Update app.py

app.py

@@ -7,7 +7,7 @@ from transformers import AutoModelForCausalLM, AutoTokenizer
 # -------------------------------------------------
 # Model setup (loaded once at startup)
 # -------------------------------------------------
-model_name = "gr0010/CustomThinker-0-8B"
+model_name = "CustomThinker-0-8B"
 
 # Load model and tokenizer globally
 print("Loading model and tokenizer...")
@@ -34,12 +34,12 @@ def generate_and_parse(messages: list, temperature: float = 0.6,
     and parses it into thinking and answer parts.
     Decorated with @spaces.GPU for Zero GPU allocation.
     """
-    # Apply chat template with enable_thinking=True for Qwen3
+    # Apply chat template WITHOUT enable_thinking to preserve thinking tags in history
     prompt_text = tokenizer.apply_chat_template(
         messages,
         tokenize=False,
         add_generation_prompt=True,
-        enable_thinking=True  # Explicitly enable thinking mode
+        enable_thinking=False  # Changed to False to preserve <think> tags in context
     )
 
     # --- CONSOLE DEBUG OUTPUT ---