[Awq] Add peft support for AWQ (#28987)

* add peft support for AWQ

* Update src/transformers/quantizers/quantizer_awq.py

Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* fix

---------

Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

src/transformers/quantizers/quantizer_awq.py

@@ -11,8 +11,11 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+import importlib.metadata
 from typing import TYPE_CHECKING
 
+from packaging import version
+
 from .base import HfQuantizer
 
 
@@ -105,6 +108,6 @@ def is_serializable(self):
 
     @property
     def is_trainable(self):
-        # AWQ does not support neither QAT (Quantization Aware Training or PEFT yet.)
-        # TODO: if this is supported in the future, do a version check here.
-        return False
+        # AWQ supports PEFT fine-tuning from version 0.2.0
+        MIN_AWQ_VERSION_FOR_PEFT = "0.2.0"
+        return version.parse(importlib.metadata.version("autoawq")) >= version.parse(MIN_AWQ_VERSION_FOR_PEFT)