try again

app.py

@@ -40,7 +40,7 @@ def predict(code):
 [/INST]
 """
 
-    print(f"Prompt: {str(prompt)}")
+    print(f"Prompt: {repr(prompt)}")
     print(f"Tokenized: {tokenizer.tokenize(prompt)}")
     pipe_out = pipe(
         prompt,
@@ -50,11 +50,13 @@ def predict(code):
         temperature=1.2,
         repetition_penalty=1.1,
     )
-    print(f"Pipe out: {str(pipe_out)}")
+    print(f"Pipe out: {repr(pipe_out)}")
 
-    return tokenizer.decode(model.generate(
-        tokenizer.encode(prompt), max_length=10**6
-    ))
+    return tokenizer.decode(
+        model.generate(
+            tokenizer.encode(prompt, return_tensors="pt"), max_length=10**6
+        )[0]
+    )
 
 
 demo = gr.Interface(