Deprecate low use models (#30781)

* Deprecate models
- graphormer
- time_series_transformer
- xlm_prophetnet
- qdqbert
- nat
- ernie_m
- tvlt
- nezha
- mega
- jukebox
- vit_hybrid
- x_clip
- deta
- speech_to_text_2
- efficientformer
- realm
- gptsan_japanese

* Fix up

* Fix speech2text2 imports

* Make sure message isn't indented

* Fix docstrings

* Correctly map for deprecated models from model_type

* Uncomment out

* Add back time series transformer and x-clip

* Import fix and fix-up

* Fix up with updated ruff

src/transformers/models/vit_hybrid/__init__.py → src/transformers/models/deprecated/vit_hybrid/__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_vit_hybrid": ["ViTHybridConfig"]}

src/transformers/models/vit_hybrid/configuration_vit_hybrid.py → src/transformers/models/deprecated/vit_hybrid/configuration_vit_hybrid.py

@@ -14,10 +14,10 @@
 # limitations under the License.
 """ViT Hybrid model configuration"""
 
-from ...configuration_utils import PretrainedConfig
-from ...utils import logging
-from ..auto.configuration_auto import CONFIG_MAPPING
-from ..bit import BitConfig
+from ....configuration_utils import PretrainedConfig
+from ....utils import logging
+from ...auto.configuration_auto import CONFIG_MAPPING
+from ...bit import BitConfig
 
 
 logger = logging.get_logger(__name__)

src/transformers/models/vit_hybrid/image_processing_vit_hybrid.py → src/transformers/models/deprecated/vit_hybrid/image_processing_vit_hybrid.py

@@ -18,14 +18,14 @@ from typing import Dict, List, Optional, Union
 
 import numpy as np
 
-from ...image_processing_utils import BaseImageProcessor, BatchFeature, get_size_dict
-from ...image_transforms import (
+from ....image_processing_utils import BaseImageProcessor, BatchFeature, get_size_dict
+from ....image_transforms import (
     convert_to_rgb,
     get_resize_output_image_size,
     resize,
     to_channel_dimension_format,
 )
-from ...image_utils import (
+from ....image_utils import (
     OPENAI_CLIP_MEAN,
     OPENAI_CLIP_STD,
     ChannelDimension,
@@ -39,7 +39,7 @@ from ...image_utils import (
     validate_kwargs,
     validate_preprocess_arguments,
 )
-from ...utils import TensorType, is_vision_available, logging
+from ....utils import TensorType, is_vision_available, logging
 
 
 logger = logging.get_logger(__name__)

src/transformers/models/vit_hybrid/modeling_vit_hybrid.py → src/transformers/models/deprecated/vit_hybrid/modeling_vit_hybrid.py

@@ -23,12 +23,12 @@ import torch.utils.checkpoint
 from torch import nn
 from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss
 
-from ...activations import ACT2FN
-from ...modeling_outputs import BaseModelOutput, BaseModelOutputWithPooling, ImageClassifierOutput
-from ...modeling_utils import PreTrainedModel
-from ...pytorch_utils import find_pruneable_heads_and_indices, prune_linear_layer
-from ...utils import add_code_sample_docstrings, add_start_docstrings, add_start_docstrings_to_model_forward, logging
-from ...utils.backbone_utils import load_backbone
+from ....activations import ACT2FN
+from ....modeling_outputs import BaseModelOutput, BaseModelOutputWithPooling, ImageClassifierOutput
+from ....modeling_utils import PreTrainedModel
+from ....pytorch_utils import find_pruneable_heads_and_indices, prune_linear_layer
+from ....utils import add_code_sample_docstrings, add_start_docstrings, add_start_docstrings_to_model_forward, logging
+from ....utils.backbone_utils import load_backbone
 from .configuration_vit_hybrid import ViTHybridConfig
 
 

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

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

src/transformers/models/xlm_prophetnet/configuration_xlm_prophetnet.py → src/transformers/models/deprecated/xlm_prophetnet/configuration_xlm_prophetnet.py

@@ -16,8 +16,8 @@
 
 from typing import Callable, Optional, Union
 
-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/xlm_prophetnet/modeling_xlm_prophetnet.py → src/transformers/models/deprecated/xlm_prophetnet/modeling_xlm_prophetnet.py

@@ -25,10 +25,10 @@ import torch.utils.checkpoint
 from torch import Tensor, nn
 from torch.nn import LayerNorm
 
-from ...activations import ACT2FN
-from ...modeling_outputs import BaseModelOutput
-from ...modeling_utils import PreTrainedModel
-from ...utils import (
+from ....activations import ACT2FN
+from ....modeling_outputs import BaseModelOutput
+from ....modeling_utils import PreTrainedModel
+from ....utils import (
     ModelOutput,
     add_start_docstrings,
     add_start_docstrings_to_model_forward,

src/transformers/models/xlm_prophetnet/tokenization_xlm_prophetnet.py → src/transformers/models/deprecated/xlm_prophetnet/tokenization_xlm_prophetnet.py

@@ -18,8 +18,8 @@ import os
 from shutil import copyfile
 from typing import Any, Dict, List, Optional, Tuple
 
-from ...tokenization_utils import PreTrainedTokenizer
-from ...utils import logging
+from ....tokenization_utils import PreTrainedTokenizer
+from ....utils import logging
 
 
 logger = logging.get_logger(__name__)

src/transformers/models/dinat/modeling_dinat.py

@@ -71,7 +71,6 @@ _IMAGE_CLASS_EXPECTED_OUTPUT = "tabby, tabby cat"
 
 
 @dataclass
-# Copied from transformers.models.nat.modeling_nat.NatEncoderOutput with Nat->Dinat
 class DinatEncoderOutput(ModelOutput):
     """
     Dinat encoder's outputs, with potential hidden states and attentions.
@@ -105,7 +104,6 @@ class DinatEncoderOutput(ModelOutput):
 
 
 @dataclass
-# Copied from transformers.models.nat.modeling_nat.NatModelOutput with Nat->Dinat
 class DinatModelOutput(ModelOutput):
     """
     Dinat model's outputs that also contains a pooling of the last hidden states.
@@ -142,7 +140,6 @@ class DinatModelOutput(ModelOutput):
 
 
 @dataclass
-# Copied from transformers.models.nat.modeling_nat.NatImageClassifierOutput with Nat->Dinat
 class DinatImageClassifierOutput(ModelOutput):
     """
     Dinat outputs for image classification.
@@ -178,7 +175,6 @@ class DinatImageClassifierOutput(ModelOutput):
     reshaped_hidden_states: Optional[Tuple[torch.FloatTensor, ...]] = None
 
 
-# Copied from transformers.models.nat.modeling_nat.NatEmbeddings with Nat->Dinat
 class DinatEmbeddings(nn.Module):
     """
     Construct the patch and position embeddings.
@@ -201,7 +197,6 @@ class DinatEmbeddings(nn.Module):
         return embeddings
 
 
-# Copied from transformers.models.nat.modeling_nat.NatPatchEmbeddings with Nat->Dinat
 class DinatPatchEmbeddings(nn.Module):
     """
     This class turns `pixel_values` of shape `(batch_size, num_channels, height, width)` into the initial
@@ -238,7 +233,6 @@ class DinatPatchEmbeddings(nn.Module):
         return embeddings
 
 
-# Copied from transformers.models.nat.modeling_nat.NatDownsampler with Nat->Dinat
 class DinatDownsampler(nn.Module):
     """
     Convolutional Downsampling Layer.
@@ -321,7 +315,6 @@ class NeighborhoodAttention(nn.Module):
 
         self.dropout = nn.Dropout(config.attention_probs_dropout_prob)
 
-    # Copied from transformers.models.nat.modeling_nat.NeighborhoodAttention.transpose_for_scores with Nat->Dinat
     def transpose_for_scores(self, x):
         new_x_shape = x.size()[:-1] + (self.num_attention_heads, self.attention_head_size)
         x = x.view(new_x_shape)
@@ -361,7 +354,6 @@ class NeighborhoodAttention(nn.Module):
         return outputs
 
 
-# Copied from transformers.models.nat.modeling_nat.NeighborhoodAttentionOutput
 class NeighborhoodAttentionOutput(nn.Module):
     def __init__(self, config, dim):
         super().__init__()
@@ -382,7 +374,6 @@ class NeighborhoodAttentionModule(nn.Module):
         self.output = NeighborhoodAttentionOutput(config, dim)
         self.pruned_heads = set()
 
-    # Copied from transformers.models.nat.modeling_nat.NeighborhoodAttentionModule.prune_heads
     def prune_heads(self, heads):
         if len(heads) == 0:
             return
@@ -401,7 +392,6 @@ class NeighborhoodAttentionModule(nn.Module):
         self.self.all_head_size = self.self.attention_head_size * self.self.num_attention_heads
         self.pruned_heads = self.pruned_heads.union(heads)
 
-    # Copied from transformers.models.nat.modeling_nat.NeighborhoodAttentionModule.forward
     def forward(
         self,
         hidden_states: torch.Tensor,
@@ -413,7 +403,6 @@ class NeighborhoodAttentionModule(nn.Module):
         return outputs
 
 
-# Copied from transformers.models.nat.modeling_nat.NatIntermediate with Nat->Dinat
 class DinatIntermediate(nn.Module):
     def __init__(self, config, dim):
         super().__init__()
@@ -429,7 +418,6 @@ class DinatIntermediate(nn.Module):
         return hidden_states
 
 
-# Copied from transformers.models.nat.modeling_nat.NatOutput with Nat->Dinat
 class DinatOutput(nn.Module):
     def __init__(self, config, dim):
         super().__init__()
@@ -539,7 +527,6 @@ class DinatStage(nn.Module):
 
         self.pointing = False
 
-    # Copied from transformers.models.nat.modeling_nat.NatStage.forward
     def forward(
         self,
         hidden_states: torch.Tensor,
@@ -582,7 +569,6 @@ class DinatEncoder(nn.Module):
             ]
         )
 
-    # Copied from transformers.models.nat.modeling_nat.NatEncoder.forward with Nat->Dinat
     def forward(
         self,
         hidden_states: torch.Tensor,
@@ -687,7 +673,6 @@ DINAT_INPUTS_DOCSTRING = r"""
     "The bare Dinat Model transformer outputting raw hidden-states without any specific head on top.",
     DINAT_START_DOCSTRING,
 )
-# Copied from transformers.models.nat.modeling_nat.NatModel with Nat->Dinat, NAT->DINAT
 class DinatModel(DinatPreTrainedModel):
     def __init__(self, config, add_pooling_layer=True):
         super().__init__(config)

src/transformers/utils/dummy_pt_objects.py

@@ -2609,6 +2609,174 @@ class DeiTPreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
+class DetaForObjectDetection(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class DetaModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class DetaPreTrainedModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class EfficientFormerForImageClassification(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class EfficientFormerForImageClassificationWithTeacher(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class EfficientFormerModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class EfficientFormerPreTrainedModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class ErnieMForInformationExtraction(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class ErnieMForMultipleChoice(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class ErnieMForQuestionAnswering(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class ErnieMForSequenceClassification(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class ErnieMForTokenClassification(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class ErnieMModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class ErnieMPreTrainedModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class GPTSanJapaneseForConditionalGeneration(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class GPTSanJapaneseModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class GPTSanJapanesePreTrainedModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class GraphormerForGraphClassification(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class GraphormerModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class GraphormerPreTrainedModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class JukeboxModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class JukeboxPreTrainedModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class JukeboxPrior(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class JukeboxVQVAE(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
 class MCTCTForCTC(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -2616,69 +2784,343 @@ class MCTCTForCTC(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-class MCTCTModel(metaclass=DummyObject):
+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 MegaForCausalLM(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class MegaForMaskedLM(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class MegaForMultipleChoice(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class MegaForQuestionAnswering(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class MegaForSequenceClassification(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class MegaForTokenClassification(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class MegaModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class MegaPreTrainedModel(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"])
+
+
+class NatBackbone(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class NatForImageClassification(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class NatModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class NatPreTrainedModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class NezhaForMaskedLM(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class NezhaForMultipleChoice(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class NezhaForNextSentencePrediction(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class NezhaForPreTraining(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class NezhaForQuestionAnswering(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class NezhaForSequenceClassification(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class NezhaForTokenClassification(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class NezhaModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class NezhaPreTrainedModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class OpenLlamaForCausalLM(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class MCTCTPreTrainedModel(metaclass=DummyObject):
+class OpenLlamaForSequenceClassification(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class MMBTForClassification(metaclass=DummyObject):
+class OpenLlamaModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class MMBTModel(metaclass=DummyObject):
+class OpenLlamaPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ModalEmbeddings(metaclass=DummyObject):
+class QDQBertForMaskedLM(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class OpenLlamaForCausalLM(metaclass=DummyObject):
+class QDQBertForMultipleChoice(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class OpenLlamaForSequenceClassification(metaclass=DummyObject):
+class QDQBertForNextSentencePrediction(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class OpenLlamaModel(metaclass=DummyObject):
+class QDQBertForQuestionAnswering(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class OpenLlamaPreTrainedModel(metaclass=DummyObject):
+class QDQBertForSequenceClassification(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class QDQBertForTokenClassification(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class QDQBertLayer(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class QDQBertLMHeadModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class QDQBertModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class QDQBertPreTrainedModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+def load_tf_weights_in_qdqbert(*args, **kwargs):
+    requires_backends(load_tf_weights_in_qdqbert, ["torch"])
+
+
+class RealmEmbedder(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class RealmForOpenQA(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class RealmKnowledgeAugEncoder(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class RealmPreTrainedModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class RealmReader(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class RealmRetriever(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class RealmScorer(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
+def load_tf_weights_in_realm(*args, **kwargs):
+    requires_backends(load_tf_weights_in_realm, ["torch"])
+
+
 class RetriBertModel(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -2693,6 +3135,20 @@ class RetriBertPreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
+class Speech2Text2ForCausalLM(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
+class Speech2Text2PreTrainedModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+
 class TrajectoryTransformerModel(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -2746,536 +3202,529 @@ def load_tf_weights_in_transfo_xl(*args, **kwargs):
     requires_backends(load_tf_weights_in_transfo_xl, ["torch"])
 
 
-class VanForImageClassification(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class VanModel(metaclass=DummyObject):
+class TvltForAudioVisualClassification(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class VanPreTrainedModel(metaclass=DummyObject):
+class TvltForPreTraining(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DepthAnythingForDepthEstimation(metaclass=DummyObject):
+class TvltModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DepthAnythingPreTrainedModel(metaclass=DummyObject):
+class TvltPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DetaForObjectDetection(metaclass=DummyObject):
+class VanForImageClassification(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DetaModel(metaclass=DummyObject):
+class VanModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DetaPreTrainedModel(metaclass=DummyObject):
+class VanPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DetrForObjectDetection(metaclass=DummyObject):
+class ViTHybridForImageClassification(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DetrForSegmentation(metaclass=DummyObject):
+class ViTHybridModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DetrModel(metaclass=DummyObject):
+class ViTHybridPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DetrPreTrainedModel(metaclass=DummyObject):
+class XLMProphetNetDecoder(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DinatBackbone(metaclass=DummyObject):
+class XLMProphetNetEncoder(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DinatForImageClassification(metaclass=DummyObject):
+class XLMProphetNetForCausalLM(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DinatModel(metaclass=DummyObject):
+class XLMProphetNetForConditionalGeneration(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DinatPreTrainedModel(metaclass=DummyObject):
+class XLMProphetNetModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class Dinov2Backbone(metaclass=DummyObject):
+class XLMProphetNetPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class Dinov2ForImageClassification(metaclass=DummyObject):
+class DepthAnythingForDepthEstimation(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class Dinov2Model(metaclass=DummyObject):
+class DepthAnythingPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class Dinov2PreTrainedModel(metaclass=DummyObject):
+class DetrForObjectDetection(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DistilBertForMaskedLM(metaclass=DummyObject):
+class DetrForSegmentation(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DistilBertForMultipleChoice(metaclass=DummyObject):
+class DetrModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DistilBertForQuestionAnswering(metaclass=DummyObject):
+class DetrPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DistilBertForSequenceClassification(metaclass=DummyObject):
+class DinatBackbone(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DistilBertForTokenClassification(metaclass=DummyObject):
+class DinatForImageClassification(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DistilBertModel(metaclass=DummyObject):
+class DinatModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DistilBertPreTrainedModel(metaclass=DummyObject):
+class DinatPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DonutSwinModel(metaclass=DummyObject):
+class Dinov2Backbone(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DonutSwinPreTrainedModel(metaclass=DummyObject):
+class Dinov2ForImageClassification(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DPRContextEncoder(metaclass=DummyObject):
+class Dinov2Model(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DPRPretrainedContextEncoder(metaclass=DummyObject):
+class Dinov2PreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DPRPreTrainedModel(metaclass=DummyObject):
+class DistilBertForMaskedLM(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DPRPretrainedQuestionEncoder(metaclass=DummyObject):
+class DistilBertForMultipleChoice(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DPRPretrainedReader(metaclass=DummyObject):
+class DistilBertForQuestionAnswering(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DPRQuestionEncoder(metaclass=DummyObject):
+class DistilBertForSequenceClassification(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DPRReader(metaclass=DummyObject):
+class DistilBertForTokenClassification(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DPTForDepthEstimation(metaclass=DummyObject):
+class DistilBertModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DPTForSemanticSegmentation(metaclass=DummyObject):
+class DistilBertPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DPTModel(metaclass=DummyObject):
+class DonutSwinModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class DPTPreTrainedModel(metaclass=DummyObject):
+class DonutSwinPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class EfficientFormerForImageClassification(metaclass=DummyObject):
+class DPRContextEncoder(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class EfficientFormerForImageClassificationWithTeacher(metaclass=DummyObject):
+class DPRPretrainedContextEncoder(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class EfficientFormerModel(metaclass=DummyObject):
+class DPRPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class EfficientFormerPreTrainedModel(metaclass=DummyObject):
+class DPRPretrainedQuestionEncoder(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class EfficientNetForImageClassification(metaclass=DummyObject):
+class DPRPretrainedReader(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class EfficientNetModel(metaclass=DummyObject):
+class DPRQuestionEncoder(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class EfficientNetPreTrainedModel(metaclass=DummyObject):
+class DPRReader(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ElectraForCausalLM(metaclass=DummyObject):
+class DPTForDepthEstimation(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ElectraForMaskedLM(metaclass=DummyObject):
+class DPTForSemanticSegmentation(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ElectraForMultipleChoice(metaclass=DummyObject):
+class DPTModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ElectraForPreTraining(metaclass=DummyObject):
+class DPTPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ElectraForQuestionAnswering(metaclass=DummyObject):
+class EfficientNetForImageClassification(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ElectraForSequenceClassification(metaclass=DummyObject):
+class EfficientNetModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ElectraForTokenClassification(metaclass=DummyObject):
+class EfficientNetPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ElectraModel(metaclass=DummyObject):
+class ElectraForCausalLM(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ElectraPreTrainedModel(metaclass=DummyObject):
+class ElectraForMaskedLM(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-def load_tf_weights_in_electra(*args, **kwargs):
-    requires_backends(load_tf_weights_in_electra, ["torch"])
-
-
-class EncodecModel(metaclass=DummyObject):
+class ElectraForMultipleChoice(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class EncodecPreTrainedModel(metaclass=DummyObject):
+class ElectraForPreTraining(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class EncoderDecoderModel(metaclass=DummyObject):
+class ElectraForQuestionAnswering(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErnieForCausalLM(metaclass=DummyObject):
+class ElectraForSequenceClassification(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErnieForMaskedLM(metaclass=DummyObject):
+class ElectraForTokenClassification(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErnieForMultipleChoice(metaclass=DummyObject):
+class ElectraModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErnieForNextSentencePrediction(metaclass=DummyObject):
+class ElectraPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErnieForPreTraining(metaclass=DummyObject):
+def load_tf_weights_in_electra(*args, **kwargs):
+    requires_backends(load_tf_weights_in_electra, ["torch"])
+
+
+class EncodecModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErnieForQuestionAnswering(metaclass=DummyObject):
+class EncodecPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErnieForSequenceClassification(metaclass=DummyObject):
+class EncoderDecoderModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErnieForTokenClassification(metaclass=DummyObject):
+class ErnieForCausalLM(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErnieModel(metaclass=DummyObject):
+class ErnieForMaskedLM(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErniePreTrainedModel(metaclass=DummyObject):
+class ErnieForMultipleChoice(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErnieMForInformationExtraction(metaclass=DummyObject):
+class ErnieForNextSentencePrediction(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErnieMForMultipleChoice(metaclass=DummyObject):
+class ErnieForPreTraining(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErnieMForQuestionAnswering(metaclass=DummyObject):
+class ErnieForQuestionAnswering(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErnieMForSequenceClassification(metaclass=DummyObject):
+class ErnieForSequenceClassification(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErnieMForTokenClassification(metaclass=DummyObject):
+class ErnieForTokenClassification(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErnieMModel(metaclass=DummyObject):
+class ErnieModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class ErnieMPreTrainedModel(metaclass=DummyObject):
+class ErniePreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
@@ -4043,48 +4492,6 @@ class GPTJPreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-class GPTSanJapaneseForConditionalGeneration(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class GPTSanJapaneseModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class GPTSanJapanesePreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class GraphormerForGraphClassification(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class GraphormerModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class GraphormerPreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
 class GroundingDinoForObjectDetection(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -4404,34 +4811,6 @@ class JetMoePreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-class JukeboxModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class JukeboxPreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class JukeboxPrior(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class JukeboxVQVAE(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
 class Kosmos2ForConditionalGeneration(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -5125,62 +5504,6 @@ class MBartPreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-class MegaForCausalLM(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MegaForMaskedLM(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MegaForMultipleChoice(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MegaForQuestionAnswering(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MegaForSequenceClassification(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MegaForTokenClassification(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MegaModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MegaPreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
 class MegatronBertForCausalLM(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -5746,175 +6069,84 @@ class MusicgenModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-class MusicgenPreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MusicgenProcessor(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MusicgenMelodyForCausalLM(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MusicgenMelodyForConditionalGeneration(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MusicgenMelodyModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MusicgenMelodyPreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MvpForCausalLM(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MvpForConditionalGeneration(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MvpForQuestionAnswering(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MvpForSequenceClassification(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MvpModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class MvpPreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class NatBackbone(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class NatForImageClassification(metaclass=DummyObject):
+class MusicgenPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class NatModel(metaclass=DummyObject):
+class MusicgenProcessor(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class NatPreTrainedModel(metaclass=DummyObject):
+class MusicgenMelodyForCausalLM(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class NezhaForMaskedLM(metaclass=DummyObject):
+class MusicgenMelodyForConditionalGeneration(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class NezhaForMultipleChoice(metaclass=DummyObject):
+class MusicgenMelodyModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class NezhaForNextSentencePrediction(metaclass=DummyObject):
+class MusicgenMelodyPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class NezhaForPreTraining(metaclass=DummyObject):
+class MvpForCausalLM(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class NezhaForQuestionAnswering(metaclass=DummyObject):
+class MvpForConditionalGeneration(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class NezhaForSequenceClassification(metaclass=DummyObject):
+class MvpForQuestionAnswering(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class NezhaForTokenClassification(metaclass=DummyObject):
+class MvpForSequenceClassification(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class NezhaModel(metaclass=DummyObject):
+class MvpModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["torch"])
 
 
-class NezhaPreTrainedModel(metaclass=DummyObject):
+class MvpPreTrainedModel(metaclass=DummyObject):
     _backends = ["torch"]
 
     def __init__(self, *args, **kwargs):
@@ -6716,80 +6948,6 @@ class PvtV2PreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-class QDQBertForMaskedLM(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class QDQBertForMultipleChoice(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class QDQBertForNextSentencePrediction(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class QDQBertForQuestionAnswering(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class QDQBertForSequenceClassification(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class QDQBertForTokenClassification(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class QDQBertLayer(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class QDQBertLMHeadModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class QDQBertModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class QDQBertPreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-def load_tf_weights_in_qdqbert(*args, **kwargs):
-    requires_backends(load_tf_weights_in_qdqbert, ["torch"])
-
-
 class Qwen2ForCausalLM(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -6888,59 +7046,6 @@ class RagTokenForGeneration(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-class RealmEmbedder(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class RealmForOpenQA(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class RealmKnowledgeAugEncoder(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class RealmPreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class RealmReader(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class RealmRetriever(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class RealmScorer(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-def load_tf_weights_in_realm(*args, **kwargs):
-    requires_backends(load_tf_weights_in_realm, ["torch"])
-
-
 class RecurrentGemmaForCausalLM(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -7716,20 +7821,6 @@ class Speech2TextPreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-class Speech2Text2ForCausalLM(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class Speech2Text2PreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
 class SpeechT5ForSpeechToSpeech(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -8277,34 +8368,6 @@ class TrOCRPreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-class TvltForAudioVisualClassification(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class TvltForPreTraining(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class TvltModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class TvltPreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
 class TvpForVideoGrounding(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -8725,27 +8788,6 @@ class ViTPreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-class ViTHybridForImageClassification(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class ViTHybridModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class ViTHybridPreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
 class ViTMAEForPreTraining(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -9194,48 +9236,6 @@ class XLMWithLMHeadModel(metaclass=DummyObject):
         requires_backends(self, ["torch"])
 
 
-class XLMProphetNetDecoder(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class XLMProphetNetEncoder(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class XLMProphetNetForCausalLM(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class XLMProphetNetForConditionalGeneration(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class XLMProphetNetModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
-class XLMProphetNetPreTrainedModel(metaclass=DummyObject):
-    _backends = ["torch"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["torch"])
-
-
 class XLMRobertaForCausalLM(metaclass=DummyObject):
     _backends = ["torch"]
 

src/transformers/utils/dummy_sentencepiece_objects.py

@@ -72,6 +72,13 @@ class ErnieMTokenizer(metaclass=DummyObject):
         requires_backends(self, ["sentencepiece"])
 
 
+class XLMProphetNetTokenizer(metaclass=DummyObject):
+    _backends = ["sentencepiece"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["sentencepiece"])
+
+
 class FNetTokenizer(metaclass=DummyObject):
     _backends = ["sentencepiece"]
 
@@ -233,13 +240,6 @@ class XGLMTokenizer(metaclass=DummyObject):
         requires_backends(self, ["sentencepiece"])
 
 
-class XLMProphetNetTokenizer(metaclass=DummyObject):
-    _backends = ["sentencepiece"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["sentencepiece"])
-
-
 class XLMRobertaTokenizer(metaclass=DummyObject):
     _backends = ["sentencepiece"]
 

src/transformers/utils/dummy_tf_objects.py

@@ -1038,168 +1038,168 @@ class TFDeiTPreTrainedModel(metaclass=DummyObject):
         requires_backends(self, ["tf"])
 
 
-class TFAdaptiveEmbedding(metaclass=DummyObject):
+class TFEfficientFormerForImageClassification(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFTransfoXLForSequenceClassification(metaclass=DummyObject):
+class TFEfficientFormerForImageClassificationWithTeacher(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFTransfoXLLMHeadModel(metaclass=DummyObject):
+class TFEfficientFormerModel(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFTransfoXLMainLayer(metaclass=DummyObject):
+class TFEfficientFormerPreTrainedModel(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFTransfoXLModel(metaclass=DummyObject):
+class TFAdaptiveEmbedding(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFTransfoXLPreTrainedModel(metaclass=DummyObject):
+class TFTransfoXLForSequenceClassification(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFDistilBertForMaskedLM(metaclass=DummyObject):
+class TFTransfoXLLMHeadModel(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFDistilBertForMultipleChoice(metaclass=DummyObject):
+class TFTransfoXLMainLayer(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFDistilBertForQuestionAnswering(metaclass=DummyObject):
+class TFTransfoXLModel(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFDistilBertForSequenceClassification(metaclass=DummyObject):
+class TFTransfoXLPreTrainedModel(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFDistilBertForTokenClassification(metaclass=DummyObject):
+class TFDistilBertForMaskedLM(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFDistilBertMainLayer(metaclass=DummyObject):
+class TFDistilBertForMultipleChoice(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFDistilBertModel(metaclass=DummyObject):
+class TFDistilBertForQuestionAnswering(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFDistilBertPreTrainedModel(metaclass=DummyObject):
+class TFDistilBertForSequenceClassification(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFDPRContextEncoder(metaclass=DummyObject):
+class TFDistilBertForTokenClassification(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFDPRPretrainedContextEncoder(metaclass=DummyObject):
+class TFDistilBertMainLayer(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFDPRPretrainedQuestionEncoder(metaclass=DummyObject):
+class TFDistilBertModel(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFDPRPretrainedReader(metaclass=DummyObject):
+class TFDistilBertPreTrainedModel(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFDPRQuestionEncoder(metaclass=DummyObject):
+class TFDPRContextEncoder(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFDPRReader(metaclass=DummyObject):
+class TFDPRPretrainedContextEncoder(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFEfficientFormerForImageClassification(metaclass=DummyObject):
+class TFDPRPretrainedQuestionEncoder(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFEfficientFormerForImageClassificationWithTeacher(metaclass=DummyObject):
+class TFDPRPretrainedReader(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFEfficientFormerModel(metaclass=DummyObject):
+class TFDPRQuestionEncoder(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["tf"])
 
 
-class TFEfficientFormerPreTrainedModel(metaclass=DummyObject):
+class TFDPRReader(metaclass=DummyObject):
     _backends = ["tf"]
 
     def __init__(self, *args, **kwargs):

src/transformers/utils/dummy_tokenizers_objects.py

@@ -121,6 +121,13 @@ class DebertaV2TokenizerFast(metaclass=DummyObject):
         requires_backends(self, ["tokenizers"])
 
 
+class RealmTokenizerFast(metaclass=DummyObject):
+    _backends = ["tokenizers"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["tokenizers"])
+
+
 class RetriBertTokenizerFast(metaclass=DummyObject):
     _backends = ["tokenizers"]
 
@@ -352,13 +359,6 @@ class Qwen2TokenizerFast(metaclass=DummyObject):
         requires_backends(self, ["tokenizers"])
 
 
-class RealmTokenizerFast(metaclass=DummyObject):
-    _backends = ["tokenizers"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["tokenizers"])
-
-
 class ReformerTokenizerFast(metaclass=DummyObject):
     _backends = ["tokenizers"]
 

src/transformers/utils/dummy_vision_objects.py

@@ -142,49 +142,63 @@ class DetaImageProcessor(metaclass=DummyObject):
         requires_backends(self, ["vision"])
 
 
-class DetrFeatureExtractor(metaclass=DummyObject):
+class EfficientFormerImageProcessor(metaclass=DummyObject):
     _backends = ["vision"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["vision"])
 
 
-class DetrImageProcessor(metaclass=DummyObject):
+class TvltImageProcessor(metaclass=DummyObject):
     _backends = ["vision"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["vision"])
 
 
-class DonutFeatureExtractor(metaclass=DummyObject):
+class ViTHybridImageProcessor(metaclass=DummyObject):
     _backends = ["vision"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["vision"])
 
 
-class DonutImageProcessor(metaclass=DummyObject):
+class DetrFeatureExtractor(metaclass=DummyObject):
     _backends = ["vision"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["vision"])
 
 
-class DPTFeatureExtractor(metaclass=DummyObject):
+class DetrImageProcessor(metaclass=DummyObject):
     _backends = ["vision"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["vision"])
 
 
-class DPTImageProcessor(metaclass=DummyObject):
+class DonutFeatureExtractor(metaclass=DummyObject):
     _backends = ["vision"]
 
     def __init__(self, *args, **kwargs):
         requires_backends(self, ["vision"])
 
 
-class EfficientFormerImageProcessor(metaclass=DummyObject):
+class DonutImageProcessor(metaclass=DummyObject):
+    _backends = ["vision"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["vision"])
+
+
+class DPTFeatureExtractor(metaclass=DummyObject):
+    _backends = ["vision"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["vision"])
+
+
+class DPTImageProcessor(metaclass=DummyObject):
     _backends = ["vision"]
 
     def __init__(self, *args, **kwargs):
@@ -520,13 +534,6 @@ class Swin2SRImageProcessor(metaclass=DummyObject):
         requires_backends(self, ["vision"])
 
 
-class TvltImageProcessor(metaclass=DummyObject):
-    _backends = ["vision"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["vision"])
-
-
 class TvpImageProcessor(metaclass=DummyObject):
     _backends = ["vision"]
 
@@ -590,13 +597,6 @@ class ViTImageProcessor(metaclass=DummyObject):
         requires_backends(self, ["vision"])
 
 
-class ViTHybridImageProcessor(metaclass=DummyObject):
-    _backends = ["vision"]
-
-    def __init__(self, *args, **kwargs):
-        requires_backends(self, ["vision"])
-
-
 class VitMatteImageProcessor(metaclass=DummyObject):
     _backends = ["vision"]
 

tests/models/deta/test_image_processing_deta.py

-# coding=utf-8
-# Copyright 2022 HuggingFace Inc.
-#
-# 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 json
-import pathlib
-import unittest
-
-from transformers.testing_utils import require_torch, require_vision, slow
-from transformers.utils import is_torch_available, is_vision_available
-
-from ...test_image_processing_common import AnnotationFormatTestMixin, ImageProcessingTestMixin, prepare_image_inputs
-
-
-if is_torch_available():
-    import torch
-
-if is_vision_available():
-    from PIL import Image
-
-    from transformers import DetaImageProcessor
-
-
-class DetaImageProcessingTester(unittest.TestCase):
-    def __init__(
-        self,
-        parent,
-        batch_size=7,
-        num_channels=3,
-        min_resolution=30,
-        max_resolution=400,
-        do_resize=True,
-        size=None,
-        do_normalize=True,
-        image_mean=[0.5, 0.5, 0.5],
-        image_std=[0.5, 0.5, 0.5],
-        do_rescale=True,
-        rescale_factor=1 / 255,
-        do_pad=True,
-    ):
-        # by setting size["longest_edge"] > max_resolution we're effectively not testing this :p
-        size = size if size is not None else {"shortest_edge": 18, "longest_edge": 1333}
-        self.parent = parent
-        self.batch_size = batch_size
-        self.num_channels = num_channels
-        self.min_resolution = min_resolution
-        self.max_resolution = max_resolution
-        self.do_resize = do_resize
-        self.size = size
-        self.do_normalize = do_normalize
-        self.image_mean = image_mean
-        self.image_std = image_std
-        self.do_rescale = do_rescale
-        self.rescale_factor = rescale_factor
-        self.do_pad = do_pad
-
-    def prepare_image_processor_dict(self):
-        return {
-            "do_resize": self.do_resize,
-            "size": self.size,
-            "do_normalize": self.do_normalize,
-            "image_mean": self.image_mean,
-            "image_std": self.image_std,
-            "do_rescale": self.do_rescale,
-            "rescale_factor": self.rescale_factor,
-            "do_pad": self.do_pad,
-        }
-
-    def get_expected_values(self, image_inputs, batched=False):
-        """
-        This function computes the expected height and width when providing images to DetaImageProcessor,
-        assuming do_resize is set to True with a scalar size.
-        """
-        if not batched:
-            image = image_inputs[0]
-            if isinstance(image, Image.Image):
-                w, h = image.size
-            else:
-                h, w = image.shape[1], image.shape[2]
-            if w < h:
-                expected_height = int(self.size["shortest_edge"] * h / w)
-                expected_width = self.size["shortest_edge"]
-            elif w > h:
-                expected_height = self.size["shortest_edge"]
-                expected_width = int(self.size["shortest_edge"] * w / h)
-            else:
-                expected_height = self.size["shortest_edge"]
-                expected_width = self.size["shortest_edge"]
-
-        else:
-            expected_values = []
-            for image in image_inputs:
-                expected_height, expected_width = self.get_expected_values([image])
-                expected_values.append((expected_height, expected_width))
-            expected_height = max(expected_values, key=lambda item: item[0])[0]
-            expected_width = max(expected_values, key=lambda item: item[1])[1]
-
-        return expected_height, expected_width
-
-    def expected_output_image_shape(self, images):
-        height, width = self.get_expected_values(images, batched=True)
-        return self.num_channels, height, width
-
-    def prepare_image_inputs(self, equal_resolution=False, numpify=False, torchify=False):
-        return prepare_image_inputs(
-            batch_size=self.batch_size,
-            num_channels=self.num_channels,
-            min_resolution=self.min_resolution,
-            max_resolution=self.max_resolution,
-            equal_resolution=equal_resolution,
-            numpify=numpify,
-            torchify=torchify,
-        )
-
-
-@require_torch
-@require_vision
-class DetaImageProcessingTest(AnnotationFormatTestMixin, ImageProcessingTestMixin, unittest.TestCase):
-    image_processing_class = DetaImageProcessor if is_vision_available() else None
-
-    def setUp(self):
-        self.image_processor_tester = DetaImageProcessingTester(self)
-
-    @property
-    def image_processor_dict(self):
-        return self.image_processor_tester.prepare_image_processor_dict()
-
-    def test_image_processor_properties(self):
-        image_processing = self.image_processing_class(**self.image_processor_dict)
-        self.assertTrue(hasattr(image_processing, "image_mean"))
-        self.assertTrue(hasattr(image_processing, "image_std"))
-        self.assertTrue(hasattr(image_processing, "do_normalize"))
-        self.assertTrue(hasattr(image_processing, "do_resize"))
-        self.assertTrue(hasattr(image_processing, "do_rescale"))
-        self.assertTrue(hasattr(image_processing, "do_pad"))
-        self.assertTrue(hasattr(image_processing, "size"))
-
-    def test_image_processor_from_dict_with_kwargs(self):
-        image_processor = self.image_processing_class.from_dict(self.image_processor_dict)
-        self.assertEqual(image_processor.size, {"shortest_edge": 18, "longest_edge": 1333})
-        self.assertEqual(image_processor.do_pad, True)
-
-    @slow
-    def test_call_pytorch_with_coco_detection_annotations(self):
-        # prepare image and target
-        image = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png")
-        with open("./tests/fixtures/tests_samples/COCO/coco_annotations.txt", "r") as f:
-            target = json.loads(f.read())
-
-        target = {"image_id": 39769, "annotations": target}
-
-        # encode them
-        image_processing = DetaImageProcessor()
-        encoding = image_processing(images=image, annotations=target, return_tensors="pt")
-
-        # verify pixel values
-        expected_shape = torch.Size([1, 3, 800, 1066])
-        self.assertEqual(encoding["pixel_values"].shape, expected_shape)
-
-        expected_slice = torch.tensor([0.2796, 0.3138, 0.3481])
-        self.assertTrue(torch.allclose(encoding["pixel_values"][0, 0, 0, :3], expected_slice, atol=1e-4))
-
-        # verify area
-        expected_area = torch.tensor([5887.9600, 11250.2061, 489353.8438, 837122.7500, 147967.5156, 165732.3438])
-        self.assertTrue(torch.allclose(encoding["labels"][0]["area"], expected_area))
-        # verify boxes
-        expected_boxes_shape = torch.Size([6, 4])
-        self.assertEqual(encoding["labels"][0]["boxes"].shape, expected_boxes_shape)
-        expected_boxes_slice = torch.tensor([0.5503, 0.2765, 0.0604, 0.2215])
-        self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"][0], expected_boxes_slice, atol=1e-3))
-        # verify image_id
-        expected_image_id = torch.tensor([39769])
-        self.assertTrue(torch.allclose(encoding["labels"][0]["image_id"], expected_image_id))
-        # verify is_crowd
-        expected_is_crowd = torch.tensor([0, 0, 0, 0, 0, 0])
-        self.assertTrue(torch.allclose(encoding["labels"][0]["iscrowd"], expected_is_crowd))
-        # verify class_labels
-        expected_class_labels = torch.tensor([75, 75, 63, 65, 17, 17])
-        self.assertTrue(torch.allclose(encoding["labels"][0]["class_labels"], expected_class_labels))
-        # verify orig_size
-        expected_orig_size = torch.tensor([480, 640])
-        self.assertTrue(torch.allclose(encoding["labels"][0]["orig_size"], expected_orig_size))
-        # verify size
-        expected_size = torch.tensor([800, 1066])
-        self.assertTrue(torch.allclose(encoding["labels"][0]["size"], expected_size))
-
-    @slow
-    def test_call_pytorch_with_coco_panoptic_annotations(self):
-        # prepare image, target and masks_path
-        image = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png")
-        with open("./tests/fixtures/tests_samples/COCO/coco_panoptic_annotations.txt", "r") as f:
-            target = json.loads(f.read())
-
-        target = {"file_name": "000000039769.png", "image_id": 39769, "segments_info": target}
-
-        masks_path = pathlib.Path("./tests/fixtures/tests_samples/COCO/coco_panoptic")
-
-        # encode them
-        image_processing = DetaImageProcessor(format="coco_panoptic")
-        encoding = image_processing(images=image, annotations=target, masks_path=masks_path, return_tensors="pt")
-
-        # verify pixel values
-        expected_shape = torch.Size([1, 3, 800, 1066])
-        self.assertEqual(encoding["pixel_values"].shape, expected_shape)
-
-        expected_slice = torch.tensor([0.2796, 0.3138, 0.3481])
-        self.assertTrue(torch.allclose(encoding["pixel_values"][0, 0, 0, :3], expected_slice, atol=1e-4))
-
-        # verify area
-        expected_area = torch.tensor([147979.6875, 165527.0469, 484638.5938, 11292.9375, 5879.6562, 7634.1147])
-        self.assertTrue(torch.allclose(encoding["labels"][0]["area"], expected_area))
-        # verify boxes
-        expected_boxes_shape = torch.Size([6, 4])
-        self.assertEqual(encoding["labels"][0]["boxes"].shape, expected_boxes_shape)
-        expected_boxes_slice = torch.tensor([0.2625, 0.5437, 0.4688, 0.8625])
-        self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"][0], expected_boxes_slice, atol=1e-3))
-        # verify image_id
-        expected_image_id = torch.tensor([39769])
-        self.assertTrue(torch.allclose(encoding["labels"][0]["image_id"], expected_image_id))
-        # verify is_crowd
-        expected_is_crowd = torch.tensor([0, 0, 0, 0, 0, 0])
-        self.assertTrue(torch.allclose(encoding["labels"][0]["iscrowd"], expected_is_crowd))
-        # verify class_labels
-        expected_class_labels = torch.tensor([17, 17, 63, 75, 75, 93])
-        self.assertTrue(torch.allclose(encoding["labels"][0]["class_labels"], expected_class_labels))
-        # verify masks
-        expected_masks_sum = 822873
-        self.assertEqual(encoding["labels"][0]["masks"].sum().item(), expected_masks_sum)
-        # verify orig_size
-        expected_orig_size = torch.tensor([480, 640])
-        self.assertTrue(torch.allclose(encoding["labels"][0]["orig_size"], expected_orig_size))
-        # verify size
-        expected_size = torch.tensor([800, 1066])
-        self.assertTrue(torch.allclose(encoding["labels"][0]["size"], expected_size))
-
-    @slow
-    # Copied from tests.models.detr.test_image_processing_detr.DetrImageProcessingTest.test_batched_coco_detection_annotations with Detr->Deta
-    def test_batched_coco_detection_annotations(self):
-        image_0 = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png")
-        image_1 = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png").resize((800, 800))
-
-        with open("./tests/fixtures/tests_samples/COCO/coco_annotations.txt", "r") as f:
-            target = json.loads(f.read())
-
-        annotations_0 = {"image_id": 39769, "annotations": target}
-        annotations_1 = {"image_id": 39769, "annotations": target}
-
-        # Adjust the bounding boxes for the resized image
-        w_0, h_0 = image_0.size
-        w_1, h_1 = image_1.size
-        for i in range(len(annotations_1["annotations"])):
-            coords = annotations_1["annotations"][i]["bbox"]
-            new_bbox = [
-                coords[0] * w_1 / w_0,
-                coords[1] * h_1 / h_0,
-                coords[2] * w_1 / w_0,
-                coords[3] * h_1 / h_0,
-            ]
-            annotations_1["annotations"][i]["bbox"] = new_bbox
-
-        images = [image_0, image_1]
-        annotations = [annotations_0, annotations_1]
-
-        image_processing = DetaImageProcessor()
-        encoding = image_processing(
-            images=images,
-            annotations=annotations,
-            return_segmentation_masks=True,
-            return_tensors="pt",  # do_convert_annotations=True
-        )
-
-        # Check the pixel values have been padded
-        postprocessed_height, postprocessed_width = 800, 1066
-        expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
-        self.assertEqual(encoding["pixel_values"].shape, expected_shape)
-
-        # Check the bounding boxes have been adjusted for padded images
-        self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
-        self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
-        expected_boxes_0 = torch.tensor(
-            [
-                [0.6879, 0.4609, 0.0755, 0.3691],
-                [0.2118, 0.3359, 0.2601, 0.1566],
-                [0.5011, 0.5000, 0.9979, 1.0000],
-                [0.5010, 0.5020, 0.9979, 0.9959],
-                [0.3284, 0.5944, 0.5884, 0.8112],
-                [0.8394, 0.5445, 0.3213, 0.9110],
-            ]
-        )
-        expected_boxes_1 = torch.tensor(
-            [
-                [0.4130, 0.2765, 0.0453, 0.2215],
-                [0.1272, 0.2016, 0.1561, 0.0940],
-                [0.3757, 0.4933, 0.7488, 0.9865],
-                [0.3759, 0.5002, 0.7492, 0.9955],
-                [0.1971, 0.5456, 0.3532, 0.8646],
-                [0.5790, 0.4115, 0.3430, 0.7161],
-            ]
-        )
-        self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
-        self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
-
-        # Check the masks have also been padded
-        self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
-        self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
-
-        # Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
-        # format and not in the range [0, 1]
-        encoding = image_processing(
-            images=images,
-            annotations=annotations,
-            return_segmentation_masks=True,
-            do_convert_annotations=False,
-            return_tensors="pt",
-        )
-        self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
-        self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
-        # Convert to absolute coordinates
-        unnormalized_boxes_0 = torch.vstack(
-            [
-                expected_boxes_0[:, 0] * postprocessed_width,
-                expected_boxes_0[:, 1] * postprocessed_height,
-                expected_boxes_0[:, 2] * postprocessed_width,
-                expected_boxes_0[:, 3] * postprocessed_height,
-            ]
-        ).T
-        unnormalized_boxes_1 = torch.vstack(
-            [
-                expected_boxes_1[:, 0] * postprocessed_width,
-                expected_boxes_1[:, 1] * postprocessed_height,
-                expected_boxes_1[:, 2] * postprocessed_width,
-                expected_boxes_1[:, 3] * postprocessed_height,
-            ]
-        ).T
-        # Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
-        expected_boxes_0 = torch.vstack(
-            [
-                unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
-                unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
-                unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
-                unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
-            ]
-        ).T
-        expected_boxes_1 = torch.vstack(
-            [
-                unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
-                unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
-                unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
-                unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
-            ]
-        ).T
-        self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
-        self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
-
-    # Copied from tests.models.detr.test_image_processing_detr.DetrImageProcessingTest.test_batched_coco_panoptic_annotations with Detr->Deta
-    def test_batched_coco_panoptic_annotations(self):
-        # prepare image, target and masks_path
-        image_0 = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png")
-        image_1 = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png").resize((800, 800))
-
-        with open("./tests/fixtures/tests_samples/COCO/coco_panoptic_annotations.txt", "r") as f:
-            target = json.loads(f.read())
-
-        annotation_0 = {"file_name": "000000039769.png", "image_id": 39769, "segments_info": target}
-        annotation_1 = {"file_name": "000000039769.png", "image_id": 39769, "segments_info": target}
-
-        w_0, h_0 = image_0.size
-        w_1, h_1 = image_1.size
-        for i in range(len(annotation_1["segments_info"])):
-            coords = annotation_1["segments_info"][i]["bbox"]
-            new_bbox = [
-                coords[0] * w_1 / w_0,
-                coords[1] * h_1 / h_0,
-                coords[2] * w_1 / w_0,
-                coords[3] * h_1 / h_0,
-            ]
-            annotation_1["segments_info"][i]["bbox"] = new_bbox
-
-        masks_path = pathlib.Path("./tests/fixtures/tests_samples/COCO/coco_panoptic")
-
-        images = [image_0, image_1]
-        annotations = [annotation_0, annotation_1]
-
-        # encode them
-        image_processing = DetaImageProcessor(format="coco_panoptic")
-        encoding = image_processing(
-            images=images,
-            annotations=annotations,
-            masks_path=masks_path,
-            return_tensors="pt",
-            return_segmentation_masks=True,
-        )
-
-        # Check the pixel values have been padded
-        postprocessed_height, postprocessed_width = 800, 1066
-        expected_shape = torch.Size([2, 3, postprocessed_height, postprocessed_width])
-        self.assertEqual(encoding["pixel_values"].shape, expected_shape)
-
-        # Check the bounding boxes have been adjusted for padded images
-        self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
-        self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
-        expected_boxes_0 = torch.tensor(
-            [
-                [0.2625, 0.5437, 0.4688, 0.8625],
-                [0.7719, 0.4104, 0.4531, 0.7125],
-                [0.5000, 0.4927, 0.9969, 0.9854],
-                [0.1688, 0.2000, 0.2063, 0.0917],
-                [0.5492, 0.2760, 0.0578, 0.2187],
-                [0.4992, 0.4990, 0.9984, 0.9979],
-            ]
-        )
-        expected_boxes_1 = torch.tensor(
-            [
-                [0.1576, 0.3262, 0.2814, 0.5175],
-                [0.4634, 0.2463, 0.2720, 0.4275],
-                [0.3002, 0.2956, 0.5985, 0.5913],
-                [0.1013, 0.1200, 0.1238, 0.0550],
-                [0.3297, 0.1656, 0.0347, 0.1312],
-                [0.2997, 0.2994, 0.5994, 0.5987],
-            ]
-        )
-        self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1e-3))
-        self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1e-3))
-
-        # Check the masks have also been padded
-        self.assertEqual(encoding["labels"][0]["masks"].shape, torch.Size([6, 800, 1066]))
-        self.assertEqual(encoding["labels"][1]["masks"].shape, torch.Size([6, 800, 1066]))
-
-        # Check if do_convert_annotations=False, then the annotations are not converted to centre_x, centre_y, width, height
-        # format and not in the range [0, 1]
-        encoding = image_processing(
-            images=images,
-            annotations=annotations,
-            masks_path=masks_path,
-            return_segmentation_masks=True,
-            do_convert_annotations=False,
-            return_tensors="pt",
-        )
-        self.assertEqual(encoding["labels"][0]["boxes"].shape, torch.Size([6, 4]))
-        self.assertEqual(encoding["labels"][1]["boxes"].shape, torch.Size([6, 4]))
-        # Convert to absolute coordinates
-        unnormalized_boxes_0 = torch.vstack(
-            [
-                expected_boxes_0[:, 0] * postprocessed_width,
-                expected_boxes_0[:, 1] * postprocessed_height,
-                expected_boxes_0[:, 2] * postprocessed_width,
-                expected_boxes_0[:, 3] * postprocessed_height,
-            ]
-        ).T
-        unnormalized_boxes_1 = torch.vstack(
-            [
-                expected_boxes_1[:, 0] * postprocessed_width,
-                expected_boxes_1[:, 1] * postprocessed_height,
-                expected_boxes_1[:, 2] * postprocessed_width,
-                expected_boxes_1[:, 3] * postprocessed_height,
-            ]
-        ).T
-        # Convert from centre_x, centre_y, width, height to x_min, y_min, x_max, y_max
-        expected_boxes_0 = torch.vstack(
-            [
-                unnormalized_boxes_0[:, 0] - unnormalized_boxes_0[:, 2] / 2,
-                unnormalized_boxes_0[:, 1] - unnormalized_boxes_0[:, 3] / 2,
-                unnormalized_boxes_0[:, 0] + unnormalized_boxes_0[:, 2] / 2,
-                unnormalized_boxes_0[:, 1] + unnormalized_boxes_0[:, 3] / 2,
-            ]
-        ).T
-        expected_boxes_1 = torch.vstack(
-            [
-                unnormalized_boxes_1[:, 0] - unnormalized_boxes_1[:, 2] / 2,
-                unnormalized_boxes_1[:, 1] - unnormalized_boxes_1[:, 3] / 2,
-                unnormalized_boxes_1[:, 0] + unnormalized_boxes_1[:, 2] / 2,
-                unnormalized_boxes_1[:, 1] + unnormalized_boxes_1[:, 3] / 2,
-            ]
-        ).T
-        self.assertTrue(torch.allclose(encoding["labels"][0]["boxes"], expected_boxes_0, rtol=1))
-        self.assertTrue(torch.allclose(encoding["labels"][1]["boxes"], expected_boxes_1, rtol=1))
-
-    # Copied from tests.models.detr.test_image_processing_detr.DetrImageProcessingTest.test_max_width_max_height_resizing_and_pad_strategy with Detr->Deta
-    def test_max_width_max_height_resizing_and_pad_strategy(self):
-        image_1 = torch.ones([200, 100, 3], dtype=torch.uint8)
-
-        # do_pad=False, max_height=100, max_width=100, image=200x100 -> 100x50
-        image_processor = DetaImageProcessor(
-            size={"max_height": 100, "max_width": 100},
-            do_pad=False,
-        )
-        inputs = image_processor(images=[image_1], return_tensors="pt")
-        self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 50]))
-
-        # do_pad=False, max_height=300, max_width=100, image=200x100 -> 200x100
-        image_processor = DetaImageProcessor(
-            size={"max_height": 300, "max_width": 100},
-            do_pad=False,
-        )
-        inputs = image_processor(images=[image_1], return_tensors="pt")
-
-        # do_pad=True, max_height=100, max_width=100, image=200x100 -> 100x100
-        image_processor = DetaImageProcessor(
-            size={"max_height": 100, "max_width": 100}, do_pad=True, pad_size={"height": 100, "width": 100}
-        )
-        inputs = image_processor(images=[image_1], return_tensors="pt")
-        self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 100, 100]))
-
-        # do_pad=True, max_height=300, max_width=100, image=200x100 -> 300x100
-        image_processor = DetaImageProcessor(
-            size={"max_height": 300, "max_width": 100},
-            do_pad=True,
-            pad_size={"height": 301, "width": 101},
-        )
-        inputs = image_processor(images=[image_1], return_tensors="pt")
-        self.assertEqual(inputs["pixel_values"].shape, torch.Size([1, 3, 301, 101]))
-
-        ### Check for batch
-        image_2 = torch.ones([100, 150, 3], dtype=torch.uint8)
-
-        # do_pad=True, max_height=150, max_width=100, images=[200x100, 100x150] -> 150x100
-        image_processor = DetaImageProcessor(
-            size={"max_height": 150, "max_width": 100},
-            do_pad=True,
-            pad_size={"height": 150, "width": 100},
-        )
-        inputs = image_processor(images=[image_1, image_2], return_tensors="pt")
-        self.assertEqual(inputs["pixel_values"].shape, torch.Size([2, 3, 150, 100]))

tests/models/deta/test_modeling_deta.py

-# coding=utf-8
-# Copyright 2022 The HuggingFace Inc. 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.
-"""Testing suite for the PyTorch DETA model."""
-
-import collections
-import inspect
-import math
-import re
-import unittest
-
-from transformers import DetaConfig, ResNetConfig, is_torch_available, is_torchvision_available, is_vision_available
-from transformers.file_utils import cached_property
-from transformers.testing_utils import require_torchvision, require_vision, slow, torch_device
-
-from ...generation.test_utils import GenerationTesterMixin
-from ...test_configuration_common import ConfigTester
-from ...test_modeling_common import ModelTesterMixin, _config_zero_init, floats_tensor
-from ...test_pipeline_mixin import PipelineTesterMixin
-
-
-if is_torch_available():
-    import torch
-
-    from transformers.pytorch_utils import id_tensor_storage
-
-if is_torchvision_available():
-    from transformers import DetaForObjectDetection, DetaModel
-
-
-if is_vision_available():
-    from PIL import Image
-
-    from transformers import AutoImageProcessor
-
-
-class DetaModelTester:
-    def __init__(
-        self,
-        parent,
-        batch_size=8,
-        is_training=True,
-        use_labels=True,
-        hidden_size=32,
-        num_hidden_layers=2,
-        num_attention_heads=8,
-        intermediate_size=4,
-        hidden_act="gelu",
-        hidden_dropout_prob=0.1,
-        attention_probs_dropout_prob=0.1,
-        num_queries=12,
-        two_stage_num_proposals=12,
-        num_channels=3,
-        image_size=224,
-        n_targets=8,
-        num_labels=91,
-        num_feature_levels=4,
-        encoder_n_points=2,
-        decoder_n_points=6,
-        two_stage=True,
-        assign_first_stage=True,
-        assign_second_stage=True,
-    ):
-        self.parent = parent
-        self.batch_size = batch_size
-        self.is_training = is_training
-        self.use_labels = use_labels
-        self.hidden_size = hidden_size
-        self.num_hidden_layers = num_hidden_layers
-        self.num_attention_heads = num_attention_heads
-        self.intermediate_size = intermediate_size
-        self.hidden_act = hidden_act
-        self.hidden_dropout_prob = hidden_dropout_prob
-        self.attention_probs_dropout_prob = attention_probs_dropout_prob
-        self.num_queries = num_queries
-        self.two_stage_num_proposals = two_stage_num_proposals
-        self.num_channels = num_channels
-        self.image_size = image_size
-        self.n_targets = n_targets
-        self.num_labels = num_labels
-        self.num_feature_levels = num_feature_levels
-        self.encoder_n_points = encoder_n_points
-        self.decoder_n_points = decoder_n_points
-        self.two_stage = two_stage
-        self.assign_first_stage = assign_first_stage
-        self.assign_second_stage = assign_second_stage
-
-        # we also set the expected seq length for both encoder and decoder
-        self.encoder_seq_length = (
-            math.ceil(self.image_size / 8) ** 2
-            + math.ceil(self.image_size / 16) ** 2
-            + math.ceil(self.image_size / 32) ** 2
-            + math.ceil(self.image_size / 64) ** 2
-        )
-        self.decoder_seq_length = self.num_queries
-
-    def prepare_config_and_inputs(self, model_class_name):
-        pixel_values = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size])
-
-        pixel_mask = torch.ones([self.batch_size, self.image_size, self.image_size], device=torch_device)
-
-        labels = None
-        if self.use_labels:
-            # labels is a list of Dict (each Dict being the labels for a given example in the batch)
-            labels = []
-            for i in range(self.batch_size):
-                target = {}
-                target["class_labels"] = torch.randint(
-                    high=self.num_labels, size=(self.n_targets,), device=torch_device
-                )
-                target["boxes"] = torch.rand(self.n_targets, 4, device=torch_device)
-                target["masks"] = torch.rand(self.n_targets, self.image_size, self.image_size, device=torch_device)
-                labels.append(target)
-
-        config = self.get_config(model_class_name)
-        return config, pixel_values, pixel_mask, labels
-
-    def get_config(self, model_class_name):
-        resnet_config = ResNetConfig(
-            num_channels=3,
-            embeddings_size=10,
-            hidden_sizes=[10, 20, 30, 40],
-            depths=[1, 1, 2, 1],
-            hidden_act="relu",
-            num_labels=3,
-            out_features=["stage2", "stage3", "stage4"],
-            out_indices=[2, 3, 4],
-        )
-        two_stage = model_class_name == "DetaForObjectDetection"
-        assign_first_stage = model_class_name == "DetaForObjectDetection"
-        assign_second_stage = model_class_name == "DetaForObjectDetection"
-        return DetaConfig(
-            d_model=self.hidden_size,
-            encoder_layers=self.num_hidden_layers,
-            decoder_layers=self.num_hidden_layers,
-            encoder_attention_heads=self.num_attention_heads,
-            decoder_attention_heads=self.num_attention_heads,
-            encoder_ffn_dim=self.intermediate_size,
-            decoder_ffn_dim=self.intermediate_size,
-            dropout=self.hidden_dropout_prob,
-            attention_dropout=self.attention_probs_dropout_prob,
-            num_queries=self.num_queries,
-            two_stage_num_proposals=self.two_stage_num_proposals,
-            num_labels=self.num_labels,
-            num_feature_levels=self.num_feature_levels,
-            encoder_n_points=self.encoder_n_points,
-            decoder_n_points=self.decoder_n_points,
-            two_stage=two_stage,
-            assign_first_stage=assign_first_stage,
-            assign_second_stage=assign_second_stage,
-            backbone_config=resnet_config,
-            backbone=None,
-        )
-
-    def prepare_config_and_inputs_for_common(self, model_class_name="DetaModel"):
-        config, pixel_values, pixel_mask, labels = self.prepare_config_and_inputs(model_class_name)
-        inputs_dict = {"pixel_values": pixel_values, "pixel_mask": pixel_mask}
-        return config, inputs_dict
-
-    def create_and_check_deta_model(self, config, pixel_values, pixel_mask, labels):
-        model = DetaModel(config=config)
-        model.to(torch_device)
-        model.eval()
-
-        result = model(pixel_values=pixel_values, pixel_mask=pixel_mask)
-        result = model(pixel_values)
-
-        self.parent.assertEqual(result.last_hidden_state.shape, (self.batch_size, self.num_queries, self.hidden_size))
-
-    def create_and_check_deta_freeze_backbone(self, config, pixel_values, pixel_mask, labels):
-        model = DetaModel(config=config)
-        model.to(torch_device)
-        model.eval()
-
-        model.freeze_backbone()
-
-        for _, param in model.backbone.model.named_parameters():
-            self.parent.assertEqual(False, param.requires_grad)
-
-    def create_and_check_deta_unfreeze_backbone(self, config, pixel_values, pixel_mask, labels):
-        model = DetaModel(config=config)
-        model.to(torch_device)
-        model.eval()
-
-        model.unfreeze_backbone()
-
-        for _, param in model.backbone.model.named_parameters():
-            self.parent.assertEqual(True, param.requires_grad)
-
-    def create_and_check_deta_object_detection_head_model(self, config, pixel_values, pixel_mask, labels):
-        model = DetaForObjectDetection(config=config)
-        model.to(torch_device)
-        model.eval()
-
-        result = model(pixel_values=pixel_values, pixel_mask=pixel_mask)
-        result = model(pixel_values)
-
-        self.parent.assertEqual(result.logits.shape, (self.batch_size, self.two_stage_num_proposals, self.num_labels))
-        self.parent.assertEqual(result.pred_boxes.shape, (self.batch_size, self.two_stage_num_proposals, 4))
-
-        result = model(pixel_values=pixel_values, pixel_mask=pixel_mask, labels=labels)
-
-        self.parent.assertEqual(result.loss.shape, ())
-        self.parent.assertEqual(result.logits.shape, (self.batch_size, self.two_stage_num_proposals, self.num_labels))
-        self.parent.assertEqual(result.pred_boxes.shape, (self.batch_size, self.two_stage_num_proposals, 4))
-
-
-@require_torchvision
-class DetaModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixin, unittest.TestCase):
-    all_model_classes = (DetaModel, DetaForObjectDetection) if is_torchvision_available() else ()
-    pipeline_model_mapping = (
-        {"image-feature-extraction": DetaModel, "object-detection": DetaForObjectDetection}
-        if is_torchvision_available()
-        else {}
-    )
-    is_encoder_decoder = True
-    test_torchscript = False
-    test_pruning = False
-    test_head_masking = False
-    test_missing_keys = False
-
-    # TODO: Fix the failed tests when this model gets more usage
-    def is_pipeline_test_to_skip(
-        self, pipeline_test_casse_name, config_class, model_architecture, tokenizer_name, processor_name
-    ):
-        if pipeline_test_casse_name == "ObjectDetectionPipelineTests":
-            return True
-
-        return False
-
-    @unittest.skip("Skip for now. PR #22437 causes some loading issue. See (not merged) #22656 for some discussions.")
-    def test_can_use_safetensors(self):
-        super().test_can_use_safetensors()
-
-    # special case for head models
-    def _prepare_for_class(self, inputs_dict, model_class, return_labels=False):
-        inputs_dict = super()._prepare_for_class(inputs_dict, model_class, return_labels=return_labels)
-
-        if return_labels:
-            if model_class.__name__ == "DetaForObjectDetection":
-                labels = []
-                for i in range(self.model_tester.batch_size):
-                    target = {}
-                    target["class_labels"] = torch.ones(
-                        size=(self.model_tester.n_targets,), device=torch_device, dtype=torch.long
-                    )
-                    target["boxes"] = torch.ones(
-                        self.model_tester.n_targets, 4, device=torch_device, dtype=torch.float
-                    )
-                    target["masks"] = torch.ones(
-                        self.model_tester.n_targets,
-                        self.model_tester.image_size,
-                        self.model_tester.image_size,
-                        device=torch_device,
-                        dtype=torch.float,
-                    )
-                    labels.append(target)
-                inputs_dict["labels"] = labels
-
-        return inputs_dict
-
-    def setUp(self):
-        self.model_tester = DetaModelTester(self)
-        self.config_tester = ConfigTester(self, config_class=DetaConfig, has_text_modality=False)
-
-    def test_config(self):
-        # we don't test common_properties and arguments_init as these don't apply for DETA
-        self.config_tester.create_and_test_config_to_json_string()
-        self.config_tester.create_and_test_config_to_json_file()
-        self.config_tester.create_and_test_config_from_and_save_pretrained()
-        self.config_tester.create_and_test_config_with_num_labels()
-        self.config_tester.check_config_can_be_init_without_params()
-
-    def test_deta_model(self):
-        config_and_inputs = self.model_tester.prepare_config_and_inputs(model_class_name="DetaModel")
-        self.model_tester.create_and_check_deta_model(*config_and_inputs)
-
-    def test_deta_freeze_backbone(self):
-        config_and_inputs = self.model_tester.prepare_config_and_inputs(model_class_name="DetaModel")
-        self.model_tester.create_and_check_deta_freeze_backbone(*config_and_inputs)
-
-    def test_deta_unfreeze_backbone(self):
-        config_and_inputs = self.model_tester.prepare_config_and_inputs(model_class_name="DetaModel")
-        self.model_tester.create_and_check_deta_unfreeze_backbone(*config_and_inputs)
-
-    def test_deta_object_detection_head_model(self):
-        config_and_inputs = self.model_tester.prepare_config_and_inputs(model_class_name="DetaForObjectDetection")
-        self.model_tester.create_and_check_deta_object_detection_head_model(*config_and_inputs)
-
-    @unittest.skip(reason="DETA does not use inputs_embeds")
-    def test_inputs_embeds(self):
-        pass
-
-    @unittest.skip(reason="DETA does not use inputs_embeds")
-    def test_inputs_embeds_matches_input_ids(self):
-        pass
-
-    @unittest.skip(reason="DETA does not have a get_input_embeddings method")
-    def test_model_common_attributes(self):
-        pass
-
-    @unittest.skip(reason="DETA is not a generative model")
-    def test_generate_without_input_ids(self):
-        pass
-
-    @unittest.skip(reason="DETA does not use token embeddings")
-    def test_resize_tokens_embeddings(self):
-        pass
-
-    @unittest.skip(reason="Feed forward chunking is not implemented")
-    def test_feed_forward_chunking(self):
-        pass
-
-    def test_attention_outputs(self):
-        config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
-        config.return_dict = True
-
-        for model_class in self.all_model_classes:
-            inputs_dict["output_attentions"] = True
-            inputs_dict["output_hidden_states"] = False
-            config.return_dict = True
-            model = model_class(config)
-            model.to(torch_device)
-            model.eval()
-            with torch.no_grad():
-                outputs = model(**self._prepare_for_class(inputs_dict, model_class))
-            attentions = outputs.encoder_attentions
-            self.assertEqual(len(attentions), self.model_tester.num_hidden_layers)
-
-            # check that output_attentions also work using config
-            del inputs_dict["output_attentions"]
-            config.output_attentions = True
-            model = model_class(config)
-            model.to(torch_device)
-            model.eval()
-            with torch.no_grad():
-                outputs = model(**self._prepare_for_class(inputs_dict, model_class))
-            attentions = outputs.encoder_attentions
-            self.assertEqual(len(attentions), self.model_tester.num_hidden_layers)
-
-            self.assertListEqual(
-                list(attentions[0].shape[-3:]),
-                [
-                    self.model_tester.num_attention_heads,
-                    self.model_tester.num_feature_levels,
-                    self.model_tester.encoder_n_points,
-                ],
-            )
-            out_len = len(outputs)
-
-            correct_outlen = 8
-
-            # loss is at first position
-            if "labels" in inputs_dict:
-                correct_outlen += 1  # loss is added to beginning
-            # Object Detection model returns pred_logits and pred_boxes
-            if model_class.__name__ == "DetaForObjectDetection":
-                correct_outlen += 2
-
-            self.assertEqual(out_len, correct_outlen)
-
-            # decoder attentions
-            decoder_attentions = outputs.decoder_attentions
-            self.assertIsInstance(decoder_attentions, (list, tuple))
-            self.assertEqual(len(decoder_attentions), self.model_tester.num_hidden_layers)
-            self.assertListEqual(
-                list(decoder_attentions[0].shape[-3:]),
-                [self.model_tester.num_attention_heads, self.model_tester.num_queries, self.model_tester.num_queries],
-            )
-
-            # cross attentions
-            cross_attentions = outputs.cross_attentions
-            self.assertIsInstance(cross_attentions, (list, tuple))
-            self.assertEqual(len(cross_attentions), self.model_tester.num_hidden_layers)
-            self.assertListEqual(
-                list(cross_attentions[0].shape[-3:]),
-                [
-                    self.model_tester.num_attention_heads,
-                    self.model_tester.num_feature_levels,
-                    self.model_tester.decoder_n_points,
-                ],
-            )
-
-            # Check attention is always last and order is fine
-            inputs_dict["output_attentions"] = True
-            inputs_dict["output_hidden_states"] = True
-            model = model_class(config)
-            model.to(torch_device)
-            model.eval()
-            with torch.no_grad():
-                outputs = model(**self._prepare_for_class(inputs_dict, model_class))
-
-            if hasattr(self.model_tester, "num_hidden_states_types"):
-                added_hidden_states = self.model_tester.num_hidden_states_types
-            elif self.is_encoder_decoder:
-                added_hidden_states = 2
-            else:
-                added_hidden_states = 1
-            self.assertEqual(out_len + added_hidden_states, len(outputs))
-
-            self_attentions = outputs.encoder_attentions
-
-            self.assertEqual(len(self_attentions), self.model_tester.num_hidden_layers)
-            self.assertListEqual(
-                list(self_attentions[0].shape[-3:]),
-                [
-                    self.model_tester.num_attention_heads,
-                    self.model_tester.num_feature_levels,
-                    self.model_tester.encoder_n_points,
-                ],
-            )
-
-    # removed retain_grad and grad on decoder_hidden_states, as queries don't require grad
-    def test_retain_grad_hidden_states_attentions(self):
-        config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
-        config.output_hidden_states = True
-        config.output_attentions = True
-
-        # no need to test all models as different heads yield the same functionality
-        model_class = self.all_model_classes[0]
-        model = model_class(config)
-        model.to(torch_device)
-
-        inputs = self._prepare_for_class(inputs_dict, model_class)
-
-        outputs = model(**inputs)
-
-        # we take the second output since last_hidden_state is the second item
-        output = outputs[1]
-
-        encoder_hidden_states = outputs.encoder_hidden_states[0]
-        encoder_attentions = outputs.encoder_attentions[0]
-        encoder_hidden_states.retain_grad()
-        encoder_attentions.retain_grad()
-
-        decoder_attentions = outputs.decoder_attentions[0]
-        decoder_attentions.retain_grad()
-
-        cross_attentions = outputs.cross_attentions[0]
-        cross_attentions.retain_grad()
-
-        output.flatten()[0].backward(retain_graph=True)
-
-        self.assertIsNotNone(encoder_hidden_states.grad)
-        self.assertIsNotNone(encoder_attentions.grad)
-        self.assertIsNotNone(decoder_attentions.grad)
-        self.assertIsNotNone(cross_attentions.grad)
-
-    def test_forward_auxiliary_loss(self):
-        config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
-        config.auxiliary_loss = True
-
-        # only test for object detection and segmentation model
-        for model_class in self.all_model_classes[1:]:
-            model = model_class(config)
-            model.to(torch_device)
-
-            inputs = self._prepare_for_class(inputs_dict, model_class, return_labels=True)
-
-            outputs = model(**inputs)
-
-            self.assertIsNotNone(outputs.auxiliary_outputs)
-            self.assertEqual(len(outputs.auxiliary_outputs), self.model_tester.num_hidden_layers - 1)
-
-    def test_forward_signature(self):
-        config, _ = self.model_tester.prepare_config_and_inputs_for_common()
-
-        for model_class in self.all_model_classes:
-            model = model_class(config)
-            signature = inspect.signature(model.forward)
-            # signature.parameters is an OrderedDict => so arg_names order is deterministic
-            arg_names = [*signature.parameters.keys()]
-
-            if model.config.is_encoder_decoder:
-                expected_arg_names = ["pixel_values", "pixel_mask"]
-                expected_arg_names.extend(
-                    ["head_mask", "decoder_head_mask", "encoder_outputs"]
-                    if "head_mask" and "decoder_head_mask" in arg_names
-                    else []
-                )
-                self.assertListEqual(arg_names[: len(expected_arg_names)], expected_arg_names)
-            else:
-                expected_arg_names = ["pixel_values", "pixel_mask"]
-                self.assertListEqual(arg_names[:1], expected_arg_names)
-
-    @unittest.skip(reason="Model doesn't use tied weights")
-    def test_tied_model_weights_key_ignore(self):
-        pass
-
-    def test_initialization(self):
-        config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
-
-        configs_no_init = _config_zero_init(config)
-        for model_class in self.all_model_classes:
-            model = model_class(config=configs_no_init)
-            # Skip the check for the backbone
-            for name, module in model.named_modules():
-                if module.__class__.__name__ == "DetaBackboneWithPositionalEncodings":
-                    backbone_params = [f"{name}.{key}" for key in module.state_dict().keys()]
-                    break
-
-            for name, param in model.named_parameters():
-                if param.requires_grad:
-                    if (
-                        "level_embed" in name
-                        or "sampling_offsets.bias" in name
-                        or "value_proj" in name
-                        or "output_proj" in name
-                        or "reference_points" in name
-                        or name in backbone_params
-                    ):
-                        continue
-                    self.assertIn(
-                        ((param.data.mean() * 1e9).round() / 1e9).item(),
-                        [0.0, 1.0],
-                        msg=f"Parameter {name} of model {model_class} seems not properly initialized",
-                    )
-
-    @unittest.skip("No support for low_cpu_mem_usage=True.")
-    def test_save_load_low_cpu_mem_usage(self):
-        pass
-
-    @unittest.skip("No support for low_cpu_mem_usage=True.")
-    def test_save_load_low_cpu_mem_usage_checkpoints(self):
-        pass
-
-    @unittest.skip("No support for low_cpu_mem_usage=True.")
-    def test_save_load_low_cpu_mem_usage_no_safetensors(self):
-        pass
-
-    # Inspired by tests.test_modeling_common.ModelTesterMixin.test_tied_weights_keys
-    def test_tied_weights_keys(self):
-        for model_class in self.all_model_classes:
-            # We need to pass model class name to correctly initialize the config.
-            # If we don't pass it, the config for `DetaForObjectDetection`` will be initialized
-            # with `two_stage=False` and the test will fail because for that case `class_embed`
-            # weights are not tied.
-            config, _ = self.model_tester.prepare_config_and_inputs_for_common(model_class_name=model_class.__name__)
-            config.tie_word_embeddings = True
-
-            model_tied = model_class(config)
-
-            ptrs = collections.defaultdict(list)
-            for name, tensor in model_tied.state_dict().items():
-                ptrs[id_tensor_storage(tensor)].append(name)
-
-            # These are all the pointers of shared tensors.
-            tied_params = [names for _, names in ptrs.items() if len(names) > 1]
-
-            tied_weight_keys = model_tied._tied_weights_keys if model_tied._tied_weights_keys is not None else []
-            # Detect we get a hit for each key
-            for key in tied_weight_keys:
-                is_tied_key = any(re.search(key, p) for group in tied_params for p in group)
-                self.assertTrue(is_tied_key, f"{key} is not a tied weight key for {model_class}.")
-
-            # Removed tied weights found from tied params -> there should only be one left after
-            for key in tied_weight_keys:
-                for i in range(len(tied_params)):
-                    tied_params[i] = [p for p in tied_params[i] if re.search(key, p) is None]
-
-            tied_params = [group for group in tied_params if len(group) > 1]
-            self.assertListEqual(
-                tied_params,
-                [],
-                f"Missing `_tied_weights_keys` for {model_class}: add all of {tied_params} except one.",
-            )
-
-
-TOLERANCE = 1e-4
-
-
-# We will verify our results on an image of cute cats
-def prepare_img():
-    image = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png")
-    return image
-
-
-@require_torchvision
-@require_vision
-@slow
-class DetaModelIntegrationTests(unittest.TestCase):
-    @cached_property
-    def default_image_processor(self):
-        return AutoImageProcessor.from_pretrained("jozhang97/deta-resnet-50") if is_vision_available() else None
-
-    def test_inference_object_detection_head(self):
-        model = DetaForObjectDetection.from_pretrained("jozhang97/deta-resnet-50").to(torch_device)
-
-        image_processor = self.default_image_processor
-        image = prepare_img()
-        inputs = image_processor(images=image, return_tensors="pt").to(torch_device)
-
-        with torch.no_grad():
-            outputs = model(**inputs)
-
-        expected_shape_logits = torch.Size((1, 300, model.config.num_labels))
-        self.assertEqual(outputs.logits.shape, expected_shape_logits)
-
-        expected_logits = torch.tensor(
-            [[-7.3978, -2.5406, -4.1668], [-8.2684, -3.9933, -3.8096], [-7.0515, -3.7973, -5.8516]]
-        ).to(torch_device)
-        expected_boxes = torch.tensor(
-            [[0.5043, 0.4973, 0.9998], [0.2542, 0.5489, 0.4748], [0.5490, 0.2765, 0.0570]]
-        ).to(torch_device)
-
-        self.assertTrue(torch.allclose(outputs.logits[0, :3, :3], expected_logits, atol=1e-4))
-
-        expected_shape_boxes = torch.Size((1, 300, 4))
-        self.assertEqual(outputs.pred_boxes.shape, expected_shape_boxes)
-        self.assertTrue(torch.allclose(outputs.pred_boxes[0, :3, :3], expected_boxes, atol=1e-4))
-
-        # verify postprocessing
-        results = image_processor.post_process_object_detection(
-            outputs, threshold=0.3, target_sizes=[image.size[::-1]]
-        )[0]
-        expected_scores = torch.tensor([0.6392, 0.6276, 0.5546, 0.5260, 0.4706], device=torch_device)
-        expected_labels = [75, 17, 17, 75, 63]
-        expected_slice_boxes = torch.tensor([40.5866, 73.2107, 176.1421, 117.1751], device=torch_device)
-
-        self.assertTrue(torch.allclose(results["scores"], expected_scores, atol=1e-4))
-        self.assertSequenceEqual(results["labels"].tolist(), expected_labels)
-        self.assertTrue(torch.allclose(results["boxes"][0, :], expected_slice_boxes))
-
-    def test_inference_object_detection_head_swin_backbone(self):
-        model = DetaForObjectDetection.from_pretrained("jozhang97/deta-swin-large").to(torch_device)
-
-        image_processor = self.default_image_processor
-        image = prepare_img()
-        inputs = image_processor(images=image, return_tensors="pt").to(torch_device)
-
-        with torch.no_grad():
-            outputs = model(**inputs)
-
-        expected_shape_logits = torch.Size((1, 300, model.config.num_labels))
-        self.assertEqual(outputs.logits.shape, expected_shape_logits)
-
-        expected_logits = torch.tensor(
-            [[-7.6308, -2.8485, -5.3737], [-7.2037, -4.5505, -4.8027], [-7.2943, -4.2611, -4.6617]]
-        ).to(torch_device)
-        expected_boxes = torch.tensor(
-            [[0.4987, 0.4969, 0.9999], [0.2549, 0.5498, 0.4805], [0.5498, 0.2757, 0.0569]]
-        ).to(torch_device)
-
-        self.assertTrue(torch.allclose(outputs.logits[0, :3, :3], expected_logits, atol=1e-4))
-
-        expected_shape_boxes = torch.Size((1, 300, 4))
-        self.assertEqual(outputs.pred_boxes.shape, expected_shape_boxes)
-        self.assertTrue(torch.allclose(outputs.pred_boxes[0, :3, :3], expected_boxes, atol=1e-4))
-
-        # verify postprocessing
-        results = image_processor.post_process_object_detection(
-            outputs, threshold=0.3, target_sizes=[image.size[::-1]]
-        )[0]
-        expected_scores = torch.tensor([0.6831, 0.6826, 0.5684, 0.5464, 0.4392], device=torch_device)
-        expected_labels = [17, 17, 75, 75, 63]
-        expected_slice_boxes = torch.tensor([345.8478, 23.6754, 639.8562, 372.8265], device=torch_device)
-
-        self.assertTrue(torch.allclose(results["scores"], expected_scores, atol=1e-4))
-        self.assertSequenceEqual(results["labels"].tolist(), expected_labels)
-        self.assertTrue(torch.allclose(results["boxes"][0, :], expected_slice_boxes))

tests/models/efficientformer/test_image_processing_efficientformer.py

-# coding=utf-8
-# Copyright 2021 HuggingFace Inc.
-#
-# 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.testing_utils import require_torch, require_vision
-from transformers.utils import is_vision_available
-
-from ...test_image_processing_common import ImageProcessingTestMixin, prepare_image_inputs
-
-
-if is_vision_available():
-    from transformers import ViTImageProcessor
-
-
-class EfficientFormerImageProcessorTester(unittest.TestCase):
-    def __init__(
-        self,
-        parent,
-        batch_size=13,
-        num_channels=3,
-        image_size=224,
-        min_resolution=30,
-        max_resolution=400,
-        do_resize=True,
-        size=None,
-        do_normalize=True,
-        image_mean=[0.5, 0.5, 0.5],
-        image_std=[0.5, 0.5, 0.5],
-    ):
-        size = size if size is not None else {"height": 18, "width": 18}
-        self.parent = parent
-        self.batch_size = batch_size
-        self.num_channels = num_channels
-        self.image_size = image_size
-        self.min_resolution = min_resolution
-        self.max_resolution = max_resolution
-        self.do_resize = do_resize
-        self.size = size
-        self.do_normalize = do_normalize
-        self.image_mean = image_mean
-        self.image_std = image_std
-
-    def prepare_image_processor_dict(self):
-        return {
-            "image_mean": self.image_mean,
-            "image_std": self.image_std,
-            "do_normalize": self.do_normalize,
-            "do_resize": self.do_resize,
-            "size": self.size,
-        }
-
-    def expected_output_image_shape(self, images):
-        return self.num_channels, self.size["height"], self.size["width"]
-
-    def prepare_image_inputs(self, equal_resolution=False, numpify=False, torchify=False):
-        return prepare_image_inputs(
-            batch_size=self.batch_size,
-            num_channels=self.num_channels,
-            min_resolution=self.min_resolution,
-            max_resolution=self.max_resolution,
-            equal_resolution=equal_resolution,
-            numpify=numpify,
-            torchify=torchify,
-        )
-
-
-@require_torch
-@require_vision
-class EfficientFormerImageProcessorTest(ImageProcessingTestMixin, unittest.TestCase):
-    image_processing_class = ViTImageProcessor if is_vision_available() else None
-
-    def setUp(self):
-        self.image_processor_tester = EfficientFormerImageProcessorTester(self)
-
-    @property
-    def image_processor_dict(self):
-        return self.image_processor_tester.prepare_image_processor_dict()
-
-    def test_image_proc_properties(self):
-        image_processor = self.image_processing_class(**self.image_processor_dict)
-        self.assertTrue(hasattr(image_processor, "image_mean"))
-        self.assertTrue(hasattr(image_processor, "image_std"))
-        self.assertTrue(hasattr(image_processor, "do_normalize"))
-        self.assertTrue(hasattr(image_processor, "do_resize"))
-        self.assertTrue(hasattr(image_processor, "size"))