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 = 512
-DEFAULT_MAX_NEW_TOKENS = 512
+MAX_MAX_NEW_TOKENS = 1024
+DEFAULT_MAX_NEW_TOKENS = 1024
 MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "4096"))
 
 #Transformers Code
@@ -52,7 +52,7 @@ def generate(
         do_sample=True,
         top_p=1.0,
         top_k=50,
-        temperature=1.0,
+        temperature=0.9,
         num_beams=1,
         repetition_penalty=1.1,
     )