Update app.py

app.py

@@ -13,8 +13,6 @@ model = PeftModel.from_pretrained(model, "phearion/bigbrain-v0.0.1")
 # Move the model to the device
 model = model.to(device)
 
-# Convert the model to TorchScript
-scripted_model = torch.jit.script(model)
 
 # Load the tokenizer
 tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-v0.1")
@@ -23,12 +21,10 @@ tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-v0.1")
 
 def greet(text):
     batch = tokenizer(f"'{text}' ->: ", return_tensors='pt')
-    # Move the batch to the device
-    batch = {k: v.to(device) for k, v in batch.items()}
-
+    
     # Use torch.no_grad to disable gradient calculation
     with torch.no_grad():
-        output_tokens = scripted_model.generate(**batch, max_new_tokens=20)
+        output_tokens = model.generate(**batch, max_new_tokens=20)
 
     return tokenizer.decode(output_tokens[0], skip_special_tokens=True)