remove print

p2m2 · Jan 20, 2025 · 4dc80b6 · 4dc80b6
1 parent b388890
commit 4dc80b6
Show file tree

Hide file tree

Showing 3 changed files with 2 additions and 8 deletions.
diff --git a/config/albert_igepp.json b/config/albert_igepp.json
@@ -1,7 +1,7 @@
 {
     "encodeur" : "sentence-transformers/all-MiniLM-L6-v2",
     "threshold_similarity_tag_chunk" : 0.60,
-    "threshold_similarity_tag" : 0.95,
+    "threshold_similarity_tag" : 0.15,
     "batch_size" : 32,
 
     "populate_owl_tag_embeddings" : {

diff --git a/config/planteome-demo-only-TO-0000394.json b/config/planteome-demo-only-TO-0000394.json
@@ -1,6 +1,6 @@
 {
     "encodeur" : "sentence-transformers/all-MiniLM-L6-v2",
-    "threshold_similarity_tag_chunk" : 0.10,
+    "threshold_similarity_tag_chunk" : 0.70,
     "threshold_similarity_tag" : 0.80,
     "batch_size" : 32,
 

diff --git a/llm_semantic_annotator/similarity/model_embedding_manager.py b/llm_semantic_annotator/similarity/model_embedding_manager.py
@@ -192,19 +192,13 @@ def compare_tags_with_chunks(self, tag_embeddings, chunks_embeddings):
 
         for doi, chunks_embedding in tqdm(list(chunks_embeddings.items())):
             chunks_matrix = torch.stack(chunks_embedding).to(device)
-            print("----------------####################################-----------")
-            print(doi,chunks_matrix)
-            print("------------TTTT---------------")
-            print(tag_embeddings_matrix)
-            print("---------------------------")
             # Calcul vectorisé des similarités sur GPU
             # Normaliser les vecteurs pour le calcul de la similarité cosinus
             chunks_norm = F.normalize(chunks_matrix, p=2, dim=1)
             tags_norm = F.normalize(tag_embeddings_matrix, p=2, dim=1)
 
             # Calcul de la similarité cosinus
             similarities = torch.mm(chunks_norm, tags_norm.t())
-            print(similarities)
             max_similarities, _ = torch.max(similarities, dim=0)
 
             # Filtrage des similarités au-dessus du seuil