[`PEFT`] Final fixes (#26559)

* fix issues with PEFT

* logger warning futurewarning issues

* fixup

* adapt from suggestions

* oops

* rm test

src/transformers/integrations/peft.py

@@ -12,6 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 import inspect
+import warnings
 from typing import TYPE_CHECKING, Any, Dict, List, Optional, Union
 
 from ..utils import (
@@ -159,6 +160,10 @@ class PeftAdapterMixin:
                 "The one in `adapter_kwargs` will be used."
             )
 
+        # Override token with adapter_kwargs' token
+        if "token" in adapter_kwargs:
+            token = adapter_kwargs.pop("token")
+
         if peft_config is None:
             adapter_config_file = find_adapter_config_file(
                 peft_model_id,
@@ -381,8 +386,8 @@ class PeftAdapterMixin:
         return active_adapters
 
     def active_adapter(self) -> str:
-        logger.warning(
-            "The `active_adapter` method is deprecated and will be removed in a future version. ", FutureWarning
+        warnings.warn(
+            "The `active_adapter` method is deprecated and will be removed in a future version.", FutureWarning
         )
 
         return self.active_adapters()[0]

src/transformers/modeling_utils.py

@@ -1933,15 +1933,21 @@ class PreTrainedModel(nn.Module, ModuleUtilsMixin, GenerationMixin, PushToHubMix
         if token is not None:
             kwargs["token"] = token
 
+        _hf_peft_config_loaded = getattr(self, "_hf_peft_config_loaded", False)
+
         # Checks if the model has been loaded in 8-bit
-        if getattr(self, "is_loaded_in_8bit", False) and getattr(self, "is_8bit_serializable", False):
-            warnings.warn(
+        if (
+            getattr(self, "is_loaded_in_8bit", False)
+            and not getattr(self, "is_8bit_serializable", False)
+            and not _hf_peft_config_loaded
+        ):
+            raise ValueError(
                 "You are calling `save_pretrained` to a 8-bit converted model you may likely encounter unexepected"
-                " behaviors. If you want to save 8-bit models, make sure to have `bitsandbytes>0.37.2` installed.",
-                UserWarning,
+                " behaviors. If you want to save 8-bit models, make sure to have `bitsandbytes>0.37.2` installed."
             )
 
-        if getattr(self, "is_loaded_in_4bit", False):
+        # If the model has adapters attached, you can save the adapters
+        if getattr(self, "is_loaded_in_4bit", False) and not _hf_peft_config_loaded:
             raise NotImplementedError(
                 "You are calling `save_pretrained` on a 4-bit converted model. This is currently not supported"
             )
@@ -1982,8 +1988,6 @@ class PreTrainedModel(nn.Module, ModuleUtilsMixin, GenerationMixin, PushToHubMix
         if self._auto_class is not None:
             custom_object_save(self, save_directory, config=self.config)
 
-        _hf_peft_config_loaded = getattr(model_to_save, "_hf_peft_config_loaded", False)
-
         # Save the config
         if is_main_process:
             if not _hf_peft_config_loaded:

tests/peft_integration/test_peft_integration.py

@@ -312,6 +312,42 @@ class PeftIntegrationTester(unittest.TestCase, PeftTesterMixin):
                 # dummy generation
                 _ = peft_model.generate(input_ids=torch.LongTensor([[0, 1, 2, 3, 4, 5, 6, 7]]).to(torch_device))
 
+    @require_torch_gpu
+    def test_peft_save_quantized(self):
+        """
+        Simple test that tests the basic usage of PEFT model save_pretrained with quantized base models
+        """
+        # 4bit
+        for model_id in self.peft_test_model_ids:
+            for transformers_class in self.transformers_test_model_classes:
+                peft_model = transformers_class.from_pretrained(model_id, load_in_4bit=True, device_map="auto")
+
+                module = peft_model.model.decoder.layers[0].self_attn.v_proj
+                self.assertTrue(module.__class__.__name__ == "Linear4bit")
+                self.assertTrue(peft_model.hf_device_map is not None)
+
+                with tempfile.TemporaryDirectory() as tmpdirname:
+                    peft_model.save_pretrained(tmpdirname)
+                    self.assertTrue("adapter_model.bin" in os.listdir(tmpdirname))
+                    self.assertTrue("adapter_config.json" in os.listdir(tmpdirname))
+                    self.assertTrue("pytorch_model.bin" not in os.listdir(tmpdirname))
+
+        # 8-bit
+        for model_id in self.peft_test_model_ids:
+            for transformers_class in self.transformers_test_model_classes:
+                peft_model = transformers_class.from_pretrained(model_id, load_in_8bit=True, device_map="auto")
+
+                module = peft_model.model.decoder.layers[0].self_attn.v_proj
+                self.assertTrue(module.__class__.__name__ == "Linear8bitLt")
+                self.assertTrue(peft_model.hf_device_map is not None)
+
+                with tempfile.TemporaryDirectory() as tmpdirname:
+                    peft_model.save_pretrained(tmpdirname)
+
+                    self.assertTrue("adapter_model.bin" in os.listdir(tmpdirname))
+                    self.assertTrue("adapter_config.json" in os.listdir(tmpdirname))
+                    self.assertTrue("pytorch_model.bin" not in os.listdir(tmpdirname))
+
     def test_peft_pipeline(self):
         """
         Simple test that tests the basic usage of PEFT model + pipeline

tests/quantization/bnb/test_mixed_int8.py

@@ -263,13 +263,6 @@ class MixedInt8Test(BaseMixedInt8Test):
 
         self.assertEqual(self.tokenizer.decode(output_sequences[0], skip_special_tokens=True), self.EXPECTED_OUTPUT)
 
-    def test_warns_save_pretrained(self):
-        r"""
-        Test whether trying to save a model after converting it in 8-bit will throw a warning.
-        """
-        with self.assertWarns(UserWarning), tempfile.TemporaryDirectory() as tmpdirname:
-            self.model_8bit.save_pretrained(tmpdirname)
-
     def test_raise_if_config_and_load_in_8bit(self):
         r"""
         Test that loading the model with the config and `load_in_8bit` raises an error