Update app.py

app.py

@@ -1,6 +1,6 @@
 #MODEL_ID = "ibm-granite/granite-3.1-2b-instruct"
 MODEL_ID = "deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B"
-
+QUANT = "Q5_K_M"
 
 import subprocess
 def run_command(command, cwd=None):
@@ -39,9 +39,9 @@ def setup_model(model_id):
     gguf_path = f"{local_dir}.gguf"
     if not os.path.exists(gguf_path):
         run_command(f'python llama.cpp/convert_hf_to_gguf.py ./{local_dir} --outfile {gguf_path}')
-    quantized_path = f"{local_dir}-IQ2_XXS.gguf"
+    quantized_path = f"{local_dir}-{QUANT}.gguf"
     if not os.path.exists(quantized_path):
-        run_command(f'./llama.cpp/build/bin/llama-quantize ./{gguf_path} {quantized_path} IQ2_XXS')
+        run_command(f'./llama.cpp/build/bin/llama-quantize ./{gguf_path} {quantized_path} {QUANT}')
     return quantized_path
 
 def chat_with_model(message, history, system_prompt, temperature, max_tokens, top_k, top_p):