Rename AdapterType -> AdapterMethod. Test fixes.

src/adapters/__init__.py

@@ -84,7 +84,7 @@
         "Seq2SeqLMHead",
         "TaggingHead",
     ],
-    "interface": ["AdapterModelInterface"],
+    "interface": ["AdapterMethod", "AdapterModelInterface"],
     "methods.adapter_layer_base": ["AdapterLayerBase", "ComposableAdapterLayerBase"],
     "model_mixin": [
         "EmbeddingAdaptersMixin",
@@ -199,7 +199,7 @@
         Seq2SeqLMHead,
         TaggingHead,
     )
-    from .interface import AdapterModelInterface
+    from .interface import AdapterMethod, AdapterModelInterface
     from .methods.adapter_layer_base import AdapterLayerBase, ComposableAdapterLayerBase
     from .model_mixin import (
         EmbeddingAdaptersMixin,

src/adapters/interface.py

@@ -2,9 +2,9 @@
 from typing import List, Optional
 
 
-class AdapterType:
+class AdapterMethod:
     """
-    Enum for the different adapter types.
+    Enum of all supported adapter method types.
     """
 
     bottleneck = "bottleneck"
@@ -25,9 +25,9 @@ def get_from_config(config) -> List[str]:
             str: The adapter type.
         """
         if config.architecture is None:
-            return [AdapterType.bottleneck]
+            return [AdapterMethod.bottleneck]
         elif config.architecture == "union":
-            return [AdapterType.get_from_config(sub_config) for sub_config in config.configs]
+            return [AdapterMethod.get_from_config(sub_config) for sub_config in config.configs]
         else:
             return [config.architecture]
 

src/adapters/model_mixin.py

@@ -19,7 +19,7 @@
 from .configuration import ADAPTER_CONFIG_MAP, AdapterConfig, AdapterFusionConfig, BnConfig
 from .context import AdapterSetup, ForwardContext
 from .hub_mixin import PushAdapterToHubMixin
-from .interface import AdapterModelInterface, AdapterType
+from .interface import AdapterMethod, AdapterModelInterface
 from .loading import AdapterFusionLoader, AdapterLoader, PredictionHeadLoader, WeightsLoader
 from .methods import METHOD_INIT_MAPPING
 from .methods.adapter_layer_base import AdapterLayerBase
@@ -484,16 +484,16 @@ def supports_adapter(self, type_or_config: Union[str, AdapterConfig]) -> bool:
             bool: True if the adapter type is supported, False otherwise.
         """
         if isinstance(type_or_config, AdapterConfig):
-            types = AdapterType.get_from_config(type_or_config)
+            types = AdapterMethod.get_from_config(type_or_config)
         else:
             types = [type_or_config]
 
         supported = []
         for _type in types:
             if getattr(self.base_model, "adapter_interface", None) is not None:
                 supported.append(_type in self.base_model.adapter_interface.adapter_types)
-            elif _type == AdapterType.prompt_tuning:
-                supported.append(self.support_prompt_tuning)
+            elif _type == AdapterMethod.prompt_tuning:
+                supported.append(self.base_model.support_prompt_tuning)
             else:
                 supported.append(True)
         return all(supported)
@@ -624,7 +624,7 @@ def add_adapter(self, adapter_name: str, config=None, overwrite_ok: bool = False
         """
         config = AdapterConfig.load(config)  # ensure config is ok and up-to-date
         # check if config is valid for this model
-        config_or_type = config or AdapterType.bottleneck
+        config_or_type = config or AdapterMethod.bottleneck
         if not self.supports_adapter(config_or_type):
             raise ValueError(f"Adapter config or type '{config_or_type}' is not supported by this model.")
         # In case adapter already exists and we allow overwriting, explicitly delete the existing one first

tests/methods/test_adapter_common.py

@@ -12,8 +12,6 @@
     DoubleSeqBnConfig,
     DoubleSeqBnInvConfig,
     Fuse,
-    InvertibleAdaptersMixin,
-    InvertibleAdaptersWrapperMixin,
     MAMConfig,
     SeqBnConfig,
     SeqBnInvConfig,
@@ -72,7 +70,7 @@ def test_delete_adapter(self):
     def test_add_adapter_with_invertible(self):
         model = self.get_model().base_model
         model.eval()
-        if not isinstance(model, InvertibleAdaptersMixin) and not isinstance(model, InvertibleAdaptersWrapperMixin):
+        if not model.supports_adapter("invertible"):
             self.skipTest("Model does not support invertible adapters.")
 
         for adapter_config in [SeqBnInvConfig(), DoubleSeqBnInvConfig()]:
@@ -120,7 +118,7 @@ def test_delete_adapter_with_invertible(self):
         """Tests if the invertible adapters are deleted correctly."""
         model = self.get_model().base_model
         model.eval()
-        if not isinstance(model, InvertibleAdaptersMixin) and not isinstance(model, InvertibleAdaptersWrapperMixin):
+        if not model.supports_adapter("invertible"):
             self.skipTest("Model does not support invertible adapters.")
 
         # iterate through all adapter invertible adapter configs
@@ -245,6 +243,8 @@ def test_model_config_serialization(self):
         """
         for k, v in ADAPTER_CONFIG_MAP.items():
             model = self.get_model()
+            if not model.supports_adapter(v):
+                continue
             # HACK: reduce the reduction factor such that
             # the small test model can have a phm_dim of 4
             if hasattr(v, "phm_layer") and v.phm_layer:
@@ -257,15 +257,19 @@ def test_model_adapter_summary(self):
         # count model parameters before
         model = self.get_model()
         model_no_params = sum(p.numel() for p in model.parameters())
+        added = []
         for k, v in ADAPTER_CONFIG_MAP.items():
+            if not model.supports_adapter(v):
+                continue
             # HACK: reduce the reduction factor such that
             # the small test model can have a phm_dim of 4
             if hasattr(v, "phm_layer") and v.phm_layer:
                 v = v.__class__(reduction_factor=4)
             model.add_adapter(k, config=v)
+            added.append(k)
         summary = model.adapter_summary(as_dict=True)
-        self.assertEqual(len(ADAPTER_CONFIG_MAP) + 1, len(summary))
-        for name in ADAPTER_CONFIG_MAP.keys():
+        self.assertEqual(len(added) + 1, len(summary))
+        for name in added:
             self.assertTrue(any([row["name"] == name for row in summary]))
         self.assertEqual(model_no_params, summary[-1]["#param"])
 

tests/methods/test_lora.py

@@ -2,7 +2,7 @@
 
 import torch
 
-from adapters import LoRAConfig
+from adapters import AdapterConfig, LoRAConfig
 from adapters.methods.lora import LoRALayer
 from transformers.testing_utils import require_torch
 
@@ -24,21 +24,30 @@ def test_merging_with_other_adapters(self):
         model.add_adapter("lora", config="lora")
 
         # Add different adapters
-        model.add_adapter("bottleneck", config="seq_bn")
-        model.add_adapter("prompt", config="prompt_tuning")
-        model.add_adapter("prefix", config="prefix_tuning")
-        model.add_adapter("ia3", config="ia3")
-        model.add_adapter("unipelt", config="unipelt")
-        model.add_adapter("mam", config="mam")
-        model.add_adapter("compacter", config="compacter[phm_dim=2, reduction_factor=8]")
+        adapter_methods = [
+            "seq_bn",
+            "prompt_tuning",
+            "prefix_tuning",
+            "ia3",
+            "unipelt",
+            "mam",
+            "compacter[phm_dim=2, reduction_factor=8]",
+        ]
+
+        for adapter_method in adapter_methods:
+            config = AdapterConfig.load(adapter_method)
+            if model.supports_adapter(config):
+                model.add_adapter(adapter_method, config=config)
 
         # Merging adapters with different architectures with LoRA should raise a ValueError
-        for adapter_architecture in ["bottleneck", "prompt", "prefix", "ia3", "unipelt", "mam", "compacter"]:
-            with self.subTest(adapter_architecture=adapter_architecture):
+        for adapter_method in adapter_methods:
+            with self.subTest(adapter_architecture=adapter_method):
+                if adapter_method not in model.adapters_config:
+                    continue
                 with self.assertRaises(ValueError):
                     model.average_adapter(
-                        adapter_name=f"average_lora_{adapter_architecture}",
-                        adapter_list=[adapter_architecture, "lora"],
+                        adapter_name=f"average_lora_{adapter_method}",
+                        adapter_list=[adapter_method, "lora"],
                         weights=[0.5, 0.5],
                         combine_strategy="linear",
                     )