method model call

app.py

@@ -2,7 +2,8 @@ from huggingface_hub import login
 from fastapi import FastAPI, Depends, HTTPException
 import logging
 from pydantic import BaseModel
-from sentence_transformers import SentenceTransformer
+from transformers import AutoTokenizer, AutoModel
+from sentence_transformers import models, SentenceTransformer
 from services.qdrant_searcher import QdrantSearcher
 from services.openai_service import generate_rag_response
 from utils.auth import token_required
@@ -30,7 +31,7 @@ logging.basicConfig(level=logging.INFO)
 huggingface_token = os.getenv('HUGGINGFACE_HUB_TOKEN')
 if huggingface_token:
     try:
-        login(token=huggingface_token, add_to_git_credential=True, write_permission=True)
+        login(token=huggingface_token, add_to_git_credential=True)
         logging.info("Successfully logged into Hugging Face Hub.")
     except Exception as e:
         logging.error(f"Failed to log into Hugging Face Hub: {e}")
@@ -45,10 +46,19 @@ access_token = os.getenv('QDRANT_ACCESS_TOKEN')
 if not qdrant_url or not access_token:
     raise ValueError("Qdrant URL or Access Token is not set. Please set the QDRANT_URL and QDRANT_ACCESS_TOKEN environment variables.")
 
-# Initialize the SentenceTransformer model with the cache directory managed by HF_HOME
+# Initialize the SentenceTransformer model with trust_remote_code using transformers
 try:
     cache_folder = os.path.join(hf_home_dir, "transformers_cache")
-    encoder = SentenceTransformer('nomic-ai/nomic-embed-text-v1.5', cache_folder=cache_folder)
+    
+    # Load the tokenizer and model with trust_remote_code=True
+    tokenizer = AutoTokenizer.from_pretrained('nomic-ai/nomic-embed-text-v1.5', trust_remote_code=True)
+    model = AutoModel.from_pretrained('nomic-ai/nomic-embed-text-v1.5', trust_remote_code=True)
+
+    # Wrap the model into a SentenceTransformer
+    word_embedding_model = models.Transformer(model_name_or_path='nomic-ai/nomic-embed-text-v1.5', model=model, tokenizer=tokenizer)
+    pooling_model = models.Pooling(word_embedding_model.get_word_embedding_dimension())
+    encoder = SentenceTransformer(modules=[word_embedding_model, pooling_model])
+    
     logging.info("Successfully loaded the SentenceTransformer model.")
 except Exception as e:
     logging.error(f"Failed to load the SentenceTransformer model: {e}")