -56,6 +56,7 @@
 
  • besskge.scoring.BaseScoreFunction
    • 
 
  • besskge.scoring.BoxE
    • 
 
  • besskge.scoring.ComplEx
    • 
+
  • besskge.scoring.ConvE
    • 
 
  • besskge.scoring.DistMult
    • 
 
  • besskge.scoring.DistanceBasedScoreFunction
    • 
 
  • besskge.scoring.InterHT
    • 
@@ -124,7 +125,7 @@ 
    besskge.scoring.RotatE
 
  • entity_initializer (Union[Tensor, List[Callable[..., Tensor]]]) – Initialization function or table for entity embeddings.
    • 
 
  • relation_initializer (Union[Tensor, List[Callable[..., Tensor]]]) – Initialization function or table for relation embeddings.
    • 
-
  • inverse_relations (bool) – If True, learn embeddings for inverse relations. Default: False
    • 
+
  • inverse_relations (bool) – If True, learn embeddings for inverse relations. Default: False.
    • 
 
 
 
diff --git a/generated/besskge.scoring.TranS.html b/generated/besskge.scoring.TranS.html
index 280e2e1..d9b3571 100644
--- a/generated/besskge.scoring.TranS.html
+++ b/generated/besskge.scoring.TranS.html
@@ -56,6 +56,7 @@
 
  • besskge.scoring.BaseScoreFunction
    • 
 
  • besskge.scoring.BoxE
    • 
 
  • besskge.scoring.ComplEx
    • 
+
  • besskge.scoring.ConvE
    • 
 
  • besskge.scoring.DistMult
    • 
 
  • besskge.scoring.DistanceBasedScoreFunction
    • 
 
  • besskge.scoring.InterHT
    • 
@@ -126,7 +127,7 @@ 
    besskge.scoring.TranS
    normalize_entities (bool) – If True, L2-normalize embeddings of head and tail entities as well as
 tilde head and tail entities before multiplying. Default: True.
    
 
  • offset (float) – Offset applied to tilde entity embeddings. Default: 1.0.
    • 
-
  • inverse_relations (bool) – If True, learn embeddings for inverse relations. Default: False
    • 
+
  • inverse_relations (bool) – If True, learn embeddings for inverse relations. Default: False.
    • 
 
 
 
diff --git a/generated/besskge.scoring.TransE.html b/generated/besskge.scoring.TransE.html
index 9046559..8a9df7f 100644
--- a/generated/besskge.scoring.TransE.html
+++ b/generated/besskge.scoring.TransE.html
@@ -56,6 +56,7 @@
 
  • besskge.scoring.BaseScoreFunction
    • 
 
  • besskge.scoring.BoxE
    • 
 
  • besskge.scoring.ComplEx
    • 
+
  • besskge.scoring.ConvE
    • 
 
  • besskge.scoring.DistMult
    • 
 
  • besskge.scoring.DistanceBasedScoreFunction
    • 
 
  • besskge.scoring.InterHT
    • 
@@ -123,7 +124,7 @@ 
    besskge.scoring.TransE
    embedding_size (int) – Size of entity and relation embeddings.
    
 
  • entity_initializer (Union[Tensor, List[Callable[..., Tensor]]]) – Initialization function or table for entity embeddings.
    • 
 
  • relation_initializer (Union[Tensor, List[Callable[..., Tensor]]]) – Initialization function or table for relation embeddings.
    • 
-
  • inverse_relations (bool) – If True, learn embeddings for inverse relations. Default: False
    • 
+
  • inverse_relations (bool) – If True, learn embeddings for inverse relations. Default: False.
    • 
 
 
 
diff --git a/generated/besskge.scoring.TripleRE.html b/generated/besskge.scoring.TripleRE.html
index 994fd02..35b2d9c 100644
--- a/generated/besskge.scoring.TripleRE.html
+++ b/generated/besskge.scoring.TripleRE.html
@@ -56,6 +56,7 @@
 
  • besskge.scoring.BaseScoreFunction
    • 
 
  • besskge.scoring.BoxE
    • 
 
  • besskge.scoring.ComplEx
    • 
+
  • besskge.scoring.ConvE
    • 
 
  • besskge.scoring.DistMult
    • 
 
  • besskge.scoring.DistanceBasedScoreFunction
    • 
 
  • besskge.scoring.InterHT
    • 
@@ -127,7 +128,7 @@ 
    besskge.scoring.TripleREfloat) – Offset factor for head/tail relation projections, as in TripleREv2.
 Default: 0.0 (no offset).
    
-
  • inverse_relations (bool) – If True, learn embeddings for inverse relations. Default: False
    • 
+
  • inverse_relations (bool) – If True, learn embeddings for inverse relations. Default: False.
    • 
 
 
 
diff --git a/generated/besskge.scoring.html b/generated/besskge.scoring.html
index 56fcc73..55e0406 100644
--- a/generated/besskge.scoring.html
+++ b/generated/besskge.scoring.html
@@ -56,6 +56,7 @@
 
  • besskge.scoring.BaseScoreFunction
    • 
 
  • besskge.scoring.BoxE
    • 
 
  • besskge.scoring.ComplEx
    • 
+
  • besskge.scoring.ConvE
    • 
 
  • besskge.scoring.DistMult
    • 
 
  • besskge.scoring.DistanceBasedScoreFunction
    • 
 
  • besskge.scoring.InterHT
    • 
@@ -118,32 +119,35 @@
 
    ComplEx(negative_sample_sharing, sharding, ...)
    
 
    ComplEx scoring function [TWR+16].
    
 
-
    DistMult(negative_sample_sharing, sharding, ...)
    
-
    DistMult scoring function [YYH+15].
    
+
    ConvE(negative_sample_sharing, sharding, ...)
    
+
    ConvE scoring function [DPPR18].
    
 
-
    DistanceBasedScoreFunction(...)
    
+
    DistMult(negative_sample_sharing, sharding, ...)
    
+
    DistMult scoring function [YYH+15].
    
+
+
    DistanceBasedScoreFunction(...)
    
 
    Base class for distance-based scoring functions.
    
 
-
    InterHT(negative_sample_sharing, ...[, ...])
    
-
    InterHT scoring function [WMW+22].
    
+
    InterHT(negative_sample_sharing, ...[, ...])
    
+
    InterHT scoring function [WMW+22].
    
 
-
    MatrixDecompositionScoreFunction(...)
    
+
    MatrixDecompositionScoreFunction(...)
    
 
    Base class for matrix-decomposition scoring functions.
    
 
-
    PairRE(negative_sample_sharing, ...[, ...])
    
-
    PairRE scoring function [CHWC21].
    
+
    PairRE(negative_sample_sharing, ...[, ...])
    
+
    PairRE scoring function [CHWC21].
    
 
-
    RotatE(negative_sample_sharing, ...[, ...])
    
-
    RotatE scoring function [SDNT19].
    
+
    RotatE(negative_sample_sharing, ...[, ...])
    
+
    RotatE scoring function [SDNT19].
    
 
-
    TranS(negative_sample_sharing, scoring_norm, ...)
    
-
    TranS scoring function [ZYX22].
    
+
    TranS(negative_sample_sharing, scoring_norm, ...)
    
+
    TranS scoring function [ZYX22].
    
 
-
    TransE(negative_sample_sharing, ...[, ...])
    
-
    TransE scoring function [BUGD+13].
    
+
    TransE(negative_sample_sharing, ...[, ...])
    
+
    TransE scoring function [BUGD+13].
    
 
-
    TripleRE(negative_sample_sharing, ...[, ...])
    
-
    TripleRE scoring function [YLL+22].
    
+
    TripleRE(negative_sample_sharing, ...[, ...])
    
+
    TripleRE scoring function [YLL+22].
    
 
 
 
diff --git a/genindex.html b/genindex.html
index e3be28c..30d510a 100644
--- a/genindex.html
+++ b/genindex.html
@@ -181,8 +181,6 @@ 
    B
    
         
  • module
 
    • 
       
-  
-  
      
       
    • 
     besskge.scoring
 
@@ -190,6 +188,8 @@ 
      B
      
         
    • module
 
      • 
       
    
+  
+  
-  
+  
      
       
    • negative_sample_sharing (besskge.scoring.BaseScoreFunction attribute)
 
       
        
         
      • (besskge.scoring.BoxE attribute)
 
        • 
         
      • (besskge.scoring.ComplEx attribute)
+
        • 
+        
      • (besskge.scoring.ConvE attribute)
 
        • 
         
      • (besskge.scoring.DistanceBasedScoreFunction attribute)
 
        • 
@@ -703,6 +715,8 @@ 
        R
        
 
       
          
         
        • (besskge.scoring.ComplEx method)
+
          • 
+        
        • (besskge.scoring.ConvE method)
 
          • 
         
        • (besskge.scoring.DistanceBasedScoreFunction method)
 
          • 
@@ -735,6 +749,8 @@ 
          R
          
         
        • (besskge.scoring.BoxE attribute)
 
          • 
         
        • (besskge.scoring.ComplEx attribute)
+
          • 
+        
        • (besskge.scoring.ConvE attribute)
 
          • 
         
        • (besskge.scoring.DistanceBasedScoreFunction attribute)
 
          • 
@@ -807,6 +823,8 @@ 
          S
          
         
        • (besskge.scoring.BoxE method)
 
          • 
         
        • (besskge.scoring.ComplEx method)
+
          • 
+        
        • (besskge.scoring.ConvE method)
 
          • 
         
        • (besskge.scoring.DistanceBasedScoreFunction method)
 
          • 
@@ -833,6 +851,8 @@ 
          S
          
         
        • (besskge.scoring.BoxE method)
 
          • 
         
        • (besskge.scoring.ComplEx method)
+
          • 
+        
        • (besskge.scoring.ConvE method)
 
          • 
         
        • (besskge.scoring.DistanceBasedScoreFunction method)
 
          • 
@@ -861,6 +881,8 @@ 
          S
          
         
        • (besskge.scoring.BoxE method)
 
          • 
         
        • (besskge.scoring.ComplEx method)
+
          • 
+        
        • (besskge.scoring.ConvE method)
 
          • 
         
        • (besskge.scoring.DistanceBasedScoreFunction method)
 
          • 
@@ -899,6 +921,8 @@ 
          S
          
         
        • (besskge.scoring.BoxE attribute)
 
          • 
         
        • (besskge.scoring.ComplEx attribute)
+
          • 
+        
        • (besskge.scoring.ConvE attribute)
 
          • 
         
        • (besskge.scoring.DistanceBasedScoreFunction attribute)
 
          • 
@@ -973,6 +997,8 @@ 
          U
          
         
        • (besskge.scoring.BoxE method)
 
          • 
         
        • (besskge.scoring.ComplEx method)
+
          • 
+        
        • (besskge.scoring.ConvE method)
 
          • 
         
        • (besskge.scoring.DistanceBasedScoreFunction method)
 
          • 
diff --git a/index.html b/index.html
index 54ee68c..3a13deb 100644
--- a/index.html
+++ b/index.html
@@ -142,6 +142,7 @@
 
        • besskge.scoring.BaseScoreFunction
          • 
 
        • besskge.scoring.BoxE
          • 
 
        • besskge.scoring.ComplEx
          • 
+
        • besskge.scoring.ConvE
          • 
 
        • besskge.scoring.DistMult
          • 
 
        • besskge.scoring.DistanceBasedScoreFunction
          • 
 
        • besskge.scoring.InterHT
          • 
@@ -173,6 +174,7 @@
 
        • besskge.embedding.init_KGE_normal
          • 
 
        • besskge.embedding.init_KGE_uniform
          • 
 
        • besskge.embedding.init_uniform_norm
          • 
+
        • besskge.embedding.init_xavier_norm
          • 
 
        • besskge.embedding.initialize_entity_embedding
          • 
 
        • besskge.embedding.initialize_relation_embedding
          • 
 
        • besskge.embedding.refactor_embedding_sharding
          • 
diff --git a/objects.inv b/objects.inv
index e2a7e02..a7331e8 100644
Binary files a/objects.inv and b/objects.inv differ
diff --git a/searchindex.js b/searchindex.js
index 8fc6ee6..41970ff 100644
--- a/searchindex.js
+++ b/searchindex.js
@@ -1 +1 @@
-Search.setIndex({"docnames": ["API_reference", "bess", "bibliography", "contrib", "dev_guide", "generated/besskge.batch_sampler", "generated/besskge.batch_sampler.RandomShardedBatchSampler", "generated/besskge.batch_sampler.RigidShardedBatchSampler", "generated/besskge.batch_sampler.ShardedBatchSampler", "generated/besskge.bess", "generated/besskge.bess.AllScoresBESS", "generated/besskge.bess.BessKGE", "generated/besskge.bess.EmbeddingMovingBessKGE", "generated/besskge.bess.ScoreMovingBessKGE", "generated/besskge.bess.TopKQueryBessKGE", "generated/besskge.dataset", "generated/besskge.dataset.KGDataset", "generated/besskge.embedding", "generated/besskge.embedding.init_KGE_normal", "generated/besskge.embedding.init_KGE_uniform", "generated/besskge.embedding.init_uniform_norm", "generated/besskge.embedding.initialize_entity_embedding", "generated/besskge.embedding.initialize_relation_embedding", "generated/besskge.embedding.refactor_embedding_sharding", "generated/besskge.loss", "generated/besskge.loss.BaseLossFunction", "generated/besskge.loss.LogSigmoidLoss", "generated/besskge.loss.MarginBasedLossFunction", "generated/besskge.loss.MarginRankingLoss", "generated/besskge.loss.SampledSoftmaxCrossEntropyLoss", "generated/besskge.metric", "generated/besskge.metric.BaseMetric", "generated/besskge.metric.Evaluation", "generated/besskge.metric.HitsAtK", "generated/besskge.metric.METRICS_DICT", "generated/besskge.metric.ReciprocalRank", "generated/besskge.negative_sampler", "generated/besskge.negative_sampler.PlaceholderNegativeSampler", "generated/besskge.negative_sampler.RandomShardedNegativeSampler", "generated/besskge.negative_sampler.ShardedNegativeSampler", "generated/besskge.negative_sampler.TripleBasedShardedNegativeSampler", "generated/besskge.negative_sampler.TypeBasedShardedNegativeSampler", "generated/besskge.pipeline", "generated/besskge.pipeline.AllScoresPipeline", "generated/besskge.scoring", "generated/besskge.scoring.BaseScoreFunction", "generated/besskge.scoring.BoxE", "generated/besskge.scoring.ComplEx", "generated/besskge.scoring.DistMult", "generated/besskge.scoring.DistanceBasedScoreFunction", "generated/besskge.scoring.InterHT", "generated/besskge.scoring.MatrixDecompositionScoreFunction", "generated/besskge.scoring.PairRE", "generated/besskge.scoring.RotatE", "generated/besskge.scoring.TranS", "generated/besskge.scoring.TransE", "generated/besskge.scoring.TripleRE", "generated/besskge.sharding", "generated/besskge.sharding.PartitionedTripleSet", "generated/besskge.sharding.Sharding", "generated/besskge.utils", "generated/besskge.utils.complex_multiplication", "generated/besskge.utils.complex_rotation", "generated/besskge.utils.gather_indices", "generated/besskge.utils.get_entity_filter", "index", "user_guide"], "filenames": ["API_reference.rst", "bess.rst", "bibliography.rst", "contrib.md", "dev_guide.rst", "generated/besskge.batch_sampler.rst", "generated/besskge.batch_sampler.RandomShardedBatchSampler.rst", "generated/besskge.batch_sampler.RigidShardedBatchSampler.rst", "generated/besskge.batch_sampler.ShardedBatchSampler.rst", "generated/besskge.bess.rst", "generated/besskge.bess.AllScoresBESS.rst", "generated/besskge.bess.BessKGE.rst", "generated/besskge.bess.EmbeddingMovingBessKGE.rst", "generated/besskge.bess.ScoreMovingBessKGE.rst", "generated/besskge.bess.TopKQueryBessKGE.rst", "generated/besskge.dataset.rst", "generated/besskge.dataset.KGDataset.rst", "generated/besskge.embedding.rst", "generated/besskge.embedding.init_KGE_normal.rst", "generated/besskge.embedding.init_KGE_uniform.rst", "generated/besskge.embedding.init_uniform_norm.rst", "generated/besskge.embedding.initialize_entity_embedding.rst", "generated/besskge.embedding.initialize_relation_embedding.rst", "generated/besskge.embedding.refactor_embedding_sharding.rst", "generated/besskge.loss.rst", "generated/besskge.loss.BaseLossFunction.rst", "generated/besskge.loss.LogSigmoidLoss.rst", "generated/besskge.loss.MarginBasedLossFunction.rst", "generated/besskge.loss.MarginRankingLoss.rst", "generated/besskge.loss.SampledSoftmaxCrossEntropyLoss.rst", "generated/besskge.metric.rst", "generated/besskge.metric.BaseMetric.rst", "generated/besskge.metric.Evaluation.rst", "generated/besskge.metric.HitsAtK.rst", "generated/besskge.metric.METRICS_DICT.rst", "generated/besskge.metric.ReciprocalRank.rst", "generated/besskge.negative_sampler.rst", "generated/besskge.negative_sampler.PlaceholderNegativeSampler.rst", "generated/besskge.negative_sampler.RandomShardedNegativeSampler.rst", "generated/besskge.negative_sampler.ShardedNegativeSampler.rst", "generated/besskge.negative_sampler.TripleBasedShardedNegativeSampler.rst", "generated/besskge.negative_sampler.TypeBasedShardedNegativeSampler.rst", "generated/besskge.pipeline.rst", "generated/besskge.pipeline.AllScoresPipeline.rst", "generated/besskge.scoring.rst", "generated/besskge.scoring.BaseScoreFunction.rst", "generated/besskge.scoring.BoxE.rst", "generated/besskge.scoring.ComplEx.rst", "generated/besskge.scoring.DistMult.rst", "generated/besskge.scoring.DistanceBasedScoreFunction.rst", "generated/besskge.scoring.InterHT.rst", "generated/besskge.scoring.MatrixDecompositionScoreFunction.rst", "generated/besskge.scoring.PairRE.rst", "generated/besskge.scoring.RotatE.rst", "generated/besskge.scoring.TranS.rst", "generated/besskge.scoring.TransE.rst", "generated/besskge.scoring.TripleRE.rst", "generated/besskge.sharding.rst", "generated/besskge.sharding.PartitionedTripleSet.rst", "generated/besskge.sharding.Sharding.rst", "generated/besskge.utils.rst", "generated/besskge.utils.complex_multiplication.rst", "generated/besskge.utils.complex_rotation.rst", "generated/besskge.utils.gather_indices.rst", "generated/besskge.utils.get_entity_filter.rst", "index.rst", "user_guide.rst"], "titles": ["BESS-KGE API Reference", "BESS overview", "Bibliography", "How to contribute to the BESS-KGE project", "How to contribute to the BESS-KGE project", "besskge.batch_sampler", "besskge.batch_sampler.RandomShardedBatchSampler", "besskge.batch_sampler.RigidShardedBatchSampler", "besskge.batch_sampler.ShardedBatchSampler", "besskge.bess", "besskge.bess.AllScoresBESS", "besskge.bess.BessKGE", "besskge.bess.EmbeddingMovingBessKGE", "besskge.bess.ScoreMovingBessKGE", "besskge.bess.TopKQueryBessKGE", "besskge.dataset", "besskge.dataset.KGDataset", "besskge.embedding", "besskge.embedding.init_KGE_normal", "besskge.embedding.init_KGE_uniform", "besskge.embedding.init_uniform_norm", "besskge.embedding.initialize_entity_embedding", "besskge.embedding.initialize_relation_embedding", "besskge.embedding.refactor_embedding_sharding", "besskge.loss", "besskge.loss.BaseLossFunction", "besskge.loss.LogSigmoidLoss", "besskge.loss.MarginBasedLossFunction", "besskge.loss.MarginRankingLoss", "besskge.loss.SampledSoftmaxCrossEntropyLoss", "besskge.metric", "besskge.metric.BaseMetric", "besskge.metric.Evaluation", "besskge.metric.HitsAtK", "besskge.metric.METRICS_DICT", "besskge.metric.ReciprocalRank", "besskge.negative_sampler", "besskge.negative_sampler.PlaceholderNegativeSampler", "besskge.negative_sampler.RandomShardedNegativeSampler", "besskge.negative_sampler.ShardedNegativeSampler", "besskge.negative_sampler.TripleBasedShardedNegativeSampler", "besskge.negative_sampler.TypeBasedShardedNegativeSampler", "besskge.pipeline", "besskge.pipeline.AllScoresPipeline", "besskge.scoring", "besskge.scoring.BaseScoreFunction", "besskge.scoring.BoxE", "besskge.scoring.ComplEx", "besskge.scoring.DistMult", "besskge.scoring.DistanceBasedScoreFunction", "besskge.scoring.InterHT", "besskge.scoring.MatrixDecompositionScoreFunction", "besskge.scoring.PairRE", "besskge.scoring.RotatE", "besskge.scoring.TranS", "besskge.scoring.TransE", "besskge.scoring.TripleRE", "besskge.sharding", "besskge.sharding.PartitionedTripleSet", "besskge.sharding.Sharding", "besskge.utils", "besskge.utils.complex_multiplication", "besskge.utils.complex_rotation", "besskge.utils.gather_indices", "besskge.utils.get_entity_filter", "BESS-KGE", "User guide"], "terms": {"when": [1, 3, 4, 11, 12, 13, 14, 22, 25, 26, 27, 28, 29, 36, 37, 40], "distribut": [1, 5, 9, 10, 11, 14, 18, 19, 20, 57, 65, 66], "workload": 1, "over": [1, 6, 7, 8, 14, 65], "n": [1, 40, 58], "worker": [1, 6, 7, 8, 65], "ipu": [1, 3, 4, 6, 7, 8, 9, 10, 43, 63, 65, 66], "randomli": [1, 16], "split": [1, 16, 40], "entiti": [1, 2, 6, 7, 8, 10, 11, 12, 13, 14, 16, 17, 21, 23, 29, 32, 33, 36, 37, 38, 39, 40, 41, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59, 64, 65, 66], "embed": [1, 2, 11, 12, 13, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 65, 66], "tabl": [1, 11, 12, 13, 17, 21, 22, 23, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 66], "shard": [1, 6, 7, 8, 10, 12, 13, 14, 21, 23, 37, 38, 39, 40, 41, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 65], "equal": [1, 12], "size": [1, 6, 7, 8, 10, 13, 14, 40, 43, 46, 47, 48, 50, 52, 53, 54, 55, 56, 66], "each": [1, 5, 6, 7, 8, 10, 11, 12, 13, 14, 16, 20, 21, 22, 32, 38, 40, 43, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59, 64], "which": [1, 12, 16, 37, 38, 39, 40, 41, 64], "i": [1, 3, 4, 6, 7, 8, 12, 13, 14, 16, 21, 22, 33, 38, 40, 43, 46, 58, 62, 63, 64, 65], "store": [1, 13, 15, 16, 43, 58, 65, 66], "": [1, 2, 3, 4], "memori": [1, 6, 7, 8, 11, 12, 13, 65], "The": [1, 3, 4, 6, 7, 8, 10, 11, 12, 13, 16, 23, 25, 26, 27, 28, 29, 32, 40, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59, 64, 65, 66], "relat": [1, 2, 10, 11, 12, 13, 14, 16, 17, 22, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 65], "type": [1, 6, 7, 8, 10, 11, 12, 13, 14, 16, 18, 19, 20, 21, 22, 23, 25, 26, 27, 28, 29, 32, 40, 41, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59, 61, 62, 63, 64, 65], "other": [1, 6, 7, 8, 11, 12, 13], "hand": 1, "replic": [1, 13], "across": [1, 3, 4], "usual": 1, "much": 1, "smaller": 1, "figur": 1, "1": [1, 2, 10, 11, 12, 13, 14, 16, 18, 19, 20, 26, 27, 28, 29, 32, 40, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 63, 66], "3": [1, 3, 4, 16, 64, 66], "induc": 1, "partit": [1, 6, 7, 8, 10, 11, 14, 37, 38, 39, 40, 41, 43, 58], "tripl": [1, 2, 5, 6, 7, 8, 10, 11, 12, 13, 14, 15, 16, 22, 25, 26, 27, 28, 29, 32, 33, 37, 38, 39, 40, 41, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 64, 65], "dataset": [1, 2, 22, 58, 65, 66], "accord": [1, 5, 11, 12, 13, 18, 19, 20, 23, 32], "pair": [1, 2, 38, 58, 64], "head": [1, 2, 10, 11, 12, 13, 14, 16, 36, 38, 40, 41, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 64], "tail": [1, 2, 10, 11, 12, 13, 14, 16, 36, 38, 40, 41, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 64], "At": [1, 7, 14], "execut": [1, 57], "time": [1, 6, 7, 8, 65], "both": [1, 3, 4, 25, 45, 65], "train": [1, 9, 11, 12, 13, 16, 42, 43, 65, 66], "infer": [1, 2, 6, 7, 8, 9, 10, 11, 14, 42, 43, 65, 66], "batch": [1, 5, 6, 7, 8, 10, 11, 12, 13, 14, 24, 25, 26, 27, 28, 29, 32, 38, 40, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 61, 62, 65, 66], "ar": [1, 3, 4, 10, 12, 13, 14, 16, 22, 25, 32, 40, 46, 58, 63, 65, 66], "construct": [1, 25, 26, 27, 28, 29, 36, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59, 64], "sampl": [1, 2, 5, 6, 7, 8, 11, 12, 13, 14, 24, 25, 26, 27, 28, 29, 36, 37, 38, 39, 40, 41, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 65, 66], "uniformli": 1, "from": [1, 2, 3, 4, 6, 7, 8, 11, 12, 13, 16, 23, 32, 37, 38, 39, 40, 41, 58, 63], "2": [1, 11, 12, 13, 41, 46, 58, 61, 62, 63, 64, 66], "neg": [1, 5, 6, 7, 8, 10, 11, 12, 13, 14, 16, 24, 25, 26, 27, 28, 29, 36, 37, 38, 39, 40, 41, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 66], "us": [1, 2, 3, 4, 6, 7, 8, 10, 11, 13, 14, 16, 25, 26, 27, 28, 29, 32, 36, 37, 38, 40, 43, 49, 58, 65, 66], "corrupt": [1, 6, 7, 8, 10, 11, 12, 13, 32, 36, 37, 38, 39, 40, 41], "order": [1, 14, 32, 40, 43, 58, 66], "also": [1, 3, 4, 59, 66], "balanc": [1, 2, 59, 65], "wai": [1, 16], "ensur": 1, "varieti": 1, "benefici": 1, "final": [1, 3, 4, 14, 46], "qualiti": [1, 16], "left": [1, 3, 4], "A": [1, 2, 32, 37, 58, 59], "made": 1, "up": [1, 3, 4, 66], "9": [1, 66], "block": [1, 10], "contain": [1, 3, 4, 16, 58], "same": [1, 7, 10, 14, 16, 21, 22, 40, 41, 58, 59, 63, 64], "number": [1, 6, 7, 8, 10, 11, 12, 13, 14, 16, 21, 22, 29, 38, 40, 43, 46, 47, 48, 50, 52, 53, 54, 55, 56, 58, 59, 66], "j": [1, 64], "0": [1, 3, 4, 6, 7, 8, 16, 18, 19, 26, 27, 28, 29, 32, 37, 46, 50, 54, 56, 58, 63, 66], "right": 1, "all": [1, 3, 4, 6, 7, 10, 13, 14, 16, 32, 37, 40, 41, 43, 58, 63], "possibli": [1, 16, 40, 43, 58], "pad": [1, 6, 7, 11, 12, 13, 14, 40, 59], "In": [1, 2, 3, 4], "thi": [1, 3, 4, 6, 7, 8, 12, 13, 14, 16, 21, 22, 43, 58, 65], "exampl": [1, 13, 14, 35], "scheme": [1, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 65], "allow": [1, 3, 4, 65], "u": [1, 56], "commun": [1, 65], "first": [1, 3, 4, 16, 38, 64], "need": [1, 16, 21, 22, 25, 26, 27, 28, 29, 43, 46, 65], "gather": [1, 10, 11, 12, 13, 14, 40, 63, 65], "its": [1, 22], "chip": [1, 65], "posit": [1, 5, 6, 7, 8, 11, 12, 13, 19, 24, 25, 26, 27, 28, 29, 32, 44], "These": 1, "includ": 1, "itself": 1, "peer": 1, "requir": [1, 3, 4, 11, 12, 13, 14, 32, 38, 63], "sram": [1, 66], "retriev": [1, 65], "triangl": 1, "colour": 1, "addit": [1, 3, 4, 22], "portion": 1, "can": [1, 3, 4, 12, 13, 14, 21, 22, 32, 43, 58, 65], "reconstruct": 1, "share": [1, 2, 6, 7, 8, 11, 12, 13, 25, 38, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 59, 65], "between": [1, 2, 12, 13, 46, 65], "through": [1, 3, 4, 11, 12, 13], "alltoal": [1, 10, 12, 13, 14], "collect": [1, 6, 7, 8, 10, 11, 12, 13, 14, 15, 58, 65], "oper": [1, 3, 4, 11, 12, 13, 65], "remain": [1, 59], "place": 1, "score": [1, 10, 11, 12, 13, 14, 24, 25, 26, 27, 28, 29, 32, 37, 40, 43, 58, 65], "where": [1, 3, 4, 10, 12, 13, 14, 16, 32, 33, 38, 40, 58], "4": [1, 2, 3, 4, 66], "exchang": 1, "an": [1, 3, 4, 10, 13, 14, 16, 21, 22, 43, 65, 66], "red": 1, "arrow": 1, "effect": [1, 37], "transpos": 1, "row": [1, 20, 32, 40, 61, 62, 63, 64], "column": [1, 16], "pictur": 1, "after": [1, 10, 43, 59], "ha": [1, 6, 7, 8, 20, 66], "correct": [1, 13, 14], "comput": [1, 2, 10, 11, 12, 13, 14, 24, 25, 26, 27, 28, 29, 30, 32, 35, 43, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 65, 66], "present": [1, 16], "abov": [1, 3, 4], "implement": [1, 9], "besskg": [1, 3, 4, 65, 66], "embeddingmovingbesskg": [1, 13, 65], "while": [1, 14], "alwai": [1, 25], "turn": 1, "out": [1, 43], "expens": 1, "mani": 1, "per": [1, 6, 7, 8, 37, 38, 39, 40, 41], "dimens": [1, 32, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "larg": [1, 2, 10, 13, 14, 43], "case": 1, "scoremovingbesskg": [1, 10, 14, 65], "increas": 1, "overal": 1, "throughput": [1, 65], "altern": 1, "work": [1, 3, 4], "well": [1, 3, 4, 50, 54], "differ": [1, 10, 14, 21, 22, 23, 32, 65], "li": [1, 2], "how": 1, "instead": [1, 37, 38, 39, 40, 41, 43, 46], "send": [1, 13], "queri": [1, 10, 13, 14, 32, 37, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58], "devic": [1, 5, 10, 11, 12, 13, 14, 37, 38, 39, 40, 41], "allgath": [1, 13], "against": [1, 10, 12, 13, 14, 37, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58], "partial": 1, "set": [1, 3, 4, 10, 11, 14, 16, 22, 32, 40, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 64, 66], "sent": 1, "via": [1, 2, 10, 14], "new": [1, 3, 4, 6, 7, 8, 23, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 65], "cheaper": 1, "although": 1, "5": 1, "correspond": [1, 13, 21, 22, 40, 63], "6": [1, 2, 3, 4, 46], "put": 1, "back": [1, 13, 59], "came": 1, "complet": [1, 2, 10, 14, 16, 43, 58], "respons": [1, 14], "acls20": [2, 46], "ralph": 2, "abboud": 2, "i\u0307smail": 2, "i\u0307lkan": 2, "ceylan": 2, "thoma": 2, "lukasiewicz": 2, "tommaso": 2, "salvatori": 2, "box": [2, 65], "model": [2, 6, 7, 8, 11, 30, 43, 44, 46, 47, 48, 50, 52, 53, 54, 55, 56, 65], "knowledg": [2, 10, 13, 14, 15, 16, 29, 37, 46, 47, 48, 50, 52, 53, 54, 55, 56, 58, 59, 65, 66], "base": [2, 6, 7, 8, 10, 11, 14, 16, 24, 25, 26, 27, 28, 29, 39, 40, 41, 43, 45, 46, 49, 51, 58], "advanc": 2, "neural": 2, "inform": 2, "process": [2, 5, 6, 7, 8, 37, 38, 39, 40, 41], "system": [2, 3, 4, 66], "33": 2, "annual": 2, "confer": 2, "2020": 2, "neurip": 2, "13": [2, 55], "asam20": [2, 16], "breit": 2, "anna": 2, "ott": 2, "simon": 2, "agibetov": 2, "asan": 2, "samwald": 2, "matthia": 2, "openbiolink": [2, 16], "benchmark": [2, 16], "framework": [2, 11, 65], "scale": [2, 25, 26, 27, 28, 29], "biomed": 2, "link": [2, 3, 4, 32, 65, 66], "predict": [2, 14, 30, 32, 33, 35, 43, 66], "bioinformat": 2, "36": 2, "4097": 2, "4098": 2, "bugd": [2, 55], "antoin": 2, "bord": 2, "nicola": 2, "usuni": 2, "alberto": 2, "garc\u00eda": 2, "dur\u00e1n": 2, "jason": 2, "weston": 2, "oksana": 2, "yakhnenko": 2, "translat": 2, "multi": 2, "data": [2, 6, 7, 8, 10, 16], "26": 2, "27th": 2, "2013": 2, "2787": 2, "2795": 2, "cjm": [2, 9, 10, 11, 14, 29], "22": [2, 9, 10, 11, 14, 29, 50, 56], "cattaneo": 2, "daniel": 2, "justu": 2, "harri": 2, "mellor": 2, "dougla": 2, "orr": 2, "jerom": 2, "maloberti": 2, "zheni": 2, "liu": 2, "thorin": 2, "farnsworth": 2, "andrew": 2, "fitzgibbon": 2, "blazej": 2, "banaszewski": 2, "carlo": 2, "luschi": 2, "bess": [2, 5, 37, 40, 42, 66], "graph": [2, 10, 13, 14, 15, 16, 29, 37, 46, 47, 48, 50, 52, 53, 54, 55, 56, 58, 59, 65, 66], "arxiv": 2, "preprint": 2, "2211": 2, "12281": 2, "2022": 2, "chwc21": [2, 52], "linlin": 2, "chao": 2, "jianshan": 2, "he": 2, "taifeng": 2, "wang": 2, "wei": 2, "chu": 2, "pairr": [2, 65], "vector": 2, "proceed": 2, "59th": 2, "meet": 2, "associ": [2, 16], "linguist": 2, "11th": 2, "intern": [2, 25, 45], "joint": 2, "natur": 2, "languag": 2, "acl": 2, "ijcnlp": 2, "2021": 2, "volum": 2, "long": 2, "paper": 2, "virtual": 2, "event": 2, "august": 2, "4360": 2, "4369": 2, "dppr18": [2, 16], "tim": 2, "dettmer": 2, "minervini": 2, "pasqual": 2, "stenetorp": 2, "pontu": 2, "sebastian": 2, "riedel": 2, "convolut": 2, "2d": 2, "32th": 2, "aaai": 2, "artifici": 2, "intellig": 2, "1811": 2, "1818": 2, "2018": 2, "hfz": [2, 16], "20": [2, 3, 4, 16, 66], "weihua": 2, "hu": 2, "fei": 2, "marinka": 2, "zitnik": 2, "yuxiao": 2, "dong": 2, "hongyu": 2, "ren": 2, "bowen": 2, "michel": 2, "catasta": 2, "jure": 2, "leskovec": 2, "open": [2, 3, 4, 65], "machin": [2, 65], "learn": [2, 22, 46, 47, 48, 50, 52, 53, 54, 55, 56], "jcmb15": [2, 29], "\u00e9": 2, "bastien": 2, "jean": 2, "kyunghyun": 2, "cho": 2, "roland": 2, "memisev": 2, "yoshua": 2, "bengio": 2, "On": 2, "veri": [2, 13], "target": 2, "vocabulari": 2, "53rd": 2, "7th": 2, "10": [2, 16, 43, 66], "2015": 2, "mbs15": [2, 16], "farzaneh": 2, "mahdisoltani": 2, "joanna": 2, "biega": 2, "fabian": 2, "m": [2, 3, 4, 66], "suchanek": 2, "yago3": [2, 16, 66], "multilingu": 2, "wikipedia": 2, "seventh": 2, "biennial": 2, "innov": 2, "research": [2, 3, 4, 66], "cidr": 2, "asilomar": 2, "ca": 2, "usa": 2, "januari": 2, "7": [2, 16, 66], "onlin": 2, "sdnt19": [2, 26, 53], "zhiqe": 2, "sun": 2, "zhi": 2, "hong": 2, "deng": 2, "jian": 2, "yun": 2, "nie": 2, "tang": 2, "rotat": [2, 62, 65], "complex": [2, 53, 61, 62, 65], "space": [2, 46], "represent": [2, 46], "iclr": 2, "2019": 2, "twr": [2, 47], "16": [2, 6, 7, 8, 47, 66], "th": 2, "o": [2, 6, 7, 8], "trouillon": 2, "johann": 2, "welbl": 2, "ric": 2, "gaussier": 2, "guillaum": 2, "bouchard": 2, "simpl": 2, "33rd": 2, "icml": 2, "2016": 2, "48": 2, "jmlr": 2, "workshop": 2, "2071": 2, "2080": 2, "wmw": [2, 50], "baoxin": 2, "qingy": 2, "meng": 2, "ziyu": 2, "honghong": 2, "zhao": 2, "dayong": 2, "wu": 2, "wanxiang": 2, "che": 2, "shijin": 2, "zhigang": 2, "chen": 2, "cong": 2, "interht": [2, 65], "interact": 2, "2202": 2, "04897": 2, "yyh": [2, 48], "15": [2, 16, 48], "bishan": 2, "yang": 2, "wen": 2, "tau": 2, "yih": 2, "xiaodong": 2, "jianfeng": 2, "gao": 2, "3rd": 2, "yll": [2, 56], "yu": 2, "zhicong": 2, "luo": 2, "huanyong": 2, "lin": 2, "hongzhu": 2, "yafeng": 2, "tripler": [2, 65], "2209": 2, "08271": 2, "zyx22": [2, 54], "xuanyu": 2, "zhang": 2, "qing": 2, "dongliang": 2, "xu": 2, "tran": [2, 65], "transit": 2, "synthet": 2, "find": [2, 64], "emnlp": 2, "1202": 2, "1208": 2, "you": [3, 4, 66], "even": [3, 4], "don": [3, 4], "t": [3, 4, 10, 12, 14, 15, 16, 37, 38, 39, 40, 41, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 64, 65], "have": [3, 4, 16, 43, 65], "access": [3, 4, 65], "ipumodel": [3, 4], "emul": [3, 4], "most": [3, 4, 14, 33, 43], "function": [3, 4, 6, 7, 8, 10, 11, 12, 13, 14, 17, 21, 22, 24, 25, 26, 27, 28, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 60, 63, 66], "physic": [3, 4, 66], "hardwar": [3, 4, 43], "tunnel": [3, 4], "web": [3, 4], "editor": [3, 4], "desktop": [3, 4], "app": [3, 4], "minimum": [3, 4], "effort": [3, 4], "excel": [3, 4], "solut": [3, 4], "test": [3, 4, 16, 66], "directli": [3, 4], "here": [3, 4, 16, 43], "do": [3, 4], "fork": [3, 4], "repositori": [3, 4], "launch": [3, 4], "hour": [3, 4], "session": [3, 4], "free": [3, 4, 65, 66], "form": [3, 4], "http": [3, 4, 16, 66], "consol": [3, 4], "com": [3, 4, 16, 66], "github": [3, 4, 16, 65, 66], "userid": [3, 4], "reponam": [3, 4], "graphcor": [3, 4, 66], "2fpytorch": [3, 4], "3a3": [3, 4], "ubuntu": [3, 4, 66], "04": [3, 4, 66], "20230703": [3, 4], "pod4": [3, 4, 66], "repopnam": [3, 4], "address": [3, 4], "e": [3, 4, 16, 61, 62, 63, 64], "g": [3, 4], "origin": [3, 4], "repo": [3, 4, 65], "start": [3, 4, 32, 65], "clone": [3, 4], "termin": [3, 4], "pane": [3, 4], "run": [3, 4, 6, 7, 8, 43, 66], "command": [3, 4], "bash": [3, 4], "gradient": [3, 4, 66], "launch_vscode_serv": [3, 4], "sh": [3, 4, 66], "name": [3, 4, 10, 12, 14, 16, 43, 58], "option": [3, 4, 6, 7, 8, 10, 11, 12, 13, 14, 16, 21, 22, 32, 40, 43, 58, 59, 66], "argument": [3, 4], "defin": [3, 4, 16, 46], "remot": [3, 4], "default": [3, 4, 6, 7, 8, 10, 11, 12, 13, 14, 18, 19, 28, 32, 38, 40, 43, 46, 47, 48, 50, 52, 53, 54, 55, 56, 58, 59], "script": [3, 4], "download": [3, 4, 16], "instal": [3, 4, 65], "depend": [3, 4, 64, 66], "ask": [3, 4], "author": [3, 4], "account": [3, 4], "write": [3, 4], "privileg": [3, 4], "provid": [3, 4, 14, 21, 22, 32, 43], "pleas": [3, 4], "refer": [3, 4, 65], "notebook": [3, 4, 66], "detail": [3, 4, 66], "step": [3, 4, 10, 11, 12, 13, 14, 43], "connect": [3, 4], "onc": [3, 4, 21, 22, 65], "dev": [3, 4], "build": [3, 4, 15, 16], "custom": [3, 4], "op": [3, 4], "now": [3, 4], "readi": [3, 4], "close": [3, 4], "stop": [3, 4], "rememb": [3, 4], "unregist": [3, 4], "explain": [3, 4], "common": [3, 4], "issu": [3, 4, 65], "paragraph": [3, 4], "To": [3, 4, 10, 11, 14, 43], "resum": [3, 4], "your": [3, 4, 66], "just": [3, 4], "section": [3, 4], "profil": [3, 4], "repeat": [3, 4, 7], "chang": [3, 4, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "extens": [3, 4], "persist": [3, 4], "poplar": [3, 4, 66], "sdk": [3, 4, 66], "follow": [3, 4, 66], "instruct": [3, 4, 66], "get": [3, 4, 65], "guid": [3, 4, 65], "Then": [3, 4], "enabl": [3, 4, 66], "creat": [3, 4, 58, 59, 66], "activ": [3, 4, 28, 65, 66], "python": [3, 4, 65, 66], "virtualenv": [3, 4, 66], "poptorch": [3, 4, 6, 7, 8, 65, 66], "wheel": [3, 4, 66], "necessari": [3, 4], "python3": [3, 4, 66], "8": [3, 4, 66], "venv": [3, 4, 66], "add": [3, 4], "bin": [3, 4, 66], "sourc": [3, 4, 6, 7, 8, 10, 11, 12, 13, 14, 16, 18, 19, 20, 21, 22, 23, 25, 26, 27, 28, 29, 31, 32, 33, 35, 37, 38, 39, 40, 41, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59, 61, 62, 63, 64, 66], "path_to_poplar_sdk": [3, 4], "pip": [3, 4, 66], "poplar_sdk_en": [3, 4, 66], "whl": [3, 4, 66], "r": [3, 4, 10, 14, 15, 16, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 62, 64, 65], "txt": [3, 4], "sever": [3, 4], "util": [3, 4, 6, 7, 8, 15, 17, 30, 65], "dure": [3, 4, 66], "check": [3, 4], "help": [3, 4], "list": [3, 4, 6, 7, 8, 16, 21, 22, 32, 40, 43, 46, 47, 48, 50, 52, 53, 54, 55, 56], "befor": [3, 4, 11, 12, 13, 14, 46, 50, 52, 54, 56], "submit": [3, 4], "pr": [3, 4], "upstream": [3, 4], "ci": [3, 4], "particular": [3, 4], "mind": [3, 4], "our": [3, 4, 66], "format": [3, 4], "error": [3, 4, 10, 14, 43, 66], "lint": [3, 4], "automat": [3, 4, 16], "insid": [3, 4, 46, 66], "unit": [3, 4], "folder": [3, 4], "individu": [3, 4], "pattern": [3, 4], "match": [3, 4], "filter": [3, 4, 11, 12, 13, 14, 43, 64], "k": [3, 4, 14, 32, 33, 34, 43, 62, 63], "cpp": [3, 4], "custom_op": [3, 4], "updat": [3, 4], "makefil": [3, 4], "ad": [3, 4, 22, 58, 65], "class": [5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59], "partitioned_triple_set": [6, 7, 8, 43], "negative_sampl": [6, 7, 8, 10, 11, 12, 13, 14, 65], "shard_b": [6, 7, 8, 10, 14], "batches_per_step": [6, 7, 8], "seed": [6, 7, 8, 16, 37, 38, 40, 41, 59], "hrt_freq_weight": [6, 7, 8], "fals": [6, 7, 8, 11, 12, 13, 14, 32, 38, 40, 43, 46, 47, 48, 50, 52, 53, 54, 55, 56, 58], "weight_smooth": [6, 7, 8], "duplicate_batch": [6, 7, 8], "return_triple_idx": [6, 7, 8], "random": [6, 16, 38, 59], "indic": [6, 7, 8, 10, 11, 12, 13, 14, 32, 40, 58, 63], "replac": 6, "No": [6, 37, 66], "appli": [6, 40, 43, 50, 54, 59, 64], "initi": [6, 7, 8, 10, 11, 12, 13, 14, 17, 18, 19, 20, 21, 22, 25, 26, 27, 28, 29, 32, 37, 38, 40, 41, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "sampler": [6, 7, 8, 10, 11, 12, 13, 14, 37, 38, 39, 40, 41, 43], "paramet": [6, 7, 8, 10, 11, 12, 13, 14, 16, 18, 19, 20, 21, 22, 23, 25, 26, 27, 28, 29, 32, 33, 35, 37, 38, 40, 41, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59, 61, 62, 63, 64, 66], "partitionedtripleset": [6, 7, 8, 65], "pre": [6, 7, 8, 16], "shardednegativesampl": [6, 7, 8, 11, 12, 13, 37, 65], "int": [6, 7, 8, 10, 11, 12, 13, 14, 16, 21, 22, 29, 33, 37, 38, 40, 41, 43, 46, 47, 48, 49, 50, 52, 53, 54, 55, 56, 59], "micro": [6, 7, 8, 11, 12, 13, 14], "call": [6, 7, 8, 37], "rng": [6, 7, 8, 37, 38, 39, 40, 41], "bool": [6, 7, 8, 11, 12, 13, 14, 18, 19, 22, 25, 26, 27, 28, 29, 32, 37, 38, 39, 40, 41, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58], "If": [6, 7, 8, 11, 12, 13, 14, 16, 21, 22, 32, 38, 40, 43, 46, 47, 48, 50, 52, 53, 54, 55, 56, 58, 63, 66], "true": [6, 7, 8, 11, 12, 13, 14, 18, 19, 22, 32, 38, 40, 43, 46, 47, 48, 50, 52, 53, 54, 55, 56, 59], "frequenc": [6, 7, 8], "weight": [6, 7, 8, 11, 12, 13, 25, 26, 27, 28, 29, 66], "float": [6, 7, 8, 16, 18, 19, 26, 27, 28, 29, 46, 50, 54, 56], "smooth": [6, 7, 8], "two": [6, 7, 8, 46, 64], "ident": [6, 7, 8], "halv": [6, 7, 8], "ht": [6, 7, 8, 12, 37, 38, 39, 40, 41], "return": [6, 7, 8, 10, 11, 12, 13, 14, 16, 18, 19, 20, 21, 22, 23, 25, 26, 27, 28, 29, 32, 33, 35, 37, 40, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59, 61, 62, 63, 64], "wrt": [6, 7, 8, 43], "get_dataload": [6, 7, 8], "shuffl": [6, 7, 8], "num_work": [6, 7, 8], "persistent_work": [6, 7, 8], "buffer_s": [6, 7, 8], "dataload": [6, 7, 8], "instanti": [6, 7, 8, 16], "appropri": [6, 7, 8], "iter": [6, 7, 8, 10, 14], "It": [6, 7, 8, 14, 43], "asynchron": [6, 7, 8], "load": [6, 7, 8, 14, 16, 59], "minim": [6, 7, 8, 65], "cpu": [6, 7, 8], "compil": [6, 7, 8, 66], "epoch": [6, 7, 8], "see": [6, 7, 8, 10, 11, 12, 13, 14, 26, 27, 28, 29, 32, 35, 37, 40, 41, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 65, 66], "torch": [6, 7, 8, 63], "__init__": [6, 7, 8, 28, 40, 41, 46, 47, 48, 50, 52, 53, 54, 55, 56], "ring": [6, 7, 8], "buffer": [6, 7, 8], "preload": [6, 7, 8], "get_dataloader_sampl": [6, 7, 8], "sample_tripl": [6, 7, 8], "idx": [6, 7, 8], "index": [6, 7, 8, 10, 40, 63, 64], "dict": [6, 7, 8, 11, 12, 13, 14, 16, 32, 43], "str": [6, 7, 8, 11, 12, 13, 14, 16, 28, 32, 34, 37, 38, 39, 40, 41, 43, 58, 64], "union": [6, 7, 8, 14, 16, 21, 22, 43, 46, 47, 48, 50, 52, 53, 54, 55, 56], "ndarrai": [6, 7, 8, 16, 40, 41, 43, 58, 59], "ani": [6, 7, 8, 11, 12, 13, 14, 16, 40, 41, 43, 58, 59, 65], "dtype": [6, 7, 8, 16, 40, 41, 43, 58, 59, 66], "int64": [6, 7, 8, 40, 58, 59], "bool_": [6, 7, 8, 40], "relev": [6, 7, 8, 11, 12, 13], "static": [6, 7, 8], "worker_init_fn": [6, 7, 8], "worker_id": [6, 7, 8], "pass": [6, 7, 8, 14, 21, 22, 23, 43, 46], "id": [6, 7, 8, 16, 22, 32, 40, 41, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59, 64], "none": [6, 7, 8, 10, 11, 12, 13, 14, 16, 21, 22, 32, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59], "specifi": [7, 10, 16, 43, 63], "shorter": 7, "ones": [7, 22], "length": [7, 21, 22, 40], "mask": [7, 11, 12, 13, 14, 32, 40], "identifi": [7, 11, 12, 13], "abstract": [8, 11, 25, 27, 45, 49, 51], "pytorch": [9, 66], "modul": [9, 10, 11, 12, 13, 14, 25, 30, 32, 43, 45], "kge": [9, 10, 11, 12, 13, 14, 30, 44, 66], "multipl": [9, 61], "candidate_sampl": [10, 14], "score_fn": [10, 11, 12, 13, 14, 43], "window_s": [10, 14, 43], "1000": [10, 43], "h": [10, 12, 14, 15, 16, 37, 38, 39, 40, 41, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 64, 65], "host": [10, 11, 12, 13, 43], "combin": [10, 11, 14, 43], "h_shard": [10, 14, 43, 58], "t_shard": [10, 14, 43, 58], "sinc": 10, "onli": [10, 14, 16, 37, 38, 39, 40, 41, 58], "part": [10, 16, 58, 61, 62], "slide": [10, 14, 43], "window": [10, 14, 43], "metric": [10, 11, 12, 13, 14, 43, 65], "should": [10, 14, 16, 43], "aggreg": 10, "pipelin": [10, 65], "allscorespipelin": [10, 65], "allscor": 10, "placeholdernegativesampl": [10, 14, 65], "basescorefunct": [10, 11, 12, 13, 14, 43, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 65], "decreas": [10, 14, 32, 43], "avoid": [10, 13, 14, 43], "oom": [10, 14, 43], "forward": [10, 11, 12, 13, 14, 25, 26, 27, 28, 29, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "similarli": [10, 14, 58], "candid": [10, 14, 32, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "thei": [10, 13, 14, 40], "togeth": [10, 14], "tensor": [10, 11, 12, 13, 14, 18, 19, 20, 21, 22, 25, 26, 27, 28, 29, 32, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 61, 62, 63, 64], "self": [10, 25, 26, 27, 28, 29], "shape": [10, 11, 12, 13, 14, 16, 21, 23, 25, 26, 27, 28, 29, 32, 40, 41, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58, 59, 61, 62, 63, 64], "known": [10, 14, 58], "loss_fn": [11, 12, 13], "evalu": [11, 12, 13, 14, 30, 43, 65], "return_scor": [11, 12, 13, 14, 43], "augment_neg": [11, 12, 13], "ht_shardpair": [11, 58], "baselossfunct": [11, 12, 13, 26, 27, 28, 29, 65], "loss": [11, 12, 13, 65], "augment": [11, 12, 13], "triple_mask": [11, 12, 13, 14, 32], "triple_weight": [11, 12, 13, 25, 26, 27, 28, 29], "negative_mask": [11, 12, 13, 14], "compris": [11, 12, 13], "four": [11, 12, 13], "phase": [11, 12, 13], "local": [11, 12, 13, 16, 58, 59, 65], "n_shard": [11, 12, 13, 14, 21, 40, 58, 59], "positive_per_partit": [11, 12, 13], "b": [11, 12, 13, 14, 19, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 63], "padded_neg": [11, 12, 13, 14, 40], "discard": [11, 12, 13, 14], "properti": [11, 12, 13, 16, 59], "n_embedding_paramet": [11, 12, 13], "trainabl": [11, 12, 13], "score_batch": [11, 12, 13], "tupl": [11, 12, 13, 16, 40, 64], "n_neg": [11, 12, 13, 25, 26, 27, 28, 29, 32, 38, 40, 41, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 58], "move": [12, 13], "done": 12, "singl": [12, 32], "total": [12, 29], "disabl": 12, "otherwis": [12, 63], "conveni": 13, "so": [13, 20], "For": [13, 14, 43, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 63, 65, 66], "valu": 13, "document": 13, "multipli": [13, 50, 54], "doe": 13, "support": [13, 19, 32, 43, 66], "100": [14, 66], "specif": [14, 16, 40, 43, 44, 58], "top": [14, 32, 43], "like": [14, 32, 33, 43, 63], "input": 14, "recommend": [14, 66], "one": [14, 22, 23, 40, 58], "want": [14, 16], "loop": 14, "topk": 14, "triplebasedshardednegativesampl": [14, 65], "unnecessari": 14, "best": 14, "respect": 14, "kept": 14, "next": 14, "rest": 14, "mask_on_gath": [14, 40], "n_entiti": [16, 21, 29, 59], "n_relation_typ": [16, 22, 46, 47, 48, 50, 52, 53, 54, 55, 56], "entity_dict": 16, "relation_dict": 16, "type_offset": [16, 59], "neg_head": [16, 58], "neg_tail": [16, 58], "repres": 16, "int32": [16, 40, 41, 43, 58, 59], "classmethod": [16, 58, 59], "build_ogbl_biokg": 16, "root": 16, "ogbl": [16, 66], "biokg": [16, 66], "ogb": 16, "stanford": 16, "edu": 16, "doc": 16, "linkprop": 16, "path": [16, 59, 66], "locat": 16, "build_ogbl_wikikg2": 16, "wikikg2": [16, 66], "build_openbiolink": 16, "high": [16, 42, 65], "version": 16, "openbiolink2020": 16, "hq": 16, "build_yago310": 16, "subgraph": 16, "least": 16, "them": [16, 40, 65, 66], "yago": 16, "org": 16, "label": 16, "from_datafram": 16, "df": 16, "head_column": 16, "relation_column": 16, "tail_column": 16, "entity_typ": 16, "1234": 16, "panda": 16, "datafram": 16, "assign": [16, 32, 58], "contigu": 16, "dictionari": [16, 32], "seri": 16, "map": [16, 59], "string": 16, "valid": 16, "instanc": 16, "from_tripl": 16, "arrai": 16, "alreadi": [16, 40], "been": [16, 65, 66], "note": 16, "manual": 16, "numpi": 16, "head_id": 16, "relation_id": [16, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "tail_id": 16, "num_tripl": 16, "offset": [16, 50, 54, 56, 59], "ht_type": 16, "n_tripl": [16, 40, 41, 58], "h_type": [16, 58], "t_type": [16, 58], "object": [16, 59], "save": [16, 59, 66], "node": 16, "edg": 16, "n_neg_head": [16, 58], "n_neg_tail": [16, 58], "out_fil": [16, 59], "pkl": 16, "output": [16, 32, 59], "file": [16, 59, 66], "h_id": 16, "r_id": 16, "t_id": 16, "assum": [16, 32, 66], "cluster": [16, 40, 59], "manag": 17, "embedding_t": [18, 19, 20], "std": 18, "divide_by_embedding_s": [18, 19], "normal": [18, 20, 46, 50, 52, 54, 56], "mean": 18, "standard": 18, "deviat": 18, "rescal": [18, 19], "row_siz": [18, 19, 21, 22, 23], "symmetr": 19, "uniform": [19, 20], "boundari": 19, "norm": [20, 46, 49, 50, 52, 53, 54, 55, 56], "callabl": [21, 22, 34, 46, 47, 48, 50, 52, 53, 54, 55, 56], "either": 21, "max_entity_per_shard": [21, 59], "unshard": 21, "alloc": [21, 22], "entri": [21, 22], "omit": [21, 22], "max_ent_per_shard": 21, "inverse_rel": [22, 46, 47, 48, 50, 52, 53, 54, 55, 56], "invers": [22, 46, 47, 48, 50, 52, 53, 54, 55, 56, 58], "direct": 22, "given": [22, 43, 64, 66], "entity_embed": [23, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "old_shard": 23, "new_shard": [23, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "refactor": 23, "n_shard_old": 23, "max_ent_per_shard_old": 23, "current": [23, 32], "n_shard_new": 23, "max_ent_per_shard_new": 23, "arg": [25, 45], "kwarg": [25, 45], "fp32": [25, 66], "state": [25, 45], "nn": [25, 45], "scriptmodul": [25, 45], "positive_scor": [25, 26, 27, 28, 29], "negative_scor": [25, 26, 27, 28, 29], "batch_siz": [25, 26, 27, 28, 29, 32, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "get_negative_weight": [25, 26, 27, 28, 29], "negative_adversarial_sampl": [25, 26, 27, 28, 29], "els": [25, 26, 27, 28, 29, 32, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "loss_scal": [25, 26, 27, 28, 29], "factor": [25, 26, 27, 28, 29, 56], "might": [25, 26, 27, 28, 29], "fp16": [25, 26, 27, 28, 29, 66], "adversari": [25, 26, 27, 28, 29], "negative_adversarial_scal": [25, 26, 27, 28, 29], "reciproc": [25, 26, 27, 28, 29, 35], "temperatur": [25, 26, 27, 28, 29], "margin": [26, 27, 28], "log": 26, "sigmoid": 26, "activation_funct": 28, "relu": 28, "rank": [28, 32, 33, 35], "pairwis": 28, "hing": 28, "marginbasedlossfunct": [28, 65], "softmax": 29, "cross": 29, "entropi": 29, "attribut": 30, "metric_list": 32, "mode": [32, 58], "averag": 32, "worst_rank_infti": 32, "reduct": [32, 35, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "return_rank": 32, "mrr": [32, 34, 35], "hit": [32, 33, 34], "optimist": 32, "pessimist": 32, "infin": 32, "worst": 32, "possibl": 32, "method": 32, "reduc": [32, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "along": [32, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 63], "sum": [32, 46, 47, 48, 51], "alongsid": 32, "dict_metrics_from_rank": 32, "batch_rank": 32, "element": [32, 63], "boolean": 32, "ranks_from_indic": 32, "ground_truth": [32, 58], "candidate_indic": 32, "ground": [32, 33, 35, 58], "truth": [32, 33, 35, 58], "n_candid": 32, "likelihood": 32, "distinct": 32, "among": [32, 33, 35], "ranks_from_scor": 32, "pos_scor": 32, "candidate_scor": 32, "stacked_metrics_from_rank": 32, "stack": 32, "n_metric": 32, "count": [33, 66], "maximum": [33, 66], "accept": 33, "hitsatk": [34, 65], "reciprocalrank": [34, 65], "basemetr": [35, 65], "corruption_schem": [37, 38, 39, 40, 41, 43], "placehold": 37, "topkquerybesskg": [37, 40, 65], "flat_negative_format": [37, 38, 39, 40, 41], "local_sampl": [37, 38, 39, 40, 41], "gener": [37, 38, 39, 40, 41, 60], "half": 38, "second": [38, 64, 66], "negative_head": 40, "negative_tail": 40, "return_sort_idx": 40, "predetermin": 40, "global": [40, 43, 58, 59, 64], "randomshardednegativesampl": [40, 41, 65], "sort": [40, 58], "recov": 40, "pad_neg": 40, "shard_count": [40, 59], "padded_shard_length": 40, "divid": 40, "view": 40, "shard_neg": 40, "shard_neg_count": 40, "sort_neg_idx": 40, "triple_typ": 41, "level": 42, "api": [42, 65, 66], "batch_sampl": [43, 65], "filter_tripl": [43, 64], "return_topk": 43, "use_ipu_model": 43, "kg": 43, "appear": [43, 64], "shardedbatchsampl": [43, 65], "whose": 43, "must": 43, "caus": 43, "go": 43, "actual": 43, "result": 43, "head_emb": [45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "tail_emb": [45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "score_tripl": [45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "negative_sample_shar": [45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "relation_embed": [45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "score_head": [45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "fix": [45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "n_head": [45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "embedding_s": [45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "broadcast": [45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "score_tail": [45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "n_tail": [45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "update_shard": [45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "scoring_norm": [46, 49, 50, 52, 53, 54, 55, 56], "entity_initi": [46, 47, 48, 50, 52, 53, 54, 55, 56], "uniform_": 46, "relation_initi": [46, 47, 48, 50, 52, 53, 54, 55, 56], "init_uniform_norm": [46, 65], "apply_tanh": 46, "dist_func_per_dim": 46, "ep": 46, "1e": 46, "06": 46, "distancebasedscorefunct": [46, 47, 48, 50, 52, 53, 54, 55, 56, 65], "center": 46, "scalar": 46, "bound": [46, 65], "bump": 46, "tanh": 46, "select": 46, "distanc": [46, 49, 50, 52, 53, 54, 55, 56], "whether": [46, 58], "outsid": 46, "make": [46, 58], "choic": 46, "separ": 46, "soften": 46, "geometr": 46, "width": 46, "boxe_scor": 46, "bumped_ht": 46, "center_ht": 46, "width_ht": 46, "box_siz": 46, "optim": [46, 66], "emb_siz": 46, "control": 46, "broadcasted_dist": [46, 49, 50, 52, 53, 54, 55, 56], "v1": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 61], "v2": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 61], "p": [46, 49, 50, 52, 53, 54, 55, 56], "reduce_embed": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "v": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 62, 65], "init_kge_norm": [47, 65], "broadcasted_dot_product": [47, 48, 51], "dot": [47, 48, 51], "product": [47, 48, 51], "init_kge_uniform": [48, 50, 52, 53, 54, 55, 56, 65], "normalize_ent": [50, 52, 54, 56], "l2": [50, 52, 54, 56], "auxiliari": 50, "matrix": 51, "decomposit": 51, "project": [52, 56], "real": [53, 61, 62], "tild": 54, "triplerev2": 56, "inverse_tripl": 58, "partition_mod": 58, "dummi": 58, "triple_count": 58, "triple_offset": 58, "triple_sort_idx": 58, "shard_h": 58, "shard_t": 58, "create_from_dataset": 58, "add_inverse_tripl": 58, "kgdataset": [58, 65], "create_from_queri": 58, "query_mod": 58, "negative_typ": 58, "n_queri": 58, "hr": 58, "rt": 58, "resp": 58, "r_inv": 58, "regular": 58, "criterion": 58, "delimit": 58, "entity_to_shard": 59, "entity_to_idx": 59, "shard_and_idx_to_ent": 59, "entity_type_count": 59, "entity_type_offset": 59, "again": 59, "n_type": 59, "npz": 59, "local_id": 59, "exclud": 59, "purpos": 60, "imaginari": [61, 62], "wise": [61, 62], "unitari": 62, "pi": 62, "x": [63, 64], "friendli": 63, "take_along_dim": 63, "dimension": 63, "dim": 63, "take": 63, "filter_mod": 64, "compar": 64, "y": 64, "determin": 64, "look": 64, "z": 64, "spars": 64, "packag": 65, "shallow": 65, "typic": 65, "littl": 65, "perform": 65, "design": 65, "maxim": 65, "bandwidth": 65, "fast": 65, "leverag": 65, "achiev": 65, "introduct": 65, "overview": 65, "librari": [65, 66], "still": 65, "develop": 65, "featur": 65, "expect": 65, "overtim": 65, "occasion": 65, "bug": 65, "mai": 65, "occur": 65, "feel": 65, "report": 65, "problem": 65, "user": 65, "usag": 65, "limit": 65, "randomshardedbatchsampl": 65, "rigidshardedbatchsampl": 65, "typebasedshardednegativesampl": 65, "allscoresbess": 65, "distmult": 65, "matrixdecompositionscorefunct": 65, "trans": 65, "logsigmoidloss": 65, "marginrankingloss": 65, "sampledsoftmaxcrossentropyloss": 65, "metrics_dict": 65, "initialize_entity_embed": 65, "initialize_relation_embed": 65, "refactor_embedding_shard": 65, "complex_multipl": 65, "complex_rot": 65, "gather_indic": 65, "get_entity_filt": 65, "code": 65, "server": 65, "paperspac": [65, 66], "setup": 65, "tip": 65, "bibliographi": 65, "popart": 66, "more": 66, "quick": 66, "git": 66, "import": 66, "1403": 66, "walkthrough": 66, "main": 66, "jupyt": 66, "we": 66, "sequenc": 66, "click": 66, "button": 66, "avail": 66, "introduc": 66, "therefor": 66, "some": 66, "approxim": 66, "estim": 66, "below": 66, "accumul": 66, "momentum": 66, "notic": 66, "cap": 66, "max": 66, "pod16": 66, "float16": 66, "sgdm": 66, "2m": 66, "2e8": 66, "13m": 66, "3e9": 66, "128": 66, "adam": 66, "4m": 66, "0e8": 66, "9m": 66, "256": 66, "ye": 66, "900k": 66, "3e8": 66, "5m": 66, "8m": 66, "2e9": 66, "512": 66, "375k": 66, "9e8": 66, "7e8": 66, "messag": 66, "about": 66, "onnx": 66, "protobuff": 66, "exceed": 66, "_popart": 66, "saveinitializerstofil": 66, "my_fil": 66}, "objects": {"": [[65, 0, 0, "-", "besskge"]], "besskge": [[5, 0, 0, "-", "batch_sampler"], [9, 0, 0, "-", "bess"], [15, 0, 0, "-", "dataset"], [17, 0, 0, "-", "embedding"], [24, 0, 0, "-", "loss"], [30, 0, 0, "-", "metric"], [36, 0, 0, "-", "negative_sampler"], [42, 0, 0, "-", "pipeline"], [44, 0, 0, "-", "scoring"], [57, 0, 0, "-", "sharding"], [60, 0, 0, "-", "utils"]], "besskge.batch_sampler": [[6, 1, 1, "", "RandomShardedBatchSampler"], [7, 1, 1, "", "RigidShardedBatchSampler"], [8, 1, 1, "", "ShardedBatchSampler"]], "besskge.batch_sampler.RandomShardedBatchSampler": [[6, 2, 1, "", "get_dataloader"], [6, 2, 1, "", "get_dataloader_sampler"], [6, 2, 1, "", "sample_triples"], [6, 2, 1, "", "worker_init_fn"]], "besskge.batch_sampler.RigidShardedBatchSampler": [[7, 2, 1, "", "get_dataloader"], [7, 2, 1, "", "get_dataloader_sampler"], [7, 2, 1, "", "sample_triples"], [7, 2, 1, "", "worker_init_fn"]], "besskge.batch_sampler.ShardedBatchSampler": [[8, 2, 1, "", "get_dataloader"], [8, 2, 1, "", "get_dataloader_sampler"], [8, 2, 1, "", "sample_triples"], [8, 2, 1, "", "worker_init_fn"]], "besskge.bess": [[10, 1, 1, "", "AllScoresBESS"], [11, 1, 1, "", "BessKGE"], [12, 1, 1, "", "EmbeddingMovingBessKGE"], [13, 1, 1, "", "ScoreMovingBessKGE"], [14, 1, 1, "", "TopKQueryBessKGE"]], "besskge.bess.AllScoresBESS": [[10, 2, 1, "", "forward"]], "besskge.bess.BessKGE": [[11, 2, 1, "", "forward"], [11, 3, 1, "", "n_embedding_parameters"], [11, 2, 1, "", "score_batch"]], "besskge.bess.EmbeddingMovingBessKGE": [[12, 2, 1, "", "forward"], [12, 3, 1, "", "n_embedding_parameters"], [12, 2, 1, "", "score_batch"]], "besskge.bess.ScoreMovingBessKGE": [[13, 2, 1, "", "forward"], [13, 3, 1, "", "n_embedding_parameters"], [13, 2, 1, "", "score_batch"]], "besskge.bess.TopKQueryBessKGE": [[14, 2, 1, "", "forward"]], "besskge.dataset": [[16, 1, 1, "", "KGDataset"]], "besskge.dataset.KGDataset": [[16, 2, 1, "", "build_ogbl_biokg"], [16, 2, 1, "", "build_ogbl_wikikg2"], [16, 2, 1, "", "build_openbiolink"], [16, 2, 1, "", "build_yago310"], [16, 4, 1, "", "entity_dict"], [16, 2, 1, "", "from_dataframe"], [16, 2, 1, "", "from_triples"], [16, 3, 1, "", "ht_types"], [16, 2, 1, "", "load"], [16, 4, 1, "", "n_entity"], [16, 4, 1, "", "n_relation_type"], [16, 4, 1, "", "neg_heads"], [16, 4, 1, "", "neg_tails"], [16, 4, 1, "", "relation_dict"], [16, 2, 1, "", "save"], [16, 4, 1, "", "triples"], [16, 4, 1, "", "type_offsets"]], "besskge.embedding": [[18, 5, 1, "", "init_KGE_normal"], [19, 5, 1, "", "init_KGE_uniform"], [20, 5, 1, "", "init_uniform_norm"], [21, 5, 1, "", "initialize_entity_embedding"], [22, 5, 1, "", "initialize_relation_embedding"], [23, 5, 1, "", "refactor_embedding_sharding"]], "besskge.loss": [[25, 1, 1, "", "BaseLossFunction"], [26, 1, 1, "", "LogSigmoidLoss"], [27, 1, 1, "", "MarginBasedLossFunction"], [28, 1, 1, "", "MarginRankingLoss"], [29, 1, 1, "", "SampledSoftmaxCrossEntropyLoss"]], "besskge.loss.BaseLossFunction": [[25, 2, 1, "", "forward"], [25, 2, 1, "", "get_negative_weights"], [25, 4, 1, "", "loss_scale"], [25, 4, 1, "", "negative_adversarial_sampling"], [25, 4, 1, "", "negative_adversarial_scale"]], "besskge.loss.LogSigmoidLoss": [[26, 2, 1, "", "forward"], [26, 2, 1, "", "get_negative_weights"], [26, 4, 1, "", "loss_scale"], [26, 4, 1, "", "negative_adversarial_sampling"], [26, 4, 1, "", "negative_adversarial_scale"]], "besskge.loss.MarginBasedLossFunction": [[27, 2, 1, "", "forward"], [27, 2, 1, "", "get_negative_weights"], [27, 4, 1, "", "loss_scale"], [27, 4, 1, "", "negative_adversarial_sampling"], [27, 4, 1, "", "negative_adversarial_scale"]], "besskge.loss.MarginRankingLoss": [[28, 2, 1, "", "forward"], [28, 2, 1, "", "get_negative_weights"], [28, 4, 1, "", "loss_scale"], [28, 4, 1, "", "negative_adversarial_sampling"], [28, 4, 1, "", "negative_adversarial_scale"]], "besskge.loss.SampledSoftmaxCrossEntropyLoss": [[29, 2, 1, "", "forward"], [29, 2, 1, "", "get_negative_weights"], [29, 4, 1, "", "loss_scale"], [29, 4, 1, "", "negative_adversarial_sampling"], [29, 4, 1, "", "negative_adversarial_scale"]], "besskge.metric": [[31, 1, 1, "", "BaseMetric"], [32, 1, 1, "", "Evaluation"], [33, 1, 1, "", "HitsAtK"], [34, 6, 1, "", "METRICS_DICT"], [35, 1, 1, "", "ReciprocalRank"]], "besskge.metric.Evaluation": [[32, 2, 1, "", "dict_metrics_from_ranks"], [32, 2, 1, "", "ranks_from_indices"], [32, 2, 1, "", "ranks_from_scores"], [32, 2, 1, "", "stacked_metrics_from_ranks"]], "besskge.negative_sampler": [[37, 1, 1, "", "PlaceholderNegativeSampler"], [38, 1, 1, "", "RandomShardedNegativeSampler"], [39, 1, 1, "", "ShardedNegativeSampler"], [40, 1, 1, "", "TripleBasedShardedNegativeSampler"], [41, 1, 1, "", "TypeBasedShardedNegativeSampler"]], "besskge.negative_sampler.PlaceholderNegativeSampler": [[37, 4, 1, "", "corruption_scheme"], [37, 4, 1, "", "flat_negative_format"], [37, 4, 1, "", "local_sampling"], [37, 4, 1, "", "rng"]], "besskge.negative_sampler.RandomShardedNegativeSampler": [[38, 4, 1, "", "corruption_scheme"], [38, 4, 1, "", "flat_negative_format"], [38, 4, 1, "", "local_sampling"], [38, 4, 1, "", "rng"]], "besskge.negative_sampler.ShardedNegativeSampler": [[39, 4, 1, "", "corruption_scheme"], [39, 4, 1, "", "flat_negative_format"], [39, 4, 1, "", "local_sampling"], [39, 4, 1, "", "rng"]], "besskge.negative_sampler.TripleBasedShardedNegativeSampler": [[40, 4, 1, "", "corruption_scheme"], [40, 4, 1, "", "flat_negative_format"], [40, 4, 1, "", "local_sampling"], [40, 2, 1, "", "pad_negatives"], [40, 4, 1, "", "rng"], [40, 2, 1, "", "shard_negatives"]], "besskge.negative_sampler.TypeBasedShardedNegativeSampler": [[41, 4, 1, "", "corruption_scheme"], [41, 4, 1, "", "flat_negative_format"], [41, 4, 1, "", "local_sampling"], [41, 4, 1, "", "rng"]], "besskge.pipeline": [[43, 1, 1, "", "AllScoresPipeline"]], "besskge.pipeline.AllScoresPipeline": [[43, 2, 1, "", "forward"]], "besskge.scoring": [[45, 1, 1, "", "BaseScoreFunction"], [46, 1, 1, "", "BoxE"], [47, 1, 1, "", "ComplEx"], [48, 1, 1, "", "DistMult"], [49, 1, 1, "", "DistanceBasedScoreFunction"], [50, 1, 1, "", "InterHT"], [51, 1, 1, "", "MatrixDecompositionScoreFunction"], [52, 1, 1, "", "PairRE"], [53, 1, 1, "", "RotatE"], [54, 1, 1, "", "TranS"], [55, 1, 1, "", "TransE"], [56, 1, 1, "", "TripleRE"]], "besskge.scoring.BaseScoreFunction": [[45, 4, 1, "", "entity_embedding"], [45, 2, 1, "", "forward"], [45, 4, 1, "", "negative_sample_sharing"], [45, 4, 1, "", "relation_embedding"], [45, 2, 1, "", "score_heads"], [45, 2, 1, "", "score_tails"], [45, 2, 1, "", "score_triple"], [45, 4, 1, "", "sharding"], [45, 2, 1, "", "update_sharding"]], "besskge.scoring.BoxE": [[46, 2, 1, "", "boxe_score"], [46, 2, 1, "", "broadcasted_distance"], [46, 4, 1, "", "entity_embedding"], [46, 2, 1, "", "forward"], [46, 4, 1, "", "negative_sample_sharing"], [46, 2, 1, "", "reduce_embedding"], [46, 4, 1, "", "relation_embedding"], [46, 2, 1, "", "score_heads"], [46, 2, 1, "", "score_tails"], [46, 2, 1, "", "score_triple"], [46, 4, 1, "", "sharding"], [46, 2, 1, "", "update_sharding"]], "besskge.scoring.ComplEx": [[47, 2, 1, "", "broadcasted_dot_product"], [47, 4, 1, "", "entity_embedding"], [47, 2, 1, "", "forward"], [47, 4, 1, "", "negative_sample_sharing"], [47, 2, 1, "", "reduce_embedding"], [47, 4, 1, "", "relation_embedding"], [47, 2, 1, "", "score_heads"], [47, 2, 1, "", "score_tails"], [47, 2, 1, "", "score_triple"], [47, 4, 1, "", "sharding"], [47, 2, 1, "", "update_sharding"]], "besskge.scoring.DistMult": [[48, 2, 1, "", "broadcasted_dot_product"], [48, 4, 1, "", "entity_embedding"], [48, 2, 1, "", "forward"], [48, 4, 1, "", "negative_sample_sharing"], [48, 2, 1, "", "reduce_embedding"], [48, 4, 1, "", "relation_embedding"], [48, 2, 1, "", "score_heads"], [48, 2, 1, "", "score_tails"], [48, 2, 1, "", "score_triple"], [48, 4, 1, "", "sharding"], [48, 2, 1, "", "update_sharding"]], "besskge.scoring.DistanceBasedScoreFunction": [[49, 2, 1, "", "broadcasted_distance"], [49, 4, 1, "", "entity_embedding"], [49, 2, 1, "", "forward"], [49, 4, 1, "", "negative_sample_sharing"], [49, 2, 1, "", "reduce_embedding"], [49, 4, 1, "", "relation_embedding"], [49, 2, 1, "", "score_heads"], [49, 2, 1, "", "score_tails"], [49, 2, 1, "", "score_triple"], [49, 4, 1, "", "sharding"], [49, 2, 1, "", "update_sharding"]], "besskge.scoring.InterHT": [[50, 2, 1, "", "broadcasted_distance"], [50, 4, 1, "", "entity_embedding"], [50, 2, 1, "", "forward"], [50, 4, 1, "", "negative_sample_sharing"], [50, 2, 1, "", "reduce_embedding"], [50, 4, 1, "", "relation_embedding"], [50, 2, 1, "", "score_heads"], [50, 2, 1, "", "score_tails"], [50, 2, 1, "", "score_triple"], [50, 4, 1, "", "sharding"], [50, 2, 1, "", "update_sharding"]], "besskge.scoring.MatrixDecompositionScoreFunction": [[51, 2, 1, "", "broadcasted_dot_product"], [51, 4, 1, "", "entity_embedding"], [51, 2, 1, "", "forward"], [51, 4, 1, "", "negative_sample_sharing"], [51, 2, 1, "", "reduce_embedding"], [51, 4, 1, "", "relation_embedding"], [51, 2, 1, "", "score_heads"], [51, 2, 1, "", "score_tails"], [51, 2, 1, "", "score_triple"], [51, 4, 1, "", "sharding"], [51, 2, 1, "", "update_sharding"]], "besskge.scoring.PairRE": [[52, 2, 1, "", "broadcasted_distance"], [52, 4, 1, "", "entity_embedding"], [52, 2, 1, "", "forward"], [52, 4, 1, "", "negative_sample_sharing"], [52, 2, 1, "", "reduce_embedding"], [52, 4, 1, "", "relation_embedding"], [52, 2, 1, "", "score_heads"], [52, 2, 1, "", "score_tails"], [52, 2, 1, "", "score_triple"], [52, 4, 1, "", "sharding"], [52, 2, 1, "", "update_sharding"]], "besskge.scoring.RotatE": [[53, 2, 1, "", "broadcasted_distance"], [53, 4, 1, "", "entity_embedding"], [53, 2, 1, "", "forward"], [53, 4, 1, "", "negative_sample_sharing"], [53, 2, 1, "", "reduce_embedding"], [53, 4, 1, "", "relation_embedding"], [53, 2, 1, "", "score_heads"], [53, 2, 1, "", "score_tails"], [53, 2, 1, "", "score_triple"], [53, 4, 1, "", "sharding"], [53, 2, 1, "", "update_sharding"]], "besskge.scoring.TranS": [[54, 2, 1, "", "broadcasted_distance"], [54, 4, 1, "", "entity_embedding"], [54, 2, 1, "", "forward"], [54, 4, 1, "", "negative_sample_sharing"], [54, 2, 1, "", "reduce_embedding"], [54, 4, 1, "", "relation_embedding"], [54, 2, 1, "", "score_heads"], [54, 2, 1, "", "score_tails"], [54, 2, 1, "", "score_triple"], [54, 4, 1, "", "sharding"], [54, 2, 1, "", "update_sharding"]], "besskge.scoring.TransE": [[55, 2, 1, "", "broadcasted_distance"], [55, 4, 1, "", "entity_embedding"], [55, 2, 1, "", "forward"], [55, 4, 1, "", "negative_sample_sharing"], [55, 2, 1, "", "reduce_embedding"], [55, 4, 1, "", "relation_embedding"], [55, 2, 1, "", "score_heads"], [55, 2, 1, "", "score_tails"], [55, 2, 1, "", "score_triple"], [55, 4, 1, "", "sharding"], [55, 2, 1, "", "update_sharding"]], "besskge.scoring.TripleRE": [[56, 2, 1, "", "broadcasted_distance"], [56, 4, 1, "", "entity_embedding"], [56, 2, 1, "", "forward"], [56, 4, 1, "", "negative_sample_sharing"], [56, 2, 1, "", "reduce_embedding"], [56, 4, 1, "", "relation_embedding"], [56, 2, 1, "", "score_heads"], [56, 2, 1, "", "score_tails"], [56, 2, 1, "", "score_triple"], [56, 4, 1, "", "sharding"], [56, 2, 1, "", "update_sharding"]], "besskge.sharding": [[58, 1, 1, "", "PartitionedTripleSet"], [59, 1, 1, "", "Sharding"]], "besskge.sharding.PartitionedTripleSet": [[58, 2, 1, "", "create_from_dataset"], [58, 2, 1, "", "create_from_queries"], [58, 4, 1, "", "dummy"], [58, 4, 1, "", "inverse_triples"], [58, 4, 1, "", "neg_heads"], [58, 4, 1, "", "neg_tails"], [58, 4, 1, "", "partition_mode"], [58, 4, 1, "", "sharding"], [58, 4, 1, "", "triple_counts"], [58, 4, 1, "", "triple_offsets"], [58, 4, 1, "", "triple_sort_idx"], [58, 4, 1, "", "triples"], [58, 4, 1, "", "types"]], "besskge.sharding.Sharding": [[59, 2, 1, "", "create"], [59, 4, 1, "", "entity_to_idx"], [59, 4, 1, "", "entity_to_shard"], [59, 4, 1, "", "entity_type_counts"], [59, 4, 1, "", "entity_type_offsets"], [59, 2, 1, "", "load"], [59, 3, 1, "", "max_entity_per_shard"], [59, 3, 1, "", "n_entity"], [59, 4, 1, "", "n_shard"], [59, 2, 1, "", "save"], [59, 4, 1, "", "shard_and_idx_to_entity"], [59, 4, 1, "", "shard_counts"]], "besskge.utils": [[61, 5, 1, "", "complex_multiplication"], [62, 5, 1, "", "complex_rotation"], [63, 5, 1, "", "gather_indices"], [64, 5, 1, "", "get_entity_filter"]]}, "objtypes": {"0": "py:module", "1": "py:class", "2": "py:method", "3": "py:property", "4": "py:attribute", "5": "py:function", "6": "py:data"}, "objnames": {"0": ["py", "module", "Python module"], "1": ["py", "class", "Python class"], "2": ["py", "method", "Python method"], "3": ["py", "property", "Python property"], "4": ["py", "attribute", "Python attribute"], "5": ["py", "function", "Python function"], "6": ["py", "data", "Python data"]}, "titleterms": {"bess": [0, 1, 3, 4, 9, 10, 11, 12, 13, 14, 65], "kge": [0, 3, 4, 65], "api": 0, "refer": 0, "overview": 1, "bibliographi": 2, "how": [3, 4], "contribut": [3, 4], "project": [3, 4], "v": [3, 4], "code": [3, 4], "server": [3, 4], "paperspac": [3, 4], "setup": [3, 4], "local": [3, 4], "machin": [3, 4], "develop": [3, 4], "tip": [3, 4], "besskg": [5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64], "batch_sampl": [5, 6, 7, 8], "randomshardedbatchsampl": 6, "rigidshardedbatchsampl": 7, "shardedbatchsampl": 8, "allscoresbess": 10, "embeddingmovingbesskg": 12, "scoremovingbesskg": 13, "topkquerybesskg": 14, "dataset": [15, 16], "kgdataset": 16, "embed": [17, 18, 19, 20, 21, 22, 23], "init_kge_norm": 18, "init_kge_uniform": 19, "init_uniform_norm": 20, "initialize_entity_embed": 21, "initialize_relation_embed": 22, "refactor_embedding_shard": 23, "loss": [24, 25, 26, 27, 28, 29], "baselossfunct": 25, "logsigmoidloss": 26, "marginbasedlossfunct": 27, "marginrankingloss": 28, "sampledsoftmaxcrossentropyloss": 29, "metric": [30, 31, 32, 33, 34, 35], "basemetr": 31, "evalu": 32, "hitsatk": 33, "metrics_dict": 34, "reciprocalrank": 35, "negative_sampl": [36, 37, 38, 39, 40, 41], "placeholdernegativesampl": 37, "randomshardednegativesampl": 38, "shardednegativesampl": 39, "triplebasedshardednegativesampl": 40, "typebasedshardednegativesampl": 41, "pipelin": [42, 43], "allscorespipelin": 43, "score": [44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56], "basescorefunct": 45, "box": 46, "complex": 47, "distmult": 48, "distancebasedscorefunct": 49, "interht": 50, "matrixdecompositionscorefunct": 51, "pairr": 52, "rotat": 53, "tran": 54, "trans": 55, "tripler": 56, "shard": [57, 58, 59], "partitionedtripleset": 58, "util": [60, 61, 62, 63, 64], "complex_multipl": 61, "complex_rot": 62, "gather_indic": 63, "get_entity_filt": 64, "content": 65, "user": 66, "guid": 66, "instal": 66, "usag": 66, "get": 66, "start": 66, "limit": 66}, "envversion": {"sphinx.domains.c": 2, "sphinx.domains.changeset": 1, "sphinx.domains.citation": 1, "sphinx.domains.cpp": 8, "sphinx.domains.index": 1, "sphinx.domains.javascript": 2, "sphinx.domains.math": 2, "sphinx.domains.python": 3, "sphinx.domains.rst": 2, "sphinx.domains.std": 2, "sphinx.ext.todo": 2, "sphinx.ext.viewcode": 1, "sphinx.ext.intersphinx": 1, "sphinxcontrib.bibtex": 9, "sphinx": 57}, "alltitles": {"BESS-KGE API Reference": [[0, "bess-kge-api-reference"]], "BESS overview": [[1, "bess-overview"]], "Bibliography": [[2, "bibliography"]], "How to contribute to the BESS-KGE project": [[3, "how-to-contribute-to-the-bess-kge-project"], [4, "how-to-contribute-to-the-bess-kge-project"]], "VS Code server on Paperspace": [[3, "vs-code-server-on-paperspace"], [4, "vs-code-server-on-paperspace"]], "Setup on local machine": [[3, "setup-on-local-machine"], [4, "setup-on-local-machine"]], "Development tips": [[3, "development-tips"], [4, "development-tips"]], "besskge.batch_sampler": [[5, "module-besskge.batch_sampler"]], "besskge.batch_sampler.RandomShardedBatchSampler": [[6, "besskge-batch-sampler-randomshardedbatchsampler"]], "besskge.batch_sampler.RigidShardedBatchSampler": [[7, "besskge-batch-sampler-rigidshardedbatchsampler"]], "besskge.batch_sampler.ShardedBatchSampler": [[8, "besskge-batch-sampler-shardedbatchsampler"]], "besskge.bess": [[9, "module-besskge.bess"]], "besskge.bess.AllScoresBESS": [[10, "besskge-bess-allscoresbess"]], "besskge.bess.BessKGE": [[11, "besskge-bess-besskge"]], "besskge.bess.EmbeddingMovingBessKGE": [[12, "besskge-bess-embeddingmovingbesskge"]], "besskge.bess.ScoreMovingBessKGE": [[13, "besskge-bess-scoremovingbesskge"]], "besskge.bess.TopKQueryBessKGE": [[14, "besskge-bess-topkquerybesskge"]], "besskge.dataset": [[15, "module-besskge.dataset"]], "besskge.dataset.KGDataset": [[16, "besskge-dataset-kgdataset"]], "besskge.embedding": [[17, "module-besskge.embedding"]], "besskge.embedding.init_KGE_normal": [[18, "besskge-embedding-init-kge-normal"]], "besskge.embedding.init_KGE_uniform": [[19, "besskge-embedding-init-kge-uniform"]], "besskge.embedding.init_uniform_norm": [[20, "besskge-embedding-init-uniform-norm"]], "besskge.embedding.initialize_entity_embedding": [[21, "besskge-embedding-initialize-entity-embedding"]], "besskge.embedding.initialize_relation_embedding": [[22, "besskge-embedding-initialize-relation-embedding"]], "besskge.embedding.refactor_embedding_sharding": [[23, "besskge-embedding-refactor-embedding-sharding"]], "besskge.loss": [[24, "module-besskge.loss"]], "besskge.loss.BaseLossFunction": [[25, "besskge-loss-baselossfunction"]], "besskge.loss.LogSigmoidLoss": [[26, "besskge-loss-logsigmoidloss"]], "besskge.loss.MarginBasedLossFunction": [[27, "besskge-loss-marginbasedlossfunction"]], "besskge.loss.MarginRankingLoss": [[28, "besskge-loss-marginrankingloss"]], "besskge.loss.SampledSoftmaxCrossEntropyLoss": [[29, "besskge-loss-sampledsoftmaxcrossentropyloss"]], "besskge.metric": [[30, "module-besskge.metric"]], "besskge.metric.BaseMetric": [[31, "besskge-metric-basemetric"]], "besskge.metric.Evaluation": [[32, "besskge-metric-evaluation"]], "besskge.metric.HitsAtK": [[33, "besskge-metric-hitsatk"]], "besskge.metric.METRICS_DICT": [[34, "besskge-metric-metrics-dict"]], "besskge.metric.ReciprocalRank": [[35, "besskge-metric-reciprocalrank"]], "besskge.negative_sampler": [[36, "module-besskge.negative_sampler"]], "besskge.negative_sampler.PlaceholderNegativeSampler": [[37, "besskge-negative-sampler-placeholdernegativesampler"]], "besskge.negative_sampler.RandomShardedNegativeSampler": [[38, "besskge-negative-sampler-randomshardednegativesampler"]], "besskge.negative_sampler.ShardedNegativeSampler": [[39, "besskge-negative-sampler-shardednegativesampler"]], "besskge.negative_sampler.TripleBasedShardedNegativeSampler": [[40, "besskge-negative-sampler-triplebasedshardednegativesampler"]], "besskge.negative_sampler.TypeBasedShardedNegativeSampler": [[41, "besskge-negative-sampler-typebasedshardednegativesampler"]], "besskge.pipeline": [[42, "module-besskge.pipeline"]], "besskge.pipeline.AllScoresPipeline": [[43, "besskge-pipeline-allscorespipeline"]], "besskge.scoring": [[44, "module-besskge.scoring"]], "besskge.scoring.BaseScoreFunction": [[45, "besskge-scoring-basescorefunction"]], "besskge.scoring.BoxE": [[46, "besskge-scoring-boxe"]], "besskge.scoring.ComplEx": [[47, "besskge-scoring-complex"]], "besskge.scoring.DistMult": [[48, "besskge-scoring-distmult"]], "besskge.scoring.DistanceBasedScoreFunction": [[49, "besskge-scoring-distancebasedscorefunction"]], "besskge.scoring.InterHT": [[50, "besskge-scoring-interht"]], "besskge.scoring.MatrixDecompositionScoreFunction": [[51, "besskge-scoring-matrixdecompositionscorefunction"]], "besskge.scoring.PairRE": [[52, "besskge-scoring-pairre"]], "besskge.scoring.RotatE": [[53, "besskge-scoring-rotate"]], "besskge.scoring.TranS": [[54, "besskge-scoring-trans"]], "besskge.scoring.TransE": [[55, "besskge-scoring-transe"]], "besskge.scoring.TripleRE": [[56, "besskge-scoring-triplere"]], "besskge.sharding": [[57, "module-besskge.sharding"]], "besskge.sharding.PartitionedTripleSet": [[58, "besskge-sharding-partitionedtripleset"]], "besskge.sharding.Sharding": [[59, "besskge-sharding-sharding"]], "besskge.utils": [[60, "module-besskge.utils"]], "besskge.utils.complex_multiplication": [[61, "besskge-utils-complex-multiplication"]], "besskge.utils.complex_rotation": [[62, "besskge-utils-complex-rotation"]], "besskge.utils.gather_indices": [[63, "besskge-utils-gather-indices"]], "besskge.utils.get_entity_filter": [[64, "besskge-utils-get-entity-filter"]], "BESS-KGE": [[65, "module-besskge"]], "Contents": [[65, null]], "User guide": [[66, "user-guide"]], "Installation and usage": [[66, "installation-and-usage"]], "Getting started": [[66, "getting-started"]], "Limitations": [[66, "limitations"]]}, "indexentries": {"besskge.batch_sampler": [[5, "module-besskge.batch_sampler"]], "module": [[5, "module-besskge.batch_sampler"], [9, "module-besskge.bess"], [15, "module-besskge.dataset"], [17, "module-besskge.embedding"], [24, "module-besskge.loss"], [30, "module-besskge.metric"], [36, "module-besskge.negative_sampler"], [42, "module-besskge.pipeline"], [44, "module-besskge.scoring"], [57, "module-besskge.sharding"], [60, "module-besskge.utils"], [65, "module-besskge"]], "randomshardedbatchsampler (class in besskge.batch_sampler)": [[6, "besskge.batch_sampler.RandomShardedBatchSampler"]], "get_dataloader() (besskge.batch_sampler.randomshardedbatchsampler method)": [[6, "besskge.batch_sampler.RandomShardedBatchSampler.get_dataloader"]], "get_dataloader_sampler() (besskge.batch_sampler.randomshardedbatchsampler method)": [[6, "besskge.batch_sampler.RandomShardedBatchSampler.get_dataloader_sampler"]], "sample_triples() (besskge.batch_sampler.randomshardedbatchsampler method)": [[6, "besskge.batch_sampler.RandomShardedBatchSampler.sample_triples"]], "worker_init_fn() (besskge.batch_sampler.randomshardedbatchsampler static method)": [[6, "besskge.batch_sampler.RandomShardedBatchSampler.worker_init_fn"]], "rigidshardedbatchsampler (class in besskge.batch_sampler)": [[7, "besskge.batch_sampler.RigidShardedBatchSampler"]], "get_dataloader() (besskge.batch_sampler.rigidshardedbatchsampler method)": [[7, "besskge.batch_sampler.RigidShardedBatchSampler.get_dataloader"]], "get_dataloader_sampler() (besskge.batch_sampler.rigidshardedbatchsampler method)": [[7, "besskge.batch_sampler.RigidShardedBatchSampler.get_dataloader_sampler"]], "sample_triples() (besskge.batch_sampler.rigidshardedbatchsampler method)": [[7, "besskge.batch_sampler.RigidShardedBatchSampler.sample_triples"]], "worker_init_fn() (besskge.batch_sampler.rigidshardedbatchsampler static method)": [[7, "besskge.batch_sampler.RigidShardedBatchSampler.worker_init_fn"]], "shardedbatchsampler (class in besskge.batch_sampler)": [[8, "besskge.batch_sampler.ShardedBatchSampler"]], "get_dataloader() (besskge.batch_sampler.shardedbatchsampler method)": [[8, "besskge.batch_sampler.ShardedBatchSampler.get_dataloader"]], "get_dataloader_sampler() (besskge.batch_sampler.shardedbatchsampler method)": [[8, "besskge.batch_sampler.ShardedBatchSampler.get_dataloader_sampler"]], "sample_triples() (besskge.batch_sampler.shardedbatchsampler method)": [[8, "besskge.batch_sampler.ShardedBatchSampler.sample_triples"]], "worker_init_fn() (besskge.batch_sampler.shardedbatchsampler static method)": [[8, "besskge.batch_sampler.ShardedBatchSampler.worker_init_fn"]], "besskge.bess": [[9, "module-besskge.bess"]], "allscoresbess (class in besskge.bess)": [[10, "besskge.bess.AllScoresBESS"]], "forward() (besskge.bess.allscoresbess method)": [[10, "besskge.bess.AllScoresBESS.forward"]], "besskge (class in besskge.bess)": [[11, "besskge.bess.BessKGE"]], "forward() (besskge.bess.besskge method)": [[11, "besskge.bess.BessKGE.forward"]], "n_embedding_parameters (besskge.bess.besskge property)": [[11, "besskge.bess.BessKGE.n_embedding_parameters"]], "score_batch() (besskge.bess.besskge method)": [[11, "besskge.bess.BessKGE.score_batch"]], "embeddingmovingbesskge (class in besskge.bess)": [[12, "besskge.bess.EmbeddingMovingBessKGE"]], "forward() (besskge.bess.embeddingmovingbesskge method)": [[12, "besskge.bess.EmbeddingMovingBessKGE.forward"]], "n_embedding_parameters (besskge.bess.embeddingmovingbesskge property)": [[12, "besskge.bess.EmbeddingMovingBessKGE.n_embedding_parameters"]], "score_batch() (besskge.bess.embeddingmovingbesskge method)": [[12, "besskge.bess.EmbeddingMovingBessKGE.score_batch"]], "scoremovingbesskge (class in besskge.bess)": [[13, "besskge.bess.ScoreMovingBessKGE"]], "forward() (besskge.bess.scoremovingbesskge method)": [[13, "besskge.bess.ScoreMovingBessKGE.forward"]], "n_embedding_parameters (besskge.bess.scoremovingbesskge property)": [[13, "besskge.bess.ScoreMovingBessKGE.n_embedding_parameters"]], "score_batch() (besskge.bess.scoremovingbesskge method)": [[13, "besskge.bess.ScoreMovingBessKGE.score_batch"]], "topkquerybesskge (class in besskge.bess)": [[14, "besskge.bess.TopKQueryBessKGE"]], "forward() (besskge.bess.topkquerybesskge method)": [[14, "besskge.bess.TopKQueryBessKGE.forward"]], "besskge.dataset": [[15, "module-besskge.dataset"]], "kgdataset (class in besskge.dataset)": [[16, "besskge.dataset.KGDataset"]], "build_ogbl_biokg() (besskge.dataset.kgdataset class method)": [[16, "besskge.dataset.KGDataset.build_ogbl_biokg"]], "build_ogbl_wikikg2() (besskge.dataset.kgdataset class method)": [[16, "besskge.dataset.KGDataset.build_ogbl_wikikg2"]], "build_openbiolink() (besskge.dataset.kgdataset class method)": [[16, "besskge.dataset.KGDataset.build_openbiolink"]], "build_yago310() (besskge.dataset.kgdataset class method)": [[16, "besskge.dataset.KGDataset.build_yago310"]], "entity_dict (besskge.dataset.kgdataset attribute)": [[16, "besskge.dataset.KGDataset.entity_dict"]], "from_dataframe() (besskge.dataset.kgdataset class method)": [[16, "besskge.dataset.KGDataset.from_dataframe"]], "from_triples() (besskge.dataset.kgdataset class method)": [[16, "besskge.dataset.KGDataset.from_triples"]], "ht_types (besskge.dataset.kgdataset property)": [[16, "besskge.dataset.KGDataset.ht_types"]], "load() (besskge.dataset.kgdataset class method)": [[16, "besskge.dataset.KGDataset.load"]], "n_entity (besskge.dataset.kgdataset attribute)": [[16, "besskge.dataset.KGDataset.n_entity"]], "n_relation_type (besskge.dataset.kgdataset attribute)": [[16, "besskge.dataset.KGDataset.n_relation_type"]], "neg_heads (besskge.dataset.kgdataset attribute)": [[16, "besskge.dataset.KGDataset.neg_heads"]], "neg_tails (besskge.dataset.kgdataset attribute)": [[16, "besskge.dataset.KGDataset.neg_tails"]], "relation_dict (besskge.dataset.kgdataset attribute)": [[16, "besskge.dataset.KGDataset.relation_dict"]], "save() (besskge.dataset.kgdataset method)": [[16, "besskge.dataset.KGDataset.save"]], "triples (besskge.dataset.kgdataset attribute)": [[16, "besskge.dataset.KGDataset.triples"]], "type_offsets (besskge.dataset.kgdataset attribute)": [[16, "besskge.dataset.KGDataset.type_offsets"]], "besskge.embedding": [[17, "module-besskge.embedding"]], "init_kge_normal() (in module besskge.embedding)": [[18, "besskge.embedding.init_KGE_normal"]], "init_kge_uniform() (in module besskge.embedding)": [[19, "besskge.embedding.init_KGE_uniform"]], "init_uniform_norm() (in module besskge.embedding)": [[20, "besskge.embedding.init_uniform_norm"]], "initialize_entity_embedding() (in module besskge.embedding)": [[21, "besskge.embedding.initialize_entity_embedding"]], "initialize_relation_embedding() (in module besskge.embedding)": [[22, "besskge.embedding.initialize_relation_embedding"]], "refactor_embedding_sharding() (in module besskge.embedding)": [[23, "besskge.embedding.refactor_embedding_sharding"]], "besskge.loss": [[24, "module-besskge.loss"]], "baselossfunction (class in besskge.loss)": [[25, "besskge.loss.BaseLossFunction"]], "forward() (besskge.loss.baselossfunction method)": [[25, "besskge.loss.BaseLossFunction.forward"]], "get_negative_weights() (besskge.loss.baselossfunction method)": [[25, "besskge.loss.BaseLossFunction.get_negative_weights"]], "loss_scale (besskge.loss.baselossfunction attribute)": [[25, "besskge.loss.BaseLossFunction.loss_scale"]], "negative_adversarial_sampling (besskge.loss.baselossfunction attribute)": [[25, "besskge.loss.BaseLossFunction.negative_adversarial_sampling"]], "negative_adversarial_scale (besskge.loss.baselossfunction attribute)": [[25, "besskge.loss.BaseLossFunction.negative_adversarial_scale"]], "logsigmoidloss (class in besskge.loss)": [[26, "besskge.loss.LogSigmoidLoss"]], "forward() (besskge.loss.logsigmoidloss method)": [[26, "besskge.loss.LogSigmoidLoss.forward"]], "get_negative_weights() (besskge.loss.logsigmoidloss method)": [[26, "besskge.loss.LogSigmoidLoss.get_negative_weights"]], "loss_scale (besskge.loss.logsigmoidloss attribute)": [[26, "besskge.loss.LogSigmoidLoss.loss_scale"]], "negative_adversarial_sampling (besskge.loss.logsigmoidloss attribute)": [[26, "besskge.loss.LogSigmoidLoss.negative_adversarial_sampling"]], "negative_adversarial_scale (besskge.loss.logsigmoidloss attribute)": [[26, "besskge.loss.LogSigmoidLoss.negative_adversarial_scale"]], "marginbasedlossfunction (class in besskge.loss)": [[27, "besskge.loss.MarginBasedLossFunction"]], "forward() (besskge.loss.marginbasedlossfunction method)": [[27, "besskge.loss.MarginBasedLossFunction.forward"]], "get_negative_weights() (besskge.loss.marginbasedlossfunction method)": [[27, "besskge.loss.MarginBasedLossFunction.get_negative_weights"]], "loss_scale (besskge.loss.marginbasedlossfunction attribute)": [[27, "besskge.loss.MarginBasedLossFunction.loss_scale"]], "negative_adversarial_sampling (besskge.loss.marginbasedlossfunction attribute)": [[27, "besskge.loss.MarginBasedLossFunction.negative_adversarial_sampling"]], "negative_adversarial_scale (besskge.loss.marginbasedlossfunction attribute)": [[27, "besskge.loss.MarginBasedLossFunction.negative_adversarial_scale"]], "marginrankingloss (class in besskge.loss)": [[28, "besskge.loss.MarginRankingLoss"]], "forward() (besskge.loss.marginrankingloss method)": [[28, "besskge.loss.MarginRankingLoss.forward"]], "get_negative_weights() (besskge.loss.marginrankingloss method)": [[28, "besskge.loss.MarginRankingLoss.get_negative_weights"]], "loss_scale (besskge.loss.marginrankingloss attribute)": [[28, "besskge.loss.MarginRankingLoss.loss_scale"]], "negative_adversarial_sampling (besskge.loss.marginrankingloss attribute)": [[28, "besskge.loss.MarginRankingLoss.negative_adversarial_sampling"]], "negative_adversarial_scale (besskge.loss.marginrankingloss attribute)": [[28, "besskge.loss.MarginRankingLoss.negative_adversarial_scale"]], "sampledsoftmaxcrossentropyloss (class in besskge.loss)": [[29, "besskge.loss.SampledSoftmaxCrossEntropyLoss"]], "forward() (besskge.loss.sampledsoftmaxcrossentropyloss method)": [[29, "besskge.loss.SampledSoftmaxCrossEntropyLoss.forward"]], "get_negative_weights() (besskge.loss.sampledsoftmaxcrossentropyloss method)": [[29, "besskge.loss.SampledSoftmaxCrossEntropyLoss.get_negative_weights"]], "loss_scale (besskge.loss.sampledsoftmaxcrossentropyloss attribute)": [[29, "besskge.loss.SampledSoftmaxCrossEntropyLoss.loss_scale"]], "negative_adversarial_sampling (besskge.loss.sampledsoftmaxcrossentropyloss attribute)": [[29, "besskge.loss.SampledSoftmaxCrossEntropyLoss.negative_adversarial_sampling"]], "negative_adversarial_scale (besskge.loss.sampledsoftmaxcrossentropyloss attribute)": [[29, "besskge.loss.SampledSoftmaxCrossEntropyLoss.negative_adversarial_scale"]], "besskge.metric": [[30, "module-besskge.metric"]], "basemetric (class in besskge.metric)": [[31, "besskge.metric.BaseMetric"]], "evaluation (class in besskge.metric)": [[32, "besskge.metric.Evaluation"]], "dict_metrics_from_ranks() (besskge.metric.evaluation method)": [[32, "besskge.metric.Evaluation.dict_metrics_from_ranks"]], "ranks_from_indices() (besskge.metric.evaluation method)": [[32, "besskge.metric.Evaluation.ranks_from_indices"]], "ranks_from_scores() (besskge.metric.evaluation method)": [[32, "besskge.metric.Evaluation.ranks_from_scores"]], "stacked_metrics_from_ranks() (besskge.metric.evaluation method)": [[32, "besskge.metric.Evaluation.stacked_metrics_from_ranks"]], "hitsatk (class in besskge.metric)": [[33, "besskge.metric.HitsAtK"]], "metrics_dict (in module besskge.metric)": [[34, "besskge.metric.METRICS_DICT"]], "reciprocalrank (class in besskge.metric)": [[35, "besskge.metric.ReciprocalRank"]], "besskge.negative_sampler": [[36, "module-besskge.negative_sampler"]], "placeholdernegativesampler (class in besskge.negative_sampler)": [[37, "besskge.negative_sampler.PlaceholderNegativeSampler"]], "corruption_scheme (besskge.negative_sampler.placeholdernegativesampler attribute)": [[37, "besskge.negative_sampler.PlaceholderNegativeSampler.corruption_scheme"]], "flat_negative_format (besskge.negative_sampler.placeholdernegativesampler attribute)": [[37, "besskge.negative_sampler.PlaceholderNegativeSampler.flat_negative_format"]], "local_sampling (besskge.negative_sampler.placeholdernegativesampler attribute)": [[37, "besskge.negative_sampler.PlaceholderNegativeSampler.local_sampling"]], "rng (besskge.negative_sampler.placeholdernegativesampler attribute)": [[37, "besskge.negative_sampler.PlaceholderNegativeSampler.rng"]], "randomshardednegativesampler (class in besskge.negative_sampler)": [[38, "besskge.negative_sampler.RandomShardedNegativeSampler"]], "corruption_scheme (besskge.negative_sampler.randomshardednegativesampler attribute)": [[38, "besskge.negative_sampler.RandomShardedNegativeSampler.corruption_scheme"]], "flat_negative_format (besskge.negative_sampler.randomshardednegativesampler attribute)": [[38, "besskge.negative_sampler.RandomShardedNegativeSampler.flat_negative_format"]], "local_sampling (besskge.negative_sampler.randomshardednegativesampler attribute)": [[38, "besskge.negative_sampler.RandomShardedNegativeSampler.local_sampling"]], "rng (besskge.negative_sampler.randomshardednegativesampler attribute)": [[38, "besskge.negative_sampler.RandomShardedNegativeSampler.rng"]], "shardednegativesampler (class in besskge.negative_sampler)": [[39, "besskge.negative_sampler.ShardedNegativeSampler"]], "corruption_scheme (besskge.negative_sampler.shardednegativesampler attribute)": [[39, "besskge.negative_sampler.ShardedNegativeSampler.corruption_scheme"]], "flat_negative_format (besskge.negative_sampler.shardednegativesampler attribute)": [[39, "besskge.negative_sampler.ShardedNegativeSampler.flat_negative_format"]], "local_sampling (besskge.negative_sampler.shardednegativesampler attribute)": [[39, "besskge.negative_sampler.ShardedNegativeSampler.local_sampling"]], "rng (besskge.negative_sampler.shardednegativesampler attribute)": [[39, "besskge.negative_sampler.ShardedNegativeSampler.rng"]], "triplebasedshardednegativesampler (class in besskge.negative_sampler)": [[40, "besskge.negative_sampler.TripleBasedShardedNegativeSampler"]], "corruption_scheme (besskge.negative_sampler.triplebasedshardednegativesampler attribute)": [[40, "besskge.negative_sampler.TripleBasedShardedNegativeSampler.corruption_scheme"]], "flat_negative_format (besskge.negative_sampler.triplebasedshardednegativesampler attribute)": [[40, "besskge.negative_sampler.TripleBasedShardedNegativeSampler.flat_negative_format"]], "local_sampling (besskge.negative_sampler.triplebasedshardednegativesampler attribute)": [[40, "besskge.negative_sampler.TripleBasedShardedNegativeSampler.local_sampling"]], "pad_negatives() (besskge.negative_sampler.triplebasedshardednegativesampler method)": [[40, "besskge.negative_sampler.TripleBasedShardedNegativeSampler.pad_negatives"]], "rng (besskge.negative_sampler.triplebasedshardednegativesampler attribute)": [[40, "besskge.negative_sampler.TripleBasedShardedNegativeSampler.rng"]], "shard_negatives() (besskge.negative_sampler.triplebasedshardednegativesampler method)": [[40, "besskge.negative_sampler.TripleBasedShardedNegativeSampler.shard_negatives"]], "typebasedshardednegativesampler (class in besskge.negative_sampler)": [[41, "besskge.negative_sampler.TypeBasedShardedNegativeSampler"]], "corruption_scheme (besskge.negative_sampler.typebasedshardednegativesampler attribute)": [[41, "besskge.negative_sampler.TypeBasedShardedNegativeSampler.corruption_scheme"]], "flat_negative_format (besskge.negative_sampler.typebasedshardednegativesampler attribute)": [[41, "besskge.negative_sampler.TypeBasedShardedNegativeSampler.flat_negative_format"]], "local_sampling (besskge.negative_sampler.typebasedshardednegativesampler attribute)": [[41, "besskge.negative_sampler.TypeBasedShardedNegativeSampler.local_sampling"]], "rng (besskge.negative_sampler.typebasedshardednegativesampler attribute)": [[41, "besskge.negative_sampler.TypeBasedShardedNegativeSampler.rng"]], "besskge.pipeline": [[42, "module-besskge.pipeline"]], "allscorespipeline (class in besskge.pipeline)": [[43, "besskge.pipeline.AllScoresPipeline"]], "forward() (besskge.pipeline.allscorespipeline method)": [[43, "besskge.pipeline.AllScoresPipeline.forward"]], "besskge.scoring": [[44, "module-besskge.scoring"]], "basescorefunction (class in besskge.scoring)": [[45, "besskge.scoring.BaseScoreFunction"]], "entity_embedding (besskge.scoring.basescorefunction attribute)": [[45, "besskge.scoring.BaseScoreFunction.entity_embedding"]], "forward() (besskge.scoring.basescorefunction method)": [[45, "besskge.scoring.BaseScoreFunction.forward"]], "negative_sample_sharing (besskge.scoring.basescorefunction attribute)": [[45, "besskge.scoring.BaseScoreFunction.negative_sample_sharing"]], "relation_embedding (besskge.scoring.basescorefunction attribute)": [[45, "besskge.scoring.BaseScoreFunction.relation_embedding"]], "score_heads() (besskge.scoring.basescorefunction method)": [[45, "besskge.scoring.BaseScoreFunction.score_heads"]], "score_tails() (besskge.scoring.basescorefunction method)": [[45, "besskge.scoring.BaseScoreFunction.score_tails"]], "score_triple() (besskge.scoring.basescorefunction method)": [[45, "besskge.scoring.BaseScoreFunction.score_triple"]], "sharding (besskge.scoring.basescorefunction attribute)": [[45, "besskge.scoring.BaseScoreFunction.sharding"]], "update_sharding() (besskge.scoring.basescorefunction method)": [[45, "besskge.scoring.BaseScoreFunction.update_sharding"]], "boxe (class in besskge.scoring)": [[46, "besskge.scoring.BoxE"]], "boxe_score() (besskge.scoring.boxe method)": [[46, "besskge.scoring.BoxE.boxe_score"]], "broadcasted_distance() (besskge.scoring.boxe method)": [[46, "besskge.scoring.BoxE.broadcasted_distance"]], "entity_embedding (besskge.scoring.boxe attribute)": [[46, "besskge.scoring.BoxE.entity_embedding"]], "forward() (besskge.scoring.boxe method)": [[46, "besskge.scoring.BoxE.forward"]], "negative_sample_sharing (besskge.scoring.boxe attribute)": [[46, "besskge.scoring.BoxE.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.boxe method)": [[46, "besskge.scoring.BoxE.reduce_embedding"]], "relation_embedding (besskge.scoring.boxe attribute)": [[46, "besskge.scoring.BoxE.relation_embedding"]], "score_heads() (besskge.scoring.boxe method)": [[46, "besskge.scoring.BoxE.score_heads"]], "score_tails() (besskge.scoring.boxe method)": [[46, "besskge.scoring.BoxE.score_tails"]], "score_triple() (besskge.scoring.boxe method)": [[46, "besskge.scoring.BoxE.score_triple"]], "sharding (besskge.scoring.boxe attribute)": [[46, "besskge.scoring.BoxE.sharding"]], "update_sharding() (besskge.scoring.boxe method)": [[46, "besskge.scoring.BoxE.update_sharding"]], "complex (class in besskge.scoring)": [[47, "besskge.scoring.ComplEx"]], "broadcasted_dot_product() (besskge.scoring.complex method)": [[47, "besskge.scoring.ComplEx.broadcasted_dot_product"]], "entity_embedding (besskge.scoring.complex attribute)": [[47, "besskge.scoring.ComplEx.entity_embedding"]], "forward() (besskge.scoring.complex method)": [[47, "besskge.scoring.ComplEx.forward"]], "negative_sample_sharing (besskge.scoring.complex attribute)": [[47, "besskge.scoring.ComplEx.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.complex method)": [[47, "besskge.scoring.ComplEx.reduce_embedding"]], "relation_embedding (besskge.scoring.complex attribute)": [[47, "besskge.scoring.ComplEx.relation_embedding"]], "score_heads() (besskge.scoring.complex method)": [[47, "besskge.scoring.ComplEx.score_heads"]], "score_tails() (besskge.scoring.complex method)": [[47, "besskge.scoring.ComplEx.score_tails"]], "score_triple() (besskge.scoring.complex method)": [[47, "besskge.scoring.ComplEx.score_triple"]], "sharding (besskge.scoring.complex attribute)": [[47, "besskge.scoring.ComplEx.sharding"]], "update_sharding() (besskge.scoring.complex method)": [[47, "besskge.scoring.ComplEx.update_sharding"]], "distmult (class in besskge.scoring)": [[48, "besskge.scoring.DistMult"]], "broadcasted_dot_product() (besskge.scoring.distmult method)": [[48, "besskge.scoring.DistMult.broadcasted_dot_product"]], "entity_embedding (besskge.scoring.distmult attribute)": [[48, "besskge.scoring.DistMult.entity_embedding"]], "forward() (besskge.scoring.distmult method)": [[48, "besskge.scoring.DistMult.forward"]], "negative_sample_sharing (besskge.scoring.distmult attribute)": [[48, "besskge.scoring.DistMult.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.distmult method)": [[48, "besskge.scoring.DistMult.reduce_embedding"]], "relation_embedding (besskge.scoring.distmult attribute)": [[48, "besskge.scoring.DistMult.relation_embedding"]], "score_heads() (besskge.scoring.distmult method)": [[48, "besskge.scoring.DistMult.score_heads"]], "score_tails() (besskge.scoring.distmult method)": [[48, "besskge.scoring.DistMult.score_tails"]], "score_triple() (besskge.scoring.distmult method)": [[48, "besskge.scoring.DistMult.score_triple"]], "sharding (besskge.scoring.distmult attribute)": [[48, "besskge.scoring.DistMult.sharding"]], "update_sharding() (besskge.scoring.distmult method)": [[48, "besskge.scoring.DistMult.update_sharding"]], "distancebasedscorefunction (class in besskge.scoring)": [[49, "besskge.scoring.DistanceBasedScoreFunction"]], "broadcasted_distance() (besskge.scoring.distancebasedscorefunction method)": [[49, "besskge.scoring.DistanceBasedScoreFunction.broadcasted_distance"]], "entity_embedding (besskge.scoring.distancebasedscorefunction attribute)": [[49, "besskge.scoring.DistanceBasedScoreFunction.entity_embedding"]], "forward() (besskge.scoring.distancebasedscorefunction method)": [[49, "besskge.scoring.DistanceBasedScoreFunction.forward"]], "negative_sample_sharing (besskge.scoring.distancebasedscorefunction attribute)": [[49, "besskge.scoring.DistanceBasedScoreFunction.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.distancebasedscorefunction method)": [[49, "besskge.scoring.DistanceBasedScoreFunction.reduce_embedding"]], "relation_embedding (besskge.scoring.distancebasedscorefunction attribute)": [[49, "besskge.scoring.DistanceBasedScoreFunction.relation_embedding"]], "score_heads() (besskge.scoring.distancebasedscorefunction method)": [[49, "besskge.scoring.DistanceBasedScoreFunction.score_heads"]], "score_tails() (besskge.scoring.distancebasedscorefunction method)": [[49, "besskge.scoring.DistanceBasedScoreFunction.score_tails"]], "score_triple() (besskge.scoring.distancebasedscorefunction method)": [[49, "besskge.scoring.DistanceBasedScoreFunction.score_triple"]], "sharding (besskge.scoring.distancebasedscorefunction attribute)": [[49, "besskge.scoring.DistanceBasedScoreFunction.sharding"]], "update_sharding() (besskge.scoring.distancebasedscorefunction method)": [[49, "besskge.scoring.DistanceBasedScoreFunction.update_sharding"]], "interht (class in besskge.scoring)": [[50, "besskge.scoring.InterHT"]], "broadcasted_distance() (besskge.scoring.interht method)": [[50, "besskge.scoring.InterHT.broadcasted_distance"]], "entity_embedding (besskge.scoring.interht attribute)": [[50, "besskge.scoring.InterHT.entity_embedding"]], "forward() (besskge.scoring.interht method)": [[50, "besskge.scoring.InterHT.forward"]], "negative_sample_sharing (besskge.scoring.interht attribute)": [[50, "besskge.scoring.InterHT.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.interht method)": [[50, "besskge.scoring.InterHT.reduce_embedding"]], "relation_embedding (besskge.scoring.interht attribute)": [[50, "besskge.scoring.InterHT.relation_embedding"]], "score_heads() (besskge.scoring.interht method)": [[50, "besskge.scoring.InterHT.score_heads"]], "score_tails() (besskge.scoring.interht method)": [[50, "besskge.scoring.InterHT.score_tails"]], "score_triple() (besskge.scoring.interht method)": [[50, "besskge.scoring.InterHT.score_triple"]], "sharding (besskge.scoring.interht attribute)": [[50, "besskge.scoring.InterHT.sharding"]], "update_sharding() (besskge.scoring.interht method)": [[50, "besskge.scoring.InterHT.update_sharding"]], "matrixdecompositionscorefunction (class in besskge.scoring)": [[51, "besskge.scoring.MatrixDecompositionScoreFunction"]], "broadcasted_dot_product() (besskge.scoring.matrixdecompositionscorefunction method)": [[51, "besskge.scoring.MatrixDecompositionScoreFunction.broadcasted_dot_product"]], "entity_embedding (besskge.scoring.matrixdecompositionscorefunction attribute)": [[51, "besskge.scoring.MatrixDecompositionScoreFunction.entity_embedding"]], "forward() (besskge.scoring.matrixdecompositionscorefunction method)": [[51, "besskge.scoring.MatrixDecompositionScoreFunction.forward"]], "negative_sample_sharing (besskge.scoring.matrixdecompositionscorefunction attribute)": [[51, "besskge.scoring.MatrixDecompositionScoreFunction.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.matrixdecompositionscorefunction method)": [[51, "besskge.scoring.MatrixDecompositionScoreFunction.reduce_embedding"]], "relation_embedding (besskge.scoring.matrixdecompositionscorefunction attribute)": [[51, "besskge.scoring.MatrixDecompositionScoreFunction.relation_embedding"]], "score_heads() (besskge.scoring.matrixdecompositionscorefunction method)": [[51, "besskge.scoring.MatrixDecompositionScoreFunction.score_heads"]], "score_tails() (besskge.scoring.matrixdecompositionscorefunction method)": [[51, "besskge.scoring.MatrixDecompositionScoreFunction.score_tails"]], "score_triple() (besskge.scoring.matrixdecompositionscorefunction method)": [[51, "besskge.scoring.MatrixDecompositionScoreFunction.score_triple"]], "sharding (besskge.scoring.matrixdecompositionscorefunction attribute)": [[51, "besskge.scoring.MatrixDecompositionScoreFunction.sharding"]], "update_sharding() (besskge.scoring.matrixdecompositionscorefunction method)": [[51, "besskge.scoring.MatrixDecompositionScoreFunction.update_sharding"]], "pairre (class in besskge.scoring)": [[52, "besskge.scoring.PairRE"]], "broadcasted_distance() (besskge.scoring.pairre method)": [[52, "besskge.scoring.PairRE.broadcasted_distance"]], "entity_embedding (besskge.scoring.pairre attribute)": [[52, "besskge.scoring.PairRE.entity_embedding"]], "forward() (besskge.scoring.pairre method)": [[52, "besskge.scoring.PairRE.forward"]], "negative_sample_sharing (besskge.scoring.pairre attribute)": [[52, "besskge.scoring.PairRE.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.pairre method)": [[52, "besskge.scoring.PairRE.reduce_embedding"]], "relation_embedding (besskge.scoring.pairre attribute)": [[52, "besskge.scoring.PairRE.relation_embedding"]], "score_heads() (besskge.scoring.pairre method)": [[52, "besskge.scoring.PairRE.score_heads"]], "score_tails() (besskge.scoring.pairre method)": [[52, "besskge.scoring.PairRE.score_tails"]], "score_triple() (besskge.scoring.pairre method)": [[52, "besskge.scoring.PairRE.score_triple"]], "sharding (besskge.scoring.pairre attribute)": [[52, "besskge.scoring.PairRE.sharding"]], "update_sharding() (besskge.scoring.pairre method)": [[52, "besskge.scoring.PairRE.update_sharding"]], "rotate (class in besskge.scoring)": [[53, "besskge.scoring.RotatE"]], "broadcasted_distance() (besskge.scoring.rotate method)": [[53, "besskge.scoring.RotatE.broadcasted_distance"]], "entity_embedding (besskge.scoring.rotate attribute)": [[53, "besskge.scoring.RotatE.entity_embedding"]], "forward() (besskge.scoring.rotate method)": [[53, "besskge.scoring.RotatE.forward"]], "negative_sample_sharing (besskge.scoring.rotate attribute)": [[53, "besskge.scoring.RotatE.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.rotate method)": [[53, "besskge.scoring.RotatE.reduce_embedding"]], "relation_embedding (besskge.scoring.rotate attribute)": [[53, "besskge.scoring.RotatE.relation_embedding"]], "score_heads() (besskge.scoring.rotate method)": [[53, "besskge.scoring.RotatE.score_heads"]], "score_tails() (besskge.scoring.rotate method)": [[53, "besskge.scoring.RotatE.score_tails"]], "score_triple() (besskge.scoring.rotate method)": [[53, "besskge.scoring.RotatE.score_triple"]], "sharding (besskge.scoring.rotate attribute)": [[53, "besskge.scoring.RotatE.sharding"]], "update_sharding() (besskge.scoring.rotate method)": [[53, "besskge.scoring.RotatE.update_sharding"]], "trans (class in besskge.scoring)": [[54, "besskge.scoring.TranS"]], "broadcasted_distance() (besskge.scoring.trans method)": [[54, "besskge.scoring.TranS.broadcasted_distance"]], "entity_embedding (besskge.scoring.trans attribute)": [[54, "besskge.scoring.TranS.entity_embedding"]], "forward() (besskge.scoring.trans method)": [[54, "besskge.scoring.TranS.forward"]], "negative_sample_sharing (besskge.scoring.trans attribute)": [[54, "besskge.scoring.TranS.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.trans method)": [[54, "besskge.scoring.TranS.reduce_embedding"]], "relation_embedding (besskge.scoring.trans attribute)": [[54, "besskge.scoring.TranS.relation_embedding"]], "score_heads() (besskge.scoring.trans method)": [[54, "besskge.scoring.TranS.score_heads"]], "score_tails() (besskge.scoring.trans method)": [[54, "besskge.scoring.TranS.score_tails"]], "score_triple() (besskge.scoring.trans method)": [[54, "besskge.scoring.TranS.score_triple"]], "sharding (besskge.scoring.trans attribute)": [[54, "besskge.scoring.TranS.sharding"]], "update_sharding() (besskge.scoring.trans method)": [[54, "besskge.scoring.TranS.update_sharding"]], "transe (class in besskge.scoring)": [[55, "besskge.scoring.TransE"]], "broadcasted_distance() (besskge.scoring.transe method)": [[55, "besskge.scoring.TransE.broadcasted_distance"]], "entity_embedding (besskge.scoring.transe attribute)": [[55, "besskge.scoring.TransE.entity_embedding"]], "forward() (besskge.scoring.transe method)": [[55, "besskge.scoring.TransE.forward"]], "negative_sample_sharing (besskge.scoring.transe attribute)": [[55, "besskge.scoring.TransE.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.transe method)": [[55, "besskge.scoring.TransE.reduce_embedding"]], "relation_embedding (besskge.scoring.transe attribute)": [[55, "besskge.scoring.TransE.relation_embedding"]], "score_heads() (besskge.scoring.transe method)": [[55, "besskge.scoring.TransE.score_heads"]], "score_tails() (besskge.scoring.transe method)": [[55, "besskge.scoring.TransE.score_tails"]], "score_triple() (besskge.scoring.transe method)": [[55, "besskge.scoring.TransE.score_triple"]], "sharding (besskge.scoring.transe attribute)": [[55, "besskge.scoring.TransE.sharding"]], "update_sharding() (besskge.scoring.transe method)": [[55, "besskge.scoring.TransE.update_sharding"]], "triplere (class in besskge.scoring)": [[56, "besskge.scoring.TripleRE"]], "broadcasted_distance() (besskge.scoring.triplere method)": [[56, "besskge.scoring.TripleRE.broadcasted_distance"]], "entity_embedding (besskge.scoring.triplere attribute)": [[56, "besskge.scoring.TripleRE.entity_embedding"]], "forward() (besskge.scoring.triplere method)": [[56, "besskge.scoring.TripleRE.forward"]], "negative_sample_sharing (besskge.scoring.triplere attribute)": [[56, "besskge.scoring.TripleRE.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.triplere method)": [[56, "besskge.scoring.TripleRE.reduce_embedding"]], "relation_embedding (besskge.scoring.triplere attribute)": [[56, "besskge.scoring.TripleRE.relation_embedding"]], "score_heads() (besskge.scoring.triplere method)": [[56, "besskge.scoring.TripleRE.score_heads"]], "score_tails() (besskge.scoring.triplere method)": [[56, "besskge.scoring.TripleRE.score_tails"]], "score_triple() (besskge.scoring.triplere method)": [[56, "besskge.scoring.TripleRE.score_triple"]], "sharding (besskge.scoring.triplere attribute)": [[56, "besskge.scoring.TripleRE.sharding"]], "update_sharding() (besskge.scoring.triplere method)": [[56, "besskge.scoring.TripleRE.update_sharding"]], "besskge.sharding": [[57, "module-besskge.sharding"]], "partitionedtripleset (class in besskge.sharding)": [[58, "besskge.sharding.PartitionedTripleSet"]], "create_from_dataset() (besskge.sharding.partitionedtripleset class method)": [[58, "besskge.sharding.PartitionedTripleSet.create_from_dataset"]], "create_from_queries() (besskge.sharding.partitionedtripleset class method)": [[58, "besskge.sharding.PartitionedTripleSet.create_from_queries"]], "dummy (besskge.sharding.partitionedtripleset attribute)": [[58, "besskge.sharding.PartitionedTripleSet.dummy"]], "inverse_triples (besskge.sharding.partitionedtripleset attribute)": [[58, "besskge.sharding.PartitionedTripleSet.inverse_triples"]], "neg_heads (besskge.sharding.partitionedtripleset attribute)": [[58, "besskge.sharding.PartitionedTripleSet.neg_heads"]], "neg_tails (besskge.sharding.partitionedtripleset attribute)": [[58, "besskge.sharding.PartitionedTripleSet.neg_tails"]], "partition_mode (besskge.sharding.partitionedtripleset attribute)": [[58, "besskge.sharding.PartitionedTripleSet.partition_mode"]], "sharding (besskge.sharding.partitionedtripleset attribute)": [[58, "besskge.sharding.PartitionedTripleSet.sharding"]], "triple_counts (besskge.sharding.partitionedtripleset attribute)": [[58, "besskge.sharding.PartitionedTripleSet.triple_counts"]], "triple_offsets (besskge.sharding.partitionedtripleset attribute)": [[58, "besskge.sharding.PartitionedTripleSet.triple_offsets"]], "triple_sort_idx (besskge.sharding.partitionedtripleset attribute)": [[58, "besskge.sharding.PartitionedTripleSet.triple_sort_idx"]], "triples (besskge.sharding.partitionedtripleset attribute)": [[58, "besskge.sharding.PartitionedTripleSet.triples"]], "types (besskge.sharding.partitionedtripleset attribute)": [[58, "besskge.sharding.PartitionedTripleSet.types"]], "sharding (class in besskge.sharding)": [[59, "besskge.sharding.Sharding"]], "create() (besskge.sharding.sharding class method)": [[59, "besskge.sharding.Sharding.create"]], "entity_to_idx (besskge.sharding.sharding attribute)": [[59, "besskge.sharding.Sharding.entity_to_idx"]], "entity_to_shard (besskge.sharding.sharding attribute)": [[59, "besskge.sharding.Sharding.entity_to_shard"]], "entity_type_counts (besskge.sharding.sharding attribute)": [[59, "besskge.sharding.Sharding.entity_type_counts"]], "entity_type_offsets (besskge.sharding.sharding attribute)": [[59, "besskge.sharding.Sharding.entity_type_offsets"]], "load() (besskge.sharding.sharding class method)": [[59, "besskge.sharding.Sharding.load"]], "max_entity_per_shard (besskge.sharding.sharding property)": [[59, "besskge.sharding.Sharding.max_entity_per_shard"]], "n_entity (besskge.sharding.sharding property)": [[59, "besskge.sharding.Sharding.n_entity"]], "n_shard (besskge.sharding.sharding attribute)": [[59, "besskge.sharding.Sharding.n_shard"]], "save() (besskge.sharding.sharding method)": [[59, "besskge.sharding.Sharding.save"]], "shard_and_idx_to_entity (besskge.sharding.sharding attribute)": [[59, "besskge.sharding.Sharding.shard_and_idx_to_entity"]], "shard_counts (besskge.sharding.sharding attribute)": [[59, "besskge.sharding.Sharding.shard_counts"]], "besskge.utils": [[60, "module-besskge.utils"]], "complex_multiplication() (in module besskge.utils)": [[61, "besskge.utils.complex_multiplication"]], "complex_rotation() (in module besskge.utils)": [[62, "besskge.utils.complex_rotation"]], "gather_indices() (in module besskge.utils)": [[63, "besskge.utils.gather_indices"]], "get_entity_filter() (in module besskge.utils)": [[64, "besskge.utils.get_entity_filter"]], "besskge": [[65, "module-besskge"]]}})
\ No newline at end of file
+Search.setIndex({"docnames": ["API_reference", "bess", "bibliography", "contrib", "dev_guide", "generated/besskge.batch_sampler", "generated/besskge.batch_sampler.RandomShardedBatchSampler", "generated/besskge.batch_sampler.RigidShardedBatchSampler", "generated/besskge.batch_sampler.ShardedBatchSampler", "generated/besskge.bess", "generated/besskge.bess.AllScoresBESS", "generated/besskge.bess.BessKGE", "generated/besskge.bess.EmbeddingMovingBessKGE", "generated/besskge.bess.ScoreMovingBessKGE", "generated/besskge.bess.TopKQueryBessKGE", "generated/besskge.dataset", "generated/besskge.dataset.KGDataset", "generated/besskge.embedding", "generated/besskge.embedding.init_KGE_normal", "generated/besskge.embedding.init_KGE_uniform", "generated/besskge.embedding.init_uniform_norm", "generated/besskge.embedding.init_xavier_norm", "generated/besskge.embedding.initialize_entity_embedding", "generated/besskge.embedding.initialize_relation_embedding", "generated/besskge.embedding.refactor_embedding_sharding", "generated/besskge.loss", "generated/besskge.loss.BaseLossFunction", "generated/besskge.loss.LogSigmoidLoss", "generated/besskge.loss.MarginBasedLossFunction", "generated/besskge.loss.MarginRankingLoss", "generated/besskge.loss.SampledSoftmaxCrossEntropyLoss", "generated/besskge.metric", "generated/besskge.metric.BaseMetric", "generated/besskge.metric.Evaluation", "generated/besskge.metric.HitsAtK", "generated/besskge.metric.METRICS_DICT", "generated/besskge.metric.ReciprocalRank", "generated/besskge.negative_sampler", "generated/besskge.negative_sampler.PlaceholderNegativeSampler", "generated/besskge.negative_sampler.RandomShardedNegativeSampler", "generated/besskge.negative_sampler.ShardedNegativeSampler", "generated/besskge.negative_sampler.TripleBasedShardedNegativeSampler", "generated/besskge.negative_sampler.TypeBasedShardedNegativeSampler", "generated/besskge.pipeline", "generated/besskge.pipeline.AllScoresPipeline", "generated/besskge.scoring", "generated/besskge.scoring.BaseScoreFunction", "generated/besskge.scoring.BoxE", "generated/besskge.scoring.ComplEx", "generated/besskge.scoring.ConvE", "generated/besskge.scoring.DistMult", "generated/besskge.scoring.DistanceBasedScoreFunction", "generated/besskge.scoring.InterHT", "generated/besskge.scoring.MatrixDecompositionScoreFunction", "generated/besskge.scoring.PairRE", "generated/besskge.scoring.RotatE", "generated/besskge.scoring.TranS", "generated/besskge.scoring.TransE", "generated/besskge.scoring.TripleRE", "generated/besskge.sharding", "generated/besskge.sharding.PartitionedTripleSet", "generated/besskge.sharding.Sharding", "generated/besskge.utils", "generated/besskge.utils.complex_multiplication", "generated/besskge.utils.complex_rotation", "generated/besskge.utils.gather_indices", "generated/besskge.utils.get_entity_filter", "index", "user_guide"], "filenames": ["API_reference.rst", "bess.rst", "bibliography.rst", "contrib.md", "dev_guide.rst", "generated/besskge.batch_sampler.rst", "generated/besskge.batch_sampler.RandomShardedBatchSampler.rst", "generated/besskge.batch_sampler.RigidShardedBatchSampler.rst", "generated/besskge.batch_sampler.ShardedBatchSampler.rst", "generated/besskge.bess.rst", "generated/besskge.bess.AllScoresBESS.rst", "generated/besskge.bess.BessKGE.rst", "generated/besskge.bess.EmbeddingMovingBessKGE.rst", "generated/besskge.bess.ScoreMovingBessKGE.rst", "generated/besskge.bess.TopKQueryBessKGE.rst", "generated/besskge.dataset.rst", "generated/besskge.dataset.KGDataset.rst", "generated/besskge.embedding.rst", "generated/besskge.embedding.init_KGE_normal.rst", "generated/besskge.embedding.init_KGE_uniform.rst", "generated/besskge.embedding.init_uniform_norm.rst", "generated/besskge.embedding.init_xavier_norm.rst", "generated/besskge.embedding.initialize_entity_embedding.rst", "generated/besskge.embedding.initialize_relation_embedding.rst", "generated/besskge.embedding.refactor_embedding_sharding.rst", "generated/besskge.loss.rst", "generated/besskge.loss.BaseLossFunction.rst", "generated/besskge.loss.LogSigmoidLoss.rst", "generated/besskge.loss.MarginBasedLossFunction.rst", "generated/besskge.loss.MarginRankingLoss.rst", "generated/besskge.loss.SampledSoftmaxCrossEntropyLoss.rst", "generated/besskge.metric.rst", "generated/besskge.metric.BaseMetric.rst", "generated/besskge.metric.Evaluation.rst", "generated/besskge.metric.HitsAtK.rst", "generated/besskge.metric.METRICS_DICT.rst", "generated/besskge.metric.ReciprocalRank.rst", "generated/besskge.negative_sampler.rst", "generated/besskge.negative_sampler.PlaceholderNegativeSampler.rst", "generated/besskge.negative_sampler.RandomShardedNegativeSampler.rst", "generated/besskge.negative_sampler.ShardedNegativeSampler.rst", "generated/besskge.negative_sampler.TripleBasedShardedNegativeSampler.rst", "generated/besskge.negative_sampler.TypeBasedShardedNegativeSampler.rst", "generated/besskge.pipeline.rst", "generated/besskge.pipeline.AllScoresPipeline.rst", "generated/besskge.scoring.rst", "generated/besskge.scoring.BaseScoreFunction.rst", "generated/besskge.scoring.BoxE.rst", "generated/besskge.scoring.ComplEx.rst", "generated/besskge.scoring.ConvE.rst", "generated/besskge.scoring.DistMult.rst", "generated/besskge.scoring.DistanceBasedScoreFunction.rst", "generated/besskge.scoring.InterHT.rst", "generated/besskge.scoring.MatrixDecompositionScoreFunction.rst", "generated/besskge.scoring.PairRE.rst", "generated/besskge.scoring.RotatE.rst", "generated/besskge.scoring.TranS.rst", "generated/besskge.scoring.TransE.rst", "generated/besskge.scoring.TripleRE.rst", "generated/besskge.sharding.rst", "generated/besskge.sharding.PartitionedTripleSet.rst", "generated/besskge.sharding.Sharding.rst", "generated/besskge.utils.rst", "generated/besskge.utils.complex_multiplication.rst", "generated/besskge.utils.complex_rotation.rst", "generated/besskge.utils.gather_indices.rst", "generated/besskge.utils.get_entity_filter.rst", "index.rst", "user_guide.rst"], "titles": ["BESS-KGE API Reference", "BESS overview", "Bibliography", "How to contribute to the BESS-KGE project", "How to contribute to the BESS-KGE project", "besskge.batch_sampler", "besskge.batch_sampler.RandomShardedBatchSampler", "besskge.batch_sampler.RigidShardedBatchSampler", "besskge.batch_sampler.ShardedBatchSampler", "besskge.bess", "besskge.bess.AllScoresBESS", "besskge.bess.BessKGE", "besskge.bess.EmbeddingMovingBessKGE", "besskge.bess.ScoreMovingBessKGE", "besskge.bess.TopKQueryBessKGE", "besskge.dataset", "besskge.dataset.KGDataset", "besskge.embedding", "besskge.embedding.init_KGE_normal", "besskge.embedding.init_KGE_uniform", "besskge.embedding.init_uniform_norm", "besskge.embedding.init_xavier_norm", "besskge.embedding.initialize_entity_embedding", "besskge.embedding.initialize_relation_embedding", "besskge.embedding.refactor_embedding_sharding", "besskge.loss", "besskge.loss.BaseLossFunction", "besskge.loss.LogSigmoidLoss", "besskge.loss.MarginBasedLossFunction", "besskge.loss.MarginRankingLoss", "besskge.loss.SampledSoftmaxCrossEntropyLoss", "besskge.metric", "besskge.metric.BaseMetric", "besskge.metric.Evaluation", "besskge.metric.HitsAtK", "besskge.metric.METRICS_DICT", "besskge.metric.ReciprocalRank", "besskge.negative_sampler", "besskge.negative_sampler.PlaceholderNegativeSampler", "besskge.negative_sampler.RandomShardedNegativeSampler", "besskge.negative_sampler.ShardedNegativeSampler", "besskge.negative_sampler.TripleBasedShardedNegativeSampler", "besskge.negative_sampler.TypeBasedShardedNegativeSampler", "besskge.pipeline", "besskge.pipeline.AllScoresPipeline", "besskge.scoring", "besskge.scoring.BaseScoreFunction", "besskge.scoring.BoxE", "besskge.scoring.ComplEx", "besskge.scoring.ConvE", "besskge.scoring.DistMult", "besskge.scoring.DistanceBasedScoreFunction", "besskge.scoring.InterHT", "besskge.scoring.MatrixDecompositionScoreFunction", "besskge.scoring.PairRE", "besskge.scoring.RotatE", "besskge.scoring.TranS", "besskge.scoring.TransE", "besskge.scoring.TripleRE", "besskge.sharding", "besskge.sharding.PartitionedTripleSet", "besskge.sharding.Sharding", "besskge.utils", "besskge.utils.complex_multiplication", "besskge.utils.complex_rotation", "besskge.utils.gather_indices", "besskge.utils.get_entity_filter", "BESS-KGE", "User guide"], "terms": {"when": [1, 3, 4, 11, 12, 13, 14, 23, 26, 27, 28, 29, 30, 37, 38, 41], "distribut": [1, 5, 9, 10, 11, 14, 18, 19, 20, 59, 67, 68], "workload": 1, "over": [1, 6, 7, 8, 14, 67], "n": [1, 41, 60], "worker": [1, 6, 7, 8, 67], "ipu": [1, 3, 4, 6, 7, 8, 9, 10, 44, 65, 67, 68], "randomli": [1, 16], "split": [1, 16, 41], "entiti": [1, 2, 6, 7, 8, 10, 11, 12, 13, 14, 16, 17, 22, 24, 30, 33, 34, 37, 38, 39, 40, 41, 42, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 61, 66, 67, 68], "embed": [1, 2, 11, 12, 13, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 67, 68], "tabl": [1, 11, 12, 13, 17, 22, 23, 24, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 68], "shard": [1, 6, 7, 8, 10, 12, 13, 14, 22, 24, 38, 39, 40, 41, 42, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 67], "equal": [1, 12], "size": [1, 6, 7, 8, 10, 13, 14, 41, 44, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58, 68], "each": [1, 5, 6, 7, 8, 10, 11, 12, 13, 14, 16, 20, 22, 23, 33, 39, 41, 44, 45, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 61, 66], "which": [1, 12, 16, 38, 39, 40, 41, 42, 66], "i": [1, 3, 4, 6, 7, 8, 12, 13, 14, 16, 22, 23, 34, 39, 41, 44, 47, 49, 60, 64, 65, 66, 67], "store": [1, 13, 15, 16, 44, 60, 67, 68], "": [1, 2, 3, 4], "memori": [1, 6, 7, 8, 11, 12, 13, 67], "The": [1, 3, 4, 6, 7, 8, 10, 11, 12, 13, 16, 24, 26, 27, 28, 29, 30, 33, 41, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 61, 66, 67, 68], "relat": [1, 2, 10, 11, 12, 13, 14, 16, 17, 23, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 67], "type": [1, 6, 7, 8, 10, 11, 12, 13, 14, 16, 18, 19, 20, 21, 22, 23, 24, 26, 27, 28, 29, 30, 33, 41, 42, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 61, 63, 64, 65, 66, 67], "other": [1, 6, 7, 8, 11, 12, 13], "hand": 1, "replic": [1, 13], "across": [1, 3, 4], "usual": 1, "much": 1, "smaller": 1, "figur": 1, "1": [1, 2, 10, 11, 12, 13, 14, 16, 18, 19, 20, 21, 27, 28, 29, 30, 33, 41, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 65, 68], "3": [1, 3, 4, 16, 49, 66, 68], "induc": 1, "partit": [1, 6, 7, 8, 10, 11, 14, 38, 39, 40, 41, 42, 44, 60], "tripl": [1, 2, 5, 6, 7, 8, 10, 11, 12, 13, 14, 15, 16, 23, 26, 27, 28, 29, 30, 33, 34, 38, 39, 40, 41, 42, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 66, 67], "dataset": [1, 2, 23, 49, 60, 67, 68], "accord": [1, 5, 11, 12, 13, 18, 19, 20, 21, 24, 33], "pair": [1, 2, 39, 60, 66], "head": [1, 2, 10, 11, 12, 13, 14, 16, 37, 39, 41, 42, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 66], "tail": [1, 2, 10, 11, 12, 13, 14, 16, 37, 39, 41, 42, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 66], "At": [1, 7, 14], "execut": [1, 59], "time": [1, 6, 7, 8, 67], "both": [1, 3, 4, 26, 46, 67], "train": [1, 9, 11, 12, 13, 16, 43, 44, 67, 68], "infer": [1, 2, 6, 7, 8, 9, 10, 11, 14, 43, 44, 67, 68], "batch": [1, 5, 6, 7, 8, 10, 11, 12, 13, 14, 25, 26, 27, 28, 29, 30, 33, 39, 41, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 63, 64, 67, 68], "ar": [1, 3, 4, 10, 12, 13, 14, 16, 23, 26, 33, 41, 47, 49, 60, 65, 67, 68], "construct": [1, 26, 27, 28, 29, 30, 37, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 61, 66], "sampl": [1, 2, 5, 6, 7, 8, 11, 12, 13, 14, 25, 26, 27, 28, 29, 30, 37, 38, 39, 40, 41, 42, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 67, 68], "uniformli": 1, "from": [1, 2, 3, 4, 6, 7, 8, 11, 12, 13, 16, 24, 33, 38, 39, 40, 41, 42, 49, 60, 65], "2": [1, 11, 12, 13, 42, 47, 49, 60, 63, 64, 65, 66, 68], "neg": [1, 5, 6, 7, 8, 10, 11, 12, 13, 14, 16, 25, 26, 27, 28, 29, 30, 37, 38, 39, 40, 41, 42, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 68], "us": [1, 2, 3, 4, 6, 7, 8, 10, 11, 13, 14, 16, 26, 27, 28, 29, 30, 33, 37, 38, 39, 41, 44, 49, 51, 60, 67, 68], "corrupt": [1, 6, 7, 8, 10, 11, 12, 13, 33, 37, 38, 39, 40, 41, 42, 49], "order": [1, 14, 33, 41, 44, 60, 68], "also": [1, 3, 4, 61, 68], "balanc": [1, 2, 61, 67], "wai": [1, 16], "ensur": 1, "varieti": 1, "benefici": 1, "final": [1, 3, 4, 14, 47, 49], "qualiti": [1, 16], "left": [1, 3, 4], "A": [1, 2, 33, 38, 60, 61], "made": 1, "up": [1, 3, 4, 68], "9": [1, 68], "block": [1, 10], "contain": [1, 3, 4, 16, 60], "same": [1, 7, 10, 14, 16, 22, 23, 41, 42, 60, 61, 65, 66], "number": [1, 6, 7, 8, 10, 11, 12, 13, 14, 16, 22, 23, 30, 39, 41, 44, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58, 60, 61, 68], "j": [1, 66], "0": [1, 3, 4, 6, 7, 8, 16, 18, 19, 21, 27, 28, 29, 30, 33, 38, 47, 49, 52, 56, 58, 60, 65, 68], "right": 1, "all": [1, 3, 4, 6, 7, 10, 13, 14, 16, 33, 38, 41, 42, 44, 49, 60, 65], "possibli": [1, 16, 41, 44, 49, 60], "pad": [1, 6, 7, 11, 12, 13, 14, 41, 61], "In": [1, 2, 3, 4], "thi": [1, 3, 4, 6, 7, 8, 12, 13, 14, 16, 22, 23, 44, 49, 60, 67], "exampl": [1, 13, 14, 36], "scheme": [1, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 21, 67], "allow": [1, 3, 4, 67], "u": [1, 58], "commun": [1, 67], "first": [1, 3, 4, 16, 39, 66], "need": [1, 16, 22, 23, 26, 27, 28, 29, 30, 44, 47, 49, 67], "gather": [1, 10, 11, 12, 13, 14, 41, 65, 67], "its": [1, 23], "chip": [1, 67], "posit": [1, 5, 6, 7, 8, 11, 12, 13, 19, 25, 26, 27, 28, 29, 30, 33, 45], "These": 1, "includ": [1, 49], "itself": 1, "peer": 1, "requir": [1, 3, 4, 11, 12, 13, 14, 33, 39, 65], "sram": [1, 68], "retriev": [1, 67], "triangl": 1, "colour": 1, "addit": [1, 3, 4, 23], "portion": 1, "can": [1, 3, 4, 12, 13, 14, 22, 23, 33, 44, 60, 67], "reconstruct": 1, "share": [1, 2, 6, 7, 8, 11, 12, 13, 26, 39, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 61, 67], "between": [1, 2, 12, 13, 47, 67], "through": [1, 3, 4, 11, 12, 13, 49], "alltoal": [1, 10, 12, 13, 14], "collect": [1, 6, 7, 8, 10, 11, 12, 13, 14, 15, 60, 67], "oper": [1, 3, 4, 11, 12, 13, 49, 67], "remain": [1, 61], "place": 1, "score": [1, 10, 11, 12, 13, 14, 25, 26, 27, 28, 29, 30, 33, 38, 41, 44, 60, 67], "where": [1, 3, 4, 10, 12, 13, 14, 16, 33, 34, 39, 41, 60], "4": [1, 2, 3, 4, 68], "exchang": 1, "an": [1, 3, 4, 10, 13, 14, 16, 22, 23, 44, 67, 68], "red": 1, "arrow": 1, "effect": [1, 38], "transpos": 1, "row": [1, 20, 33, 41, 63, 64, 65, 66], "column": [1, 16], "pictur": 1, "after": [1, 10, 44, 49, 61], "ha": [1, 6, 7, 8, 20, 68], "correct": [1, 13, 14], "comput": [1, 2, 10, 11, 12, 13, 14, 25, 26, 27, 28, 29, 30, 31, 33, 36, 44, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 67, 68], "present": [1, 16], "abov": [1, 3, 4], "implement": [1, 9], "besskg": [1, 3, 4, 67, 68], "embeddingmovingbesskg": [1, 13, 67], "while": [1, 14], "alwai": [1, 26], "turn": 1, "out": [1, 44], "expens": 1, "mani": 1, "per": [1, 6, 7, 8, 38, 39, 40, 41, 42], "dimens": [1, 33, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "larg": [1, 2, 10, 13, 14, 44], "case": 1, "scoremovingbesskg": [1, 10, 14, 67], "increas": 1, "overal": 1, "throughput": [1, 67], "altern": 1, "work": [1, 3, 4], "well": [1, 3, 4, 52, 56], "differ": [1, 10, 14, 22, 23, 24, 33, 49, 67], "li": [1, 2], "how": 1, "instead": [1, 38, 39, 40, 41, 42, 44, 47], "send": [1, 13], "queri": [1, 10, 13, 14, 33, 38, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60], "devic": [1, 5, 10, 11, 12, 13, 14, 38, 39, 40, 41, 42], "allgath": [1, 13], "against": [1, 10, 12, 13, 14, 38, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60], "partial": 1, "set": [1, 3, 4, 10, 11, 14, 16, 23, 33, 41, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 66, 68], "sent": 1, "via": [1, 2, 10, 14], "new": [1, 3, 4, 6, 7, 8, 24, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 67], "cheaper": 1, "although": 1, "5": 1, "correspond": [1, 13, 22, 23, 41, 65], "6": [1, 2, 3, 4, 47], "put": 1, "back": [1, 13, 61], "came": 1, "complet": [1, 2, 10, 14, 16, 44, 60], "respons": [1, 14], "acls20": [2, 47], "ralph": 2, "abboud": 2, "i\u0307smail": 2, "i\u0307lkan": 2, "ceylan": 2, "thoma": 2, "lukasiewicz": 2, "tommaso": 2, "salvatori": 2, "box": [2, 67], "model": [2, 6, 7, 8, 11, 31, 44, 45, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58, 67], "knowledg": [2, 10, 13, 14, 15, 16, 30, 38, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58, 60, 61, 67, 68], "base": [2, 6, 7, 8, 10, 11, 14, 16, 25, 26, 27, 28, 29, 30, 40, 41, 42, 44, 46, 47, 51, 53, 60], "advanc": 2, "neural": 2, "inform": 2, "process": [2, 5, 6, 7, 8, 38, 39, 40, 41, 42], "system": [2, 3, 4, 68], "33": 2, "annual": 2, "confer": 2, "2020": 2, "neurip": 2, "13": [2, 57], "asam20": [2, 16], "breit": 2, "anna": 2, "ott": 2, "simon": 2, "agibetov": 2, "asan": 2, "samwald": 2, "matthia": 2, "openbiolink": [2, 16], "benchmark": [2, 16], "framework": [2, 11, 67], "scale": [2, 21, 26, 27, 28, 29, 30], "biomed": 2, "link": [2, 3, 4, 33, 67, 68], "predict": [2, 14, 31, 33, 34, 36, 44, 68], "bioinformat": 2, "36": 2, "4097": 2, "4098": 2, "bugd": [2, 57], "antoin": 2, "bord": 2, "nicola": 2, "usuni": 2, "alberto": 2, "garc\u00eda": 2, "dur\u00e1n": 2, "jason": 2, "weston": 2, "oksana": 2, "yakhnenko": 2, "translat": 2, "multi": 2, "data": [2, 6, 7, 8, 10, 16], "26": 2, "27th": 2, "2013": 2, "2787": 2, "2795": 2, "cjm": [2, 9, 10, 11, 14, 30], "22": [2, 9, 10, 11, 14, 30, 52, 58], "cattaneo": 2, "daniel": 2, "justu": 2, "harri": 2, "mellor": 2, "dougla": 2, "orr": 2, "jerom": 2, "maloberti": 2, "zheni": 2, "liu": 2, "thorin": 2, "farnsworth": 2, "andrew": 2, "fitzgibbon": 2, "blazej": 2, "banaszewski": 2, "carlo": 2, "luschi": 2, "bess": [2, 5, 38, 41, 43, 68], "graph": [2, 10, 13, 14, 15, 16, 30, 38, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58, 60, 61, 67, 68], "arxiv": 2, "preprint": 2, "2211": 2, "12281": 2, "2022": 2, "chwc21": [2, 54], "linlin": 2, "chao": 2, "jianshan": 2, "he": 2, "taifeng": 2, "wang": 2, "wei": 2, "chu": 2, "pairr": [2, 67], "vector": 2, "proceed": 2, "59th": 2, "meet": 2, "associ": [2, 16], "linguist": 2, "11th": 2, "intern": [2, 26, 46], "joint": 2, "natur": 2, "languag": 2, "acl": 2, "ijcnlp": 2, "2021": 2, "volum": 2, "long": 2, "paper": 2, "virtual": 2, "event": 2, "august": 2, "4360": 2, "4369": 2, "dppr18": [2, 16, 49], "tim": 2, "dettmer": 2, "minervini": 2, "pasqual": 2, "stenetorp": 2, "pontu": 2, "sebastian": 2, "riedel": 2, "convolut": [2, 49], "2d": [2, 49], "32th": 2, "aaai": 2, "artifici": 2, "intellig": 2, "1811": 2, "1818": 2, "2018": 2, "hfz": [2, 16], "20": [2, 3, 4, 16, 68], "weihua": 2, "hu": 2, "fei": 2, "marinka": 2, "zitnik": 2, "yuxiao": 2, "dong": 2, "hongyu": 2, "ren": 2, "bowen": 2, "michel": 2, "catasta": 2, "jure": 2, "leskovec": 2, "open": [2, 3, 4, 67], "machin": [2, 67], "learn": [2, 23, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58], "jcmb15": [2, 30], "\u00e9": 2, "bastien": 2, "jean": 2, "kyunghyun": 2, "cho": 2, "roland": 2, "memisev": 2, "yoshua": 2, "bengio": 2, "On": 2, "veri": [2, 13], "target": 2, "vocabulari": 2, "53rd": 2, "7th": 2, "10": [2, 16, 44, 68], "2015": 2, "mbs15": [2, 16], "farzaneh": 2, "mahdisoltani": 2, "joanna": 2, "biega": 2, "fabian": 2, "m": [2, 3, 4, 68], "suchanek": 2, "yago3": [2, 16, 68], "multilingu": 2, "wikipedia": 2, "seventh": 2, "biennial": 2, "innov": 2, "research": [2, 3, 4, 68], "cidr": 2, "asilomar": 2, "ca": 2, "usa": 2, "januari": 2, "7": [2, 16, 68], "onlin": 2, "sdnt19": [2, 27, 55], "zhiqe": 2, "sun": 2, "zhi": 2, "hong": 2, "deng": 2, "jian": 2, "yun": 2, "nie": 2, "tang": 2, "rotat": [2, 64, 67], "complex": [2, 55, 63, 64, 67], "space": [2, 47], "represent": [2, 47], "iclr": 2, "2019": 2, "twr": [2, 48], "16": [2, 6, 7, 8, 48, 68], "th": 2, "o": [2, 6, 7, 8], "trouillon": 2, "johann": 2, "welbl": 2, "ric": 2, "gaussier": 2, "guillaum": 2, "bouchard": 2, "simpl": 2, "33rd": 2, "icml": 2, "2016": 2, "48": 2, "jmlr": 2, "workshop": 2, "2071": 2, "2080": 2, "wmw": [2, 52], "baoxin": 2, "qingy": 2, "meng": 2, "ziyu": 2, "honghong": 2, "zhao": 2, "dayong": 2, "wu": 2, "wanxiang": 2, "che": 2, "shijin": 2, "zhigang": 2, "chen": 2, "cong": 2, "interht": [2, 67], "interact": 2, "2202": 2, "04897": 2, "yyh": [2, 50], "15": [2, 16, 50], "bishan": 2, "yang": 2, "wen": 2, "tau": 2, "yih": 2, "xiaodong": 2, "jianfeng": 2, "gao": 2, "3rd": 2, "yll": [2, 58], "yu": 2, "zhicong": 2, "luo": 2, "huanyong": 2, "lin": 2, "hongzhu": 2, "yafeng": 2, "tripler": [2, 67], "2209": 2, "08271": 2, "zyx22": [2, 56], "xuanyu": 2, "zhang": 2, "qing": 2, "dongliang": 2, "xu": 2, "tran": [2, 67], "transit": 2, "synthet": 2, "find": [2, 66], "emnlp": 2, "1202": 2, "1208": 2, "you": [3, 4, 68], "even": [3, 4], "don": [3, 4], "t": [3, 4, 10, 12, 14, 15, 16, 38, 39, 40, 41, 42, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 66, 67], "have": [3, 4, 16, 44, 49, 67], "access": [3, 4, 67], "ipumodel": [3, 4], "emul": [3, 4], "most": [3, 4, 14, 34, 44], "function": [3, 4, 6, 7, 8, 10, 11, 12, 13, 14, 17, 22, 23, 25, 26, 27, 28, 29, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 62, 65, 68], "physic": [3, 4, 68], "hardwar": [3, 4, 44], "tunnel": [3, 4], "web": [3, 4], "editor": [3, 4], "desktop": [3, 4], "app": [3, 4], "minimum": [3, 4], "effort": [3, 4], "excel": [3, 4], "solut": [3, 4], "test": [3, 4, 16, 68], "directli": [3, 4], "here": [3, 4, 16, 44], "do": [3, 4], "fork": [3, 4], "repositori": [3, 4], "launch": [3, 4], "hour": [3, 4], "session": [3, 4], "free": [3, 4, 67, 68], "form": [3, 4], "http": [3, 4, 16, 68], "consol": [3, 4], "com": [3, 4, 16, 68], "github": [3, 4, 16, 67, 68], "userid": [3, 4], "reponam": [3, 4], "graphcor": [3, 4, 68], "2fpytorch": [3, 4], "3a3": [3, 4], "ubuntu": [3, 4, 68], "04": [3, 4, 68], "20230703": [3, 4], "pod4": [3, 4, 68], "repopnam": [3, 4], "address": [3, 4], "e": [3, 4, 16, 63, 64, 65, 66], "g": [3, 4], "origin": [3, 4], "repo": [3, 4, 67], "start": [3, 4, 33, 67], "clone": [3, 4], "termin": [3, 4], "pane": [3, 4], "run": [3, 4, 6, 7, 8, 44, 68], "command": [3, 4], "bash": [3, 4], "gradient": [3, 4, 68], "launch_vscode_serv": [3, 4], "sh": [3, 4, 68], "name": [3, 4, 10, 12, 14, 16, 44, 60], "option": [3, 4, 6, 7, 8, 10, 11, 12, 13, 14, 16, 22, 23, 33, 41, 44, 60, 61, 68], "argument": [3, 4, 49], "defin": [3, 4, 16, 47], "remot": [3, 4], "default": [3, 4, 6, 7, 8, 10, 11, 12, 13, 14, 18, 19, 21, 29, 33, 39, 41, 44, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58, 60, 61], "script": [3, 4], "download": [3, 4, 16], "instal": [3, 4, 67], "depend": [3, 4, 66, 68], "ask": [3, 4], "author": [3, 4], "account": [3, 4], "write": [3, 4], "privileg": [3, 4], "provid": [3, 4, 14, 22, 23, 33, 44], "pleas": [3, 4], "refer": [3, 4, 67], "notebook": [3, 4, 68], "detail": [3, 4, 68], "step": [3, 4, 10, 11, 12, 13, 14, 44], "connect": [3, 4], "onc": [3, 4, 22, 23, 67], "dev": [3, 4], "build": [3, 4, 15, 16], "custom": [3, 4], "op": [3, 4], "now": [3, 4], "readi": [3, 4], "close": [3, 4], "stop": [3, 4], "rememb": [3, 4], "unregist": [3, 4], "explain": [3, 4], "common": [3, 4], "issu": [3, 4, 67], "paragraph": [3, 4], "To": [3, 4, 10, 11, 14, 44], "resum": [3, 4], "your": [3, 4, 68], "just": [3, 4], "section": [3, 4], "profil": [3, 4], "repeat": [3, 4, 7], "chang": [3, 4, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "extens": [3, 4], "persist": [3, 4], "poplar": [3, 4, 68], "sdk": [3, 4, 68], "follow": [3, 4, 68], "instruct": [3, 4, 68], "get": [3, 4, 67], "guid": [3, 4, 67], "Then": [3, 4], "enabl": [3, 4, 68], "creat": [3, 4, 60, 61, 68], "activ": [3, 4, 29, 67, 68], "python": [3, 4, 67, 68], "virtualenv": [3, 4, 68], "poptorch": [3, 4, 6, 7, 8, 67, 68], "wheel": [3, 4, 68], "necessari": [3, 4], "python3": [3, 4, 68], "8": [3, 4, 68], "venv": [3, 4, 68], "add": [3, 4], "bin": [3, 4, 68], "sourc": [3, 4, 6, 7, 8, 10, 11, 12, 13, 14, 16, 18, 19, 20, 21, 22, 23, 24, 26, 27, 28, 29, 30, 32, 33, 34, 36, 38, 39, 40, 41, 42, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 61, 63, 64, 65, 66, 68], "path_to_poplar_sdk": [3, 4], "pip": [3, 4, 68], "poplar_sdk_en": [3, 4, 68], "whl": [3, 4, 68], "r": [3, 4, 10, 14, 15, 16, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 64, 66, 67], "txt": [3, 4], "sever": [3, 4], "util": [3, 4, 6, 7, 8, 15, 17, 31, 67], "dure": [3, 4, 68], "check": [3, 4], "help": [3, 4], "list": [3, 4, 6, 7, 8, 16, 22, 23, 33, 41, 44, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58], "befor": [3, 4, 11, 12, 13, 14, 47, 49, 52, 54, 56, 58], "submit": [3, 4], "pr": [3, 4], "upstream": [3, 4], "ci": [3, 4], "particular": [3, 4], "mind": [3, 4], "our": [3, 4, 68], "format": [3, 4], "error": [3, 4, 10, 14, 44, 68], "lint": [3, 4], "automat": [3, 4, 16], "insid": [3, 4, 47, 68], "unit": [3, 4], "folder": [3, 4], "individu": [3, 4], "pattern": [3, 4], "match": [3, 4], "filter": [3, 4, 11, 12, 13, 14, 44, 66], "k": [3, 4, 14, 33, 34, 35, 44, 64, 65], "cpp": [3, 4], "custom_op": [3, 4], "updat": [3, 4], "makefil": [3, 4], "ad": [3, 4, 23, 60, 67], "class": [5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61], "partitioned_triple_set": [6, 7, 8, 44], "negative_sampl": [6, 7, 8, 10, 11, 12, 13, 14, 67], "shard_b": [6, 7, 8, 10, 14], "batches_per_step": [6, 7, 8], "seed": [6, 7, 8, 16, 38, 39, 41, 42, 61], "hrt_freq_weight": [6, 7, 8], "fals": [6, 7, 8, 11, 12, 13, 14, 33, 39, 41, 44, 47, 48, 50, 52, 54, 55, 56, 57, 58, 60], "weight_smooth": [6, 7, 8], "duplicate_batch": [6, 7, 8], "return_triple_idx": [6, 7, 8], "random": [6, 16, 39, 61], "indic": [6, 7, 8, 10, 11, 12, 13, 14, 33, 41, 60, 65], "replac": 6, "No": [6, 38, 68], "appli": [6, 41, 44, 49, 52, 56, 61, 66], "initi": [6, 7, 8, 10, 11, 12, 13, 14, 17, 18, 19, 20, 21, 22, 23, 26, 27, 28, 29, 30, 33, 38, 39, 41, 42, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "sampler": [6, 7, 8, 10, 11, 12, 13, 14, 38, 39, 40, 41, 42, 44, 49], "paramet": [6, 7, 8, 10, 11, 12, 13, 14, 16, 18, 19, 20, 21, 22, 23, 24, 26, 27, 28, 29, 30, 33, 34, 36, 38, 39, 41, 42, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 61, 63, 64, 65, 66, 68], "partitionedtripleset": [6, 7, 8, 49, 67], "pre": [6, 7, 8, 16], "shardednegativesampl": [6, 7, 8, 11, 12, 13, 38, 67], "int": [6, 7, 8, 10, 11, 12, 13, 14, 16, 22, 23, 30, 34, 38, 39, 41, 42, 44, 47, 48, 49, 50, 51, 52, 54, 55, 56, 57, 58, 61], "micro": [6, 7, 8, 11, 12, 13, 14], "call": [6, 7, 8, 38], "rng": [6, 7, 8, 38, 39, 40, 41, 42], "bool": [6, 7, 8, 11, 12, 13, 14, 18, 19, 23, 26, 27, 28, 29, 30, 33, 38, 39, 40, 41, 42, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60], "If": [6, 7, 8, 11, 12, 13, 14, 16, 22, 23, 33, 39, 41, 44, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58, 60, 65, 68], "true": [6, 7, 8, 11, 12, 13, 14, 18, 19, 23, 33, 39, 41, 44, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58, 61], "frequenc": [6, 7, 8], "weight": [6, 7, 8, 11, 12, 13, 26, 27, 28, 29, 30, 68], "float": [6, 7, 8, 16, 18, 19, 21, 27, 28, 29, 30, 47, 49, 52, 56, 58], "smooth": [6, 7, 8], "two": [6, 7, 8, 47, 49, 66], "ident": [6, 7, 8], "halv": [6, 7, 8], "ht": [6, 7, 8, 12, 38, 39, 40, 41, 42], "return": [6, 7, 8, 10, 11, 12, 13, 14, 16, 18, 19, 20, 21, 22, 23, 24, 26, 27, 28, 29, 30, 33, 34, 36, 38, 41, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 61, 63, 64, 65, 66], "wrt": [6, 7, 8, 44], "get_dataload": [6, 7, 8], "shuffl": [6, 7, 8], "num_work": [6, 7, 8], "persistent_work": [6, 7, 8], "buffer_s": [6, 7, 8], "dataload": [6, 7, 8], "instanti": [6, 7, 8, 16], "appropri": [6, 7, 8], "iter": [6, 7, 8, 10, 14], "It": [6, 7, 8, 14, 44], "asynchron": [6, 7, 8], "load": [6, 7, 8, 14, 16, 61], "minim": [6, 7, 8, 67], "cpu": [6, 7, 8], "compil": [6, 7, 8, 68], "epoch": [6, 7, 8], "see": [6, 7, 8, 10, 11, 12, 13, 14, 27, 28, 29, 30, 33, 36, 38, 41, 42, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 67, 68], "torch": [6, 7, 8, 65], "__init__": [6, 7, 8, 29, 41, 42, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58], "ring": [6, 7, 8], "buffer": [6, 7, 8], "preload": [6, 7, 8], "get_dataloader_sampl": [6, 7, 8], "sample_tripl": [6, 7, 8], "idx": [6, 7, 8], "index": [6, 7, 8, 10, 41, 65, 66], "dict": [6, 7, 8, 11, 12, 13, 14, 16, 33, 44], "str": [6, 7, 8, 11, 12, 13, 14, 16, 29, 33, 35, 38, 39, 40, 41, 42, 44, 60, 66], "union": [6, 7, 8, 14, 16, 22, 23, 44, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58], "ndarrai": [6, 7, 8, 16, 41, 42, 44, 60, 61], "ani": [6, 7, 8, 11, 12, 13, 14, 16, 41, 42, 44, 60, 61, 67], "dtype": [6, 7, 8, 16, 41, 42, 44, 60, 61, 68], "int64": [6, 7, 8, 41, 60, 61], "bool_": [6, 7, 8, 41], "relev": [6, 7, 8, 11, 12, 13], "static": [6, 7, 8], "worker_init_fn": [6, 7, 8], "worker_id": [6, 7, 8], "pass": [6, 7, 8, 14, 22, 23, 24, 44, 47, 49], "id": [6, 7, 8, 16, 23, 33, 41, 42, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 61, 66], "none": [6, 7, 8, 10, 11, 12, 13, 14, 16, 22, 23, 33, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 61], "specifi": [7, 10, 16, 44, 65], "shorter": 7, "ones": [7, 23], "length": [7, 22, 23, 41], "mask": [7, 11, 12, 13, 14, 33, 41], "identifi": [7, 11, 12, 13], "abstract": [8, 11, 26, 28, 46, 51, 53], "pytorch": [9, 68], "modul": [9, 10, 11, 12, 13, 14, 26, 31, 33, 44, 46], "kge": [9, 10, 11, 12, 13, 14, 31, 45, 68], "multipl": [9, 63], "candidate_sampl": [10, 14], "score_fn": [10, 11, 12, 13, 14, 44], "window_s": [10, 14, 44], "1000": [10, 44], "h": [10, 12, 14, 15, 16, 38, 39, 40, 41, 42, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 66, 67], "host": [10, 11, 12, 13, 44], "combin": [10, 11, 14, 44, 49], "h_shard": [10, 14, 44, 60], "t_shard": [10, 14, 44, 60], "sinc": 10, "onli": [10, 14, 16, 38, 39, 40, 41, 42, 49, 60], "part": [10, 16, 60, 63, 64], "slide": [10, 14, 44], "window": [10, 14, 44], "metric": [10, 11, 12, 13, 14, 44, 67], "should": [10, 14, 16, 44, 49], "aggreg": 10, "pipelin": [10, 67], "allscorespipelin": [10, 67], "allscor": 10, "placeholdernegativesampl": [10, 14, 67], "basescorefunct": [10, 11, 12, 13, 14, 44, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 67], "decreas": [10, 14, 33, 44], "avoid": [10, 13, 14, 44], "oom": [10, 14, 44], "forward": [10, 11, 12, 13, 14, 26, 27, 28, 29, 30, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "similarli": [10, 14, 60], "candid": [10, 14, 33, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "thei": [10, 13, 14, 41], "togeth": [10, 14], "tensor": [10, 11, 12, 13, 14, 18, 19, 20, 21, 22, 23, 26, 27, 28, 29, 30, 33, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 63, 64, 65, 66], "self": [10, 26, 27, 28, 29, 30], "shape": [10, 11, 12, 13, 14, 16, 22, 24, 26, 27, 28, 29, 30, 33, 41, 42, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 61, 63, 64, 65, 66], "known": [10, 14, 60], "loss_fn": [11, 12, 13], "evalu": [11, 12, 13, 14, 31, 44, 67], "return_scor": [11, 12, 13, 14, 44], "augment_neg": [11, 12, 13], "ht_shardpair": [11, 60], "baselossfunct": [11, 12, 13, 27, 28, 29, 30, 67], "loss": [11, 12, 13, 49, 67], "augment": [11, 12, 13], "triple_mask": [11, 12, 13, 14, 33], "triple_weight": [11, 12, 13, 26, 27, 28, 29, 30], "negative_mask": [11, 12, 13, 14], "compris": [11, 12, 13], "four": [11, 12, 13], "phase": [11, 12, 13], "local": [11, 12, 13, 16, 60, 61, 67], "n_shard": [11, 12, 13, 14, 22, 41, 60, 61], "positive_per_partit": [11, 12, 13], "b": [11, 12, 13, 14, 19, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 65], "padded_neg": [11, 12, 13, 14, 41], "discard": [11, 12, 13, 14], "properti": [11, 12, 13, 16, 61], "n_embedding_paramet": [11, 12, 13], "trainabl": [11, 12, 13], "score_batch": [11, 12, 13], "tupl": [11, 12, 13, 16, 41, 66], "n_neg": [11, 12, 13, 26, 27, 28, 29, 30, 33, 39, 41, 42, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60], "move": [12, 13], "done": 12, "singl": [12, 33], "total": [12, 30], "disabl": 12, "otherwis": [12, 65], "conveni": 13, "so": [13, 20], "For": [13, 14, 44, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 65, 67, 68], "valu": 13, "document": 13, "multipli": [13, 52, 56], "doe": 13, "support": [13, 19, 33, 44, 68], "100": [14, 68], "specif": [14, 16, 41, 44, 45, 60], "top": [14, 33, 44], "like": [14, 33, 34, 44, 65], "input": [14, 49], "recommend": [14, 68], "one": [14, 23, 24, 41, 60], "want": [14, 16], "loop": 14, "topk": 14, "triplebasedshardednegativesampl": [14, 67], "unnecessari": 14, "best": 14, "respect": 14, "kept": 14, "next": 14, "rest": 14, "mask_on_gath": [14, 41], "n_entiti": [16, 22, 30, 61], "n_relation_typ": [16, 23, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58], "entity_dict": 16, "relation_dict": 16, "type_offset": [16, 61], "neg_head": [16, 60], "neg_tail": [16, 60], "repres": 16, "int32": [16, 41, 42, 44, 60, 61], "classmethod": [16, 60, 61], "build_ogbl_biokg": 16, "root": 16, "ogbl": [16, 68], "biokg": [16, 68], "ogb": 16, "stanford": 16, "edu": 16, "doc": 16, "linkprop": 16, "path": [16, 61, 68], "locat": 16, "build_ogbl_wikikg2": 16, "wikikg2": [16, 68], "build_openbiolink": 16, "high": [16, 43, 67], "version": 16, "openbiolink2020": 16, "hq": 16, "build_yago310": 16, "subgraph": 16, "least": 16, "them": [16, 41, 67, 68], "yago": 16, "org": 16, "label": 16, "from_datafram": 16, "df": 16, "head_column": 16, "relation_column": 16, "tail_column": 16, "entity_typ": 16, "1234": 16, "panda": 16, "datafram": 16, "assign": [16, 33, 60], "contigu": 16, "dictionari": [16, 33], "seri": 16, "map": [16, 61], "string": 16, "valid": 16, "instanc": 16, "from_tripl": 16, "arrai": 16, "alreadi": [16, 41], "been": [16, 49, 67, 68], "note": [16, 49], "manual": 16, "numpi": 16, "head_id": 16, "relation_id": [16, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "tail_id": 16, "num_tripl": 16, "offset": [16, 52, 56, 58, 61], "ht_type": 16, "n_tripl": [16, 41, 42, 60], "h_type": [16, 60], "t_type": [16, 60], "object": [16, 61], "save": [16, 61, 68], "node": 16, "edg": 16, "n_neg_head": [16, 60], "n_neg_tail": [16, 60], "out_fil": [16, 61], "pkl": 16, "output": [16, 33, 49, 61], "file": [16, 61, 68], "h_id": 16, "r_id": 16, "t_id": 16, "assum": [16, 33, 49, 68], "cluster": [16, 41, 61], "manag": 17, "embedding_t": [18, 19, 20, 21], "std": 18, "divide_by_embedding_s": [18, 19], "normal": [18, 20, 21, 47, 49, 52, 54, 56, 58], "mean": 18, "standard": [18, 21], "deviat": [18, 21], "rescal": [18, 19], "row_siz": [18, 19, 21, 22, 23, 24], "symmetr": 19, "uniform": [19, 20], "boundari": 19, "norm": [20, 47, 51, 52, 54, 55, 56, 57, 58], "gain": 21, "xavier": 21, "fan_in": 21, "fan_out": 21, "factor": [21, 26, 27, 28, 29, 30, 58], "callabl": [22, 23, 35, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58], "either": 22, "max_entity_per_shard": [22, 61], "unshard": 22, "alloc": [22, 23], "entri": [22, 23], "omit": [22, 23], "max_ent_per_shard": 22, "inverse_rel": [23, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58], "invers": [23, 47, 48, 49, 50, 52, 54, 55, 56, 57, 58, 60], "direct": 23, "given": [23, 44, 66, 68], "entity_embed": [24, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "old_shard": 24, "new_shard": [24, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "refactor": 24, "n_shard_old": 24, "max_ent_per_shard_old": 24, "current": [24, 33], "n_shard_new": 24, "max_ent_per_shard_new": 24, "arg": [26, 46], "kwarg": [26, 46], "fp32": [26, 68], "state": [26, 46], "nn": [26, 46], "scriptmodul": [26, 46], "positive_scor": [26, 27, 28, 29, 30], "negative_scor": [26, 27, 28, 29, 30], "batch_siz": [26, 27, 28, 29, 30, 33, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "get_negative_weight": [26, 27, 28, 29, 30], "negative_adversarial_sampl": [26, 27, 28, 29, 30], "els": [26, 27, 28, 29, 30, 33, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "loss_scal": [26, 27, 28, 29, 30], "might": [26, 27, 28, 29, 30], "fp16": [26, 27, 28, 29, 30, 68], "adversari": [26, 27, 28, 29, 30], "negative_adversarial_scal": [26, 27, 28, 29, 30], "reciproc": [26, 27, 28, 29, 30, 36], "temperatur": [26, 27, 28, 29, 30], "margin": [27, 28, 29], "log": 27, "sigmoid": [27, 49], "activation_funct": 29, "relu": 29, "rank": [29, 33, 34, 36], "pairwis": 29, "hing": 29, "marginbasedlossfunct": [29, 67], "softmax": 30, "cross": 30, "entropi": 30, "attribut": 31, "metric_list": 33, "mode": [33, 60], "averag": 33, "worst_rank_infti": 33, "reduct": [33, 36, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "return_rank": 33, "mrr": [33, 35, 36], "hit": [33, 34, 35], "optimist": 33, "pessimist": 33, "infin": 33, "worst": 33, "possibl": 33, "method": 33, "reduc": [33, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "along": [33, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 65], "sum": [33, 47, 48, 49, 50, 53], "alongsid": 33, "dict_metrics_from_rank": 33, "batch_rank": 33, "element": [33, 65], "boolean": 33, "ranks_from_indic": 33, "ground_truth": [33, 60], "candidate_indic": 33, "ground": [33, 34, 36, 60], "truth": [33, 34, 36, 60], "n_candid": 33, "likelihood": 33, "distinct": 33, "among": [33, 34, 36], "ranks_from_scor": 33, "pos_scor": 33, "candidate_scor": 33, "stacked_metrics_from_rank": 33, "stack": 33, "n_metric": 33, "count": [34, 68], "maximum": [34, 68], "accept": 34, "hitsatk": [35, 67], "reciprocalrank": [35, 67], "basemetr": [36, 67], "corruption_schem": [38, 39, 40, 41, 42, 44], "placehold": 38, "topkquerybesskg": [38, 41, 67], "flat_negative_format": [38, 39, 40, 41, 42], "local_sampl": [38, 39, 40, 41, 42], "gener": [38, 39, 40, 41, 42, 62], "half": 39, "second": [39, 66, 68], "negative_head": 41, "negative_tail": 41, "return_sort_idx": 41, "predetermin": 41, "global": [41, 44, 60, 61, 66], "randomshardednegativesampl": [41, 42, 67], "sort": [41, 60], "recov": 41, "pad_neg": 41, "shard_count": [41, 61], "padded_shard_length": 41, "divid": 41, "view": 41, "shard_neg": 41, "shard_neg_count": 41, "sort_neg_idx": 41, "triple_typ": 42, "level": 43, "api": [43, 67, 68], "batch_sampl": [44, 67], "filter_tripl": [44, 66], "return_topk": 44, "use_ipu_model": 44, "kg": 44, "appear": [44, 66], "shardedbatchsampl": [44, 67], "whose": 44, "must": 44, "caus": 44, "go": 44, "actual": 44, "result": 44, "head_emb": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "tail_emb": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "score_tripl": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "negative_sample_shar": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "relation_embed": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "score_head": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "fix": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "n_head": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "embedding_s": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "broadcast": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "score_tail": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "n_tail": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "update_shard": [46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "scoring_norm": [47, 51, 52, 54, 55, 56, 57, 58], "entity_initi": [47, 48, 49, 50, 52, 54, 55, 56, 57, 58], "uniform_": 47, "relation_initi": [47, 48, 49, 50, 52, 54, 55, 56, 57, 58], "init_uniform_norm": [47, 67], "apply_tanh": 47, "dist_func_per_dim": 47, "ep": 47, "1e": 47, "06": 47, "distancebasedscorefunct": [47, 48, 49, 50, 52, 54, 55, 56, 57, 58, 67], "center": 47, "scalar": [47, 49], "bound": [47, 67], "bump": 47, "tanh": 47, "select": 47, "distanc": [47, 51, 52, 54, 55, 56, 57, 58], "whether": [47, 60], "outsid": 47, "make": [47, 60], "choic": 47, "separ": 47, "soften": 47, "geometr": 47, "width": [47, 49], "boxe_scor": 47, "bumped_ht": 47, "center_ht": 47, "width_ht": 47, "box_siz": 47, "optim": [47, 68], "emb_siz": 47, "control": 47, "broadcasted_dist": [47, 51, 52, 54, 55, 56, 57, 58], "v1": [47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 63], "v2": [47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 63], "p": [47, 51, 52, 54, 55, 56, 57, 58], "reduce_embed": [47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "v": [47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 64, 67], "init_kge_norm": [48, 67], "broadcasted_dot_product": [48, 49, 50, 53], "dot": [48, 49, 50, 53], "product": [48, 49, 50, 53], "embedding_height": 49, "embedding_width": 49, "init_xavier_norm": [49, 67], "zeros_": 49, "input_channel": 49, "output_channel": 49, "32": 49, "kernel_height": 49, "kernel_width": 49, "input_dropout": 49, "feature_map_dropout": 49, "hidden_dropout": 49, "batch_norm": 49, "layer": 49, "we": [49, 68], "By": 49, "design": [49, 67], "add_inverse_tripl": [49, 60], "create_from_dataset": [49, 60], "height": 49, "reshap": 49, "concaten": 49, "bias": 49, "channel": 49, "conv2d": 49, "kernel": 49, "rate": 49, "dropout": 49, "linear": 49, "init_kge_uniform": [50, 52, 54, 55, 56, 57, 58, 67], "normalize_ent": [52, 54, 56, 58], "l2": [52, 54, 56, 58], "auxiliari": 52, "matrix": 53, "decomposit": 53, "project": [54, 58], "real": [55, 63, 64], "tild": 56, "triplerev2": 58, "inverse_tripl": 60, "partition_mod": 60, "dummi": 60, "triple_count": 60, "triple_offset": 60, "triple_sort_idx": 60, "shard_h": 60, "shard_t": 60, "kgdataset": [60, 67], "create_from_queri": 60, "query_mod": 60, "negative_typ": 60, "n_queri": 60, "hr": 60, "rt": 60, "resp": 60, "r_inv": 60, "regular": 60, "criterion": 60, "delimit": 60, "entity_to_shard": 61, "entity_to_idx": 61, "shard_and_idx_to_ent": 61, "entity_type_count": 61, "entity_type_offset": 61, "again": 61, "n_type": 61, "npz": 61, "local_id": 61, "exclud": 61, "purpos": 62, "imaginari": [63, 64], "wise": [63, 64], "unitari": 64, "pi": 64, "x": [65, 66], "friendli": 65, "take_along_dim": 65, "dimension": 65, "dim": 65, "take": 65, "filter_mod": 66, "compar": 66, "y": 66, "determin": 66, "look": 66, "z": 66, "spars": 66, "packag": 67, "shallow": 67, "typic": 67, "littl": 67, "perform": 67, "maxim": 67, "bandwidth": 67, "fast": 67, "leverag": 67, "achiev": 67, "introduct": 67, "overview": 67, "librari": [67, 68], "still": 67, "develop": 67, "featur": 67, "expect": 67, "overtim": 67, "occasion": 67, "bug": 67, "mai": 67, "occur": 67, "feel": 67, "report": 67, "problem": 67, "user": 67, "usag": 67, "limit": 67, "randomshardedbatchsampl": 67, "rigidshardedbatchsampl": 67, "typebasedshardednegativesampl": 67, "allscoresbess": 67, "conv": 67, "distmult": 67, "matrixdecompositionscorefunct": 67, "trans": 67, "logsigmoidloss": 67, "marginrankingloss": 67, "sampledsoftmaxcrossentropyloss": 67, "metrics_dict": 67, "initialize_entity_embed": 67, "initialize_relation_embed": 67, "refactor_embedding_shard": 67, "complex_multipl": 67, "complex_rot": 67, "gather_indic": 67, "get_entity_filt": 67, "code": 67, "server": 67, "paperspac": [67, 68], "setup": 67, "tip": 67, "bibliographi": 67, "popart": 68, "more": 68, "quick": 68, "git": 68, "import": 68, "1403": 68, "walkthrough": 68, "main": 68, "jupyt": 68, "sequenc": 68, "click": 68, "button": 68, "avail": 68, "introduc": 68, "therefor": 68, "some": 68, "approxim": 68, "estim": 68, "below": 68, "accumul": 68, "momentum": 68, "notic": 68, "cap": 68, "max": 68, "pod16": 68, "float16": 68, "sgdm": 68, "2m": 68, "2e8": 68, "13m": 68, "3e9": 68, "128": 68, "adam": 68, "4m": 68, "0e8": 68, "9m": 68, "256": 68, "ye": 68, "900k": 68, "3e8": 68, "5m": 68, "8m": 68, "2e9": 68, "512": 68, "375k": 68, "9e8": 68, "7e8": 68, "messag": 68, "about": 68, "onnx": 68, "protobuff": 68, "exceed": 68, "_popart": 68, "saveinitializerstofil": 68, "my_fil": 68}, "objects": {"": [[67, 0, 0, "-", "besskge"]], "besskge": [[5, 0, 0, "-", "batch_sampler"], [9, 0, 0, "-", "bess"], [15, 0, 0, "-", "dataset"], [17, 0, 0, "-", "embedding"], [25, 0, 0, "-", "loss"], [31, 0, 0, "-", "metric"], [37, 0, 0, "-", "negative_sampler"], [43, 0, 0, "-", "pipeline"], [45, 0, 0, "-", "scoring"], [59, 0, 0, "-", "sharding"], [62, 0, 0, "-", "utils"]], "besskge.batch_sampler": [[6, 1, 1, "", "RandomShardedBatchSampler"], [7, 1, 1, "", "RigidShardedBatchSampler"], [8, 1, 1, "", "ShardedBatchSampler"]], "besskge.batch_sampler.RandomShardedBatchSampler": [[6, 2, 1, "", "get_dataloader"], [6, 2, 1, "", "get_dataloader_sampler"], [6, 2, 1, "", "sample_triples"], [6, 2, 1, "", "worker_init_fn"]], "besskge.batch_sampler.RigidShardedBatchSampler": [[7, 2, 1, "", "get_dataloader"], [7, 2, 1, "", "get_dataloader_sampler"], [7, 2, 1, "", "sample_triples"], [7, 2, 1, "", "worker_init_fn"]], "besskge.batch_sampler.ShardedBatchSampler": [[8, 2, 1, "", "get_dataloader"], [8, 2, 1, "", "get_dataloader_sampler"], [8, 2, 1, "", "sample_triples"], [8, 2, 1, "", "worker_init_fn"]], "besskge.bess": [[10, 1, 1, "", "AllScoresBESS"], [11, 1, 1, "", "BessKGE"], [12, 1, 1, "", "EmbeddingMovingBessKGE"], [13, 1, 1, "", "ScoreMovingBessKGE"], [14, 1, 1, "", "TopKQueryBessKGE"]], "besskge.bess.AllScoresBESS": [[10, 2, 1, "", "forward"]], "besskge.bess.BessKGE": [[11, 2, 1, "", "forward"], [11, 3, 1, "", "n_embedding_parameters"], [11, 2, 1, "", "score_batch"]], "besskge.bess.EmbeddingMovingBessKGE": [[12, 2, 1, "", "forward"], [12, 3, 1, "", "n_embedding_parameters"], [12, 2, 1, "", "score_batch"]], "besskge.bess.ScoreMovingBessKGE": [[13, 2, 1, "", "forward"], [13, 3, 1, "", "n_embedding_parameters"], [13, 2, 1, "", "score_batch"]], "besskge.bess.TopKQueryBessKGE": [[14, 2, 1, "", "forward"]], "besskge.dataset": [[16, 1, 1, "", "KGDataset"]], "besskge.dataset.KGDataset": [[16, 2, 1, "", "build_ogbl_biokg"], [16, 2, 1, "", "build_ogbl_wikikg2"], [16, 2, 1, "", "build_openbiolink"], [16, 2, 1, "", "build_yago310"], [16, 4, 1, "", "entity_dict"], [16, 2, 1, "", "from_dataframe"], [16, 2, 1, "", "from_triples"], [16, 3, 1, "", "ht_types"], [16, 2, 1, "", "load"], [16, 4, 1, "", "n_entity"], [16, 4, 1, "", "n_relation_type"], [16, 4, 1, "", "neg_heads"], [16, 4, 1, "", "neg_tails"], [16, 4, 1, "", "relation_dict"], [16, 2, 1, "", "save"], [16, 4, 1, "", "triples"], [16, 4, 1, "", "type_offsets"]], "besskge.embedding": [[18, 5, 1, "", "init_KGE_normal"], [19, 5, 1, "", "init_KGE_uniform"], [20, 5, 1, "", "init_uniform_norm"], [21, 5, 1, "", "init_xavier_norm"], [22, 5, 1, "", "initialize_entity_embedding"], [23, 5, 1, "", "initialize_relation_embedding"], [24, 5, 1, "", "refactor_embedding_sharding"]], "besskge.loss": [[26, 1, 1, "", "BaseLossFunction"], [27, 1, 1, "", "LogSigmoidLoss"], [28, 1, 1, "", "MarginBasedLossFunction"], [29, 1, 1, "", "MarginRankingLoss"], [30, 1, 1, "", "SampledSoftmaxCrossEntropyLoss"]], "besskge.loss.BaseLossFunction": [[26, 2, 1, "", "forward"], [26, 2, 1, "", "get_negative_weights"], [26, 4, 1, "", "loss_scale"], [26, 4, 1, "", "negative_adversarial_sampling"], [26, 4, 1, "", "negative_adversarial_scale"]], "besskge.loss.LogSigmoidLoss": [[27, 2, 1, "", "forward"], [27, 2, 1, "", "get_negative_weights"], [27, 4, 1, "", "loss_scale"], [27, 4, 1, "", "negative_adversarial_sampling"], [27, 4, 1, "", "negative_adversarial_scale"]], "besskge.loss.MarginBasedLossFunction": [[28, 2, 1, "", "forward"], [28, 2, 1, "", "get_negative_weights"], [28, 4, 1, "", "loss_scale"], [28, 4, 1, "", "negative_adversarial_sampling"], [28, 4, 1, "", "negative_adversarial_scale"]], "besskge.loss.MarginRankingLoss": [[29, 2, 1, "", "forward"], [29, 2, 1, "", "get_negative_weights"], [29, 4, 1, "", "loss_scale"], [29, 4, 1, "", "negative_adversarial_sampling"], [29, 4, 1, "", "negative_adversarial_scale"]], "besskge.loss.SampledSoftmaxCrossEntropyLoss": [[30, 2, 1, "", "forward"], [30, 2, 1, "", "get_negative_weights"], [30, 4, 1, "", "loss_scale"], [30, 4, 1, "", "negative_adversarial_sampling"], [30, 4, 1, "", "negative_adversarial_scale"]], "besskge.metric": [[32, 1, 1, "", "BaseMetric"], [33, 1, 1, "", "Evaluation"], [34, 1, 1, "", "HitsAtK"], [35, 6, 1, "", "METRICS_DICT"], [36, 1, 1, "", "ReciprocalRank"]], "besskge.metric.Evaluation": [[33, 2, 1, "", "dict_metrics_from_ranks"], [33, 2, 1, "", "ranks_from_indices"], [33, 2, 1, "", "ranks_from_scores"], [33, 2, 1, "", "stacked_metrics_from_ranks"]], "besskge.negative_sampler": [[38, 1, 1, "", "PlaceholderNegativeSampler"], [39, 1, 1, "", "RandomShardedNegativeSampler"], [40, 1, 1, "", "ShardedNegativeSampler"], [41, 1, 1, "", "TripleBasedShardedNegativeSampler"], [42, 1, 1, "", "TypeBasedShardedNegativeSampler"]], "besskge.negative_sampler.PlaceholderNegativeSampler": [[38, 4, 1, "", "corruption_scheme"], [38, 4, 1, "", "flat_negative_format"], [38, 4, 1, "", "local_sampling"], [38, 4, 1, "", "rng"]], "besskge.negative_sampler.RandomShardedNegativeSampler": [[39, 4, 1, "", "corruption_scheme"], [39, 4, 1, "", "flat_negative_format"], [39, 4, 1, "", "local_sampling"], [39, 4, 1, "", "rng"]], "besskge.negative_sampler.ShardedNegativeSampler": [[40, 4, 1, "", "corruption_scheme"], [40, 4, 1, "", "flat_negative_format"], [40, 4, 1, "", "local_sampling"], [40, 4, 1, "", "rng"]], "besskge.negative_sampler.TripleBasedShardedNegativeSampler": [[41, 4, 1, "", "corruption_scheme"], [41, 4, 1, "", "flat_negative_format"], [41, 4, 1, "", "local_sampling"], [41, 2, 1, "", "pad_negatives"], [41, 4, 1, "", "rng"], [41, 2, 1, "", "shard_negatives"]], "besskge.negative_sampler.TypeBasedShardedNegativeSampler": [[42, 4, 1, "", "corruption_scheme"], [42, 4, 1, "", "flat_negative_format"], [42, 4, 1, "", "local_sampling"], [42, 4, 1, "", "rng"]], "besskge.pipeline": [[44, 1, 1, "", "AllScoresPipeline"]], "besskge.pipeline.AllScoresPipeline": [[44, 2, 1, "", "forward"]], "besskge.scoring": [[46, 1, 1, "", "BaseScoreFunction"], [47, 1, 1, "", "BoxE"], [48, 1, 1, "", "ComplEx"], [49, 1, 1, "", "ConvE"], [50, 1, 1, "", "DistMult"], [51, 1, 1, "", "DistanceBasedScoreFunction"], [52, 1, 1, "", "InterHT"], [53, 1, 1, "", "MatrixDecompositionScoreFunction"], [54, 1, 1, "", "PairRE"], [55, 1, 1, "", "RotatE"], [56, 1, 1, "", "TranS"], [57, 1, 1, "", "TransE"], [58, 1, 1, "", "TripleRE"]], "besskge.scoring.BaseScoreFunction": [[46, 4, 1, "", "entity_embedding"], [46, 2, 1, "", "forward"], [46, 4, 1, "", "negative_sample_sharing"], [46, 4, 1, "", "relation_embedding"], [46, 2, 1, "", "score_heads"], [46, 2, 1, "", "score_tails"], [46, 2, 1, "", "score_triple"], [46, 4, 1, "", "sharding"], [46, 2, 1, "", "update_sharding"]], "besskge.scoring.BoxE": [[47, 2, 1, "", "boxe_score"], [47, 2, 1, "", "broadcasted_distance"], [47, 4, 1, "", "entity_embedding"], [47, 2, 1, "", "forward"], [47, 4, 1, "", "negative_sample_sharing"], [47, 2, 1, "", "reduce_embedding"], [47, 4, 1, "", "relation_embedding"], [47, 2, 1, "", "score_heads"], [47, 2, 1, "", "score_tails"], [47, 2, 1, "", "score_triple"], [47, 4, 1, "", "sharding"], [47, 2, 1, "", "update_sharding"]], "besskge.scoring.ComplEx": [[48, 2, 1, "", "broadcasted_dot_product"], [48, 4, 1, "", "entity_embedding"], [48, 2, 1, "", "forward"], [48, 4, 1, "", "negative_sample_sharing"], [48, 2, 1, "", "reduce_embedding"], [48, 4, 1, "", "relation_embedding"], [48, 2, 1, "", "score_heads"], [48, 2, 1, "", "score_tails"], [48, 2, 1, "", "score_triple"], [48, 4, 1, "", "sharding"], [48, 2, 1, "", "update_sharding"]], "besskge.scoring.ConvE": [[49, 2, 1, "", "broadcasted_dot_product"], [49, 4, 1, "", "entity_embedding"], [49, 2, 1, "", "forward"], [49, 4, 1, "", "negative_sample_sharing"], [49, 2, 1, "", "reduce_embedding"], [49, 4, 1, "", "relation_embedding"], [49, 2, 1, "", "score_heads"], [49, 2, 1, "", "score_tails"], [49, 2, 1, "", "score_triple"], [49, 4, 1, "", "sharding"], [49, 2, 1, "", "update_sharding"]], "besskge.scoring.DistMult": [[50, 2, 1, "", "broadcasted_dot_product"], [50, 4, 1, "", "entity_embedding"], [50, 2, 1, "", "forward"], [50, 4, 1, "", "negative_sample_sharing"], [50, 2, 1, "", "reduce_embedding"], [50, 4, 1, "", "relation_embedding"], [50, 2, 1, "", "score_heads"], [50, 2, 1, "", "score_tails"], [50, 2, 1, "", "score_triple"], [50, 4, 1, "", "sharding"], [50, 2, 1, "", "update_sharding"]], "besskge.scoring.DistanceBasedScoreFunction": [[51, 2, 1, "", "broadcasted_distance"], [51, 4, 1, "", "entity_embedding"], [51, 2, 1, "", "forward"], [51, 4, 1, "", "negative_sample_sharing"], [51, 2, 1, "", "reduce_embedding"], [51, 4, 1, "", "relation_embedding"], [51, 2, 1, "", "score_heads"], [51, 2, 1, "", "score_tails"], [51, 2, 1, "", "score_triple"], [51, 4, 1, "", "sharding"], [51, 2, 1, "", "update_sharding"]], "besskge.scoring.InterHT": [[52, 2, 1, "", "broadcasted_distance"], [52, 4, 1, "", "entity_embedding"], [52, 2, 1, "", "forward"], [52, 4, 1, "", "negative_sample_sharing"], [52, 2, 1, "", "reduce_embedding"], [52, 4, 1, "", "relation_embedding"], [52, 2, 1, "", "score_heads"], [52, 2, 1, "", "score_tails"], [52, 2, 1, "", "score_triple"], [52, 4, 1, "", "sharding"], [52, 2, 1, "", "update_sharding"]], "besskge.scoring.MatrixDecompositionScoreFunction": [[53, 2, 1, "", "broadcasted_dot_product"], [53, 4, 1, "", "entity_embedding"], [53, 2, 1, "", "forward"], [53, 4, 1, "", "negative_sample_sharing"], [53, 2, 1, "", "reduce_embedding"], [53, 4, 1, "", "relation_embedding"], [53, 2, 1, "", "score_heads"], [53, 2, 1, "", "score_tails"], [53, 2, 1, "", "score_triple"], [53, 4, 1, "", "sharding"], [53, 2, 1, "", "update_sharding"]], "besskge.scoring.PairRE": [[54, 2, 1, "", "broadcasted_distance"], [54, 4, 1, "", "entity_embedding"], [54, 2, 1, "", "forward"], [54, 4, 1, "", "negative_sample_sharing"], [54, 2, 1, "", "reduce_embedding"], [54, 4, 1, "", "relation_embedding"], [54, 2, 1, "", "score_heads"], [54, 2, 1, "", "score_tails"], [54, 2, 1, "", "score_triple"], [54, 4, 1, "", "sharding"], [54, 2, 1, "", "update_sharding"]], "besskge.scoring.RotatE": [[55, 2, 1, "", "broadcasted_distance"], [55, 4, 1, "", "entity_embedding"], [55, 2, 1, "", "forward"], [55, 4, 1, "", "negative_sample_sharing"], [55, 2, 1, "", "reduce_embedding"], [55, 4, 1, "", "relation_embedding"], [55, 2, 1, "", "score_heads"], [55, 2, 1, "", "score_tails"], [55, 2, 1, "", "score_triple"], [55, 4, 1, "", "sharding"], [55, 2, 1, "", "update_sharding"]], "besskge.scoring.TranS": [[56, 2, 1, "", "broadcasted_distance"], [56, 4, 1, "", "entity_embedding"], [56, 2, 1, "", "forward"], [56, 4, 1, "", "negative_sample_sharing"], [56, 2, 1, "", "reduce_embedding"], [56, 4, 1, "", "relation_embedding"], [56, 2, 1, "", "score_heads"], [56, 2, 1, "", "score_tails"], [56, 2, 1, "", "score_triple"], [56, 4, 1, "", "sharding"], [56, 2, 1, "", "update_sharding"]], "besskge.scoring.TransE": [[57, 2, 1, "", "broadcasted_distance"], [57, 4, 1, "", "entity_embedding"], [57, 2, 1, "", "forward"], [57, 4, 1, "", "negative_sample_sharing"], [57, 2, 1, "", "reduce_embedding"], [57, 4, 1, "", "relation_embedding"], [57, 2, 1, "", "score_heads"], [57, 2, 1, "", "score_tails"], [57, 2, 1, "", "score_triple"], [57, 4, 1, "", "sharding"], [57, 2, 1, "", "update_sharding"]], "besskge.scoring.TripleRE": [[58, 2, 1, "", "broadcasted_distance"], [58, 4, 1, "", "entity_embedding"], [58, 2, 1, "", "forward"], [58, 4, 1, "", "negative_sample_sharing"], [58, 2, 1, "", "reduce_embedding"], [58, 4, 1, "", "relation_embedding"], [58, 2, 1, "", "score_heads"], [58, 2, 1, "", "score_tails"], [58, 2, 1, "", "score_triple"], [58, 4, 1, "", "sharding"], [58, 2, 1, "", "update_sharding"]], "besskge.sharding": [[60, 1, 1, "", "PartitionedTripleSet"], [61, 1, 1, "", "Sharding"]], "besskge.sharding.PartitionedTripleSet": [[60, 2, 1, "", "create_from_dataset"], [60, 2, 1, "", "create_from_queries"], [60, 4, 1, "", "dummy"], [60, 4, 1, "", "inverse_triples"], [60, 4, 1, "", "neg_heads"], [60, 4, 1, "", "neg_tails"], [60, 4, 1, "", "partition_mode"], [60, 4, 1, "", "sharding"], [60, 4, 1, "", "triple_counts"], [60, 4, 1, "", "triple_offsets"], [60, 4, 1, "", "triple_sort_idx"], [60, 4, 1, "", "triples"], [60, 4, 1, "", "types"]], "besskge.sharding.Sharding": [[61, 2, 1, "", "create"], [61, 4, 1, "", "entity_to_idx"], [61, 4, 1, "", "entity_to_shard"], [61, 4, 1, "", "entity_type_counts"], [61, 4, 1, "", "entity_type_offsets"], [61, 2, 1, "", "load"], [61, 3, 1, "", "max_entity_per_shard"], [61, 3, 1, "", "n_entity"], [61, 4, 1, "", "n_shard"], [61, 2, 1, "", "save"], [61, 4, 1, "", "shard_and_idx_to_entity"], [61, 4, 1, "", "shard_counts"]], "besskge.utils": [[63, 5, 1, "", "complex_multiplication"], [64, 5, 1, "", "complex_rotation"], [65, 5, 1, "", "gather_indices"], [66, 5, 1, "", "get_entity_filter"]]}, "objtypes": {"0": "py:module", "1": "py:class", "2": "py:method", "3": "py:property", "4": "py:attribute", "5": "py:function", "6": "py:data"}, "objnames": {"0": ["py", "module", "Python module"], "1": ["py", "class", "Python class"], "2": ["py", "method", "Python method"], "3": ["py", "property", "Python property"], "4": ["py", "attribute", "Python attribute"], "5": ["py", "function", "Python function"], "6": ["py", "data", "Python data"]}, "titleterms": {"bess": [0, 1, 3, 4, 9, 10, 11, 12, 13, 14, 67], "kge": [0, 3, 4, 67], "api": 0, "refer": 0, "overview": 1, "bibliographi": 2, "how": [3, 4], "contribut": [3, 4], "project": [3, 4], "v": [3, 4], "code": [3, 4], "server": [3, 4], "paperspac": [3, 4], "setup": [3, 4], "local": [3, 4], "machin": [3, 4], "develop": [3, 4], "tip": [3, 4], "besskg": [5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66], "batch_sampl": [5, 6, 7, 8], "randomshardedbatchsampl": 6, "rigidshardedbatchsampl": 7, "shardedbatchsampl": 8, "allscoresbess": 10, "embeddingmovingbesskg": 12, "scoremovingbesskg": 13, "topkquerybesskg": 14, "dataset": [15, 16], "kgdataset": 16, "embed": [17, 18, 19, 20, 21, 22, 23, 24], "init_kge_norm": 18, "init_kge_uniform": 19, "init_uniform_norm": 20, "init_xavier_norm": 21, "initialize_entity_embed": 22, "initialize_relation_embed": 23, "refactor_embedding_shard": 24, "loss": [25, 26, 27, 28, 29, 30], "baselossfunct": 26, "logsigmoidloss": 27, "marginbasedlossfunct": 28, "marginrankingloss": 29, "sampledsoftmaxcrossentropyloss": 30, "metric": [31, 32, 33, 34, 35, 36], "basemetr": 32, "evalu": 33, "hitsatk": 34, "metrics_dict": 35, "reciprocalrank": 36, "negative_sampl": [37, 38, 39, 40, 41, 42], "placeholdernegativesampl": 38, "randomshardednegativesampl": 39, "shardednegativesampl": 40, "triplebasedshardednegativesampl": 41, "typebasedshardednegativesampl": 42, "pipelin": [43, 44], "allscorespipelin": 44, "score": [45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58], "basescorefunct": 46, "box": 47, "complex": 48, "conv": 49, "distmult": 50, "distancebasedscorefunct": 51, "interht": 52, "matrixdecompositionscorefunct": 53, "pairr": 54, "rotat": 55, "tran": 56, "trans": 57, "tripler": 58, "shard": [59, 60, 61], "partitionedtripleset": 60, "util": [62, 63, 64, 65, 66], "complex_multipl": 63, "complex_rot": 64, "gather_indic": 65, "get_entity_filt": 66, "content": 67, "user": 68, "guid": 68, "instal": 68, "usag": 68, "get": 68, "start": 68, "limit": 68}, "envversion": {"sphinx.domains.c": 2, "sphinx.domains.changeset": 1, "sphinx.domains.citation": 1, "sphinx.domains.cpp": 8, "sphinx.domains.index": 1, "sphinx.domains.javascript": 2, "sphinx.domains.math": 2, "sphinx.domains.python": 3, "sphinx.domains.rst": 2, "sphinx.domains.std": 2, "sphinx.ext.todo": 2, "sphinx.ext.viewcode": 1, "sphinx.ext.intersphinx": 1, "sphinxcontrib.bibtex": 9, "sphinx": 57}, "alltitles": {"BESS-KGE API Reference": [[0, "bess-kge-api-reference"]], "BESS overview": [[1, "bess-overview"]], "Bibliography": [[2, "bibliography"]], "How to contribute to the BESS-KGE project": [[3, "how-to-contribute-to-the-bess-kge-project"], [4, "how-to-contribute-to-the-bess-kge-project"]], "VS Code server on Paperspace": [[3, "vs-code-server-on-paperspace"], [4, "vs-code-server-on-paperspace"]], "Setup on local machine": [[3, "setup-on-local-machine"], [4, "setup-on-local-machine"]], "Development tips": [[3, "development-tips"], [4, "development-tips"]], "besskge.batch_sampler": [[5, "module-besskge.batch_sampler"]], "besskge.batch_sampler.RandomShardedBatchSampler": [[6, "besskge-batch-sampler-randomshardedbatchsampler"]], "besskge.batch_sampler.RigidShardedBatchSampler": [[7, "besskge-batch-sampler-rigidshardedbatchsampler"]], "besskge.batch_sampler.ShardedBatchSampler": [[8, "besskge-batch-sampler-shardedbatchsampler"]], "besskge.bess": [[9, "module-besskge.bess"]], "besskge.bess.AllScoresBESS": [[10, "besskge-bess-allscoresbess"]], "besskge.bess.BessKGE": [[11, "besskge-bess-besskge"]], "besskge.bess.EmbeddingMovingBessKGE": [[12, "besskge-bess-embeddingmovingbesskge"]], "besskge.bess.ScoreMovingBessKGE": [[13, "besskge-bess-scoremovingbesskge"]], "besskge.bess.TopKQueryBessKGE": [[14, "besskge-bess-topkquerybesskge"]], "besskge.dataset": [[15, "module-besskge.dataset"]], "besskge.dataset.KGDataset": [[16, "besskge-dataset-kgdataset"]], "besskge.embedding": [[17, "module-besskge.embedding"]], "besskge.embedding.init_KGE_normal": [[18, "besskge-embedding-init-kge-normal"]], "besskge.embedding.init_KGE_uniform": [[19, "besskge-embedding-init-kge-uniform"]], "besskge.embedding.init_uniform_norm": [[20, "besskge-embedding-init-uniform-norm"]], "besskge.embedding.init_xavier_norm": [[21, "besskge-embedding-init-xavier-norm"]], "besskge.embedding.initialize_entity_embedding": [[22, "besskge-embedding-initialize-entity-embedding"]], "besskge.embedding.initialize_relation_embedding": [[23, "besskge-embedding-initialize-relation-embedding"]], "besskge.embedding.refactor_embedding_sharding": [[24, "besskge-embedding-refactor-embedding-sharding"]], "besskge.loss": [[25, "module-besskge.loss"]], "besskge.loss.BaseLossFunction": [[26, "besskge-loss-baselossfunction"]], "besskge.loss.LogSigmoidLoss": [[27, "besskge-loss-logsigmoidloss"]], "besskge.loss.MarginBasedLossFunction": [[28, "besskge-loss-marginbasedlossfunction"]], "besskge.loss.MarginRankingLoss": [[29, "besskge-loss-marginrankingloss"]], "besskge.loss.SampledSoftmaxCrossEntropyLoss": [[30, "besskge-loss-sampledsoftmaxcrossentropyloss"]], "besskge.metric": [[31, "module-besskge.metric"]], "besskge.metric.BaseMetric": [[32, "besskge-metric-basemetric"]], "besskge.metric.Evaluation": [[33, "besskge-metric-evaluation"]], "besskge.metric.HitsAtK": [[34, "besskge-metric-hitsatk"]], "besskge.metric.METRICS_DICT": [[35, "besskge-metric-metrics-dict"]], "besskge.metric.ReciprocalRank": [[36, "besskge-metric-reciprocalrank"]], "besskge.negative_sampler": [[37, "module-besskge.negative_sampler"]], "besskge.negative_sampler.PlaceholderNegativeSampler": [[38, "besskge-negative-sampler-placeholdernegativesampler"]], "besskge.negative_sampler.RandomShardedNegativeSampler": [[39, "besskge-negative-sampler-randomshardednegativesampler"]], "besskge.negative_sampler.ShardedNegativeSampler": [[40, "besskge-negative-sampler-shardednegativesampler"]], "besskge.negative_sampler.TripleBasedShardedNegativeSampler": [[41, "besskge-negative-sampler-triplebasedshardednegativesampler"]], "besskge.negative_sampler.TypeBasedShardedNegativeSampler": [[42, "besskge-negative-sampler-typebasedshardednegativesampler"]], "besskge.pipeline": [[43, "module-besskge.pipeline"]], "besskge.pipeline.AllScoresPipeline": [[44, "besskge-pipeline-allscorespipeline"]], "besskge.scoring": [[45, "module-besskge.scoring"]], "besskge.scoring.BaseScoreFunction": [[46, "besskge-scoring-basescorefunction"]], "besskge.scoring.BoxE": [[47, "besskge-scoring-boxe"]], "besskge.scoring.ComplEx": [[48, "besskge-scoring-complex"]], "besskge.scoring.ConvE": [[49, "besskge-scoring-conve"]], "besskge.scoring.DistMult": [[50, "besskge-scoring-distmult"]], "besskge.scoring.DistanceBasedScoreFunction": [[51, "besskge-scoring-distancebasedscorefunction"]], "besskge.scoring.InterHT": [[52, "besskge-scoring-interht"]], "besskge.scoring.MatrixDecompositionScoreFunction": [[53, "besskge-scoring-matrixdecompositionscorefunction"]], "besskge.scoring.PairRE": [[54, "besskge-scoring-pairre"]], "besskge.scoring.RotatE": [[55, "besskge-scoring-rotate"]], "besskge.scoring.TranS": [[56, "besskge-scoring-trans"]], "besskge.scoring.TransE": [[57, "besskge-scoring-transe"]], "besskge.scoring.TripleRE": [[58, "besskge-scoring-triplere"]], "besskge.sharding": [[59, "module-besskge.sharding"]], "besskge.sharding.PartitionedTripleSet": [[60, "besskge-sharding-partitionedtripleset"]], "besskge.sharding.Sharding": [[61, "besskge-sharding-sharding"]], "besskge.utils": [[62, "module-besskge.utils"]], "besskge.utils.complex_multiplication": [[63, "besskge-utils-complex-multiplication"]], "besskge.utils.complex_rotation": [[64, "besskge-utils-complex-rotation"]], "besskge.utils.gather_indices": [[65, "besskge-utils-gather-indices"]], "besskge.utils.get_entity_filter": [[66, "besskge-utils-get-entity-filter"]], "BESS-KGE": [[67, "module-besskge"]], "Contents": [[67, null]], "User guide": [[68, "user-guide"]], "Installation and usage": [[68, "installation-and-usage"]], "Getting started": [[68, "getting-started"]], "Limitations": [[68, "limitations"]]}, "indexentries": {"besskge.batch_sampler": [[5, "module-besskge.batch_sampler"]], "module": [[5, "module-besskge.batch_sampler"], [9, "module-besskge.bess"], [15, "module-besskge.dataset"], [17, "module-besskge.embedding"], [25, "module-besskge.loss"], [31, "module-besskge.metric"], [37, "module-besskge.negative_sampler"], [43, "module-besskge.pipeline"], [45, "module-besskge.scoring"], [59, "module-besskge.sharding"], [62, "module-besskge.utils"], [67, "module-besskge"]], "randomshardedbatchsampler (class in besskge.batch_sampler)": [[6, "besskge.batch_sampler.RandomShardedBatchSampler"]], "get_dataloader() (besskge.batch_sampler.randomshardedbatchsampler method)": [[6, "besskge.batch_sampler.RandomShardedBatchSampler.get_dataloader"]], "get_dataloader_sampler() (besskge.batch_sampler.randomshardedbatchsampler method)": [[6, "besskge.batch_sampler.RandomShardedBatchSampler.get_dataloader_sampler"]], "sample_triples() (besskge.batch_sampler.randomshardedbatchsampler method)": [[6, "besskge.batch_sampler.RandomShardedBatchSampler.sample_triples"]], "worker_init_fn() (besskge.batch_sampler.randomshardedbatchsampler static method)": [[6, "besskge.batch_sampler.RandomShardedBatchSampler.worker_init_fn"]], "rigidshardedbatchsampler (class in besskge.batch_sampler)": [[7, "besskge.batch_sampler.RigidShardedBatchSampler"]], "get_dataloader() (besskge.batch_sampler.rigidshardedbatchsampler method)": [[7, "besskge.batch_sampler.RigidShardedBatchSampler.get_dataloader"]], "get_dataloader_sampler() (besskge.batch_sampler.rigidshardedbatchsampler method)": [[7, "besskge.batch_sampler.RigidShardedBatchSampler.get_dataloader_sampler"]], "sample_triples() (besskge.batch_sampler.rigidshardedbatchsampler method)": [[7, "besskge.batch_sampler.RigidShardedBatchSampler.sample_triples"]], "worker_init_fn() (besskge.batch_sampler.rigidshardedbatchsampler static method)": [[7, "besskge.batch_sampler.RigidShardedBatchSampler.worker_init_fn"]], "shardedbatchsampler (class in besskge.batch_sampler)": [[8, "besskge.batch_sampler.ShardedBatchSampler"]], "get_dataloader() (besskge.batch_sampler.shardedbatchsampler method)": [[8, "besskge.batch_sampler.ShardedBatchSampler.get_dataloader"]], "get_dataloader_sampler() (besskge.batch_sampler.shardedbatchsampler method)": [[8, "besskge.batch_sampler.ShardedBatchSampler.get_dataloader_sampler"]], "sample_triples() (besskge.batch_sampler.shardedbatchsampler method)": [[8, "besskge.batch_sampler.ShardedBatchSampler.sample_triples"]], "worker_init_fn() (besskge.batch_sampler.shardedbatchsampler static method)": [[8, "besskge.batch_sampler.ShardedBatchSampler.worker_init_fn"]], "besskge.bess": [[9, "module-besskge.bess"]], "allscoresbess (class in besskge.bess)": [[10, "besskge.bess.AllScoresBESS"]], "forward() (besskge.bess.allscoresbess method)": [[10, "besskge.bess.AllScoresBESS.forward"]], "besskge (class in besskge.bess)": [[11, "besskge.bess.BessKGE"]], "forward() (besskge.bess.besskge method)": [[11, "besskge.bess.BessKGE.forward"]], "n_embedding_parameters (besskge.bess.besskge property)": [[11, "besskge.bess.BessKGE.n_embedding_parameters"]], "score_batch() (besskge.bess.besskge method)": [[11, "besskge.bess.BessKGE.score_batch"]], "embeddingmovingbesskge (class in besskge.bess)": [[12, "besskge.bess.EmbeddingMovingBessKGE"]], "forward() (besskge.bess.embeddingmovingbesskge method)": [[12, "besskge.bess.EmbeddingMovingBessKGE.forward"]], "n_embedding_parameters (besskge.bess.embeddingmovingbesskge property)": [[12, "besskge.bess.EmbeddingMovingBessKGE.n_embedding_parameters"]], "score_batch() (besskge.bess.embeddingmovingbesskge method)": [[12, "besskge.bess.EmbeddingMovingBessKGE.score_batch"]], "scoremovingbesskge (class in besskge.bess)": [[13, "besskge.bess.ScoreMovingBessKGE"]], "forward() (besskge.bess.scoremovingbesskge method)": [[13, "besskge.bess.ScoreMovingBessKGE.forward"]], "n_embedding_parameters (besskge.bess.scoremovingbesskge property)": [[13, "besskge.bess.ScoreMovingBessKGE.n_embedding_parameters"]], "score_batch() (besskge.bess.scoremovingbesskge method)": [[13, "besskge.bess.ScoreMovingBessKGE.score_batch"]], "topkquerybesskge (class in besskge.bess)": [[14, "besskge.bess.TopKQueryBessKGE"]], "forward() (besskge.bess.topkquerybesskge method)": [[14, "besskge.bess.TopKQueryBessKGE.forward"]], "besskge.dataset": [[15, "module-besskge.dataset"]], "kgdataset (class in besskge.dataset)": [[16, "besskge.dataset.KGDataset"]], "build_ogbl_biokg() (besskge.dataset.kgdataset class method)": [[16, "besskge.dataset.KGDataset.build_ogbl_biokg"]], "build_ogbl_wikikg2() (besskge.dataset.kgdataset class method)": [[16, "besskge.dataset.KGDataset.build_ogbl_wikikg2"]], "build_openbiolink() (besskge.dataset.kgdataset class method)": [[16, "besskge.dataset.KGDataset.build_openbiolink"]], "build_yago310() (besskge.dataset.kgdataset class method)": [[16, "besskge.dataset.KGDataset.build_yago310"]], "entity_dict (besskge.dataset.kgdataset attribute)": [[16, "besskge.dataset.KGDataset.entity_dict"]], "from_dataframe() (besskge.dataset.kgdataset class method)": [[16, "besskge.dataset.KGDataset.from_dataframe"]], "from_triples() (besskge.dataset.kgdataset class method)": [[16, "besskge.dataset.KGDataset.from_triples"]], "ht_types (besskge.dataset.kgdataset property)": [[16, "besskge.dataset.KGDataset.ht_types"]], "load() (besskge.dataset.kgdataset class method)": [[16, "besskge.dataset.KGDataset.load"]], "n_entity (besskge.dataset.kgdataset attribute)": [[16, "besskge.dataset.KGDataset.n_entity"]], "n_relation_type (besskge.dataset.kgdataset attribute)": [[16, "besskge.dataset.KGDataset.n_relation_type"]], "neg_heads (besskge.dataset.kgdataset attribute)": [[16, "besskge.dataset.KGDataset.neg_heads"]], "neg_tails (besskge.dataset.kgdataset attribute)": [[16, "besskge.dataset.KGDataset.neg_tails"]], "relation_dict (besskge.dataset.kgdataset attribute)": [[16, "besskge.dataset.KGDataset.relation_dict"]], "save() (besskge.dataset.kgdataset method)": [[16, "besskge.dataset.KGDataset.save"]], "triples (besskge.dataset.kgdataset attribute)": [[16, "besskge.dataset.KGDataset.triples"]], "type_offsets (besskge.dataset.kgdataset attribute)": [[16, "besskge.dataset.KGDataset.type_offsets"]], "besskge.embedding": [[17, "module-besskge.embedding"]], "init_kge_normal() (in module besskge.embedding)": [[18, "besskge.embedding.init_KGE_normal"]], "init_kge_uniform() (in module besskge.embedding)": [[19, "besskge.embedding.init_KGE_uniform"]], "init_uniform_norm() (in module besskge.embedding)": [[20, "besskge.embedding.init_uniform_norm"]], "init_xavier_norm() (in module besskge.embedding)": [[21, "besskge.embedding.init_xavier_norm"]], "initialize_entity_embedding() (in module besskge.embedding)": [[22, "besskge.embedding.initialize_entity_embedding"]], "initialize_relation_embedding() (in module besskge.embedding)": [[23, "besskge.embedding.initialize_relation_embedding"]], "refactor_embedding_sharding() (in module besskge.embedding)": [[24, "besskge.embedding.refactor_embedding_sharding"]], "besskge.loss": [[25, "module-besskge.loss"]], "baselossfunction (class in besskge.loss)": [[26, "besskge.loss.BaseLossFunction"]], "forward() (besskge.loss.baselossfunction method)": [[26, "besskge.loss.BaseLossFunction.forward"]], "get_negative_weights() (besskge.loss.baselossfunction method)": [[26, "besskge.loss.BaseLossFunction.get_negative_weights"]], "loss_scale (besskge.loss.baselossfunction attribute)": [[26, "besskge.loss.BaseLossFunction.loss_scale"]], "negative_adversarial_sampling (besskge.loss.baselossfunction attribute)": [[26, "besskge.loss.BaseLossFunction.negative_adversarial_sampling"]], "negative_adversarial_scale (besskge.loss.baselossfunction attribute)": [[26, "besskge.loss.BaseLossFunction.negative_adversarial_scale"]], "logsigmoidloss (class in besskge.loss)": [[27, "besskge.loss.LogSigmoidLoss"]], "forward() (besskge.loss.logsigmoidloss method)": [[27, "besskge.loss.LogSigmoidLoss.forward"]], "get_negative_weights() (besskge.loss.logsigmoidloss method)": [[27, "besskge.loss.LogSigmoidLoss.get_negative_weights"]], "loss_scale (besskge.loss.logsigmoidloss attribute)": [[27, "besskge.loss.LogSigmoidLoss.loss_scale"]], "negative_adversarial_sampling (besskge.loss.logsigmoidloss attribute)": [[27, "besskge.loss.LogSigmoidLoss.negative_adversarial_sampling"]], "negative_adversarial_scale (besskge.loss.logsigmoidloss attribute)": [[27, "besskge.loss.LogSigmoidLoss.negative_adversarial_scale"]], "marginbasedlossfunction (class in besskge.loss)": [[28, "besskge.loss.MarginBasedLossFunction"]], "forward() (besskge.loss.marginbasedlossfunction method)": [[28, "besskge.loss.MarginBasedLossFunction.forward"]], "get_negative_weights() (besskge.loss.marginbasedlossfunction method)": [[28, "besskge.loss.MarginBasedLossFunction.get_negative_weights"]], "loss_scale (besskge.loss.marginbasedlossfunction attribute)": [[28, "besskge.loss.MarginBasedLossFunction.loss_scale"]], "negative_adversarial_sampling (besskge.loss.marginbasedlossfunction attribute)": [[28, "besskge.loss.MarginBasedLossFunction.negative_adversarial_sampling"]], "negative_adversarial_scale (besskge.loss.marginbasedlossfunction attribute)": [[28, "besskge.loss.MarginBasedLossFunction.negative_adversarial_scale"]], "marginrankingloss (class in besskge.loss)": [[29, "besskge.loss.MarginRankingLoss"]], "forward() (besskge.loss.marginrankingloss method)": [[29, "besskge.loss.MarginRankingLoss.forward"]], "get_negative_weights() (besskge.loss.marginrankingloss method)": [[29, "besskge.loss.MarginRankingLoss.get_negative_weights"]], "loss_scale (besskge.loss.marginrankingloss attribute)": [[29, "besskge.loss.MarginRankingLoss.loss_scale"]], "negative_adversarial_sampling (besskge.loss.marginrankingloss attribute)": [[29, "besskge.loss.MarginRankingLoss.negative_adversarial_sampling"]], "negative_adversarial_scale (besskge.loss.marginrankingloss attribute)": [[29, "besskge.loss.MarginRankingLoss.negative_adversarial_scale"]], "sampledsoftmaxcrossentropyloss (class in besskge.loss)": [[30, "besskge.loss.SampledSoftmaxCrossEntropyLoss"]], "forward() (besskge.loss.sampledsoftmaxcrossentropyloss method)": [[30, "besskge.loss.SampledSoftmaxCrossEntropyLoss.forward"]], "get_negative_weights() (besskge.loss.sampledsoftmaxcrossentropyloss method)": [[30, "besskge.loss.SampledSoftmaxCrossEntropyLoss.get_negative_weights"]], "loss_scale (besskge.loss.sampledsoftmaxcrossentropyloss attribute)": [[30, "besskge.loss.SampledSoftmaxCrossEntropyLoss.loss_scale"]], "negative_adversarial_sampling (besskge.loss.sampledsoftmaxcrossentropyloss attribute)": [[30, "besskge.loss.SampledSoftmaxCrossEntropyLoss.negative_adversarial_sampling"]], "negative_adversarial_scale (besskge.loss.sampledsoftmaxcrossentropyloss attribute)": [[30, "besskge.loss.SampledSoftmaxCrossEntropyLoss.negative_adversarial_scale"]], "besskge.metric": [[31, "module-besskge.metric"]], "basemetric (class in besskge.metric)": [[32, "besskge.metric.BaseMetric"]], "evaluation (class in besskge.metric)": [[33, "besskge.metric.Evaluation"]], "dict_metrics_from_ranks() (besskge.metric.evaluation method)": [[33, "besskge.metric.Evaluation.dict_metrics_from_ranks"]], "ranks_from_indices() (besskge.metric.evaluation method)": [[33, "besskge.metric.Evaluation.ranks_from_indices"]], "ranks_from_scores() (besskge.metric.evaluation method)": [[33, "besskge.metric.Evaluation.ranks_from_scores"]], "stacked_metrics_from_ranks() (besskge.metric.evaluation method)": [[33, "besskge.metric.Evaluation.stacked_metrics_from_ranks"]], "hitsatk (class in besskge.metric)": [[34, "besskge.metric.HitsAtK"]], "metrics_dict (in module besskge.metric)": [[35, "besskge.metric.METRICS_DICT"]], "reciprocalrank (class in besskge.metric)": [[36, "besskge.metric.ReciprocalRank"]], "besskge.negative_sampler": [[37, "module-besskge.negative_sampler"]], "placeholdernegativesampler (class in besskge.negative_sampler)": [[38, "besskge.negative_sampler.PlaceholderNegativeSampler"]], "corruption_scheme (besskge.negative_sampler.placeholdernegativesampler attribute)": [[38, "besskge.negative_sampler.PlaceholderNegativeSampler.corruption_scheme"]], "flat_negative_format (besskge.negative_sampler.placeholdernegativesampler attribute)": [[38, "besskge.negative_sampler.PlaceholderNegativeSampler.flat_negative_format"]], "local_sampling (besskge.negative_sampler.placeholdernegativesampler attribute)": [[38, "besskge.negative_sampler.PlaceholderNegativeSampler.local_sampling"]], "rng (besskge.negative_sampler.placeholdernegativesampler attribute)": [[38, "besskge.negative_sampler.PlaceholderNegativeSampler.rng"]], "randomshardednegativesampler (class in besskge.negative_sampler)": [[39, "besskge.negative_sampler.RandomShardedNegativeSampler"]], "corruption_scheme (besskge.negative_sampler.randomshardednegativesampler attribute)": [[39, "besskge.negative_sampler.RandomShardedNegativeSampler.corruption_scheme"]], "flat_negative_format (besskge.negative_sampler.randomshardednegativesampler attribute)": [[39, "besskge.negative_sampler.RandomShardedNegativeSampler.flat_negative_format"]], "local_sampling (besskge.negative_sampler.randomshardednegativesampler attribute)": [[39, "besskge.negative_sampler.RandomShardedNegativeSampler.local_sampling"]], "rng (besskge.negative_sampler.randomshardednegativesampler attribute)": [[39, "besskge.negative_sampler.RandomShardedNegativeSampler.rng"]], "shardednegativesampler (class in besskge.negative_sampler)": [[40, "besskge.negative_sampler.ShardedNegativeSampler"]], "corruption_scheme (besskge.negative_sampler.shardednegativesampler attribute)": [[40, "besskge.negative_sampler.ShardedNegativeSampler.corruption_scheme"]], "flat_negative_format (besskge.negative_sampler.shardednegativesampler attribute)": [[40, "besskge.negative_sampler.ShardedNegativeSampler.flat_negative_format"]], "local_sampling (besskge.negative_sampler.shardednegativesampler attribute)": [[40, "besskge.negative_sampler.ShardedNegativeSampler.local_sampling"]], "rng (besskge.negative_sampler.shardednegativesampler attribute)": [[40, "besskge.negative_sampler.ShardedNegativeSampler.rng"]], "triplebasedshardednegativesampler (class in besskge.negative_sampler)": [[41, "besskge.negative_sampler.TripleBasedShardedNegativeSampler"]], "corruption_scheme (besskge.negative_sampler.triplebasedshardednegativesampler attribute)": [[41, "besskge.negative_sampler.TripleBasedShardedNegativeSampler.corruption_scheme"]], "flat_negative_format (besskge.negative_sampler.triplebasedshardednegativesampler attribute)": [[41, "besskge.negative_sampler.TripleBasedShardedNegativeSampler.flat_negative_format"]], "local_sampling (besskge.negative_sampler.triplebasedshardednegativesampler attribute)": [[41, "besskge.negative_sampler.TripleBasedShardedNegativeSampler.local_sampling"]], "pad_negatives() (besskge.negative_sampler.triplebasedshardednegativesampler method)": [[41, "besskge.negative_sampler.TripleBasedShardedNegativeSampler.pad_negatives"]], "rng (besskge.negative_sampler.triplebasedshardednegativesampler attribute)": [[41, "besskge.negative_sampler.TripleBasedShardedNegativeSampler.rng"]], "shard_negatives() (besskge.negative_sampler.triplebasedshardednegativesampler method)": [[41, "besskge.negative_sampler.TripleBasedShardedNegativeSampler.shard_negatives"]], "typebasedshardednegativesampler (class in besskge.negative_sampler)": [[42, "besskge.negative_sampler.TypeBasedShardedNegativeSampler"]], "corruption_scheme (besskge.negative_sampler.typebasedshardednegativesampler attribute)": [[42, "besskge.negative_sampler.TypeBasedShardedNegativeSampler.corruption_scheme"]], "flat_negative_format (besskge.negative_sampler.typebasedshardednegativesampler attribute)": [[42, "besskge.negative_sampler.TypeBasedShardedNegativeSampler.flat_negative_format"]], "local_sampling (besskge.negative_sampler.typebasedshardednegativesampler attribute)": [[42, "besskge.negative_sampler.TypeBasedShardedNegativeSampler.local_sampling"]], "rng (besskge.negative_sampler.typebasedshardednegativesampler attribute)": [[42, "besskge.negative_sampler.TypeBasedShardedNegativeSampler.rng"]], "besskge.pipeline": [[43, "module-besskge.pipeline"]], "allscorespipeline (class in besskge.pipeline)": [[44, "besskge.pipeline.AllScoresPipeline"]], "forward() (besskge.pipeline.allscorespipeline method)": [[44, "besskge.pipeline.AllScoresPipeline.forward"]], "besskge.scoring": [[45, "module-besskge.scoring"]], "basescorefunction (class in besskge.scoring)": [[46, "besskge.scoring.BaseScoreFunction"]], "entity_embedding (besskge.scoring.basescorefunction attribute)": [[46, "besskge.scoring.BaseScoreFunction.entity_embedding"]], "forward() (besskge.scoring.basescorefunction method)": [[46, "besskge.scoring.BaseScoreFunction.forward"]], "negative_sample_sharing (besskge.scoring.basescorefunction attribute)": [[46, "besskge.scoring.BaseScoreFunction.negative_sample_sharing"]], "relation_embedding (besskge.scoring.basescorefunction attribute)": [[46, "besskge.scoring.BaseScoreFunction.relation_embedding"]], "score_heads() (besskge.scoring.basescorefunction method)": [[46, "besskge.scoring.BaseScoreFunction.score_heads"]], "score_tails() (besskge.scoring.basescorefunction method)": [[46, "besskge.scoring.BaseScoreFunction.score_tails"]], "score_triple() (besskge.scoring.basescorefunction method)": [[46, "besskge.scoring.BaseScoreFunction.score_triple"]], "sharding (besskge.scoring.basescorefunction attribute)": [[46, "besskge.scoring.BaseScoreFunction.sharding"]], "update_sharding() (besskge.scoring.basescorefunction method)": [[46, "besskge.scoring.BaseScoreFunction.update_sharding"]], "boxe (class in besskge.scoring)": [[47, "besskge.scoring.BoxE"]], "boxe_score() (besskge.scoring.boxe method)": [[47, "besskge.scoring.BoxE.boxe_score"]], "broadcasted_distance() (besskge.scoring.boxe method)": [[47, "besskge.scoring.BoxE.broadcasted_distance"]], "entity_embedding (besskge.scoring.boxe attribute)": [[47, "besskge.scoring.BoxE.entity_embedding"]], "forward() (besskge.scoring.boxe method)": [[47, "besskge.scoring.BoxE.forward"]], "negative_sample_sharing (besskge.scoring.boxe attribute)": [[47, "besskge.scoring.BoxE.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.boxe method)": [[47, "besskge.scoring.BoxE.reduce_embedding"]], "relation_embedding (besskge.scoring.boxe attribute)": [[47, "besskge.scoring.BoxE.relation_embedding"]], "score_heads() (besskge.scoring.boxe method)": [[47, "besskge.scoring.BoxE.score_heads"]], "score_tails() (besskge.scoring.boxe method)": [[47, "besskge.scoring.BoxE.score_tails"]], "score_triple() (besskge.scoring.boxe method)": [[47, "besskge.scoring.BoxE.score_triple"]], "sharding (besskge.scoring.boxe attribute)": [[47, "besskge.scoring.BoxE.sharding"]], "update_sharding() (besskge.scoring.boxe method)": [[47, "besskge.scoring.BoxE.update_sharding"]], "complex (class in besskge.scoring)": [[48, "besskge.scoring.ComplEx"]], "broadcasted_dot_product() (besskge.scoring.complex method)": [[48, "besskge.scoring.ComplEx.broadcasted_dot_product"]], "entity_embedding (besskge.scoring.complex attribute)": [[48, "besskge.scoring.ComplEx.entity_embedding"]], "forward() (besskge.scoring.complex method)": [[48, "besskge.scoring.ComplEx.forward"]], "negative_sample_sharing (besskge.scoring.complex attribute)": [[48, "besskge.scoring.ComplEx.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.complex method)": [[48, "besskge.scoring.ComplEx.reduce_embedding"]], "relation_embedding (besskge.scoring.complex attribute)": [[48, "besskge.scoring.ComplEx.relation_embedding"]], "score_heads() (besskge.scoring.complex method)": [[48, "besskge.scoring.ComplEx.score_heads"]], "score_tails() (besskge.scoring.complex method)": [[48, "besskge.scoring.ComplEx.score_tails"]], "score_triple() (besskge.scoring.complex method)": [[48, "besskge.scoring.ComplEx.score_triple"]], "sharding (besskge.scoring.complex attribute)": [[48, "besskge.scoring.ComplEx.sharding"]], "update_sharding() (besskge.scoring.complex method)": [[48, "besskge.scoring.ComplEx.update_sharding"]], "conve (class in besskge.scoring)": [[49, "besskge.scoring.ConvE"]], "broadcasted_dot_product() (besskge.scoring.conve method)": [[49, "besskge.scoring.ConvE.broadcasted_dot_product"]], "entity_embedding (besskge.scoring.conve attribute)": [[49, "besskge.scoring.ConvE.entity_embedding"]], "forward() (besskge.scoring.conve method)": [[49, "besskge.scoring.ConvE.forward"]], "negative_sample_sharing (besskge.scoring.conve attribute)": [[49, "besskge.scoring.ConvE.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.conve method)": [[49, "besskge.scoring.ConvE.reduce_embedding"]], "relation_embedding (besskge.scoring.conve attribute)": [[49, "besskge.scoring.ConvE.relation_embedding"]], "score_heads() (besskge.scoring.conve method)": [[49, "besskge.scoring.ConvE.score_heads"]], "score_tails() (besskge.scoring.conve method)": [[49, "besskge.scoring.ConvE.score_tails"]], "score_triple() (besskge.scoring.conve method)": [[49, "besskge.scoring.ConvE.score_triple"]], "sharding (besskge.scoring.conve attribute)": [[49, "besskge.scoring.ConvE.sharding"]], "update_sharding() (besskge.scoring.conve method)": [[49, "besskge.scoring.ConvE.update_sharding"]], "distmult (class in besskge.scoring)": [[50, "besskge.scoring.DistMult"]], "broadcasted_dot_product() (besskge.scoring.distmult method)": [[50, "besskge.scoring.DistMult.broadcasted_dot_product"]], "entity_embedding (besskge.scoring.distmult attribute)": [[50, "besskge.scoring.DistMult.entity_embedding"]], "forward() (besskge.scoring.distmult method)": [[50, "besskge.scoring.DistMult.forward"]], "negative_sample_sharing (besskge.scoring.distmult attribute)": [[50, "besskge.scoring.DistMult.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.distmult method)": [[50, "besskge.scoring.DistMult.reduce_embedding"]], "relation_embedding (besskge.scoring.distmult attribute)": [[50, "besskge.scoring.DistMult.relation_embedding"]], "score_heads() (besskge.scoring.distmult method)": [[50, "besskge.scoring.DistMult.score_heads"]], "score_tails() (besskge.scoring.distmult method)": [[50, "besskge.scoring.DistMult.score_tails"]], "score_triple() (besskge.scoring.distmult method)": [[50, "besskge.scoring.DistMult.score_triple"]], "sharding (besskge.scoring.distmult attribute)": [[50, "besskge.scoring.DistMult.sharding"]], "update_sharding() (besskge.scoring.distmult method)": [[50, "besskge.scoring.DistMult.update_sharding"]], "distancebasedscorefunction (class in besskge.scoring)": [[51, "besskge.scoring.DistanceBasedScoreFunction"]], "broadcasted_distance() (besskge.scoring.distancebasedscorefunction method)": [[51, "besskge.scoring.DistanceBasedScoreFunction.broadcasted_distance"]], "entity_embedding (besskge.scoring.distancebasedscorefunction attribute)": [[51, "besskge.scoring.DistanceBasedScoreFunction.entity_embedding"]], "forward() (besskge.scoring.distancebasedscorefunction method)": [[51, "besskge.scoring.DistanceBasedScoreFunction.forward"]], "negative_sample_sharing (besskge.scoring.distancebasedscorefunction attribute)": [[51, "besskge.scoring.DistanceBasedScoreFunction.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.distancebasedscorefunction method)": [[51, "besskge.scoring.DistanceBasedScoreFunction.reduce_embedding"]], "relation_embedding (besskge.scoring.distancebasedscorefunction attribute)": [[51, "besskge.scoring.DistanceBasedScoreFunction.relation_embedding"]], "score_heads() (besskge.scoring.distancebasedscorefunction method)": [[51, "besskge.scoring.DistanceBasedScoreFunction.score_heads"]], "score_tails() (besskge.scoring.distancebasedscorefunction method)": [[51, "besskge.scoring.DistanceBasedScoreFunction.score_tails"]], "score_triple() (besskge.scoring.distancebasedscorefunction method)": [[51, "besskge.scoring.DistanceBasedScoreFunction.score_triple"]], "sharding (besskge.scoring.distancebasedscorefunction attribute)": [[51, "besskge.scoring.DistanceBasedScoreFunction.sharding"]], "update_sharding() (besskge.scoring.distancebasedscorefunction method)": [[51, "besskge.scoring.DistanceBasedScoreFunction.update_sharding"]], "interht (class in besskge.scoring)": [[52, "besskge.scoring.InterHT"]], "broadcasted_distance() (besskge.scoring.interht method)": [[52, "besskge.scoring.InterHT.broadcasted_distance"]], "entity_embedding (besskge.scoring.interht attribute)": [[52, "besskge.scoring.InterHT.entity_embedding"]], "forward() (besskge.scoring.interht method)": [[52, "besskge.scoring.InterHT.forward"]], "negative_sample_sharing (besskge.scoring.interht attribute)": [[52, "besskge.scoring.InterHT.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.interht method)": [[52, "besskge.scoring.InterHT.reduce_embedding"]], "relation_embedding (besskge.scoring.interht attribute)": [[52, "besskge.scoring.InterHT.relation_embedding"]], "score_heads() (besskge.scoring.interht method)": [[52, "besskge.scoring.InterHT.score_heads"]], "score_tails() (besskge.scoring.interht method)": [[52, "besskge.scoring.InterHT.score_tails"]], "score_triple() (besskge.scoring.interht method)": [[52, "besskge.scoring.InterHT.score_triple"]], "sharding (besskge.scoring.interht attribute)": [[52, "besskge.scoring.InterHT.sharding"]], "update_sharding() (besskge.scoring.interht method)": [[52, "besskge.scoring.InterHT.update_sharding"]], "matrixdecompositionscorefunction (class in besskge.scoring)": [[53, "besskge.scoring.MatrixDecompositionScoreFunction"]], "broadcasted_dot_product() (besskge.scoring.matrixdecompositionscorefunction method)": [[53, "besskge.scoring.MatrixDecompositionScoreFunction.broadcasted_dot_product"]], "entity_embedding (besskge.scoring.matrixdecompositionscorefunction attribute)": [[53, "besskge.scoring.MatrixDecompositionScoreFunction.entity_embedding"]], "forward() (besskge.scoring.matrixdecompositionscorefunction method)": [[53, "besskge.scoring.MatrixDecompositionScoreFunction.forward"]], "negative_sample_sharing (besskge.scoring.matrixdecompositionscorefunction attribute)": [[53, "besskge.scoring.MatrixDecompositionScoreFunction.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.matrixdecompositionscorefunction method)": [[53, "besskge.scoring.MatrixDecompositionScoreFunction.reduce_embedding"]], "relation_embedding (besskge.scoring.matrixdecompositionscorefunction attribute)": [[53, "besskge.scoring.MatrixDecompositionScoreFunction.relation_embedding"]], "score_heads() (besskge.scoring.matrixdecompositionscorefunction method)": [[53, "besskge.scoring.MatrixDecompositionScoreFunction.score_heads"]], "score_tails() (besskge.scoring.matrixdecompositionscorefunction method)": [[53, "besskge.scoring.MatrixDecompositionScoreFunction.score_tails"]], "score_triple() (besskge.scoring.matrixdecompositionscorefunction method)": [[53, "besskge.scoring.MatrixDecompositionScoreFunction.score_triple"]], "sharding (besskge.scoring.matrixdecompositionscorefunction attribute)": [[53, "besskge.scoring.MatrixDecompositionScoreFunction.sharding"]], "update_sharding() (besskge.scoring.matrixdecompositionscorefunction method)": [[53, "besskge.scoring.MatrixDecompositionScoreFunction.update_sharding"]], "pairre (class in besskge.scoring)": [[54, "besskge.scoring.PairRE"]], "broadcasted_distance() (besskge.scoring.pairre method)": [[54, "besskge.scoring.PairRE.broadcasted_distance"]], "entity_embedding (besskge.scoring.pairre attribute)": [[54, "besskge.scoring.PairRE.entity_embedding"]], "forward() (besskge.scoring.pairre method)": [[54, "besskge.scoring.PairRE.forward"]], "negative_sample_sharing (besskge.scoring.pairre attribute)": [[54, "besskge.scoring.PairRE.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.pairre method)": [[54, "besskge.scoring.PairRE.reduce_embedding"]], "relation_embedding (besskge.scoring.pairre attribute)": [[54, "besskge.scoring.PairRE.relation_embedding"]], "score_heads() (besskge.scoring.pairre method)": [[54, "besskge.scoring.PairRE.score_heads"]], "score_tails() (besskge.scoring.pairre method)": [[54, "besskge.scoring.PairRE.score_tails"]], "score_triple() (besskge.scoring.pairre method)": [[54, "besskge.scoring.PairRE.score_triple"]], "sharding (besskge.scoring.pairre attribute)": [[54, "besskge.scoring.PairRE.sharding"]], "update_sharding() (besskge.scoring.pairre method)": [[54, "besskge.scoring.PairRE.update_sharding"]], "rotate (class in besskge.scoring)": [[55, "besskge.scoring.RotatE"]], "broadcasted_distance() (besskge.scoring.rotate method)": [[55, "besskge.scoring.RotatE.broadcasted_distance"]], "entity_embedding (besskge.scoring.rotate attribute)": [[55, "besskge.scoring.RotatE.entity_embedding"]], "forward() (besskge.scoring.rotate method)": [[55, "besskge.scoring.RotatE.forward"]], "negative_sample_sharing (besskge.scoring.rotate attribute)": [[55, "besskge.scoring.RotatE.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.rotate method)": [[55, "besskge.scoring.RotatE.reduce_embedding"]], "relation_embedding (besskge.scoring.rotate attribute)": [[55, "besskge.scoring.RotatE.relation_embedding"]], "score_heads() (besskge.scoring.rotate method)": [[55, "besskge.scoring.RotatE.score_heads"]], "score_tails() (besskge.scoring.rotate method)": [[55, "besskge.scoring.RotatE.score_tails"]], "score_triple() (besskge.scoring.rotate method)": [[55, "besskge.scoring.RotatE.score_triple"]], "sharding (besskge.scoring.rotate attribute)": [[55, "besskge.scoring.RotatE.sharding"]], "update_sharding() (besskge.scoring.rotate method)": [[55, "besskge.scoring.RotatE.update_sharding"]], "trans (class in besskge.scoring)": [[56, "besskge.scoring.TranS"]], "broadcasted_distance() (besskge.scoring.trans method)": [[56, "besskge.scoring.TranS.broadcasted_distance"]], "entity_embedding (besskge.scoring.trans attribute)": [[56, "besskge.scoring.TranS.entity_embedding"]], "forward() (besskge.scoring.trans method)": [[56, "besskge.scoring.TranS.forward"]], "negative_sample_sharing (besskge.scoring.trans attribute)": [[56, "besskge.scoring.TranS.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.trans method)": [[56, "besskge.scoring.TranS.reduce_embedding"]], "relation_embedding (besskge.scoring.trans attribute)": [[56, "besskge.scoring.TranS.relation_embedding"]], "score_heads() (besskge.scoring.trans method)": [[56, "besskge.scoring.TranS.score_heads"]], "score_tails() (besskge.scoring.trans method)": [[56, "besskge.scoring.TranS.score_tails"]], "score_triple() (besskge.scoring.trans method)": [[56, "besskge.scoring.TranS.score_triple"]], "sharding (besskge.scoring.trans attribute)": [[56, "besskge.scoring.TranS.sharding"]], "update_sharding() (besskge.scoring.trans method)": [[56, "besskge.scoring.TranS.update_sharding"]], "transe (class in besskge.scoring)": [[57, "besskge.scoring.TransE"]], "broadcasted_distance() (besskge.scoring.transe method)": [[57, "besskge.scoring.TransE.broadcasted_distance"]], "entity_embedding (besskge.scoring.transe attribute)": [[57, "besskge.scoring.TransE.entity_embedding"]], "forward() (besskge.scoring.transe method)": [[57, "besskge.scoring.TransE.forward"]], "negative_sample_sharing (besskge.scoring.transe attribute)": [[57, "besskge.scoring.TransE.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.transe method)": [[57, "besskge.scoring.TransE.reduce_embedding"]], "relation_embedding (besskge.scoring.transe attribute)": [[57, "besskge.scoring.TransE.relation_embedding"]], "score_heads() (besskge.scoring.transe method)": [[57, "besskge.scoring.TransE.score_heads"]], "score_tails() (besskge.scoring.transe method)": [[57, "besskge.scoring.TransE.score_tails"]], "score_triple() (besskge.scoring.transe method)": [[57, "besskge.scoring.TransE.score_triple"]], "sharding (besskge.scoring.transe attribute)": [[57, "besskge.scoring.TransE.sharding"]], "update_sharding() (besskge.scoring.transe method)": [[57, "besskge.scoring.TransE.update_sharding"]], "triplere (class in besskge.scoring)": [[58, "besskge.scoring.TripleRE"]], "broadcasted_distance() (besskge.scoring.triplere method)": [[58, "besskge.scoring.TripleRE.broadcasted_distance"]], "entity_embedding (besskge.scoring.triplere attribute)": [[58, "besskge.scoring.TripleRE.entity_embedding"]], "forward() (besskge.scoring.triplere method)": [[58, "besskge.scoring.TripleRE.forward"]], "negative_sample_sharing (besskge.scoring.triplere attribute)": [[58, "besskge.scoring.TripleRE.negative_sample_sharing"]], "reduce_embedding() (besskge.scoring.triplere method)": [[58, "besskge.scoring.TripleRE.reduce_embedding"]], "relation_embedding (besskge.scoring.triplere attribute)": [[58, "besskge.scoring.TripleRE.relation_embedding"]], "score_heads() (besskge.scoring.triplere method)": [[58, "besskge.scoring.TripleRE.score_heads"]], "score_tails() (besskge.scoring.triplere method)": [[58, "besskge.scoring.TripleRE.score_tails"]], "score_triple() (besskge.scoring.triplere method)": [[58, "besskge.scoring.TripleRE.score_triple"]], "sharding (besskge.scoring.triplere attribute)": [[58, "besskge.scoring.TripleRE.sharding"]], "update_sharding() (besskge.scoring.triplere method)": [[58, "besskge.scoring.TripleRE.update_sharding"]], "besskge.sharding": [[59, "module-besskge.sharding"]], "partitionedtripleset (class in besskge.sharding)": [[60, "besskge.sharding.PartitionedTripleSet"]], "create_from_dataset() (besskge.sharding.partitionedtripleset class method)": [[60, "besskge.sharding.PartitionedTripleSet.create_from_dataset"]], "create_from_queries() (besskge.sharding.partitionedtripleset class method)": [[60, "besskge.sharding.PartitionedTripleSet.create_from_queries"]], "dummy (besskge.sharding.partitionedtripleset attribute)": [[60, "besskge.sharding.PartitionedTripleSet.dummy"]], "inverse_triples (besskge.sharding.partitionedtripleset attribute)": [[60, "besskge.sharding.PartitionedTripleSet.inverse_triples"]], "neg_heads (besskge.sharding.partitionedtripleset attribute)": [[60, "besskge.sharding.PartitionedTripleSet.neg_heads"]], "neg_tails (besskge.sharding.partitionedtripleset attribute)": [[60, "besskge.sharding.PartitionedTripleSet.neg_tails"]], "partition_mode (besskge.sharding.partitionedtripleset attribute)": [[60, "besskge.sharding.PartitionedTripleSet.partition_mode"]], "sharding (besskge.sharding.partitionedtripleset attribute)": [[60, "besskge.sharding.PartitionedTripleSet.sharding"]], "triple_counts (besskge.sharding.partitionedtripleset attribute)": [[60, "besskge.sharding.PartitionedTripleSet.triple_counts"]], "triple_offsets (besskge.sharding.partitionedtripleset attribute)": [[60, "besskge.sharding.PartitionedTripleSet.triple_offsets"]], "triple_sort_idx (besskge.sharding.partitionedtripleset attribute)": [[60, "besskge.sharding.PartitionedTripleSet.triple_sort_idx"]], "triples (besskge.sharding.partitionedtripleset attribute)": [[60, "besskge.sharding.PartitionedTripleSet.triples"]], "types (besskge.sharding.partitionedtripleset attribute)": [[60, "besskge.sharding.PartitionedTripleSet.types"]], "sharding (class in besskge.sharding)": [[61, "besskge.sharding.Sharding"]], "create() (besskge.sharding.sharding class method)": [[61, "besskge.sharding.Sharding.create"]], "entity_to_idx (besskge.sharding.sharding attribute)": [[61, "besskge.sharding.Sharding.entity_to_idx"]], "entity_to_shard (besskge.sharding.sharding attribute)": [[61, "besskge.sharding.Sharding.entity_to_shard"]], "entity_type_counts (besskge.sharding.sharding attribute)": [[61, "besskge.sharding.Sharding.entity_type_counts"]], "entity_type_offsets (besskge.sharding.sharding attribute)": [[61, "besskge.sharding.Sharding.entity_type_offsets"]], "load() (besskge.sharding.sharding class method)": [[61, "besskge.sharding.Sharding.load"]], "max_entity_per_shard (besskge.sharding.sharding property)": [[61, "besskge.sharding.Sharding.max_entity_per_shard"]], "n_entity (besskge.sharding.sharding property)": [[61, "besskge.sharding.Sharding.n_entity"]], "n_shard (besskge.sharding.sharding attribute)": [[61, "besskge.sharding.Sharding.n_shard"]], "save() (besskge.sharding.sharding method)": [[61, "besskge.sharding.Sharding.save"]], "shard_and_idx_to_entity (besskge.sharding.sharding attribute)": [[61, "besskge.sharding.Sharding.shard_and_idx_to_entity"]], "shard_counts (besskge.sharding.sharding attribute)": [[61, "besskge.sharding.Sharding.shard_counts"]], "besskge.utils": [[62, "module-besskge.utils"]], "complex_multiplication() (in module besskge.utils)": [[63, "besskge.utils.complex_multiplication"]], "complex_rotation() (in module besskge.utils)": [[64, "besskge.utils.complex_rotation"]], "gather_indices() (in module besskge.utils)": [[65, "besskge.utils.gather_indices"]], "get_entity_filter() (in module besskge.utils)": [[66, "besskge.utils.get_entity_filter"]], "besskge": [[67, "module-besskge"]]}})
\ No newline at end of file