Update app.py

app.py

@@ -16,32 +16,29 @@ tokenizer = MT5TokenizerFast.from_pretrained(
 )
 
 def predict(text):
-    with torch.no_grad():
-        input_ids = tokenizer.encode(text, return_tensors="pt", add_special_tokens=True)
+# with torch.no_grad():
+    input_ids = tokenizer.encode(text, return_tensors="pt", add_special_tokens=True)
+    generated_ids = model.generate(
+        input_ids=input_ids,
+        num_beams=5,
+        max_length=1000,
+        repetition_penalty=3.0, #default = 2.5
+        length_penalty=1.0,
+        early_stopping=True,
+        top_p=50, #default 50 
+        top_k=20, #default 20
+        num_return_sequences=3,
+    )
 
-        input_ids = input_ids.cuda()
-
-        generated_ids = model.generate(
-            input_ids=input_ids,
-            num_beams=5,
-            max_length=1000,
-            repetition_penalty=3.0, #default = 2.5
-            length_penalty=1.0,
-            early_stopping=True,
-            top_p=50, #default 50 
-            top_k=20, #default 20
-            num_return_sequences=3,
+    preds = [
+        tokenizer.decode(
+            g,
+            skip_special_tokens=True,
+            clean_up_tokenization_spaces=True,
         )
-
-        preds = [
-            tokenizer.decode(
-                g,
-                skip_special_tokens=True,
-                clean_up_tokenization_spaces=True,
-            )
-            for g in generated_ids
-        ]
-    return preds
+        for g in generated_ids
+    ]
+return preds
 
 # text_to_predict = predict(text)
 # predicted = ['Q: ' + text for text in predict(text_to_predict)]