From 79666c298dd7e2a51881b3d0e3f6ed420b374339 Mon Sep 17 00:00:00 2001
From: hiyouga <hiyouga@buaa.edu.cn>
Date: Sun, 21 Apr 2024 19:20:18 +0800
Subject: [PATCH] fix #3365

Former-commit-id: a1d31ffc8cb7a6a477704efe779d485d83b8b9fb
---
 src/llmtuner/model/patcher.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/src/llmtuner/model/patcher.py b/src/llmtuner/model/patcher.py
index a1b19fb1..53616dd9 100644
--- a/src/llmtuner/model/patcher.py
+++ b/src/llmtuner/model/patcher.py
@@ -139,6 +139,7 @@ def _configure_quantization(
 
         if quant_method == QuantizationMethod.GPTQ:
             require_version("auto_gptq>=0.5.0", "To fix: pip install auto_gptq>=0.5.0")
+            quantization_config.pop("disable_exllama", None)  # remove deprecated args
             quantization_config["use_exllama"] = False  # disable exllama
 
         if quant_method == QuantizationMethod.AWQ: