Release 0.16.0

- Bump versions
- Update a comment to poin to new PR
- Remove a test skip that is obsolete after huggingface#2579

Author: BenjaminBossan

setup.py

@@ -15,7 +15,7 @@
 from setuptools import find_packages, setup
 
 
-VERSION = "0.15.2.dev0"
+VERSION = "0.16.0"
 
 extras = {}
 extras["quality"] = [

src/peft/__init__.py

@@ -12,7 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-__version__ = "0.15.2.dev0"
+__version__ = "0.16.0"
 
 from .auto import (
     MODEL_TYPE_TO_PEFT_MODEL_MAPPING,

tests/test_gpu_examples.py

@@ -3807,7 +3807,7 @@ def setUp(self):
         # torchao breaks with fp16 and if a previous test uses fp16, transformers will set this env var, which affects
         # subsequent tests, therefore the env var needs to be cleared explicitly
         #
-        # TODO: remove this once https://github.com/huggingface/transformers/pull/34886 is merged
+        # TODO: remove this once https://github.com/huggingface/transformers/pull/37259 is merged
         os.environ.pop("ACCELERATE_MIXED_PRECISION", None)
 
     def tearDown(self):

tests/testing_common.py

@@ -1393,9 +1393,6 @@ def _test_peft_model_device_map(self, model_id, config_cls, config_kwargs):
     def _test_training_prompt_learning_tasks(self, model_id, config_cls, config_kwargs):
         if not issubclass(config_cls, PromptLearningConfig):
             return pytest.skip(f"Test not applicable for {config_cls}")
-        if ("gemma" in model_id.lower()) and (config_cls == PrefixTuningConfig):
-            # TODO might be caused by the 4d causal attention mask of gemma
-            return pytest.skip("Prefix tuning + gemma is currently failing")
 
         with hub_online_once(model_id):
             model = self.transformers_class.from_pretrained(model_id)