Update app.py

app.py

@@ -10,6 +10,10 @@ from datetime import timedelta
 from datetime import timezone
 import io
 import requests
+import torch.nn.functional as F
+import torch
+from torch import Tensor
+from transformers import AutoTokenizer, AutoModel
 
 app = Flask(__name__)
 
@@ -25,6 +29,15 @@ qdrant_url = os.environ.get("qdrant_url")
 client = QdrantClient(url=qdrant_url, port=443, api_key=qdrant_api_key, prefer_grpc=False)
 
 # device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+
+def average_pool(last_hidden_states: Tensor,
+                 attention_mask: Tensor) -> Tensor:
+    last_hidden = last_hidden_states.masked_fill(~attention_mask[..., None].bool(), 0.0)
+    return last_hidden.sum(dim=1) / attention_mask.sum(dim=1)[..., None]
+
+tokenizer = AutoTokenizer.from_pretrained('intfloat/e5-base-v2')
+model = AutoModel.from_pretrained('intfloat/e5-base-v2').to(device)
 
 def e5embed(query):
   batch_dict = tokenizer(query, max_length=512, padding=True, truncation=True, return_tensors='pt')