From 3e8d17e66d223e681a08272351368ab077560d1d Mon Sep 17 00:00:00 2001
From: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
Date: Tue, 16 Nov 2021 09:24:40 -0500
Subject: [PATCH] Add forward method to dummy models (#14419)

* Add forward method to dummy models

* Fix quality
---
 src/transformers/utils/dummy_flax_objects.py  |  291 +++++
 src/transformers/utils/dummy_pt_objects.py    | 1092 +++++++++++++++++
 .../utils/dummy_scatter_objects.py            |   15 +
 src/transformers/utils/dummy_tf_objects.py    |  624 ++++++++++
 .../utils/dummy_timm_and_vision_objects.py    |   12 +
 utils/check_dummies.py                        |   41 +-
 6 files changed, 2071 insertions(+), 4 deletions(-)

diff --git a/src/transformers/utils/dummy_flax_objects.py b/src/transformers/utils/dummy_flax_objects.py
index 655f6bd566d..cf54fab8119 100644
--- a/src/transformers/utils/dummy_flax_objects.py
+++ b/src/transformers/utils/dummy_flax_objects.py
@@ -75,6 +75,9 @@ class FlaxPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAlbertForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -84,6 +87,9 @@ class FlaxAlbertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAlbertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -93,6 +99,9 @@ class FlaxAlbertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAlbertForPreTraining:
     def __init__(self, *args, **kwargs):
@@ -107,6 +116,9 @@ class FlaxAlbertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAlbertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -116,6 +128,9 @@ class FlaxAlbertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAlbertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -125,6 +140,9 @@ class FlaxAlbertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAlbertModel:
     def __init__(self, *args, **kwargs):
@@ -134,6 +152,9 @@ class FlaxAlbertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAlbertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -143,6 +164,9 @@ class FlaxAlbertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 FLAX_MODEL_FOR_CAUSAL_LM_MAPPING = None
 
@@ -188,6 +212,9 @@ class FlaxAutoModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAutoModelForCausalLM:
     def __init__(self, *args, **kwargs):
@@ -197,6 +224,9 @@ class FlaxAutoModelForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAutoModelForImageClassification:
     def __init__(self, *args, **kwargs):
@@ -206,6 +236,9 @@ class FlaxAutoModelForImageClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAutoModelForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -215,6 +248,9 @@ class FlaxAutoModelForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAutoModelForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -224,6 +260,9 @@ class FlaxAutoModelForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAutoModelForNextSentencePrediction:
     def __init__(self, *args, **kwargs):
@@ -233,6 +272,9 @@ class FlaxAutoModelForNextSentencePrediction:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAutoModelForPreTraining:
     def __init__(self, *args, **kwargs):
@@ -242,6 +284,9 @@ class FlaxAutoModelForPreTraining:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAutoModelForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -251,6 +296,9 @@ class FlaxAutoModelForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAutoModelForSeq2SeqLM:
     def __init__(self, *args, **kwargs):
@@ -260,6 +308,9 @@ class FlaxAutoModelForSeq2SeqLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAutoModelForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -269,6 +320,9 @@ class FlaxAutoModelForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAutoModelForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -278,6 +332,9 @@ class FlaxAutoModelForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxAutoModelForVision2Seq:
     def __init__(self, *args, **kwargs):
@@ -287,6 +344,9 @@ class FlaxAutoModelForVision2Seq:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBartForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -296,6 +356,9 @@ class FlaxBartForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBartForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -305,6 +368,9 @@ class FlaxBartForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBartForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -314,6 +380,9 @@ class FlaxBartForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBartModel:
     def __init__(self, *args, **kwargs):
@@ -323,6 +392,9 @@ class FlaxBartModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBartPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -332,6 +404,9 @@ class FlaxBartPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBeitForImageClassification:
     def __init__(self, *args, **kwargs):
@@ -346,6 +421,9 @@ class FlaxBeitForMaskedImageModeling:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBeitModel:
     def __init__(self, *args, **kwargs):
@@ -355,6 +433,9 @@ class FlaxBeitModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBeitPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -364,6 +445,9 @@ class FlaxBeitPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBertForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -373,6 +457,9 @@ class FlaxBertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -382,6 +469,9 @@ class FlaxBertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBertForNextSentencePrediction:
     def __init__(self, *args, **kwargs):
@@ -401,6 +491,9 @@ class FlaxBertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -410,6 +503,9 @@ class FlaxBertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -419,6 +515,9 @@ class FlaxBertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBertModel:
     def __init__(self, *args, **kwargs):
@@ -428,6 +527,9 @@ class FlaxBertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -437,6 +539,9 @@ class FlaxBertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBigBirdForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -446,6 +551,9 @@ class FlaxBigBirdForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBigBirdForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -455,6 +563,9 @@ class FlaxBigBirdForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBigBirdForPreTraining:
     def __init__(self, *args, **kwargs):
@@ -469,6 +580,9 @@ class FlaxBigBirdForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBigBirdForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -478,6 +592,9 @@ class FlaxBigBirdForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBigBirdForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -487,6 +604,9 @@ class FlaxBigBirdForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBigBirdModel:
     def __init__(self, *args, **kwargs):
@@ -496,6 +616,9 @@ class FlaxBigBirdModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxBigBirdPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -505,6 +628,9 @@ class FlaxBigBirdPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxCLIPModel:
     def __init__(self, *args, **kwargs):
@@ -514,6 +640,9 @@ class FlaxCLIPModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxCLIPPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -523,6 +652,9 @@ class FlaxCLIPPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxCLIPTextModel:
     def __init__(self, *args, **kwargs):
@@ -532,6 +664,9 @@ class FlaxCLIPTextModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxCLIPTextPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -541,6 +676,9 @@ class FlaxCLIPTextPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxCLIPVisionModel:
     def __init__(self, *args, **kwargs):
@@ -550,6 +688,9 @@ class FlaxCLIPVisionModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxCLIPVisionPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -559,6 +700,9 @@ class FlaxCLIPVisionPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxDistilBertForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -568,6 +712,9 @@ class FlaxDistilBertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxDistilBertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -577,6 +724,9 @@ class FlaxDistilBertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxDistilBertForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -586,6 +736,9 @@ class FlaxDistilBertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxDistilBertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -595,6 +748,9 @@ class FlaxDistilBertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxDistilBertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -604,6 +760,9 @@ class FlaxDistilBertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxDistilBertModel:
     def __init__(self, *args, **kwargs):
@@ -613,6 +772,9 @@ class FlaxDistilBertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxDistilBertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -622,6 +784,9 @@ class FlaxDistilBertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxElectraForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -631,6 +796,9 @@ class FlaxElectraForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxElectraForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -640,6 +808,9 @@ class FlaxElectraForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxElectraForPreTraining:
     def __init__(self, *args, **kwargs):
@@ -654,6 +825,9 @@ class FlaxElectraForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxElectraForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -663,6 +837,9 @@ class FlaxElectraForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxElectraForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -672,6 +849,9 @@ class FlaxElectraForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxElectraModel:
     def __init__(self, *args, **kwargs):
@@ -681,6 +861,9 @@ class FlaxElectraModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxElectraPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -690,6 +873,9 @@ class FlaxElectraPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxEncoderDecoderModel:
     def __init__(self, *args, **kwargs):
@@ -699,6 +885,9 @@ class FlaxEncoderDecoderModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxGPT2LMHeadModel:
     def __init__(self, *args, **kwargs):
@@ -708,6 +897,9 @@ class FlaxGPT2LMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxGPT2Model:
     def __init__(self, *args, **kwargs):
@@ -717,6 +909,9 @@ class FlaxGPT2Model:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxGPT2PreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -726,6 +921,9 @@ class FlaxGPT2PreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxGPTNeoForCausalLM:
     def __init__(self, *args, **kwargs):
@@ -735,6 +933,9 @@ class FlaxGPTNeoForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxGPTNeoModel:
     def __init__(self, *args, **kwargs):
@@ -744,6 +945,9 @@ class FlaxGPTNeoModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxGPTNeoPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -753,6 +957,9 @@ class FlaxGPTNeoPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxMarianModel:
     def __init__(self, *args, **kwargs):
@@ -762,6 +969,9 @@ class FlaxMarianModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxMarianMTModel:
     def __init__(self, *args, **kwargs):
@@ -771,6 +981,9 @@ class FlaxMarianMTModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxMarianPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -780,6 +993,9 @@ class FlaxMarianPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxMBartForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -789,6 +1005,9 @@ class FlaxMBartForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxMBartForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -798,6 +1017,9 @@ class FlaxMBartForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxMBartForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -807,6 +1029,9 @@ class FlaxMBartForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxMBartModel:
     def __init__(self, *args, **kwargs):
@@ -816,6 +1041,9 @@ class FlaxMBartModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxMBartPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -825,6 +1053,9 @@ class FlaxMBartPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxMT5ForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -834,6 +1065,9 @@ class FlaxMT5ForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxMT5Model:
     def __init__(self, *args, **kwargs):
@@ -843,6 +1077,9 @@ class FlaxMT5Model:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxPegasusForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -852,6 +1089,9 @@ class FlaxPegasusForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxPegasusModel:
     def __init__(self, *args, **kwargs):
@@ -861,6 +1101,9 @@ class FlaxPegasusModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxPegasusPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -870,6 +1113,9 @@ class FlaxPegasusPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxRobertaForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -879,6 +1125,9 @@ class FlaxRobertaForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxRobertaForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -888,6 +1137,9 @@ class FlaxRobertaForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxRobertaForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -897,6 +1149,9 @@ class FlaxRobertaForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxRobertaForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -906,6 +1161,9 @@ class FlaxRobertaForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxRobertaForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -915,6 +1173,9 @@ class FlaxRobertaForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxRobertaModel:
     def __init__(self, *args, **kwargs):
@@ -924,6 +1185,9 @@ class FlaxRobertaModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxRobertaPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -933,6 +1197,9 @@ class FlaxRobertaPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxT5ForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -942,6 +1209,9 @@ class FlaxT5ForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxT5Model:
     def __init__(self, *args, **kwargs):
@@ -951,6 +1221,9 @@ class FlaxT5Model:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxT5PreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -960,6 +1233,9 @@ class FlaxT5PreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxVisionEncoderDecoderModel:
     def __init__(self, *args, **kwargs):
@@ -969,6 +1245,9 @@ class FlaxVisionEncoderDecoderModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxViTForImageClassification:
     def __init__(self, *args, **kwargs):
@@ -983,6 +1262,9 @@ class FlaxViTModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxViTPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -992,6 +1274,9 @@ class FlaxViTPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxWav2Vec2ForCTC:
     def __init__(self, *args, **kwargs):
@@ -1011,6 +1296,9 @@ class FlaxWav2Vec2Model:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
 
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
+
 
 class FlaxWav2Vec2PreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1019,3 +1307,6 @@ class FlaxWav2Vec2PreTrainedModel:
     @classmethod
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["flax"])
+
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, ["flax"])
diff --git a/src/transformers/utils/dummy_pt_objects.py b/src/transformers/utils/dummy_pt_objects.py
index 7f15a016842..9e2d466dab1 100644
--- a/src/transformers/utils/dummy_pt_objects.py
+++ b/src/transformers/utils/dummy_pt_objects.py
@@ -223,6 +223,9 @@ class PreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def apply_chunking_to_forward(*args, **kwargs):
     requires_backends(apply_chunking_to_forward, ["torch"])
@@ -243,6 +246,9 @@ class AlbertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AlbertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -252,6 +258,9 @@ class AlbertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AlbertForPreTraining:
     def __init__(self, *args, **kwargs):
@@ -266,6 +275,9 @@ class AlbertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AlbertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -275,6 +287,9 @@ class AlbertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AlbertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -284,6 +299,9 @@ class AlbertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AlbertModel:
     def __init__(self, *args, **kwargs):
@@ -293,6 +311,9 @@ class AlbertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AlbertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -302,6 +323,9 @@ class AlbertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_albert(*args, **kwargs):
     requires_backends(load_tf_weights_in_albert, ["torch"])
@@ -372,6 +396,9 @@ class AutoModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForAudioClassification:
     def __init__(self, *args, **kwargs):
@@ -381,6 +408,9 @@ class AutoModelForAudioClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForCausalLM:
     def __init__(self, *args, **kwargs):
@@ -390,6 +420,9 @@ class AutoModelForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForCTC:
     def __init__(self, *args, **kwargs):
@@ -399,6 +432,9 @@ class AutoModelForCTC:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForImageClassification:
     def __init__(self, *args, **kwargs):
@@ -408,6 +444,9 @@ class AutoModelForImageClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForImageSegmentation:
     def __init__(self, *args, **kwargs):
@@ -417,6 +456,9 @@ class AutoModelForImageSegmentation:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -426,6 +468,9 @@ class AutoModelForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -435,6 +480,9 @@ class AutoModelForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForNextSentencePrediction:
     def __init__(self, *args, **kwargs):
@@ -444,6 +492,9 @@ class AutoModelForNextSentencePrediction:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForObjectDetection:
     def __init__(self, *args, **kwargs):
@@ -453,6 +504,9 @@ class AutoModelForObjectDetection:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForPreTraining:
     def __init__(self, *args, **kwargs):
@@ -462,6 +516,9 @@ class AutoModelForPreTraining:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -471,6 +528,9 @@ class AutoModelForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForSeq2SeqLM:
     def __init__(self, *args, **kwargs):
@@ -480,6 +540,9 @@ class AutoModelForSeq2SeqLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -489,6 +552,9 @@ class AutoModelForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForSpeechSeq2Seq:
     def __init__(self, *args, **kwargs):
@@ -498,6 +564,9 @@ class AutoModelForSpeechSeq2Seq:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForTableQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -507,6 +576,9 @@ class AutoModelForTableQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -516,6 +588,9 @@ class AutoModelForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelForVision2Seq:
     def __init__(self, *args, **kwargs):
@@ -525,6 +600,9 @@ class AutoModelForVision2Seq:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class AutoModelWithLMHead:
     def __init__(self, *args, **kwargs):
@@ -534,6 +612,9 @@ class AutoModelWithLMHead:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 BART_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -546,6 +627,9 @@ class BartForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BartForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -555,6 +639,9 @@ class BartForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BartForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -564,6 +651,9 @@ class BartForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BartForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -573,6 +663,9 @@ class BartForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BartModel:
     def __init__(self, *args, **kwargs):
@@ -582,6 +675,9 @@ class BartModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BartPretrainedModel:
     def __init__(self, *args, **kwargs):
@@ -591,6 +687,9 @@ class BartPretrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class PretrainedBartModel:
     def __init__(self, *args, **kwargs):
@@ -600,6 +699,9 @@ class PretrainedBartModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 BEIT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -617,6 +719,9 @@ class BeitForMaskedImageModeling:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BeitForSemanticSegmentation:
     def __init__(self, *args, **kwargs):
@@ -631,6 +736,9 @@ class BeitModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BeitPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -640,6 +748,9 @@ class BeitPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 BERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -652,6 +763,9 @@ class BertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -661,6 +775,9 @@ class BertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BertForNextSentencePrediction:
     def __init__(self, *args, **kwargs):
@@ -680,6 +797,9 @@ class BertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -689,6 +809,9 @@ class BertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -698,6 +821,9 @@ class BertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BertLayer:
     def __init__(self, *args, **kwargs):
@@ -712,6 +838,9 @@ class BertLMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BertModel:
     def __init__(self, *args, **kwargs):
@@ -721,6 +850,9 @@ class BertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -730,6 +862,9 @@ class BertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_bert(*args, **kwargs):
     requires_backends(load_tf_weights_in_bert, ["torch"])
@@ -753,6 +888,9 @@ class BertGenerationPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_bert_generation(*args, **kwargs):
     requires_backends(load_tf_weights_in_bert_generation, ["torch"])
@@ -769,6 +907,9 @@ class BigBirdForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BigBirdForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -778,6 +919,9 @@ class BigBirdForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BigBirdForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -787,6 +931,9 @@ class BigBirdForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BigBirdForPreTraining:
     def __init__(self, *args, **kwargs):
@@ -801,6 +948,9 @@ class BigBirdForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BigBirdForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -810,6 +960,9 @@ class BigBirdForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BigBirdForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -819,6 +972,9 @@ class BigBirdForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BigBirdLayer:
     def __init__(self, *args, **kwargs):
@@ -833,6 +989,9 @@ class BigBirdModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BigBirdPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -842,6 +1001,9 @@ class BigBirdPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_big_bird(*args, **kwargs):
     requires_backends(load_tf_weights_in_big_bird, ["torch"])
@@ -858,6 +1020,9 @@ class BigBirdPegasusForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BigBirdPegasusForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -867,6 +1032,9 @@ class BigBirdPegasusForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BigBirdPegasusForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -876,6 +1044,9 @@ class BigBirdPegasusForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BigBirdPegasusForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -885,6 +1056,9 @@ class BigBirdPegasusForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BigBirdPegasusModel:
     def __init__(self, *args, **kwargs):
@@ -894,6 +1068,9 @@ class BigBirdPegasusModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BigBirdPegasusPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -903,6 +1080,9 @@ class BigBirdPegasusPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 BLENDERBOT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -915,6 +1095,9 @@ class BlenderbotForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BlenderbotForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -924,6 +1107,9 @@ class BlenderbotForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BlenderbotModel:
     def __init__(self, *args, **kwargs):
@@ -933,6 +1119,9 @@ class BlenderbotModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BlenderbotPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -942,6 +1131,9 @@ class BlenderbotPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 BLENDERBOT_SMALL_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -954,6 +1146,9 @@ class BlenderbotSmallForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BlenderbotSmallForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -963,6 +1158,9 @@ class BlenderbotSmallForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BlenderbotSmallModel:
     def __init__(self, *args, **kwargs):
@@ -972,6 +1170,9 @@ class BlenderbotSmallModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class BlenderbotSmallPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -981,6 +1182,9 @@ class BlenderbotSmallPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 CAMEMBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -993,6 +1197,9 @@ class CamembertForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CamembertForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -1002,6 +1209,9 @@ class CamembertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CamembertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -1011,6 +1221,9 @@ class CamembertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CamembertForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -1020,6 +1233,9 @@ class CamembertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CamembertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1029,6 +1245,9 @@ class CamembertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CamembertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1038,6 +1257,9 @@ class CamembertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CamembertModel:
     def __init__(self, *args, **kwargs):
@@ -1047,6 +1269,9 @@ class CamembertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 CANINE_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1059,6 +1284,9 @@ class CanineForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CanineForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -1068,6 +1296,9 @@ class CanineForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CanineForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1077,6 +1308,9 @@ class CanineForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CanineForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1086,6 +1320,9 @@ class CanineForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CanineLayer:
     def __init__(self, *args, **kwargs):
@@ -1100,6 +1337,9 @@ class CanineModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CaninePreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1109,6 +1349,9 @@ class CaninePreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_canine(*args, **kwargs):
     requires_backends(load_tf_weights_in_canine, ["torch"])
@@ -1125,6 +1368,9 @@ class CLIPModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CLIPPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1134,6 +1380,9 @@ class CLIPPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CLIPTextModel:
     def __init__(self, *args, **kwargs):
@@ -1143,6 +1392,9 @@ class CLIPTextModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CLIPVisionModel:
     def __init__(self, *args, **kwargs):
@@ -1152,6 +1404,9 @@ class CLIPVisionModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 CONVBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1164,6 +1419,9 @@ class ConvBertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ConvBertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -1173,6 +1431,9 @@ class ConvBertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ConvBertForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -1182,6 +1443,9 @@ class ConvBertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ConvBertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1191,6 +1455,9 @@ class ConvBertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ConvBertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1200,6 +1467,9 @@ class ConvBertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ConvBertLayer:
     def __init__(self, *args, **kwargs):
@@ -1214,6 +1484,9 @@ class ConvBertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ConvBertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1223,6 +1496,9 @@ class ConvBertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_convbert(*args, **kwargs):
     requires_backends(load_tf_weights_in_convbert, ["torch"])
@@ -1239,6 +1515,9 @@ class CTRLForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CTRLLMHeadModel:
     def __init__(self, *args, **kwargs):
@@ -1248,6 +1527,9 @@ class CTRLLMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CTRLModel:
     def __init__(self, *args, **kwargs):
@@ -1257,6 +1539,9 @@ class CTRLModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class CTRLPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1266,6 +1551,9 @@ class CTRLPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 DEBERTA_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1278,6 +1566,9 @@ class DebertaForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DebertaForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -1287,6 +1578,9 @@ class DebertaForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DebertaForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1296,6 +1590,9 @@ class DebertaForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DebertaForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1305,6 +1602,9 @@ class DebertaForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DebertaModel:
     def __init__(self, *args, **kwargs):
@@ -1314,6 +1614,9 @@ class DebertaModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DebertaPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1323,6 +1626,9 @@ class DebertaPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 DEBERTA_V2_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1335,6 +1641,9 @@ class DebertaV2ForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DebertaV2ForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -1344,6 +1653,9 @@ class DebertaV2ForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DebertaV2ForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1353,6 +1665,9 @@ class DebertaV2ForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DebertaV2ForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1362,6 +1677,9 @@ class DebertaV2ForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DebertaV2Model:
     def __init__(self, *args, **kwargs):
@@ -1371,6 +1689,9 @@ class DebertaV2Model:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DebertaV2PreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1380,6 +1701,9 @@ class DebertaV2PreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 DEIT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1402,6 +1726,9 @@ class DeiTModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DeiTPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1411,6 +1738,9 @@ class DeiTPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 DISTILBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1423,6 +1753,9 @@ class DistilBertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DistilBertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -1432,6 +1765,9 @@ class DistilBertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DistilBertForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -1441,6 +1777,9 @@ class DistilBertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DistilBertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1450,6 +1789,9 @@ class DistilBertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DistilBertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1459,6 +1801,9 @@ class DistilBertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DistilBertModel:
     def __init__(self, *args, **kwargs):
@@ -1468,6 +1813,9 @@ class DistilBertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DistilBertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1477,6 +1825,9 @@ class DistilBertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 DPR_CONTEXT_ENCODER_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1505,6 +1856,9 @@ class DPRPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class DPRPretrainedQuestionEncoder:
     def __init__(self, *args, **kwargs):
@@ -1537,6 +1891,9 @@ class ElectraForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ElectraForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -1546,6 +1903,9 @@ class ElectraForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ElectraForPreTraining:
     def __init__(self, *args, **kwargs):
@@ -1560,6 +1920,9 @@ class ElectraForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ElectraForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1569,6 +1932,9 @@ class ElectraForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ElectraForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1578,6 +1944,9 @@ class ElectraForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ElectraModel:
     def __init__(self, *args, **kwargs):
@@ -1587,6 +1956,9 @@ class ElectraModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ElectraPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1596,6 +1968,9 @@ class ElectraPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_electra(*args, **kwargs):
     requires_backends(load_tf_weights_in_electra, ["torch"])
@@ -1609,6 +1984,9 @@ class EncoderDecoderModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 FLAUBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1621,6 +1999,9 @@ class FlaubertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FlaubertForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -1630,6 +2011,9 @@ class FlaubertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FlaubertForQuestionAnsweringSimple:
     def __init__(self, *args, **kwargs):
@@ -1639,6 +2023,9 @@ class FlaubertForQuestionAnsweringSimple:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FlaubertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1648,6 +2035,9 @@ class FlaubertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FlaubertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1657,6 +2047,9 @@ class FlaubertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FlaubertModel:
     def __init__(self, *args, **kwargs):
@@ -1666,6 +2059,9 @@ class FlaubertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FlaubertWithLMHeadModel:
     def __init__(self, *args, **kwargs):
@@ -1675,6 +2071,9 @@ class FlaubertWithLMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 FNET_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1687,6 +2086,9 @@ class FNetForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FNetForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -1696,6 +2098,9 @@ class FNetForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FNetForNextSentencePrediction:
     def __init__(self, *args, **kwargs):
@@ -1715,6 +2120,9 @@ class FNetForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FNetForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1724,6 +2132,9 @@ class FNetForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FNetForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1733,6 +2144,9 @@ class FNetForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FNetLayer:
     def __init__(self, *args, **kwargs):
@@ -1747,6 +2161,9 @@ class FNetModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FNetPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1756,6 +2173,9 @@ class FNetPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FSMTForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -1765,6 +2185,9 @@ class FSMTForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FSMTModel:
     def __init__(self, *args, **kwargs):
@@ -1774,6 +2197,9 @@ class FSMTModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class PretrainedFSMTModel:
     def __init__(self, *args, **kwargs):
@@ -1783,6 +2209,9 @@ class PretrainedFSMTModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 FUNNEL_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1795,6 +2224,9 @@ class FunnelBaseModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FunnelForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -1804,6 +2236,9 @@ class FunnelForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FunnelForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -1813,6 +2248,9 @@ class FunnelForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FunnelForPreTraining:
     def __init__(self, *args, **kwargs):
@@ -1827,6 +2265,9 @@ class FunnelForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FunnelForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1836,6 +2277,9 @@ class FunnelForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FunnelForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1845,6 +2289,9 @@ class FunnelForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FunnelModel:
     def __init__(self, *args, **kwargs):
@@ -1854,6 +2301,9 @@ class FunnelModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class FunnelPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1863,6 +2313,9 @@ class FunnelPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_funnel(*args, **kwargs):
     requires_backends(load_tf_weights_in_funnel, ["torch"])
@@ -1879,6 +2332,9 @@ class GPT2DoubleHeadsModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class GPT2ForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1888,6 +2344,9 @@ class GPT2ForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class GPT2ForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1897,6 +2356,9 @@ class GPT2ForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class GPT2LMHeadModel:
     def __init__(self, *args, **kwargs):
@@ -1906,6 +2368,9 @@ class GPT2LMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class GPT2Model:
     def __init__(self, *args, **kwargs):
@@ -1915,6 +2380,9 @@ class GPT2Model:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class GPT2PreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1924,6 +2392,9 @@ class GPT2PreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_gpt2(*args, **kwargs):
     requires_backends(load_tf_weights_in_gpt2, ["torch"])
@@ -1940,6 +2411,9 @@ class GPTNeoForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class GPTNeoForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1949,6 +2423,9 @@ class GPTNeoForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class GPTNeoModel:
     def __init__(self, *args, **kwargs):
@@ -1958,6 +2435,9 @@ class GPTNeoModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class GPTNeoPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1967,6 +2447,9 @@ class GPTNeoPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_gpt_neo(*args, **kwargs):
     requires_backends(load_tf_weights_in_gpt_neo, ["torch"])
@@ -1983,6 +2466,9 @@ class GPTJForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class GPTJForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1992,6 +2478,9 @@ class GPTJForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class GPTJModel:
     def __init__(self, *args, **kwargs):
@@ -2001,6 +2490,9 @@ class GPTJModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class GPTJPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2010,6 +2502,9 @@ class GPTJPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 HUBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -2027,6 +2522,9 @@ class HubertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class HubertModel:
     def __init__(self, *args, **kwargs):
@@ -2036,6 +2534,9 @@ class HubertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class HubertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2045,6 +2546,9 @@ class HubertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 IBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -2057,6 +2561,9 @@ class IBertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class IBertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -2066,6 +2573,9 @@ class IBertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class IBertForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -2075,6 +2585,9 @@ class IBertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class IBertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -2084,6 +2597,9 @@ class IBertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class IBertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -2093,6 +2609,9 @@ class IBertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class IBertModel:
     def __init__(self, *args, **kwargs):
@@ -2102,6 +2621,9 @@ class IBertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class IBertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2111,6 +2633,9 @@ class IBertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 LAYOUTLM_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -2123,6 +2648,9 @@ class LayoutLMForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LayoutLMForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -2132,6 +2660,9 @@ class LayoutLMForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LayoutLMForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -2141,6 +2672,9 @@ class LayoutLMForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LayoutLMModel:
     def __init__(self, *args, **kwargs):
@@ -2150,6 +2684,9 @@ class LayoutLMModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LayoutLMPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2159,6 +2696,9 @@ class LayoutLMPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 LAYOUTLMV2_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -2171,6 +2711,9 @@ class LayoutLMv2ForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LayoutLMv2ForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -2180,6 +2723,9 @@ class LayoutLMv2ForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LayoutLMv2ForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -2189,6 +2735,9 @@ class LayoutLMv2ForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LayoutLMv2Model:
     def __init__(self, *args, **kwargs):
@@ -2198,6 +2747,9 @@ class LayoutLMv2Model:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LayoutLMv2PreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2207,6 +2759,9 @@ class LayoutLMv2PreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 LED_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -2219,6 +2774,9 @@ class LEDForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LEDForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -2228,6 +2786,9 @@ class LEDForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LEDForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -2237,6 +2798,9 @@ class LEDForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LEDModel:
     def __init__(self, *args, **kwargs):
@@ -2246,6 +2810,9 @@ class LEDModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LEDPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2255,6 +2822,9 @@ class LEDPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 LONGFORMER_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -2267,6 +2837,9 @@ class LongformerForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LongformerForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -2276,6 +2849,9 @@ class LongformerForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LongformerForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -2285,6 +2861,9 @@ class LongformerForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LongformerForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -2294,6 +2873,9 @@ class LongformerForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LongformerForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -2303,6 +2885,9 @@ class LongformerForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LongformerModel:
     def __init__(self, *args, **kwargs):
@@ -2312,6 +2897,9 @@ class LongformerModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LongformerPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2321,6 +2909,9 @@ class LongformerPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LongformerSelfAttention:
     def __init__(self, *args, **kwargs):
@@ -2353,6 +2944,9 @@ class LukeModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LukePreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2362,6 +2956,9 @@ class LukePreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LxmertEncoder:
     def __init__(self, *args, **kwargs):
@@ -2381,6 +2978,9 @@ class LxmertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LxmertModel:
     def __init__(self, *args, **kwargs):
@@ -2390,6 +2990,9 @@ class LxmertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LxmertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2399,6 +3002,9 @@ class LxmertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class LxmertVisualFeatureEncoder:
     def __init__(self, *args, **kwargs):
@@ -2421,6 +3027,9 @@ class M2M100ForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class M2M100Model:
     def __init__(self, *args, **kwargs):
@@ -2430,6 +3039,9 @@ class M2M100Model:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class M2M100PreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2439,6 +3051,9 @@ class M2M100PreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MarianForCausalLM:
     def __init__(self, *args, **kwargs):
@@ -2448,6 +3063,9 @@ class MarianForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MarianModel:
     def __init__(self, *args, **kwargs):
@@ -2457,6 +3075,9 @@ class MarianModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MarianMTModel:
     def __init__(self, *args, **kwargs):
@@ -2466,6 +3087,9 @@ class MarianMTModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MBartForCausalLM:
     def __init__(self, *args, **kwargs):
@@ -2475,6 +3099,9 @@ class MBartForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MBartForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -2484,6 +3111,9 @@ class MBartForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MBartForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -2493,6 +3123,9 @@ class MBartForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MBartForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -2502,6 +3135,9 @@ class MBartForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MBartModel:
     def __init__(self, *args, **kwargs):
@@ -2511,6 +3147,9 @@ class MBartModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MBartPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2520,6 +3159,9 @@ class MBartPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 MEGATRON_BERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -2532,6 +3174,9 @@ class MegatronBertForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MegatronBertForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -2541,6 +3186,9 @@ class MegatronBertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MegatronBertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -2550,6 +3198,9 @@ class MegatronBertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MegatronBertForNextSentencePrediction:
     def __init__(self, *args, **kwargs):
@@ -2569,6 +3220,9 @@ class MegatronBertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MegatronBertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -2578,6 +3232,9 @@ class MegatronBertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MegatronBertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -2587,6 +3244,9 @@ class MegatronBertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MegatronBertModel:
     def __init__(self, *args, **kwargs):
@@ -2596,6 +3256,9 @@ class MegatronBertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MegatronBertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2605,6 +3268,9 @@ class MegatronBertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MMBTForClassification:
     def __init__(self, *args, **kwargs):
@@ -2619,6 +3285,9 @@ class MMBTModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ModalEmbeddings:
     def __init__(self, *args, **kwargs):
@@ -2636,6 +3305,9 @@ class MobileBertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MobileBertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -2645,6 +3317,9 @@ class MobileBertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MobileBertForNextSentencePrediction:
     def __init__(self, *args, **kwargs):
@@ -2664,6 +3339,9 @@ class MobileBertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MobileBertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -2673,6 +3351,9 @@ class MobileBertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MobileBertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -2682,6 +3363,9 @@ class MobileBertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MobileBertLayer:
     def __init__(self, *args, **kwargs):
@@ -2696,6 +3380,9 @@ class MobileBertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MobileBertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2705,6 +3392,9 @@ class MobileBertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_mobilebert(*args, **kwargs):
     requires_backends(load_tf_weights_in_mobilebert, ["torch"])
@@ -2721,6 +3411,9 @@ class MPNetForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MPNetForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -2730,6 +3423,9 @@ class MPNetForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MPNetForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -2739,6 +3435,9 @@ class MPNetForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MPNetForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -2748,6 +3447,9 @@ class MPNetForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MPNetForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -2757,6 +3459,9 @@ class MPNetForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MPNetLayer:
     def __init__(self, *args, **kwargs):
@@ -2771,6 +3476,9 @@ class MPNetModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MPNetPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2780,6 +3488,9 @@ class MPNetPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MT5EncoderModel:
     def __init__(self, *args, **kwargs):
@@ -2789,6 +3500,9 @@ class MT5EncoderModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MT5ForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -2798,6 +3512,9 @@ class MT5ForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class MT5Model:
     def __init__(self, *args, **kwargs):
@@ -2807,6 +3524,9 @@ class MT5Model:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 OPENAI_GPT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -2819,6 +3539,9 @@ class OpenAIGPTDoubleHeadsModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class OpenAIGPTForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -2828,6 +3551,9 @@ class OpenAIGPTForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class OpenAIGPTLMHeadModel:
     def __init__(self, *args, **kwargs):
@@ -2837,6 +3563,9 @@ class OpenAIGPTLMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class OpenAIGPTModel:
     def __init__(self, *args, **kwargs):
@@ -2846,6 +3575,9 @@ class OpenAIGPTModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class OpenAIGPTPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2855,6 +3587,9 @@ class OpenAIGPTPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_openai_gpt(*args, **kwargs):
     requires_backends(load_tf_weights_in_openai_gpt, ["torch"])
@@ -2868,6 +3603,9 @@ class PegasusForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class PegasusForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -2877,6 +3615,9 @@ class PegasusForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class PegasusModel:
     def __init__(self, *args, **kwargs):
@@ -2886,6 +3627,9 @@ class PegasusModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class PegasusPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2895,6 +3639,9 @@ class PegasusPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 PROPHETNET_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -2917,6 +3664,9 @@ class ProphetNetForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ProphetNetForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -2926,6 +3676,9 @@ class ProphetNetForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ProphetNetModel:
     def __init__(self, *args, **kwargs):
@@ -2935,6 +3688,9 @@ class ProphetNetModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ProphetNetPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2944,6 +3700,9 @@ class ProphetNetPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RagModel:
     def __init__(self, *args, **kwargs):
@@ -2953,6 +3712,9 @@ class RagModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RagPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2962,6 +3724,9 @@ class RagPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RagSequenceForGeneration:
     def __init__(self, *args, **kwargs):
@@ -2989,6 +3754,9 @@ class ReformerForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ReformerForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -2998,6 +3766,9 @@ class ReformerForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ReformerForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -3007,6 +3778,9 @@ class ReformerForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ReformerLayer:
     def __init__(self, *args, **kwargs):
@@ -3021,6 +3795,9 @@ class ReformerModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ReformerModelWithLMHead:
     def __init__(self, *args, **kwargs):
@@ -3030,6 +3807,9 @@ class ReformerModelWithLMHead:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ReformerPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3039,6 +3819,9 @@ class ReformerPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 REMBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -3051,6 +3834,9 @@ class RemBertForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RemBertForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -3060,6 +3846,9 @@ class RemBertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RemBertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -3069,6 +3858,9 @@ class RemBertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RemBertForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -3078,6 +3870,9 @@ class RemBertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RemBertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -3087,6 +3882,9 @@ class RemBertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RemBertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -3096,6 +3894,9 @@ class RemBertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RemBertLayer:
     def __init__(self, *args, **kwargs):
@@ -3110,6 +3911,9 @@ class RemBertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RemBertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3119,6 +3923,9 @@ class RemBertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_rembert(*args, **kwargs):
     requires_backends(load_tf_weights_in_rembert, ["torch"])
@@ -3135,6 +3942,9 @@ class RetriBertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RetriBertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3144,6 +3954,9 @@ class RetriBertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 ROBERTA_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -3156,6 +3969,9 @@ class RobertaForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RobertaForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -3165,6 +3981,9 @@ class RobertaForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RobertaForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -3174,6 +3993,9 @@ class RobertaForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RobertaForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -3183,6 +4005,9 @@ class RobertaForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RobertaForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -3192,6 +4017,9 @@ class RobertaForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RobertaForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -3201,6 +4029,9 @@ class RobertaForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RobertaModel:
     def __init__(self, *args, **kwargs):
@@ -3210,6 +4041,9 @@ class RobertaModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RobertaPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3219,6 +4053,9 @@ class RobertaPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 ROFORMER_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -3231,6 +4068,9 @@ class RoFormerForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RoFormerForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -3240,6 +4080,9 @@ class RoFormerForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RoFormerForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -3249,6 +4092,9 @@ class RoFormerForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RoFormerForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -3258,6 +4104,9 @@ class RoFormerForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RoFormerForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -3267,6 +4116,9 @@ class RoFormerForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RoFormerForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -3276,6 +4128,9 @@ class RoFormerForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RoFormerLayer:
     def __init__(self, *args, **kwargs):
@@ -3290,6 +4145,9 @@ class RoFormerModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class RoFormerPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3299,6 +4157,9 @@ class RoFormerPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_roformer(*args, **kwargs):
     requires_backends(load_tf_weights_in_roformer, ["torch"])
@@ -3335,6 +4196,9 @@ class SegformerModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class SegformerPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3344,6 +4208,9 @@ class SegformerPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 SEW_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -3361,6 +4228,9 @@ class SEWForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class SEWModel:
     def __init__(self, *args, **kwargs):
@@ -3370,6 +4240,9 @@ class SEWModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class SEWPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3379,6 +4252,9 @@ class SEWPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 SEW_D_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -3396,6 +4272,9 @@ class SEWDForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class SEWDModel:
     def __init__(self, *args, **kwargs):
@@ -3405,6 +4284,9 @@ class SEWDModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class SEWDPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3414,6 +4296,9 @@ class SEWDPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class SpeechEncoderDecoderModel:
     def __init__(self, *args, **kwargs):
@@ -3423,6 +4308,9 @@ class SpeechEncoderDecoderModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 SPEECH_TO_TEXT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -3435,6 +4323,9 @@ class Speech2TextForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class Speech2TextModel:
     def __init__(self, *args, **kwargs):
@@ -3444,6 +4335,9 @@ class Speech2TextModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class Speech2TextPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3453,6 +4347,9 @@ class Speech2TextPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class Speech2Text2ForCausalLM:
     def __init__(self, *args, **kwargs):
@@ -3462,6 +4359,9 @@ class Speech2Text2ForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class Speech2Text2PreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3471,6 +4371,9 @@ class Speech2Text2PreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 SPLINTER_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -3483,6 +4386,9 @@ class SplinterForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class SplinterLayer:
     def __init__(self, *args, **kwargs):
@@ -3497,6 +4403,9 @@ class SplinterModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class SplinterPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3506,6 +4415,9 @@ class SplinterPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 SQUEEZEBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -3518,6 +4430,9 @@ class SqueezeBertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class SqueezeBertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -3527,6 +4442,9 @@ class SqueezeBertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class SqueezeBertForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -3536,6 +4454,9 @@ class SqueezeBertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class SqueezeBertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -3545,6 +4466,9 @@ class SqueezeBertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class SqueezeBertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -3554,6 +4478,9 @@ class SqueezeBertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class SqueezeBertModel:
     def __init__(self, *args, **kwargs):
@@ -3563,6 +4490,9 @@ class SqueezeBertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class SqueezeBertModule:
     def __init__(self, *args, **kwargs):
@@ -3577,6 +4507,9 @@ class SqueezeBertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 T5_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -3589,6 +4522,9 @@ class T5EncoderModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class T5ForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -3598,6 +4534,9 @@ class T5ForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class T5Model:
     def __init__(self, *args, **kwargs):
@@ -3607,6 +4546,9 @@ class T5Model:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class T5PreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3616,6 +4558,9 @@ class T5PreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_t5(*args, **kwargs):
     requires_backends(load_tf_weights_in_t5, ["torch"])
@@ -3637,6 +4582,9 @@ class TransfoXLForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class TransfoXLLMHeadModel:
     def __init__(self, *args, **kwargs):
@@ -3646,6 +4594,9 @@ class TransfoXLLMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class TransfoXLModel:
     def __init__(self, *args, **kwargs):
@@ -3655,6 +4606,9 @@ class TransfoXLModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class TransfoXLPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3664,6 +4618,9 @@ class TransfoXLPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_transfo_xl(*args, **kwargs):
     requires_backends(load_tf_weights_in_transfo_xl, ["torch"])
@@ -3680,6 +4637,9 @@ class TrOCRForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class TrOCRPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3689,6 +4649,9 @@ class TrOCRPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 UNISPEECH_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -3711,6 +4674,9 @@ class UniSpeechForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class UniSpeechModel:
     def __init__(self, *args, **kwargs):
@@ -3720,6 +4686,9 @@ class UniSpeechModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class UniSpeechPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3729,6 +4698,9 @@ class UniSpeechPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 UNISPEECH_SAT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -3751,6 +4723,9 @@ class UniSpeechSatForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class UniSpeechSatModel:
     def __init__(self, *args, **kwargs):
@@ -3760,6 +4735,9 @@ class UniSpeechSatModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class UniSpeechSatPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3769,6 +4747,9 @@ class UniSpeechSatPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class VisionEncoderDecoderModel:
     def __init__(self, *args, **kwargs):
@@ -3778,6 +4759,9 @@ class VisionEncoderDecoderModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 VISUAL_BERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -3790,6 +4774,9 @@ class VisualBertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class VisualBertForPreTraining:
     def __init__(self, *args, **kwargs):
@@ -3804,6 +4791,9 @@ class VisualBertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class VisualBertForRegionToPhraseAlignment:
     def __init__(self, *args, **kwargs):
@@ -3828,6 +4818,9 @@ class VisualBertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class VisualBertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3837,6 +4830,9 @@ class VisualBertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 VIT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -3854,6 +4850,9 @@ class ViTModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class ViTPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3863,6 +4862,9 @@ class ViTPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 WAV_2_VEC_2_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -3880,6 +4882,9 @@ class Wav2Vec2ForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class Wav2Vec2ForPreTraining:
     def __init__(self, *args, **kwargs):
@@ -3894,6 +4899,9 @@ class Wav2Vec2ForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class Wav2Vec2Model:
     def __init__(self, *args, **kwargs):
@@ -3903,6 +4911,9 @@ class Wav2Vec2Model:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class Wav2Vec2PreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3912,6 +4923,9 @@ class Wav2Vec2PreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 XLM_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -3924,6 +4938,9 @@ class XLMForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLMForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -3933,6 +4950,9 @@ class XLMForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLMForQuestionAnsweringSimple:
     def __init__(self, *args, **kwargs):
@@ -3942,6 +4962,9 @@ class XLMForQuestionAnsweringSimple:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLMForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -3951,6 +4974,9 @@ class XLMForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLMForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -3960,6 +4986,9 @@ class XLMForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLMModel:
     def __init__(self, *args, **kwargs):
@@ -3969,6 +4998,9 @@ class XLMModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLMPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -3978,6 +5010,9 @@ class XLMPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLMWithLMHeadModel:
     def __init__(self, *args, **kwargs):
@@ -3987,6 +5022,9 @@ class XLMWithLMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 XLM_PROPHETNET_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -4009,6 +5047,9 @@ class XLMProphetNetForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLMProphetNetForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -4018,6 +5059,9 @@ class XLMProphetNetForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLMProphetNetModel:
     def __init__(self, *args, **kwargs):
@@ -4027,6 +5071,9 @@ class XLMProphetNetModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 XLM_ROBERTA_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -4039,6 +5086,9 @@ class XLMRobertaForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLMRobertaForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -4048,6 +5098,9 @@ class XLMRobertaForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLMRobertaForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -4057,6 +5110,9 @@ class XLMRobertaForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLMRobertaForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -4066,6 +5122,9 @@ class XLMRobertaForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLMRobertaForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -4075,6 +5134,9 @@ class XLMRobertaForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLMRobertaForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -4084,6 +5146,9 @@ class XLMRobertaForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLMRobertaModel:
     def __init__(self, *args, **kwargs):
@@ -4093,6 +5158,9 @@ class XLMRobertaModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 XLNET_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -4105,6 +5173,9 @@ class XLNetForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLNetForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -4114,6 +5185,9 @@ class XLNetForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLNetForQuestionAnsweringSimple:
     def __init__(self, *args, **kwargs):
@@ -4123,6 +5197,9 @@ class XLNetForQuestionAnsweringSimple:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLNetForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -4132,6 +5209,9 @@ class XLNetForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLNetForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -4141,6 +5221,9 @@ class XLNetForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLNetLMHeadModel:
     def __init__(self, *args, **kwargs):
@@ -4150,6 +5233,9 @@ class XLNetLMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLNetModel:
     def __init__(self, *args, **kwargs):
@@ -4159,6 +5245,9 @@ class XLNetModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 class XLNetPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -4168,6 +5257,9 @@ class XLNetPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["torch"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
 
 def load_tf_weights_in_xlnet(*args, **kwargs):
     requires_backends(load_tf_weights_in_xlnet, ["torch"])
diff --git a/src/transformers/utils/dummy_scatter_objects.py b/src/transformers/utils/dummy_scatter_objects.py
index 639ced6f5d5..6a53b0f9635 100644
--- a/src/transformers/utils/dummy_scatter_objects.py
+++ b/src/transformers/utils/dummy_scatter_objects.py
@@ -13,6 +13,9 @@ class TapasForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["scatter"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["scatter"])
+
 
 class TapasForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -22,6 +25,9 @@ class TapasForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["scatter"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["scatter"])
+
 
 class TapasForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -31,6 +37,9 @@ class TapasForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["scatter"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["scatter"])
+
 
 class TapasModel:
     def __init__(self, *args, **kwargs):
@@ -40,6 +49,9 @@ class TapasModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["scatter"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["scatter"])
+
 
 class TapasPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -49,6 +61,9 @@ class TapasPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["scatter"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["scatter"])
+
 
 def load_tf_weights_in_tapas(*args, **kwargs):
     requires_backends(load_tf_weights_in_tapas, ["scatter"])
diff --git a/src/transformers/utils/dummy_tf_objects.py b/src/transformers/utils/dummy_tf_objects.py
index c3762125b5d..f92ee0ad9d6 100644
--- a/src/transformers/utils/dummy_tf_objects.py
+++ b/src/transformers/utils/dummy_tf_objects.py
@@ -32,6 +32,9 @@ class TFLayoutLMForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFLayoutLMForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -41,6 +44,9 @@ class TFLayoutLMForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFLayoutLMForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -50,6 +56,9 @@ class TFLayoutLMForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFLayoutLMMainLayer:
     def __init__(self, *args, **kwargs):
@@ -64,6 +73,9 @@ class TFLayoutLMModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFLayoutLMPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -73,6 +85,9 @@ class TFLayoutLMPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -82,6 +97,9 @@ class TFPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFSequenceSummary:
     def __init__(self, *args, **kwargs):
@@ -108,6 +126,9 @@ class TFAlbertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFAlbertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -117,6 +138,9 @@ class TFAlbertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFAlbertForPreTraining:
     def __init__(self, *args, **kwargs):
@@ -131,6 +155,9 @@ class TFAlbertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFAlbertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -140,6 +167,9 @@ class TFAlbertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFAlbertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -149,6 +179,9 @@ class TFAlbertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFAlbertMainLayer:
     def __init__(self, *args, **kwargs):
@@ -163,6 +196,9 @@ class TFAlbertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFAlbertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -172,6 +208,9 @@ class TFAlbertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_MODEL_FOR_CAUSAL_LM_MAPPING = None
 
@@ -217,6 +256,9 @@ class TFAutoModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFAutoModelForCausalLM:
     def __init__(self, *args, **kwargs):
@@ -226,6 +268,9 @@ class TFAutoModelForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFAutoModelForImageClassification:
     def __init__(self, *args, **kwargs):
@@ -235,6 +280,9 @@ class TFAutoModelForImageClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFAutoModelForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -244,6 +292,9 @@ class TFAutoModelForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFAutoModelForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -253,6 +304,9 @@ class TFAutoModelForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFAutoModelForPreTraining:
     def __init__(self, *args, **kwargs):
@@ -262,6 +316,9 @@ class TFAutoModelForPreTraining:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFAutoModelForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -271,6 +328,9 @@ class TFAutoModelForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFAutoModelForSeq2SeqLM:
     def __init__(self, *args, **kwargs):
@@ -280,6 +340,9 @@ class TFAutoModelForSeq2SeqLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFAutoModelForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -289,6 +352,9 @@ class TFAutoModelForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFAutoModelForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -298,6 +364,9 @@ class TFAutoModelForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFAutoModelWithLMHead:
     def __init__(self, *args, **kwargs):
@@ -307,6 +376,9 @@ class TFAutoModelWithLMHead:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFBartForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -316,6 +388,9 @@ class TFBartForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFBartModel:
     def __init__(self, *args, **kwargs):
@@ -325,6 +400,9 @@ class TFBartModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFBartPretrainedModel:
     def __init__(self, *args, **kwargs):
@@ -334,6 +412,9 @@ class TFBartPretrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -351,6 +432,9 @@ class TFBertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFBertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -360,6 +444,9 @@ class TFBertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFBertForNextSentencePrediction:
     def __init__(self, *args, **kwargs):
@@ -379,6 +466,9 @@ class TFBertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFBertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -388,6 +478,9 @@ class TFBertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFBertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -397,6 +490,9 @@ class TFBertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFBertLMHeadModel:
     def __init__(self, *args, **kwargs):
@@ -406,6 +502,9 @@ class TFBertLMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFBertMainLayer:
     def __init__(self, *args, **kwargs):
@@ -420,6 +519,9 @@ class TFBertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFBertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -429,6 +531,9 @@ class TFBertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFBlenderbotForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -438,6 +543,9 @@ class TFBlenderbotForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFBlenderbotModel:
     def __init__(self, *args, **kwargs):
@@ -447,6 +555,9 @@ class TFBlenderbotModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFBlenderbotPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -456,6 +567,9 @@ class TFBlenderbotPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFBlenderbotSmallForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -465,6 +579,9 @@ class TFBlenderbotSmallForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFBlenderbotSmallModel:
     def __init__(self, *args, **kwargs):
@@ -474,6 +591,9 @@ class TFBlenderbotSmallModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFBlenderbotSmallPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -483,6 +603,9 @@ class TFBlenderbotSmallPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_CAMEMBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -495,6 +618,9 @@ class TFCamembertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFCamembertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -504,6 +630,9 @@ class TFCamembertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFCamembertForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -513,6 +642,9 @@ class TFCamembertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFCamembertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -522,6 +654,9 @@ class TFCamembertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFCamembertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -531,6 +666,9 @@ class TFCamembertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFCamembertModel:
     def __init__(self, *args, **kwargs):
@@ -540,6 +678,9 @@ class TFCamembertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_CONVBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -552,6 +693,9 @@ class TFConvBertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFConvBertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -561,6 +705,9 @@ class TFConvBertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFConvBertForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -570,6 +717,9 @@ class TFConvBertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFConvBertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -579,6 +729,9 @@ class TFConvBertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFConvBertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -588,6 +741,9 @@ class TFConvBertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFConvBertLayer:
     def __init__(self, *args, **kwargs):
@@ -602,6 +758,9 @@ class TFConvBertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFConvBertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -611,6 +770,9 @@ class TFConvBertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_CTRL_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -623,6 +785,9 @@ class TFCTRLForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFCTRLLMHeadModel:
     def __init__(self, *args, **kwargs):
@@ -632,6 +797,9 @@ class TFCTRLLMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFCTRLModel:
     def __init__(self, *args, **kwargs):
@@ -641,6 +809,9 @@ class TFCTRLModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFCTRLPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -650,6 +821,9 @@ class TFCTRLPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_DEBERTA_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -662,6 +836,9 @@ class TFDebertaForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFDebertaForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -671,6 +848,9 @@ class TFDebertaForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFDebertaForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -680,6 +860,9 @@ class TFDebertaForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFDebertaForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -689,6 +872,9 @@ class TFDebertaForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFDebertaModel:
     def __init__(self, *args, **kwargs):
@@ -698,6 +884,9 @@ class TFDebertaModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFDebertaPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -707,6 +896,9 @@ class TFDebertaPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_DEBERTA_V2_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -719,6 +911,9 @@ class TFDebertaV2ForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFDebertaV2ForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -728,6 +923,9 @@ class TFDebertaV2ForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFDebertaV2ForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -737,6 +935,9 @@ class TFDebertaV2ForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFDebertaV2ForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -746,6 +947,9 @@ class TFDebertaV2ForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFDebertaV2Model:
     def __init__(self, *args, **kwargs):
@@ -755,6 +959,9 @@ class TFDebertaV2Model:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFDebertaV2PreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -764,6 +971,9 @@ class TFDebertaV2PreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_DISTILBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -776,6 +986,9 @@ class TFDistilBertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFDistilBertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -785,6 +998,9 @@ class TFDistilBertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFDistilBertForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -794,6 +1010,9 @@ class TFDistilBertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFDistilBertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -803,6 +1022,9 @@ class TFDistilBertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFDistilBertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -812,6 +1034,9 @@ class TFDistilBertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFDistilBertMainLayer:
     def __init__(self, *args, **kwargs):
@@ -826,6 +1051,9 @@ class TFDistilBertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFDistilBertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -835,6 +1063,9 @@ class TFDistilBertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_DPR_CONTEXT_ENCODER_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -886,6 +1117,9 @@ class TFElectraForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFElectraForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -895,6 +1129,9 @@ class TFElectraForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFElectraForPreTraining:
     def __init__(self, *args, **kwargs):
@@ -909,6 +1146,9 @@ class TFElectraForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFElectraForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -918,6 +1158,9 @@ class TFElectraForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFElectraForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -927,6 +1170,9 @@ class TFElectraForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFElectraModel:
     def __init__(self, *args, **kwargs):
@@ -936,6 +1182,9 @@ class TFElectraModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFElectraPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -945,6 +1194,9 @@ class TFElectraPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFEncoderDecoderModel:
     def __init__(self, *args, **kwargs):
@@ -954,6 +1206,9 @@ class TFEncoderDecoderModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_FLAUBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -966,6 +1221,9 @@ class TFFlaubertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFFlaubertForQuestionAnsweringSimple:
     def __init__(self, *args, **kwargs):
@@ -975,6 +1233,9 @@ class TFFlaubertForQuestionAnsweringSimple:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFFlaubertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -984,6 +1245,9 @@ class TFFlaubertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFFlaubertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -993,6 +1257,9 @@ class TFFlaubertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFFlaubertModel:
     def __init__(self, *args, **kwargs):
@@ -1002,6 +1269,9 @@ class TFFlaubertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFFlaubertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1011,6 +1281,9 @@ class TFFlaubertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFFlaubertWithLMHeadModel:
     def __init__(self, *args, **kwargs):
@@ -1020,6 +1293,9 @@ class TFFlaubertWithLMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_FUNNEL_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1032,6 +1308,9 @@ class TFFunnelBaseModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFFunnelForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -1041,6 +1320,9 @@ class TFFunnelForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFFunnelForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -1050,6 +1332,9 @@ class TFFunnelForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFFunnelForPreTraining:
     def __init__(self, *args, **kwargs):
@@ -1064,6 +1349,9 @@ class TFFunnelForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFFunnelForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1073,6 +1361,9 @@ class TFFunnelForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFFunnelForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1082,6 +1373,9 @@ class TFFunnelForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFFunnelModel:
     def __init__(self, *args, **kwargs):
@@ -1091,6 +1385,9 @@ class TFFunnelModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFFunnelPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1100,6 +1397,9 @@ class TFFunnelPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_GPT2_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1112,6 +1412,9 @@ class TFGPT2DoubleHeadsModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFGPT2ForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1121,6 +1424,9 @@ class TFGPT2ForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFGPT2LMHeadModel:
     def __init__(self, *args, **kwargs):
@@ -1130,6 +1436,9 @@ class TFGPT2LMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFGPT2MainLayer:
     def __init__(self, *args, **kwargs):
@@ -1144,6 +1453,9 @@ class TFGPT2Model:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFGPT2PreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1153,6 +1465,9 @@ class TFGPT2PreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_HUBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1170,6 +1485,9 @@ class TFHubertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFHubertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1179,6 +1497,9 @@ class TFHubertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFLEDForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -1188,6 +1509,9 @@ class TFLEDForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFLEDModel:
     def __init__(self, *args, **kwargs):
@@ -1197,6 +1521,9 @@ class TFLEDModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFLEDPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1206,6 +1533,9 @@ class TFLEDPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_LONGFORMER_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1218,6 +1548,9 @@ class TFLongformerForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFLongformerForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -1227,6 +1560,9 @@ class TFLongformerForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFLongformerForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -1236,6 +1572,9 @@ class TFLongformerForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFLongformerForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1245,6 +1584,9 @@ class TFLongformerForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFLongformerForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1254,6 +1596,9 @@ class TFLongformerForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFLongformerModel:
     def __init__(self, *args, **kwargs):
@@ -1263,6 +1608,9 @@ class TFLongformerModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFLongformerPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1272,6 +1620,9 @@ class TFLongformerPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFLongformerSelfAttention:
     def __init__(self, *args, **kwargs):
@@ -1299,6 +1650,9 @@ class TFLxmertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFLxmertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1308,6 +1662,9 @@ class TFLxmertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFLxmertVisualFeatureEncoder:
     def __init__(self, *args, **kwargs):
@@ -1322,6 +1679,9 @@ class TFMarianModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMarianMTModel:
     def __init__(self, *args, **kwargs):
@@ -1331,6 +1691,9 @@ class TFMarianMTModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMarianPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1340,6 +1703,9 @@ class TFMarianPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMBartForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -1349,6 +1715,9 @@ class TFMBartForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMBartModel:
     def __init__(self, *args, **kwargs):
@@ -1358,6 +1727,9 @@ class TFMBartModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMBartPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1367,6 +1739,9 @@ class TFMBartPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1379,6 +1754,9 @@ class TFMobileBertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMobileBertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -1388,6 +1766,9 @@ class TFMobileBertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMobileBertForNextSentencePrediction:
     def __init__(self, *args, **kwargs):
@@ -1407,6 +1788,9 @@ class TFMobileBertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMobileBertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1416,6 +1800,9 @@ class TFMobileBertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMobileBertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1425,6 +1812,9 @@ class TFMobileBertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMobileBertMainLayer:
     def __init__(self, *args, **kwargs):
@@ -1439,6 +1829,9 @@ class TFMobileBertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMobileBertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1448,6 +1841,9 @@ class TFMobileBertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_MPNET_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1460,6 +1856,9 @@ class TFMPNetForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMPNetForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -1469,6 +1868,9 @@ class TFMPNetForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMPNetForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -1478,6 +1880,9 @@ class TFMPNetForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMPNetForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1487,6 +1892,9 @@ class TFMPNetForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMPNetForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1496,6 +1904,9 @@ class TFMPNetForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMPNetMainLayer:
     def __init__(self, *args, **kwargs):
@@ -1510,6 +1921,9 @@ class TFMPNetModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMPNetPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1519,6 +1933,9 @@ class TFMPNetPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMT5EncoderModel:
     def __init__(self, *args, **kwargs):
@@ -1528,6 +1945,9 @@ class TFMT5EncoderModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMT5ForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -1537,6 +1957,9 @@ class TFMT5ForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFMT5Model:
     def __init__(self, *args, **kwargs):
@@ -1546,6 +1969,9 @@ class TFMT5Model:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_OPENAI_GPT_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1558,6 +1984,9 @@ class TFOpenAIGPTDoubleHeadsModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFOpenAIGPTForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1567,6 +1996,9 @@ class TFOpenAIGPTForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFOpenAIGPTLMHeadModel:
     def __init__(self, *args, **kwargs):
@@ -1576,6 +2008,9 @@ class TFOpenAIGPTLMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFOpenAIGPTMainLayer:
     def __init__(self, *args, **kwargs):
@@ -1590,6 +2025,9 @@ class TFOpenAIGPTModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFOpenAIGPTPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1599,6 +2037,9 @@ class TFOpenAIGPTPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFPegasusForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -1608,6 +2049,9 @@ class TFPegasusForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFPegasusModel:
     def __init__(self, *args, **kwargs):
@@ -1617,6 +2061,9 @@ class TFPegasusModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFPegasusPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1626,6 +2073,9 @@ class TFPegasusPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRagModel:
     def __init__(self, *args, **kwargs):
@@ -1635,6 +2085,9 @@ class TFRagModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRagPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1644,6 +2097,9 @@ class TFRagPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRagSequenceForGeneration:
     def __init__(self, *args, **kwargs):
@@ -1666,6 +2122,9 @@ class TFRemBertForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRemBertForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -1675,6 +2134,9 @@ class TFRemBertForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRemBertForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -1684,6 +2146,9 @@ class TFRemBertForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRemBertForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -1693,6 +2158,9 @@ class TFRemBertForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRemBertForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1702,6 +2170,9 @@ class TFRemBertForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRemBertForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1711,6 +2182,9 @@ class TFRemBertForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRemBertLayer:
     def __init__(self, *args, **kwargs):
@@ -1725,6 +2199,9 @@ class TFRemBertModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRemBertPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1734,6 +2211,9 @@ class TFRemBertPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_ROBERTA_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1746,6 +2226,9 @@ class TFRobertaForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRobertaForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -1755,6 +2238,9 @@ class TFRobertaForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRobertaForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -1764,6 +2250,9 @@ class TFRobertaForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRobertaForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -1773,6 +2262,9 @@ class TFRobertaForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRobertaForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1782,6 +2274,9 @@ class TFRobertaForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRobertaForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1791,6 +2286,9 @@ class TFRobertaForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRobertaMainLayer:
     def __init__(self, *args, **kwargs):
@@ -1805,6 +2303,9 @@ class TFRobertaModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRobertaPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1814,6 +2315,9 @@ class TFRobertaPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_ROFORMER_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1826,6 +2330,9 @@ class TFRoFormerForCausalLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRoFormerForMaskedLM:
     def __init__(self, *args, **kwargs):
@@ -1835,6 +2342,9 @@ class TFRoFormerForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRoFormerForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -1844,6 +2354,9 @@ class TFRoFormerForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRoFormerForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -1853,6 +2366,9 @@ class TFRoFormerForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRoFormerForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -1862,6 +2378,9 @@ class TFRoFormerForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRoFormerForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -1871,6 +2390,9 @@ class TFRoFormerForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRoFormerLayer:
     def __init__(self, *args, **kwargs):
@@ -1885,6 +2407,9 @@ class TFRoFormerModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFRoFormerPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1894,6 +2419,9 @@ class TFRoFormerPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_T5_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1906,6 +2434,9 @@ class TFT5EncoderModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFT5ForConditionalGeneration:
     def __init__(self, *args, **kwargs):
@@ -1915,6 +2446,9 @@ class TFT5ForConditionalGeneration:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFT5Model:
     def __init__(self, *args, **kwargs):
@@ -1924,6 +2458,9 @@ class TFT5Model:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFT5PreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1933,6 +2470,9 @@ class TFT5PreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_TRANSFO_XL_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -1950,6 +2490,9 @@ class TFTransfoXLForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFTransfoXLLMHeadModel:
     def __init__(self, *args, **kwargs):
@@ -1959,6 +2502,9 @@ class TFTransfoXLLMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFTransfoXLMainLayer:
     def __init__(self, *args, **kwargs):
@@ -1973,6 +2519,9 @@ class TFTransfoXLModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFTransfoXLPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -1982,6 +2531,9 @@ class TFTransfoXLPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFViTForImageClassification:
     def __init__(self, *args, **kwargs):
@@ -1996,6 +2548,9 @@ class TFViTModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFViTPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2005,6 +2560,9 @@ class TFViTPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_WAV_2_VEC_2_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -2022,6 +2580,9 @@ class TFWav2Vec2Model:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFWav2Vec2PreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2031,6 +2592,9 @@ class TFWav2Vec2PreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_XLM_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -2043,6 +2607,9 @@ class TFXLMForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLMForQuestionAnsweringSimple:
     def __init__(self, *args, **kwargs):
@@ -2052,6 +2619,9 @@ class TFXLMForQuestionAnsweringSimple:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLMForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -2061,6 +2631,9 @@ class TFXLMForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLMForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -2070,6 +2643,9 @@ class TFXLMForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLMMainLayer:
     def __init__(self, *args, **kwargs):
@@ -2084,6 +2660,9 @@ class TFXLMModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLMPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2093,6 +2672,9 @@ class TFXLMPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLMWithLMHeadModel:
     def __init__(self, *args, **kwargs):
@@ -2102,6 +2684,9 @@ class TFXLMWithLMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_XLM_ROBERTA_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -2114,6 +2699,9 @@ class TFXLMRobertaForMaskedLM:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLMRobertaForMultipleChoice:
     def __init__(self, *args, **kwargs):
@@ -2123,6 +2711,9 @@ class TFXLMRobertaForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLMRobertaForQuestionAnswering:
     def __init__(self, *args, **kwargs):
@@ -2132,6 +2723,9 @@ class TFXLMRobertaForQuestionAnswering:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLMRobertaForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -2141,6 +2735,9 @@ class TFXLMRobertaForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLMRobertaForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -2150,6 +2747,9 @@ class TFXLMRobertaForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLMRobertaModel:
     def __init__(self, *args, **kwargs):
@@ -2159,6 +2759,9 @@ class TFXLMRobertaModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 TF_XLNET_PRETRAINED_MODEL_ARCHIVE_LIST = None
 
@@ -2171,6 +2774,9 @@ class TFXLNetForMultipleChoice:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLNetForQuestionAnsweringSimple:
     def __init__(self, *args, **kwargs):
@@ -2180,6 +2786,9 @@ class TFXLNetForQuestionAnsweringSimple:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLNetForSequenceClassification:
     def __init__(self, *args, **kwargs):
@@ -2189,6 +2798,9 @@ class TFXLNetForSequenceClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLNetForTokenClassification:
     def __init__(self, *args, **kwargs):
@@ -2198,6 +2810,9 @@ class TFXLNetForTokenClassification:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLNetLMHeadModel:
     def __init__(self, *args, **kwargs):
@@ -2207,6 +2822,9 @@ class TFXLNetLMHeadModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLNetMainLayer:
     def __init__(self, *args, **kwargs):
@@ -2221,6 +2839,9 @@ class TFXLNetModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class TFXLNetPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -2230,6 +2851,9 @@ class TFXLNetPreTrainedModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["tf"])
 
+    def call(self, *args, **kwargs):
+        requires_backends(self, ["tf"])
+
 
 class AdamWeightDecay:
     def __init__(self, *args, **kwargs):
diff --git a/src/transformers/utils/dummy_timm_and_vision_objects.py b/src/transformers/utils/dummy_timm_and_vision_objects.py
index 6a92c8dc275..baa0563cc2e 100644
--- a/src/transformers/utils/dummy_timm_and_vision_objects.py
+++ b/src/transformers/utils/dummy_timm_and_vision_objects.py
@@ -13,6 +13,9 @@ class DetrForObjectDetection:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["timm", "vision"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["timm", "vision"])
+
 
 class DetrForSegmentation:
     def __init__(self, *args, **kwargs):
@@ -22,6 +25,9 @@ class DetrForSegmentation:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["timm", "vision"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["timm", "vision"])
+
 
 class DetrModel:
     def __init__(self, *args, **kwargs):
@@ -31,6 +37,9 @@ class DetrModel:
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["timm", "vision"])
 
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["timm", "vision"])
+
 
 class DetrPreTrainedModel:
     def __init__(self, *args, **kwargs):
@@ -39,3 +48,6 @@ class DetrPreTrainedModel:
     @classmethod
     def from_pretrained(cls, *args, **kwargs):
         requires_backends(cls, ["timm", "vision"])
+
+    def forward(self, *args, **kwargs):
+        requires_backends(self, ["timm", "vision"])
diff --git a/utils/check_dummies.py b/utils/check_dummies.py
index 0c98908968e..336f38d75a6 100644
--- a/utils/check_dummies.py
+++ b/utils/check_dummies.py
@@ -43,6 +43,30 @@ class {0}:
         requires_backends(cls, {1})
 """
 
+PT_DUMMY_PRETRAINED_CLASS = (
+    DUMMY_PRETRAINED_CLASS
+    + """
+    def forward(self, *args, **kwargs):
+        requires_backends(self, {1})
+"""
+)
+
+TF_DUMMY_PRETRAINED_CLASS = (
+    DUMMY_PRETRAINED_CLASS
+    + """
+    def call(self, *args, **kwargs):
+        requires_backends(self, {1})
+"""
+)
+
+FLAX_DUMMY_PRETRAINED_CLASS = (
+    DUMMY_PRETRAINED_CLASS
+    + """
+    def __call__(self, *args, **kwargs):
+        requires_backends(self, {1})
+"""
+)
+
 DUMMY_CLASS = """
 class {0}:
     def __init__(self, *args, **kwargs):
@@ -102,8 +126,7 @@ def read_init():
 
 def create_dummy_object(name, backend_name):
     """Create the code for the dummy object corresponding to `name`."""
-    _pretrained = [
-        "Config",
+    _models = [
         "ForCausalLM",
         "ForConditionalGeneration",
         "ForMaskedLM",
@@ -114,14 +137,24 @@ def create_dummy_object(name, backend_name):
         "ForSequenceClassification",
         "ForTokenClassification",
         "Model",
-        "Tokenizer",
-        "Processor",
     ]
+    _pretrained = ["Config", "Tokenizer", "Processor"]
     if name.isupper():
         return DUMMY_CONSTANT.format(name)
     elif name.islower():
         return DUMMY_FUNCTION.format(name, backend_name)
     else:
+        is_model = False
+        for part in _models:
+            if part in name:
+                is_model = True
+                break
+        if is_model:
+            if name.startswith("TF"):
+                return TF_DUMMY_PRETRAINED_CLASS.format(name, backend_name)
+            if name.startswith("Flax"):
+                return FLAX_DUMMY_PRETRAINED_CLASS.format(name, backend_name)
+            return PT_DUMMY_PRETRAINED_CLASS.format(name, backend_name)
         is_pretrained = False
         for part in _pretrained:
             if part in name: