feat: update model

load_models.py

@@ -64,7 +64,7 @@ def load_quantized_model_gguf_ggml(model_id, model_basename, device_type, loggin
             kwargs["n_gpu_layers"] = 1
         if device_type.lower() == "cuda":
             kwargs["n_gpu_layers"] = N_GPU_LAYERS
-            kwargs["n_batch"] = MAX_NEW_TOKENS  # set this based on your GPU
+            kwargs["n_batch"] = N_BATCH  # set this based on your GPU
 
         # kwargs["stream"] = stream