Update app.py

app.py

@@ -8,8 +8,8 @@ import torch
 from huggingface_hub import InferenceClient
 from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
 
-MAX_MAX_NEW_TOKENS = 2048
-DEFAULT_MAX_NEW_TOKENS = 1024
+MAX_MAX_NEW_TOKENS = 1024
+DEFAULT_MAX_NEW_TOKENS = 512
 MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "4096"))
 
 #Inference API Code