Update chatbot_rag.py

chatbot_rag.py

@@ -52,8 +52,8 @@ def build_qa():
     "text-generation",
     model=model,
     tokenizer=tokenizer,
-    max_new_tokens=128,
-    temperature=0.4,      # keeps answers deterministic but less rigid than 0
+    max_new_tokens=256,
+    temperature=0.2,      # keeps answers deterministic but less rigid than 0
     do_sample=True,       # allow some randomness
     top_p=0.9,            # nucleus sampling to avoid loops
     repetition_penalty=1.2,  # 🚀 penalize repeats