Simplify registering "built-in" presets (#1818)

Instead of registering them with every class a preset should work with,
we just register them with the associated backbone. We can use that
to build `cls.preset` accessors for all library classes. E.g.

```python
keras_nlp.models.PaliGemmaTokenizer.presets
keras_nlp.models.Gpt2Backbone.presets
keras_nlp.models.TextClassifier.presets
```

keras_nlp/src/layers/preprocessing/audio_converter.py

@@ -16,10 +16,9 @@
     PreprocessingLayer,
 )
 from keras_nlp.src.utils.preset_utils import AUDIO_CONVERTER_CONFIG_FILE
+from keras_nlp.src.utils.preset_utils import builtin_presets
 from keras_nlp.src.utils.preset_utils import find_subclass
 from keras_nlp.src.utils.preset_utils import get_preset_loader
-from keras_nlp.src.utils.preset_utils import list_presets
-from keras_nlp.src.utils.preset_utils import list_subclasses
 from keras_nlp.src.utils.preset_utils import save_serialized_object
 from keras_nlp.src.utils.python_utils import classproperty
 
@@ -52,11 +51,8 @@ class AudioConverter(PreprocessingLayer):
 
     @classproperty
     def presets(cls):
-        """List built-in presets for a `Task` subclass."""
-        presets = list_presets(cls)
-        for subclass in list_subclasses(cls):
-            presets.update(subclass.presets)
-        return presets
+        """List built-in presets for an `AudioConverter` subclass."""
+        return builtin_presets(cls)
 
     @classmethod
     def from_preset(

keras_nlp/src/layers/preprocessing/audio_converter_test.py

@@ -28,9 +28,11 @@
 
 class AudioConverterTest(TestCase):
     def test_preset_accessors(self):
-        pali_gemma_presets = set(WhisperAudioConverter.presets.keys())
+        whisper_presets = set(WhisperAudioConverter.presets.keys())
         all_presets = set(AudioConverter.presets.keys())
-        self.assertContainsSubset(pali_gemma_presets, all_presets)
+        self.assertContainsSubset(whisper_presets, all_presets)
+        self.assertIn("whisper_tiny_en", whisper_presets)
+        self.assertIn("whisper_tiny_en", all_presets)
 
     @pytest.mark.large
     def test_from_preset(self):

keras_nlp/src/layers/preprocessing/image_converter.py

@@ -16,10 +16,9 @@
     PreprocessingLayer,
 )
 from keras_nlp.src.utils.preset_utils import IMAGE_CONVERTER_CONFIG_FILE
+from keras_nlp.src.utils.preset_utils import builtin_presets
 from keras_nlp.src.utils.preset_utils import find_subclass
 from keras_nlp.src.utils.preset_utils import get_preset_loader
-from keras_nlp.src.utils.preset_utils import list_presets
-from keras_nlp.src.utils.preset_utils import list_subclasses
 from keras_nlp.src.utils.preset_utils import save_serialized_object
 from keras_nlp.src.utils.python_utils import classproperty
 
@@ -55,11 +54,8 @@ class ImageConverter(PreprocessingLayer):
 
     @classproperty
     def presets(cls):
-        """List built-in presets for a `Task` subclass."""
-        presets = list_presets(cls)
-        for subclass in list_subclasses(cls):
-            presets.update(subclass.presets)
-        return presets
+        """List built-in presets for an `ImageConverter` subclass."""
+        return builtin_presets(cls)
 
     @classmethod
     def from_preset(

keras_nlp/src/layers/preprocessing/image_converter_test.py

@@ -33,6 +33,8 @@ def test_preset_accessors(self):
         pali_gemma_presets = set(PaliGemmaImageConverter.presets.keys())
         all_presets = set(ImageConverter.presets.keys())
         self.assertContainsSubset(pali_gemma_presets, all_presets)
+        self.assertIn("pali_gemma_3b_mix_224", pali_gemma_presets)
+        self.assertIn("pali_gemma_3b_mix_224", all_presets)
 
     @pytest.mark.large
     def test_from_preset(self):

keras_nlp/src/models/albert/__init__.py

@@ -14,7 +14,6 @@
 
 from keras_nlp.src.models.albert.albert_backbone import AlbertBackbone
 from keras_nlp.src.models.albert.albert_presets import backbone_presets
-from keras_nlp.src.models.albert.albert_tokenizer import AlbertTokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (AlbertBackbone, AlbertTokenizer))
+register_presets(backbone_presets, AlbertBackbone)

keras_nlp/src/models/backbone.py

@@ -20,9 +20,8 @@
 from keras_nlp.src.utils.keras_utils import assert_quantization_support
 from keras_nlp.src.utils.preset_utils import CONFIG_FILE
 from keras_nlp.src.utils.preset_utils import MODEL_WEIGHTS_FILE
+from keras_nlp.src.utils.preset_utils import builtin_presets
 from keras_nlp.src.utils.preset_utils import get_preset_loader
-from keras_nlp.src.utils.preset_utils import list_presets
-from keras_nlp.src.utils.preset_utils import list_subclasses
 from keras_nlp.src.utils.preset_utils import save_metadata
 from keras_nlp.src.utils.preset_utils import save_serialized_object
 from keras_nlp.src.utils.python_utils import classproperty
@@ -141,11 +140,8 @@ def from_config(cls, config):
 
     @classproperty
     def presets(cls):
-        """List built-in presets for a `Task` subclass."""
-        presets = list_presets(cls)
-        for subclass in list_subclasses(cls):
-            presets.update(subclass.presets)
-        return presets
+        """List built-in presets for a `Backbone` subclass."""
+        return builtin_presets(cls)
 
     @classmethod
     def from_preset(

keras_nlp/src/models/backbone_test.py

@@ -34,6 +34,12 @@ def test_preset_accessors(self):
         all_presets = set(Backbone.presets.keys())
         self.assertContainsSubset(bert_presets, all_presets)
         self.assertContainsSubset(gpt2_presets, all_presets)
+        self.assertIn("bert_tiny_en_uncased", bert_presets)
+        self.assertNotIn("bert_tiny_en_uncased", gpt2_presets)
+        self.assertIn("gpt2_base_en", gpt2_presets)
+        self.assertNotIn("gpt2_base_en", bert_presets)
+        self.assertIn("bert_tiny_en_uncased", all_presets)
+        self.assertIn("gpt2_base_en", all_presets)
 
     @pytest.mark.large
     def test_from_preset(self):

keras_nlp/src/models/bart/__init__.py

@@ -14,7 +14,6 @@
 
 from keras_nlp.src.models.bart.bart_backbone import BartBackbone
 from keras_nlp.src.models.bart.bart_presets import backbone_presets
-from keras_nlp.src.models.bart.bart_tokenizer import BartTokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (BartBackbone, BartTokenizer))
+register_presets(backbone_presets, BartBackbone)

keras_nlp/src/models/bert/__init__.py

@@ -14,10 +14,6 @@
 
 from keras_nlp.src.models.bert.bert_backbone import BertBackbone
 from keras_nlp.src.models.bert.bert_presets import backbone_presets
-from keras_nlp.src.models.bert.bert_presets import classifier_presets
-from keras_nlp.src.models.bert.bert_text_classifier import BertTextClassifier
-from keras_nlp.src.models.bert.bert_tokenizer import BertTokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (BertBackbone, BertTokenizer))
-register_presets(classifier_presets, (BertTextClassifier, BertTokenizer))
+register_presets(backbone_presets, BertBackbone)

keras_nlp/src/models/bert/bert_presets.py

@@ -129,9 +129,6 @@
         },
         "kaggle_handle": "kaggle://keras/bert/keras/bert_large_en/2",
     },
-}
-
-classifier_presets = {
     "bert_tiny_en_uncased_sst2": {
         "metadata": {
             "description": (
@@ -143,5 +140,5 @@
             "model_card": "https://github.com/google-research/bert/blob/master/README.md",
         },
         "kaggle_handle": "kaggle://keras/bert/keras/bert_tiny_en_uncased_sst2/4",
-    }
+    },
 }

keras_nlp/src/models/bloom/__init__.py

@@ -14,7 +14,6 @@
 
 from keras_nlp.src.models.bloom.bloom_backbone import BloomBackbone
 from keras_nlp.src.models.bloom.bloom_presets import backbone_presets
-from keras_nlp.src.models.bloom.bloom_tokenizer import BloomTokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (BloomBackbone, BloomTokenizer))
+register_presets(backbone_presets, BloomBackbone)

keras_nlp/src/models/deberta_v3/__init__.py

@@ -16,9 +16,6 @@
     DebertaV3Backbone,
 )
 from keras_nlp.src.models.deberta_v3.deberta_v3_presets import backbone_presets
-from keras_nlp.src.models.deberta_v3.deberta_v3_tokenizer import (
-    DebertaV3Tokenizer,
-)
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (DebertaV3Backbone, DebertaV3Tokenizer))
+register_presets(backbone_presets, DebertaV3Backbone)

keras_nlp/src/models/distil_bert/__init__.py

@@ -18,9 +18,6 @@
 from keras_nlp.src.models.distil_bert.distil_bert_presets import (
     backbone_presets,
 )
-from keras_nlp.src.models.distil_bert.distil_bert_tokenizer import (
-    DistilBertTokenizer,
-)
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (DistilBertBackbone, DistilBertTokenizer))
+register_presets(backbone_presets, DistilBertBackbone)

keras_nlp/src/models/electra/__init__.py

@@ -14,7 +14,6 @@
 
 from keras_nlp.src.models.electra.electra_backbone import ElectraBackbone
 from keras_nlp.src.models.electra.electra_presets import backbone_presets
-from keras_nlp.src.models.electra.electra_tokenizer import ElectraTokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (ElectraBackbone, ElectraTokenizer))
+register_presets(backbone_presets, ElectraBackbone)

keras_nlp/src/models/f_net/__init__.py

@@ -14,7 +14,6 @@
 
 from keras_nlp.src.models.f_net.f_net_backbone import FNetBackbone
 from keras_nlp.src.models.f_net.f_net_presets import backbone_presets
-from keras_nlp.src.models.f_net.f_net_tokenizer import FNetTokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (FNetBackbone, FNetTokenizer))
+register_presets(backbone_presets, FNetBackbone)

keras_nlp/src/models/falcon/__init__.py

@@ -14,7 +14,6 @@
 
 from keras_nlp.src.models.falcon.falcon_backbone import FalconBackbone
 from keras_nlp.src.models.falcon.falcon_presets import backbone_presets
-from keras_nlp.src.models.falcon.falcon_tokenizer import FalconTokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (FalconBackbone, FalconTokenizer))
+register_presets(backbone_presets, FalconBackbone)

keras_nlp/src/models/gemma/__init__.py

@@ -14,7 +14,6 @@
 
 from keras_nlp.src.models.gemma.gemma_backbone import GemmaBackbone
 from keras_nlp.src.models.gemma.gemma_presets import backbone_presets
-from keras_nlp.src.models.gemma.gemma_tokenizer import GemmaTokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (GemmaBackbone, GemmaTokenizer))
+register_presets(backbone_presets, GemmaBackbone)

keras_nlp/src/models/gpt2/__init__.py

@@ -14,7 +14,6 @@
 
 from keras_nlp.src.models.gpt2.gpt2_backbone import GPT2Backbone
 from keras_nlp.src.models.gpt2.gpt2_presets import backbone_presets
-from keras_nlp.src.models.gpt2.gpt2_tokenizer import GPT2Tokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (GPT2Backbone, GPT2Tokenizer))
+register_presets(backbone_presets, GPT2Backbone)

keras_nlp/src/models/llama/__init__.py

@@ -14,7 +14,6 @@
 
 from keras_nlp.src.models.llama.llama_backbone import LlamaBackbone
 from keras_nlp.src.models.llama.llama_presets import backbone_presets
-from keras_nlp.src.models.llama.llama_tokenizer import LlamaTokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (LlamaBackbone, LlamaTokenizer))
+register_presets(backbone_presets, LlamaBackbone)

keras_nlp/src/models/llama3/__init__.py

@@ -14,7 +14,6 @@
 
 from keras_nlp.src.models.llama3.llama3_backbone import Llama3Backbone
 from keras_nlp.src.models.llama3.llama3_presets import backbone_presets
-from keras_nlp.src.models.llama3.llama3_tokenizer import Llama3Tokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (Llama3Backbone, Llama3Tokenizer))
+register_presets(backbone_presets, Llama3Backbone)

keras_nlp/src/models/mistral/__init__.py

@@ -14,7 +14,6 @@
 
 from keras_nlp.src.models.mistral.mistral_backbone import MistralBackbone
 from keras_nlp.src.models.mistral.mistral_presets import backbone_presets
-from keras_nlp.src.models.mistral.mistral_tokenizer import MistralTokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (MistralBackbone, MistralTokenizer))
+register_presets(backbone_presets, MistralBackbone)

keras_nlp/src/models/opt/__init__.py

@@ -14,7 +14,6 @@
 
 from keras_nlp.src.models.opt.opt_backbone import OPTBackbone
 from keras_nlp.src.models.opt.opt_presets import backbone_presets
-from keras_nlp.src.models.opt.opt_tokenizer import OPTTokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (OPTBackbone, OPTTokenizer))
+register_presets(backbone_presets, OPTBackbone)

keras_nlp/src/models/pali_gemma/__init__.py

@@ -15,9 +15,6 @@
     PaliGemmaBackbone,
 )
 from keras_nlp.src.models.pali_gemma.pali_gemma_presets import backbone_presets
-from keras_nlp.src.models.pali_gemma.pali_gemma_tokenizer import (
-    PaliGemmaTokenizer,
-)
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (PaliGemmaBackbone, PaliGemmaTokenizer))
+register_presets(backbone_presets, PaliGemmaBackbone)

keras_nlp/src/models/phi3/__init__.py

@@ -14,7 +14,6 @@
 
 from keras_nlp.src.models.phi3.phi3_backbone import Phi3Backbone
 from keras_nlp.src.models.phi3.phi3_presets import backbone_presets
-from keras_nlp.src.models.phi3.phi3_tokenizer import Phi3Tokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (Phi3Backbone, Phi3Tokenizer))
+register_presets(backbone_presets, Phi3Backbone)

keras_nlp/src/models/preprocessor.py

@@ -19,10 +19,9 @@
     PreprocessingLayer,
 )
 from keras_nlp.src.utils.preset_utils import PREPROCESSOR_CONFIG_FILE
+from keras_nlp.src.utils.preset_utils import builtin_presets
 from keras_nlp.src.utils.preset_utils import find_subclass
 from keras_nlp.src.utils.preset_utils import get_preset_loader
-from keras_nlp.src.utils.preset_utils import list_presets
-from keras_nlp.src.utils.preset_utils import list_subclasses
 from keras_nlp.src.utils.preset_utils import save_serialized_object
 from keras_nlp.src.utils.python_utils import classproperty
 
@@ -120,13 +119,8 @@ def from_config(cls, config):
 
     @classproperty
     def presets(cls):
-        presets = list_presets(cls)
-        # We can also load backbone presets.
-        if cls.tokenizer_cls is not None:
-            presets.update(cls.tokenizer_cls.presets)
-        for subclass in list_subclasses(cls):
-            presets.update(subclass.presets)
-        return presets
+        """List built-in presets for a `Preprocessor` subclass."""
+        return builtin_presets(cls)
 
     @classmethod
     def from_preset(

keras_nlp/src/models/preprocessor_test.py

@@ -46,6 +46,12 @@ def test_preset_accessors(self):
         all_presets = set(Preprocessor.presets.keys())
         self.assertContainsSubset(bert_presets, all_presets)
         self.assertContainsSubset(gpt2_presets, all_presets)
+        self.assertIn("bert_tiny_en_uncased", bert_presets)
+        self.assertNotIn("bert_tiny_en_uncased", gpt2_presets)
+        self.assertIn("gpt2_base_en", gpt2_presets)
+        self.assertNotIn("gpt2_base_en", bert_presets)
+        self.assertIn("bert_tiny_en_uncased", all_presets)
+        self.assertIn("gpt2_base_en", all_presets)
 
     @pytest.mark.large
     def test_from_preset(self):

keras_nlp/src/models/roberta/__init__.py

@@ -14,7 +14,6 @@
 
 from keras_nlp.src.models.roberta.roberta_backbone import RobertaBackbone
 from keras_nlp.src.models.roberta.roberta_presets import backbone_presets
-from keras_nlp.src.models.roberta.roberta_tokenizer import RobertaTokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (RobertaBackbone, RobertaTokenizer))
+register_presets(backbone_presets, RobertaBackbone)

keras_nlp/src/models/t5/__init__.py

@@ -14,7 +14,6 @@
 
 from keras_nlp.src.models.t5.t5_backbone import T5Backbone
 from keras_nlp.src.models.t5.t5_presets import backbone_presets
-from keras_nlp.src.models.t5.t5_tokenizer import T5Tokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (T5Backbone, T5Tokenizer))
+register_presets(backbone_presets, T5Backbone)

keras_nlp/src/models/task.py

@@ -24,10 +24,9 @@
 from keras_nlp.src.utils.pipeline_model import PipelineModel
 from keras_nlp.src.utils.preset_utils import TASK_CONFIG_FILE
 from keras_nlp.src.utils.preset_utils import TASK_WEIGHTS_FILE
+from keras_nlp.src.utils.preset_utils import builtin_presets
 from keras_nlp.src.utils.preset_utils import find_subclass
 from keras_nlp.src.utils.preset_utils import get_preset_loader
-from keras_nlp.src.utils.preset_utils import list_presets
-from keras_nlp.src.utils.preset_utils import list_subclasses
 from keras_nlp.src.utils.preset_utils import save_serialized_object
 from keras_nlp.src.utils.python_utils import classproperty
 
@@ -133,13 +132,7 @@ def from_config(cls, config):
     @classproperty
     def presets(cls):
         """List built-in presets for a `Task` subclass."""
-        presets = list_presets(cls)
-        # We can also load backbone presets.
-        if cls.backbone_cls is not None:
-            presets.update(cls.backbone_cls.presets)
-        for subclass in list_subclasses(cls):
-            presets.update(subclass.presets)
-        return presets
+        return builtin_presets(cls)
 
     @classmethod
     def from_preset(

keras_nlp/src/models/task_test.py

@@ -62,6 +62,12 @@ def test_preset_accessors(self):
         all_presets = set(Task.presets.keys())
         self.assertContainsSubset(bert_presets, all_presets)
         self.assertContainsSubset(gpt2_presets, all_presets)
+        self.assertIn("bert_tiny_en_uncased", bert_presets)
+        self.assertNotIn("bert_tiny_en_uncased", gpt2_presets)
+        self.assertIn("gpt2_base_en", gpt2_presets)
+        self.assertNotIn("gpt2_base_en", bert_presets)
+        self.assertIn("bert_tiny_en_uncased", all_presets)
+        self.assertIn("gpt2_base_en", all_presets)
 
     @pytest.mark.large
     def test_from_preset(self):

keras_nlp/src/models/whisper/__init__.py

@@ -14,7 +14,6 @@
 
 from keras_nlp.src.models.whisper.whisper_backbone import WhisperBackbone
 from keras_nlp.src.models.whisper.whisper_presets import backbone_presets
-from keras_nlp.src.models.whisper.whisper_tokenizer import WhisperTokenizer
 from keras_nlp.src.utils.preset_utils import register_presets
 
-register_presets(backbone_presets, (WhisperBackbone, WhisperTokenizer))
+register_presets(backbone_presets, WhisperBackbone)