Reorganize repo (#8580)

* Put models in subfolders * Styling * Fix imports in tests * More fixes in test imports * Sneaky hidden imports * Fix imports in doc files * More sneaky imports * Finish fixing tests * Fix examples * Fix path for copies * More fixes for examples * Fix dummy files * More fixes for example * More model import fixes * Is this why you're unhappy GitHub? * Fix imports in conver command
huggingface · Nov 17, 2020 · c89bdfb · c89bdfb
1 parent 9015073
commit c89bdfb
Show file tree

Hide file tree

Showing 381 changed files with 2,555 additions and 1,475 deletions.
diff --git a/.github/workflows/github-torch-hub.yml b/.github/workflows/github-torch-hub.yml
@@ -8,6 +8,9 @@ on:
 jobs:
   torch_hub_integration:
     runs-on: ubuntu-latest
+    env:
+      # TODO quickfix but may need more investigation
+      ACTIONS_ALLOW_UNSECURE_COMMANDS: True
     steps:
     # no checkout necessary here.
     - name: Extract branch name

diff --git a/.gitignore b/.gitignore
@@ -133,7 +133,6 @@ dmypy.json
 tensorflow_code
 
 # Models
-models
 proc_data
 
 # examples

diff --git a/docs/source/model_doc/albert.rst b/docs/source/model_doc/albert.rst
@@ -51,10 +51,10 @@ AlbertTokenizer
 Albert specific outputs
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. autoclass:: transformers.modeling_albert.AlbertForPreTrainingOutput
+.. autoclass:: transformers.models.albert.modeling_albert.AlbertForPreTrainingOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_albert.TFAlbertForPreTrainingOutput
+.. autoclass:: transformers.models.albert.modeling_tf_albert.TFAlbertForPreTrainingOutput
     :members:
 
 

diff --git a/docs/source/model_doc/bart.rst b/docs/source/model_doc/bart.rst
@@ -44,8 +44,8 @@ Implementation Notes
 - Bart doesn't use :obj:`token_type_ids` for sequence classification. Use :class:`~transformers.BartTokenizer` or
   :meth:`~transformers.BartTokenizer.encode` to get the proper splitting.
 - The forward pass of :class:`~transformers.BartModel` will create decoder inputs (using the helper function
-  :func:`transformers.modeling_bart._prepare_bart_decoder_inputs`) if they are not passed. This is different than some
-  other modeling APIs.
+  :func:`transformers.models.bart.modeling_bart._prepare_bart_decoder_inputs`) if they are not passed. This is
+  different than some other modeling APIs.
 - Model predictions are intended to be identical to the original implementation when
   :obj:`force_bos_token_to_be_generated=True`. This only works, however, if the string you pass to
   :func:`fairseq.encode` starts with a space.
@@ -93,7 +93,7 @@ BartModel
 .. autoclass:: transformers.BartModel
     :members: forward
 
-.. autofunction:: transformers.modeling_bart._prepare_bart_decoder_inputs
+.. autofunction:: transformers.models.bart.modeling_bart._prepare_bart_decoder_inputs
 
 
 BartForConditionalGeneration

diff --git a/docs/source/model_doc/bert.rst b/docs/source/model_doc/bert.rst
@@ -57,10 +57,10 @@ BertTokenizerFast
 Bert specific outputs
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. autoclass:: transformers.modeling_bert.BertForPreTrainingOutput
+.. autoclass:: transformers.models.bert.modeling_bert.BertForPreTrainingOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_bert.TFBertForPreTrainingOutput
+.. autoclass:: transformers.models.bert.modeling_tf_bert.TFBertForPreTrainingOutput
     :members:
 
 

diff --git a/docs/source/model_doc/dpr.rst b/docs/source/model_doc/dpr.rst
@@ -71,13 +71,13 @@ DPRReaderTokenizerFast
 DPR specific outputs
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. autoclass:: transformers.modeling_dpr.DPRContextEncoderOutput
+.. autoclass:: transformers.models.dpr.modeling_dpr.DPRContextEncoderOutput
     :members:
 
-.. autoclass:: transformers.modeling_dpr.DPRQuestionEncoderOutput
+.. autoclass:: transformers.models.dpr.modeling_dpr.DPRQuestionEncoderOutput
     :members:
 
-.. autoclass:: transformers.modeling_dpr.DPRReaderOutput
+.. autoclass:: transformers.models.dpr.modeling_dpr.DPRReaderOutput
     :members:
 
 

diff --git a/docs/source/model_doc/electra.rst b/docs/source/model_doc/electra.rst
@@ -69,10 +69,10 @@ ElectraTokenizerFast
 Electra specific outputs
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. autoclass:: transformers.modeling_electra.ElectraForPreTrainingOutput
+.. autoclass:: transformers.models.electra.modeling_electra.ElectraForPreTrainingOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_electra.TFElectraForPreTrainingOutput
+.. autoclass:: transformers.models.electra.modeling_tf_electra.TFElectraForPreTrainingOutput
     :members:
 
 

diff --git a/docs/source/model_doc/funnel.rst b/docs/source/model_doc/funnel.rst
@@ -65,10 +65,10 @@ FunnelTokenizerFast
 Funnel specific outputs
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. autoclass:: transformers.modeling_funnel.FunnelForPreTrainingOutput
+.. autoclass:: transformers.models.funnel.modeling_funnel.FunnelForPreTrainingOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_funnel.TFFunnelForPreTrainingOutput
+.. autoclass:: transformers.models.funnel.modeling_tf_funnel.TFFunnelForPreTrainingOutput
     :members:
 
 

diff --git a/docs/source/model_doc/gpt.rst b/docs/source/model_doc/gpt.rst
@@ -72,10 +72,10 @@ OpenAIGPTTokenizerFast
 OpenAI specific outputs
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. autoclass:: transformers.modeling_openai.OpenAIGPTDoubleHeadsModelOutput
+.. autoclass:: transformers.models.openai.modeling_openai.OpenAIGPTDoubleHeadsModelOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_openai.TFOpenAIGPTDoubleHeadsModelOutput
+.. autoclass:: transformers.models.openai.modeling_tf_openai.TFOpenAIGPTDoubleHeadsModelOutput
     :members:
 
 

diff --git a/docs/source/model_doc/gpt2.rst b/docs/source/model_doc/gpt2.rst
@@ -60,10 +60,10 @@ GPT2TokenizerFast
 GPT2 specific outputs
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. autoclass:: transformers.modeling_gpt2.GPT2DoubleHeadsModelOutput
+.. autoclass:: transformers.models.gpt2.modeling_gpt2.GPT2DoubleHeadsModelOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_gpt2.TFGPT2DoubleHeadsModelOutput
+.. autoclass:: transformers.models.gpt2.modeling_tf_gpt2.TFGPT2DoubleHeadsModelOutput
     :members:
 
 

diff --git a/docs/source/model_doc/longformer.rst b/docs/source/model_doc/longformer.rst
@@ -93,29 +93,27 @@ LongformerTokenizerFast
 Longformer specific outputs
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. autoclass:: transformers.modeling_longformer.LongformerBaseModelOutput
+.. autoclass:: transformers.models.longformer.modeling_longformer.LongformerBaseModelOutput
     :members: 
 
-.. autoclass:: transformers.modeling_longformer.LongformerBaseModelOutputWithPooling
+.. autoclass:: transformers.models.longformer.modeling_longformer.LongformerBaseModelOutputWithPooling
     :members: 
 
-.. autoclass:: transformers.modeling_longformer.LongformerMultipleChoiceModelOutput
+.. autoclass:: transformers.models.longformer.modeling_longformer.LongformerMultipleChoiceModelOutput
     :members: 
 
-.. autoclass:: transformers.modeling_longformer.LongformerQuestionAnsweringModelOutput
+.. autoclass:: transformers.models.longformer.modeling_longformer.LongformerQuestionAnsweringModelOutput
     :members: 
 
-.. autoclass:: transformers.modeling_tf_longformer.TFLongformerBaseModelOutput
+.. autoclass:: transformers.models.longformer.modeling_tf_longformer.TFLongformerBaseModelOutput
     :members: 
 
-.. autoclass:: transformers.modeling_tf_longformer.TFLongformerBaseModelOutputWithPooling
+.. autoclass:: transformers.models.longformer.modeling_tf_longformer.TFLongformerBaseModelOutputWithPooling
     :members: 
 
-.. autoclass:: transformers.modeling_tf_longformer.TFLongformerQuestionAnsweringModelOutput
+.. autoclass:: transformers.models.longformer.modeling_tf_longformer.TFLongformerQuestionAnsweringModelOutput
     :members: 
 
-LongformerModel
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 LongformerModel
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

diff --git a/docs/source/model_doc/lxmert.rst b/docs/source/model_doc/lxmert.rst
@@ -67,19 +67,19 @@ LxmertTokenizerFast
 Lxmert specific outputs
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. autoclass:: transformers.modeling_lxmert.LxmertModelOutput
+.. autoclass:: transformers.models.lxmert.modeling_lxmert.LxmertModelOutput
     :members:
 
-.. autoclass:: transformers.modeling_lxmert.LxmertForPreTrainingOutput
+.. autoclass:: transformers.models.lxmert.modeling_lxmert.LxmertForPreTrainingOutput
     :members:
 
-.. autoclass:: transformers.modeling_lxmert.LxmertForQuestionAnsweringOutput
+.. autoclass:: transformers.models.lxmert.modeling_lxmert.LxmertForQuestionAnsweringOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_lxmert.TFLxmertModelOutput
+.. autoclass:: transformers.models.lxmert.modeling_tf_lxmert.TFLxmertModelOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_lxmert.TFLxmertForPreTrainingOutput
+.. autoclass:: transformers.models.lxmert.modeling_tf_lxmert.TFLxmertForPreTrainingOutput
     :members:
 
 

diff --git a/docs/source/model_doc/mobilebert.rst b/docs/source/model_doc/mobilebert.rst
@@ -58,10 +58,10 @@ MobileBertTokenizerFast
 MobileBert specific outputs
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. autoclass:: transformers.modeling_mobilebert.MobileBertForPreTrainingOutput
+.. autoclass:: transformers.models.mobilebert.modeling_mobilebert.MobileBertForPreTrainingOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_mobilebert.TFMobileBertForPreTrainingOutput
+.. autoclass:: transformers.models.mobilebert.modeling_tf_mobilebert.TFMobileBertForPreTrainingOutput
     :members:
 
 

diff --git a/docs/source/model_doc/prophetnet.rst b/docs/source/model_doc/prophetnet.rst
@@ -47,16 +47,16 @@ ProphetNetTokenizer
 ProphetNet specific outputs
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. autoclass:: transformers.modeling_prophetnet.ProphetNetSeq2SeqLMOutput
+.. autoclass:: transformers.models.prophetnet.modeling_prophetnet.ProphetNetSeq2SeqLMOutput
     :members:
 
-.. autoclass:: transformers.modeling_prophetnet.ProphetNetSeq2SeqModelOutput
+.. autoclass:: transformers.models.prophetnet.modeling_prophetnet.ProphetNetSeq2SeqModelOutput
     :members:
 
-.. autoclass:: transformers.modeling_prophetnet.ProphetNetDecoderModelOutput
+.. autoclass:: transformers.models.prophetnet.modeling_prophetnet.ProphetNetDecoderModelOutput
     :members:
 
-.. autoclass:: transformers.modeling_prophetnet.ProphetNetDecoderLMOutput
+.. autoclass:: transformers.models.prophetnet.modeling_prophetnet.ProphetNetDecoderLMOutput
     :members:
 
 ProphetNetModel

diff --git a/docs/source/model_doc/rag.rst b/docs/source/model_doc/rag.rst
@@ -50,10 +50,10 @@ RagTokenizer
 Rag specific outputs
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. autoclass:: transformers.modeling_rag.RetrievAugLMMarginOutput
+.. autoclass:: transformers.models.rag.modeling_rag.RetrievAugLMMarginOutput
     :members:
 
-.. autoclass:: transformers.modeling_rag.RetrievAugLMOutput
+.. autoclass:: transformers.models.rag.modeling_rag.RetrievAugLMOutput
     :members:
 
 RagRetriever

diff --git a/docs/source/model_doc/transformerxl.rst b/docs/source/model_doc/transformerxl.rst
@@ -49,16 +49,16 @@ TransfoXLTokenizer
 TransfoXL specific outputs
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. autoclass:: transformers.modeling_transfo_xl.TransfoXLModelOutput
+.. autoclass:: transformers.models.transfo_xl.modeling_transfo_xl.TransfoXLModelOutput
     :members:
 
-.. autoclass:: transformers.modeling_transfo_xl.TransfoXLLMHeadModelOutput
+.. autoclass:: transformers.models.transfo_xl.modeling_transfo_xl.TransfoXLLMHeadModelOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_transfo_xl.TFTransfoXLModelOutput
+.. autoclass:: transformers.models.transfo_xl.modeling_tf_transfo_xl.TFTransfoXLModelOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_transfo_xl.TFTransfoXLLMHeadModelOutput
+.. autoclass:: transformers.models.transfo_xl.modeling_tf_transfo_xl.TFTransfoXLLMHeadModelOutput
     :members:
 
 

diff --git a/docs/source/model_doc/xlm.rst b/docs/source/model_doc/xlm.rst
@@ -50,7 +50,7 @@ XLMTokenizer
 XLM specific outputs
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. autoclass:: transformers.modeling_xlm.XLMForQuestionAnsweringOutput
+.. autoclass:: transformers.models.xlm.modeling_xlm.XLMForQuestionAnsweringOutput
     :members:
 
 

diff --git a/docs/source/model_doc/xlnet.rst b/docs/source/model_doc/xlnet.rst
@@ -53,43 +53,43 @@ XLNetTokenizer
 XLNet specific outputs
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-.. autoclass:: transformers.modeling_xlnet.XLNetModelOutput
+.. autoclass:: transformers.models.xlnet.modeling_xlnet.XLNetModelOutput
     :members:
 
-.. autoclass:: transformers.modeling_xlnet.XLNetLMHeadModelOutput
+.. autoclass:: transformers.models.xlnet.modeling_xlnet.XLNetLMHeadModelOutput
     :members:
 
-.. autoclass:: transformers.modeling_xlnet.XLNetForSequenceClassificationOutput
+.. autoclass:: transformers.models.xlnet.modeling_xlnet.XLNetForSequenceClassificationOutput
     :members:
 
-.. autoclass:: transformers.modeling_xlnet.XLNetForMultipleChoiceOutput
+.. autoclass:: transformers.models.xlnet.modeling_xlnet.XLNetForMultipleChoiceOutput
     :members:
 
-.. autoclass:: transformers.modeling_xlnet.XLNetForTokenClassificationOutput
+.. autoclass:: transformers.models.xlnet.modeling_xlnet.XLNetForTokenClassificationOutput
     :members:
 
-.. autoclass:: transformers.modeling_xlnet.XLNetForQuestionAnsweringSimpleOutput
+.. autoclass:: transformers.models.xlnet.modeling_xlnet.XLNetForQuestionAnsweringSimpleOutput
     :members:
 
-.. autoclass:: transformers.modeling_xlnet.XLNetForQuestionAnsweringOutput
+.. autoclass:: transformers.models.xlnet.modeling_xlnet.XLNetForQuestionAnsweringOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_xlnet.TFXLNetModelOutput
+.. autoclass:: transformers.models.xlnet.modeling_tf_xlnet.TFXLNetModelOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_xlnet.TFXLNetLMHeadModelOutput
+.. autoclass:: transformers.models.xlnet.modeling_tf_xlnet.TFXLNetLMHeadModelOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_xlnet.TFXLNetForSequenceClassificationOutput
+.. autoclass:: transformers.models.xlnet.modeling_tf_xlnet.TFXLNetForSequenceClassificationOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_xlnet.TFXLNetForMultipleChoiceOutput
+.. autoclass:: transformers.models.xlnet.modeling_tf_xlnet.TFXLNetForMultipleChoiceOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_xlnet.TFXLNetForTokenClassificationOutput
+.. autoclass:: transformers.models.xlnet.modeling_tf_xlnet.TFXLNetForTokenClassificationOutput
     :members:
 
-.. autoclass:: transformers.modeling_tf_xlnet.TFXLNetForQuestionAnsweringSimpleOutput
+.. autoclass:: transformers.models.xlnet.modeling_tf_xlnet.TFXLNetForQuestionAnsweringSimpleOutput
     :members:
 
 

diff --git a/examples/bert-loses-patience/pabee/modeling_pabee_albert.py b/examples/bert-loses-patience/pabee/modeling_pabee_albert.py
@@ -21,7 +21,7 @@
 from torch.nn import CrossEntropyLoss, MSELoss
 
 from transformers.file_utils import add_start_docstrings, add_start_docstrings_to_model_forward
-from transformers.modeling_albert import (
+from transformers.models.albert.modeling_albert import (
     ALBERT_INPUTS_DOCSTRING,
     ALBERT_START_DOCSTRING,
     AlbertModel,

diff --git a/examples/bert-loses-patience/pabee/modeling_pabee_bert.py b/examples/bert-loses-patience/pabee/modeling_pabee_bert.py
@@ -23,7 +23,7 @@
 from torch.nn import CrossEntropyLoss, MSELoss
 
 from transformers.file_utils import add_start_docstrings, add_start_docstrings_to_model_forward
-from transformers.modeling_bert import (
+from transformers.models.bert.modeling_bert import (
     BERT_INPUTS_DOCSTRING,
     BERT_START_DOCSTRING,
     BertEncoder,

diff --git a/examples/contrib/run_camembert.py b/examples/contrib/run_camembert.py
@@ -1,7 +1,6 @@
 import torch
 
-from transformers.modeling_camembert import CamembertForMaskedLM
-from transformers.tokenization_camembert import CamembertTokenizer
+from transformers import CamembertForMaskedLM, CamembertTokenizer
 
 
 def fill_mask(masked_input, model, tokenizer, topk=5):

diff --git a/examples/contrib/run_swag.py b/examples/contrib/run_swag.py
@@ -32,8 +32,14 @@
 from tqdm import tqdm, trange
 
 import transformers
-from transformers import WEIGHTS_NAME, AdamW, AutoConfig, AutoTokenizer, get_linear_schedule_with_warmup
-from transformers.modeling_auto import AutoModelForMultipleChoice
+from transformers import (
+    WEIGHTS_NAME,
+    AdamW,
+    AutoConfig,
+    AutoModelForMultipleChoice,
+    AutoTokenizer,
+    get_linear_schedule_with_warmup,
+)
 from transformers.trainer_utils import is_main_process
 
 

diff --git a/examples/deebert/src/modeling_highway_bert.py b/examples/deebert/src/modeling_highway_bert.py
@@ -3,7 +3,7 @@
 from torch.nn import CrossEntropyLoss, MSELoss
 
 from transformers.file_utils import add_start_docstrings, add_start_docstrings_to_model_forward
-from transformers.modeling_bert import (
+from transformers.models.bert.modeling_bert import (
     BERT_INPUTS_DOCSTRING,
     BERT_START_DOCSTRING,
     BertEmbeddings,

diff --git a/examples/deebert/src/modeling_highway_roberta.py b/examples/deebert/src/modeling_highway_roberta.py
@@ -3,9 +3,13 @@
 import torch.nn as nn
 from torch.nn import CrossEntropyLoss, MSELoss
 
-from transformers.configuration_roberta import RobertaConfig
+from transformers import RobertaConfig
 from transformers.file_utils import add_start_docstrings, add_start_docstrings_to_model_forward
-from transformers.modeling_roberta import ROBERTA_INPUTS_DOCSTRING, ROBERTA_START_DOCSTRING, RobertaEmbeddings
+from transformers.models.roberta.modeling_roberta import (
+    ROBERTA_INPUTS_DOCSTRING,
+    ROBERTA_START_DOCSTRING,
+    RobertaEmbeddings,
+)
 
 from .modeling_highway_bert import BertPreTrainedModel, DeeBertModel, HighwayException, entropy