diff --git a/pyproject.toml b/pyproject.toml
index 480b944fd..cb219ee9c 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -53,7 +53,7 @@ dependencies = [
     "pyarrow>=21.0",
     "dill>=0.3.8", # datasets requirements
     "pypcre>=0.2.4",
-    "torchao>=0.14.0", # fix bad transformers 4.57.1 breaking torchao compat
+    "torchao>=0.14.1", # fix bad transformers 4.57.1 breaking torchao compat
     # "cython>=3.1.4", # required by hf-xet/hf-transfer
 # "flash-attn>=2.8.3", <-- install for lower vram usage
 ]
diff --git a/requirements.txt b/requirements.txt
index 697b032cb..128a701a3 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -19,4 +19,4 @@ datasets>=3.6.0
 pyarrow>=21.0
 dill>=0.3.8
 pypcre>=0.2.4
-torchao>=0.14.0
+torchao>=0.14.1
diff --git a/tests/models/model_test.py b/tests/models/model_test.py
index 7d0ee4892..e2b75776c 100644
--- a/tests/models/model_test.py
+++ b/tests/models/model_test.py
@@ -109,6 +109,7 @@ class ModelTest(unittest.TestCase):
     ACT_GROUP_AWARE = True
     FAIL_SAFE = True
     EORA = None
+    DAMP_PERCENT = 0.05
 
     SAVE_PATH = None  # default is temp folder
 
@@ -719,6 +720,7 @@ def quantModel(self, model_id_or_path, trust_remote_code=False, dtype="auto", ne
             adapter=self.EORA,
             pack_impl="cpu",
             vram_strategy=self.VRAM_STRATEGY,
+            damp_percent=self.DAMP_PERCENT,
         )
 
         log.info(f"Quant config: {quantize_config}")