Track/llm kg

.gitignore

@@ -1,3 +1,6 @@
 checkpoint
 *.npy
 *.log
+models
+data/wikidata_5m/*.dict
+data/wikidata_5m/*.json

commands/command1.txt

@@ -0,0 +1,2 @@
+set CUDA_VISIBLE_DEVICES=2,3 & python "C:\E Drive\AI DRG Code Base\KnowledgeGraphEmbedding\codes\run.py" --do_train --cuda --data_path "C:\E Drive\AI DRG Code Base\KnowledgeGraphEmbedding\data\countries_S1" --model RotatE -n 256 -b 1024 -d 1000 -g 24.0 -a 1.0 -adv -lr 0.0001 --max_steps 200 -save "C:\E Drive\AI DRG Code Base\KnowledgeGraphEmbedding\models\countries_S1" -de --save_checkpoint_steps 10 --log_steps 1
+python "C:\E Drive\AI DRG Code Base\KnowledgeGraphEmbedding\codes\run.py" --do_train --cuda --data_path "C:\E Drive\AI DRG Code Base\KnowledgeGraphEmbedding\data\countries_S1" --model RotatE -n 128 -b 1024 -d 256 -g 24.0 -a 1.0 -adv -lr 0.0001 --max_steps 100 -save "C:\E Drive\AI DRG Code Base\KnowledgeGraphEmbedding\models\countries_S1" -de --save_checkpoint_steps 10 --log_steps 5

data/wikidata_5m/add_graph.py

@@ -0,0 +1,68 @@
+import os
+import json
+
+path = os.path.join(os.getcwd(), 'KnowledgeGraphEmbedding\data\wikidata_5m')
+print(path)
+
+class Graph():
+    def __init__(self, graph):
+        self.graph = graph
+
+    def getGraph(self):
+        return self.graph
+
+    def getId2Entity(self):
+        return self.graph['id2entity']
+
+    def getEntity2Id(self):
+        return self.graph['entity2id']
+
+    def getId2Relation(self):
+        return self.graph['id2relation']
+
+    def getRelation2Id(self):
+        return self.graph['relation2id']
+
+def readId2Entity(graph: Graph, path=path):
+    id_entity = graph.getId2Entity()
+    print(len(id_entity))
+
+    entity_file = open(os.path.join(path, 'entities.dict'), 'a+')
+    # counter = 10
+    i = 0
+    for entity in id_entity:
+        entity_file.write(f'{i}\t{entity}\n')
+        i = i+1
+
+def readEntity2Id(graph: Graph, path=path):
+    entity_id = graph.getEntity2Id()
+    print(type(entity_id))
+
+def readId2Relation(graph: Graph, path=path):
+    id_relation = graph.getId2Relation()
+    print(len(id_relation))
+
+    relation_file = open(os.path.join(path, 'relations.dict'), 'a+')
+    # counter = 10
+    i = 0
+    for relation in id_relation:
+        relation_file.write(f'{i}\t{relation}\n')
+        i = i+1
+
+def readRelation2Id(graph: Graph, path=path):
+    relation_id = graph.getRelation2Id()
+    print(type(relation_id))
+
+def run(path):
+    graph_path = os.path.join(path, 'graph_id.json')
+    graph_file = open(graph_path, 'r')
+    graph: Graph = Graph(json.load(graph_file))
+
+    print(graph.getGraph().keys())
+    readId2Entity(graph, path)
+    # readEntity2Id(graph, path)
+    readId2Relation(graph, path)
+    # readRelation2Id(graph, path)
+
+if __name__=='__main__':
+    run(path=path)