Update app.py

app.py

@@ -25,13 +25,12 @@ def last_token_pool(last_hidden_states: Tensor, attention_mask: Tensor) -> Tenso
 def get_detailed_instruct(task_description: str, query: str) -> str:
     return f'Instruct: {task_description}\nQuery: {query}'
 
-tokenizer = AutoTokenizer.from_pretrained('intfloat/e5-mistral-7b-instruct')
-model = AutoModel.from_pretrained('intfloat/e5-mistral-7b-instruct')
 
 @spaces.GPU
 def compute_embeddings(*input_texts):
+    tokenizer = AutoTokenizer.from_pretrained('intfloat/e5-mistral-7b-instruct')
+    model = AutoModel.from_pretrained('intfloat/e5-mistral-7b-instruct')
     device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
-
     model.to(device)
     max_length = 4096
     task = 'Given a web search query, retrieve relevant passages that answer the query'