change embedding settings and fix bug

Signed-off-by: charliehuang09 <charliehuang09@gmail.com>
frc971 · Mar 26, 2024 · d70fc93 · d70fc93
1 parent 1631acf
commit d70fc93
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 2 deletions.
diff --git a/scripts/encode_data.py b/scripts/encode_data.py
@@ -35,7 +35,7 @@ def __init__(self,
         self.do_parent_document = do_parent_document
 
         self.embeddings_model = OpenAIEmbeddings(
-            model="text-embedding-3-large", dimensions=1536)
+            model="text-embedding-3-large", dimensions=3072)
 
         self.text_splitter = TokenTextSplitter(chunk_size=chunk_size,
                                                chunk_overlap=0)

diff --git a/scripts/qdrant_upsert.py b/scripts/qdrant_upsert.py
@@ -13,7 +13,7 @@ def qdrant_upsert(path, collection_name, port=6333):
         pass
     client.create_collection(collection_name=collection_name,
                              vectors_config=models.VectorParams(
-                                 size=1536, distance=models.Distance.COSINE))
+                                 size=3072, distance=models.Distance.EUCLID))
 
     data = np.load(path, allow_pickle=True)