Deprecate models (#24787)

* Deprecate some models

* Fix imports

* Fix inits too

* Remove tests

* Add deprecated banner to documentation

* Remove from init

* Fix auto classes

* Style

* Remote upgrade strategy 1

* Remove site package cache

* Revert this part

* Fix typo...

* Update utils

* Update docs/source/en/model_doc/bort.md
Co-authored-by: Lysandre Debut <lysandre.debut@reseau.eseo.fr>

* Address review comments

* With all files saved

---------
Co-authored-by: Lysandre Debut <lysandre.debut@reseau.eseo.fr>

src/transformers/models/retribert/__init__.py → src/transformers/models/deprecated/retribert/__init__.py

@@ -14,7 +14,7 @@
 
 from typing import TYPE_CHECKING
 
-from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tokenizers_available, is_torch_available
+from ....utils import OptionalDependencyNotAvailable, _LazyModule, is_tokenizers_available, is_torch_available
 
 
 _import_structure = {

src/transformers/models/retribert/configuration_retribert.py → src/transformers/models/deprecated/retribert/configuration_retribert.py

@@ -14,8 +14,8 @@
 # limitations under the License.
 """ RetriBERT model configuration"""
 
-from ...configuration_utils import PretrainedConfig
-from ...utils import logging
+from ....configuration_utils import PretrainedConfig
+from ....utils import logging
 
 
 logger = logging.get_logger(__name__)

src/transformers/models/retribert/modeling_retribert.py → src/transformers/models/deprecated/retribert/modeling_retribert.py

@@ -24,9 +24,9 @@ import torch
 import torch.utils.checkpoint as checkpoint
 from torch import nn
 
-from ...modeling_utils import PreTrainedModel
-from ...utils import add_start_docstrings, logging
-from ..bert.modeling_bert import BertModel
+from ....modeling_utils import PreTrainedModel
+from ....utils import add_start_docstrings, logging
+from ...bert.modeling_bert import BertModel
 from .configuration_retribert import RetriBertConfig
 
 

src/transformers/models/retribert/tokenization_retribert.py → src/transformers/models/deprecated/retribert/tokenization_retribert.py

@@ -19,8 +19,8 @@ import os
 import unicodedata
 from typing import List, Optional, Tuple
 
-from ...tokenization_utils import PreTrainedTokenizer, _is_control, _is_punctuation, _is_whitespace
-from ...utils import logging
+from ....tokenization_utils import PreTrainedTokenizer, _is_control, _is_punctuation, _is_whitespace
+from ....utils import logging
 
 
 logger = logging.get_logger(__name__)

src/transformers/models/retribert/tokenization_retribert_fast.py → src/transformers/models/deprecated/retribert/tokenization_retribert_fast.py

@@ -19,8 +19,8 @@ from typing import List, Optional, Tuple
 
 from tokenizers import normalizers
 
-from ...tokenization_utils_fast import PreTrainedTokenizerFast
-from ...utils import logging
+from ....tokenization_utils_fast import PreTrainedTokenizerFast
+from ....utils import logging
 from .tokenization_retribert import RetriBertTokenizer
 
 

src/transformers/models/tapex/__init__.py → src/transformers/models/deprecated/tapex/__init__.py

@@ -13,7 +13,7 @@
 # limitations under the License.
 from typing import TYPE_CHECKING
 
-from ...file_utils import _LazyModule
+from ....utils import _LazyModule
 
 
 _import_structure = {"tokenization_tapex": ["TapexTokenizer"]}

src/transformers/models/tapex/tokenization_tapex.py → src/transformers/models/deprecated/tapex/tokenization_tapex.py

@@ -22,10 +22,10 @@ from typing import Dict, List, Optional, Tuple, Union
 
 import regex as re
 
-from ...file_utils import ExplicitEnum, PaddingStrategy, TensorType, add_end_docstrings, is_pandas_available
-from ...tokenization_utils import AddedToken, PreTrainedTokenizer
-from ...tokenization_utils_base import ENCODE_KWARGS_DOCSTRING, BatchEncoding, TextInput, TruncationStrategy
-from ...utils import logging
+from ....file_utils import ExplicitEnum, PaddingStrategy, TensorType, add_end_docstrings, is_pandas_available
+from ....tokenization_utils import AddedToken, PreTrainedTokenizer
+from ....tokenization_utils_base import ENCODE_KWARGS_DOCSTRING, BatchEncoding, TextInput, TruncationStrategy
+from ....utils import logging
 
 
 if is_pandas_available():

src/transformers/models/trajectory_transformer/__init__.py → src/transformers/models/deprecated/trajectory_transformer/__init__.py

@@ -13,7 +13,7 @@
 # limitations under the License.
 from typing import TYPE_CHECKING
 
-from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available
+from ....utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available
 
 
 _import_structure = {

src/transformers/models/trajectory_transformer/configuration_trajectory_transformer.py → src/transformers/models/deprecated/trajectory_transformer/configuration_trajectory_transformer.py

@@ -14,8 +14,8 @@
 # limitations under the License.
 """ TrajectoryTransformer model configuration"""
 
-from ...configuration_utils import PretrainedConfig
-from ...utils import logging
+from ....configuration_utils import PretrainedConfig
+from ....utils import logging
 
 
 logger = logging.get_logger(__name__)

src/transformers/models/trajectory_transformer/modeling_trajectory_transformer.py → src/transformers/models/deprecated/trajectory_transformer/modeling_trajectory_transformer.py

@@ -25,8 +25,8 @@ import torch.utils.checkpoint
 from torch import nn
 from torch.nn import functional as F
 
-from ...modeling_utils import PreTrainedModel
-from ...utils import (
+from ....modeling_utils import PreTrainedModel
+from ....utils import (
     ModelOutput,
     add_start_docstrings,
     add_start_docstrings_to_model_forward,

src/transformers/models/van/__init__.py → src/transformers/models/deprecated/van/__init__.py

@@ -13,7 +13,7 @@
 # limitations under the License.
 from typing import TYPE_CHECKING
 
-from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available, is_vision_available
+from ....utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available, is_vision_available
 
 
 _import_structure = {"configuration_van": ["VAN_PRETRAINED_CONFIG_ARCHIVE_MAP", "VanConfig"]}

src/transformers/models/van/configuration_van.py → src/transformers/models/deprecated/van/configuration_van.py

@@ -14,8 +14,8 @@
 # limitations under the License.
 """ VAN model configuration"""
 
-from ...configuration_utils import PretrainedConfig
-from ...utils import logging
+from ....configuration_utils import PretrainedConfig
+from ....utils import logging
 
 
 logger = logging.get_logger(__name__)

src/transformers/models/van/modeling_van.py → src/transformers/models/deprecated/van/modeling_van.py

@@ -23,14 +23,14 @@ import torch.utils.checkpoint
 from torch import nn
 from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss
 
-from ...activations import ACT2FN
-from ...modeling_outputs import (
+from ....activations import ACT2FN
+from ....modeling_outputs import (
     BaseModelOutputWithNoAttention,
     BaseModelOutputWithPoolingAndNoAttention,
     ImageClassifierOutputWithNoAttention,
 )
-from ...modeling_utils import PreTrainedModel
-from ...utils import add_code_sample_docstrings, add_start_docstrings, add_start_docstrings_to_model_forward, logging
+from ....modeling_utils import PreTrainedModel
+from ....utils import add_code_sample_docstrings, add_start_docstrings, add_start_docstrings_to_model_forward, logging
 from .configuration_van import VanConfig
 
 

src/transformers/utils/dummy_pt_objects.py

@@ -2306,6 +2306,109 @@ class DeiTPreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
+MCTCT_PRETRAINED_MODEL_ARCHIVE_LIST = None
+
+
+class MCTCTForCTC(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class MCTCTModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class MCTCTPreTrainedModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class MMBTForClassification(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class MMBTModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class ModalEmbeddings(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+RETRIBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
+
+
+class RetriBertModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class RetriBertPreTrainedModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+TRAJECTORY_TRANSFORMER_PRETRAINED_MODEL_ARCHIVE_LIST = None
+
+
+class TrajectoryTransformerModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class TrajectoryTransformerPreTrainedModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+VAN_PRETRAINED_MODEL_ARCHIVE_LIST = None
+
+
+class VanForImageClassification(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class VanModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class VanPreTrainedModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
 DETA_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
 
@@ -4484,30 +4587,6 @@ class MBartPreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-MCTCT_PRETRAINED_MODEL_ARCHIVE_LIST = None
-
-
-class MCTCTForCTC(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MCTCTModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MCTCTPreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
 MEGA_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
 
@@ -4664,27 +4743,6 @@ class MgpstrPreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-class MMBTForClassification(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MMBTModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class ModalEmbeddings(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
 MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
 
@@ -6088,23 +6146,6 @@ class ResNetPreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-RETRIBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
-
-
-class RetriBertModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class RetriBertPreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
 ROBERTA_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
 
@@ -7030,23 +7071,6 @@ class TimmBackbone(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-TRAJECTORY_TRANSFORMER_PRETRAINED_MODEL_ARCHIVE_LIST = None
-
-
-class TrajectoryTransformerModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class TrajectoryTransformerPreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
 TRANSFO_XL_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
 
@@ -7276,30 +7300,6 @@ class UperNetPreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-VAN_PRETRAINED_MODEL_ARCHIVE_LIST = None
-
-
-class VanForImageClassification(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class VanModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class VanPreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
 VIDEOMAE_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
 

src/transformers/utils/dummy_tokenizers_objects.py

@@ -107,6 +107,13 @@ class DebertaV2TokenizerFast(metaclass=DummyObject):
         requires_backends(self, ["tokenizers"])
 
 
+class RetriBertTokenizerFast(metaclass=DummyObject):
+    _backends = ["tokenizers"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["tokenizers"])
+
+
 class DistilBertTokenizerFast(metaclass=DummyObject):
     _backends = ["tokenizers"]
 
@@ -331,13 +338,6 @@ class RemBertTokenizerFast(metaclass=DummyObject):
         requires_backends(self, ["tokenizers"])
 
 
-class RetriBertTokenizerFast(metaclass=DummyObject):
-    _backends = ["tokenizers"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["tokenizers"])
-
-
 class RobertaTokenizerFast(metaclass=DummyObject):
     _backends = ["tokenizers"]
 

tests/models/bort/test_modeling_bort.py

-# coding=utf-8
-# Copyright 2020 The HuggingFace Team. All rights reserved.
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#     http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# 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 unittest
-
-from transformers import is_torch_available
-from transformers.testing_utils import require_sentencepiece, require_tokenizers, require_torch, slow, torch_device
-
-
-if is_torch_available():
-    import torch
-
-    from transformers import AutoModel
-
-
-@require_torch
-@require_sentencepiece
-@require_tokenizers
-class BortIntegrationTest(unittest.TestCase):
-    @slow
-    def test_output_embeds_base_model(self):
-        model = AutoModel.from_pretrained("amazon/bort")
-        model.to(torch_device)
-
-        input_ids = torch.tensor(
-            [[0, 18077, 4082, 7804, 8606, 6195, 2457, 3321, 11, 10489, 16, 269, 2579, 328, 2]],
-            device=torch_device,
-            dtype=torch.long,
-        )  # Schloß Nymphenburg in Munich is really nice!
-        output = model(input_ids)["last_hidden_state"]
-        expected_shape = torch.Size((1, 15, 1024))
-        self.assertEqual(output.shape, expected_shape)
-        # compare the actual values for a slice.
-        expected_slice = torch.tensor(
-            [[[-0.0349, 0.0436, -1.8654], [-0.6964, 0.0835, -1.7393], [-0.9819, 0.2956, -0.2868]]],
-            device=torch_device,
-            dtype=torch.float,
-        )
-        self.assertTrue(torch.allclose(output[:, :3, :3], expected_slice, atol=1e-4))

tests/models/bort/test_modeling_tf_bort.py

-# coding=utf-8
-# Copyright 2020 The HuggingFace Team. All rights reserved.
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#     http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# 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.
-
-from __future__ import annotations
-
-import unittest
-
-from transformers import is_tf_available
-from transformers.testing_utils import require_sentencepiece, require_tf, require_tokenizers, slow
-
-
-if is_tf_available():
-    import numpy as np
-    import tensorflow as tf
-
-    from transformers import TFAutoModel
-
-
-@require_tf
-@require_sentencepiece
-@require_tokenizers
-class TFBortIntegrationTest(unittest.TestCase):
-    @slow
-    def test_output_embeds_base_model(self):
-        model = TFAutoModel.from_pretrained("amazon/bort")
-
-        input_ids = tf.convert_to_tensor(
-            [[0, 18077, 4082, 7804, 8606, 6195, 2457, 3321, 11, 10489, 16, 269, 2579, 328, 2]],
-            dtype=tf.int32,
-        )  # Schloß Nymphenburg in Munich is really nice!
-
-        output = model(input_ids)["last_hidden_state"]
-        expected_shape = tf.TensorShape((1, 15, 1024))
-        self.assertEqual(output.shape, expected_shape)
-        # compare the actual values for a slice.
-        expected_slice = tf.convert_to_tensor(
-            [[[-0.0349, 0.0436, -1.8654], [-0.6964, 0.0835, -1.7393], [-0.9819, 0.2956, -0.2868]]],
-            dtype=tf.float32,
-        )
-
-        self.assertTrue(np.allclose(output[:, :3, :3].numpy(), expected_slice.numpy(), atol=1e-4))