Update app.py

app.py

@@ -9,24 +9,19 @@ model_path = r'ssocean/NAIP'
 device = 'cuda:0'
 
 global model, tokenizer
-model = AutoModelForSequenceClassification.from_pretrained(
-    model_path,
-    num_labels=1,
-    load_in_8bit=True,
-)
-tokenizer = AutoTokenizer.from_pretrained(model_path)
-model.eval()
-
+model = None
+tokenizer = None
 
 
 @spaces.GPU(duration=60, enable_queue=True)
 def predict(title, abstract):
-    model = AutoModelForSequenceClassification.from_pretrained(
-    model_path,
-    num_labels=1,
-    load_in_8bit=True,
-    device_map={'': device(type='cuda', index=0)}
-)
+    if model is None:
+        model = AutoModelForSequenceClassification.from_pretrained(
+        model_path,
+        num_labels=1,
+        load_in_8bit=True,)
+        tokenizer = AutoTokenizer.from_pretrained(model_path)
+        model.eval()
     text = f'''Given a certain paper, Title: {title}\n Abstract: {abstract}. \n Predict its normalized academic impact (between 0 and 1):'''
     inputs = tokenizer(text, return_tensors="pt").to(device)
     with torch.no_grad():