flan

question_paper.py

@@ -31,7 +31,7 @@ def llama():
     # tresponse = TOKENIZER.decode(tokens[0], skip_special_tokens=False)
     # print(tresponse)
     input = "Generate 5 jokes"
-    output = pipe_flan(input)
+    output = pipe_flan(input, {"max_new_tokens": 1024,})
     response_message = {"message": output[0]["generated_text"]}
     json_response = jsonable_encoder(response_message)
     return JSONResponse(content=json_response)