update embedding model

load_data.py

@@ -24,10 +24,10 @@ load_dotenv()
 
 
 HF_TOKEN = os.getenv("HF_TOKEN")
-EMBEDDING_MODEL_NAME = "Snowflake/snowflake-arctic-embed-m-long"
-EMBEDDING_MODEL_REVISION = "ac9d0cb43661ee1f7d67b3aa63614d65a6c86463"
+EMBEDDING_MODEL_NAME = "Alibaba-NLP/gte-large-en-v1.5"
+EMBEDDING_MODEL_REVISION = "104333d6af6f97649377c2afbde10a7704870c7b"
 INFERENCE_MODEL_URL = (
-    "https://pqzap00ebpl1ydt4.us-east-1.aws.endpoints.huggingface.cloud"
+    "https://spwy1g6626yhjhpr.us-east-1.aws.endpoints.huggingface.cloud"
 )
 DATASET_PARQUET_URL = (
     "hf://datasets/librarian-bots/dataset_cards_with_metadata/data/train-*.parquet"
@@ -62,7 +62,10 @@ def get_embedding_function():
 def get_collection(chroma_client, embedding_function):
     logger.info(f"Getting or creating collection: {COLLECTION_NAME}")
     return chroma_client.create_collection(
-        name=COLLECTION_NAME, get_or_create=True, embedding_function=embedding_function
+        name=COLLECTION_NAME,
+        get_or_create=True,
+        embedding_function=embedding_function,
+        metadata={"hnsw:space": "cosine"},
     )