Result of new doc style with fixes (#17015)

* Result of new doc style with fixes

* Add last two files

* Bump hf-doc-builder

docs/source/en/model_doc/bert-generation.mdx

@@ -49,7 +49,7 @@ Usage:
 
 >>> input_ids = tokenizer(
 ...     "This is a long article to summarize", add_special_tokens=False, return_tensors="pt"
->>> ).input_ids
+... ).input_ids
 >>> labels = tokenizer("This is a short summary", return_tensors="pt").input_ids
 
 >>> # train...
@@ -67,7 +67,7 @@ Usage:
 
 >>> input_ids = tokenizer(
 ...     "This is the first sentence. This is the second sentence.", add_special_tokens=False, return_tensors="pt"
->>> ).input_ids
+... ).input_ids
 
 >>> outputs = sentence_fuser.generate(input_ids)
 

docs/source/en/model_doc/luke.mdx

@@ -97,7 +97,7 @@ Example:
 >>> entities = [
 ...     "Beyoncé",
 ...     "Los Angeles",
->>> ]  # Wikipedia entity titles corresponding to the entity mentions "Beyoncé" and "Los Angeles"
+... ]  # Wikipedia entity titles corresponding to the entity mentions "Beyoncé" and "Los Angeles"
 >>> entity_spans = [(0, 7), (17, 28)]  # character-based entity spans corresponding to "Beyoncé" and "Los Angeles"
 >>> inputs = tokenizer(text, entities=entities, entity_spans=entity_spans, add_prefix_space=True, return_tensors="pt")
 >>> outputs = model(**inputs)

setup.py

@@ -111,7 +111,7 @@ _deps = [
     "ftfy",
     "fugashi>=1.0",
     "GitPython<3.1.19",
-    "hf-doc-builder>=0.2.0",
+    "hf-doc-builder>=0.3.0",
     "huggingface-hub>=0.1.0,<1.0",
     "importlib_metadata",
     "ipadic>=1.0.0,<2.0",

src/transformers/dependency_versions_table.py

@@ -18,7 +18,7 @@ deps = {
     "ftfy": "ftfy",
     "fugashi": "fugashi>=1.0",
     "GitPython": "GitPython<3.1.19",
-    "hf-doc-builder": "hf-doc-builder>=0.2.0",
+    "hf-doc-builder": "hf-doc-builder>=0.3.0",
     "huggingface-hub": "huggingface-hub>=0.1.0,<1.0",
     "importlib_metadata": "importlib_metadata",
     "ipadic": "ipadic>=1.0.0,<2.0",

src/transformers/models/encoder_decoder/modeling_encoder_decoder.py

@@ -457,7 +457,7 @@ class EncoderDecoderModel(PreTrainedModel):
         >>> tokenizer = BertTokenizer.from_pretrained("bert-base-uncased")
         >>> model = EncoderDecoderModel.from_encoder_decoder_pretrained(
         ...     "bert-base-uncased", "bert-base-uncased"
-        >>> )  # initialize Bert2Bert from pre-trained checkpoints
+        ... )  # initialize Bert2Bert from pre-trained checkpoints
 
         >>> # training
         >>> model.config.decoder_start_token_id = tokenizer.cls_token_id

src/transformers/models/encoder_decoder/modeling_tf_encoder_decoder.py

@@ -528,7 +528,7 @@ class TFEncoderDecoderModel(TFPreTrainedModel, TFCausalLanguageModelingLoss):
         >>> # forward
         >>> input_ids = tokenizer.encode(
         ...     "Hello, my dog is cute", add_special_tokens=True, return_tensors="tf"
-        >>> )  # Batch size 1
+        ... )  # Batch size 1
         >>> outputs = model(input_ids=input_ids, decoder_input_ids=input_ids)
 
         >>> # training

src/transformers/models/gpt2/modeling_tf_gpt2.py

@@ -1061,7 +1061,7 @@ class TFGPT2DoubleHeadsModel(TFGPT2PreTrainedModel):
 
         >>> embedding_layer = model.resize_token_embeddings(
         ...     len(tokenizer)
-        >>> )  # Update the model embeddings with the new vocabulary size
+        ... )  # Update the model embeddings with the new vocabulary size
 
         >>> choices = ["Hello, my dog is cute [CLS]", "Hello, my cat is cute [CLS]"]
         >>> encoded_choices = [tokenizer.encode(s) for s in choices]

src/transformers/models/imagegpt/modeling_imagegpt.py

@@ -1000,7 +1000,7 @@ class ImageGPTForCausalImageModeling(ImageGPTPreTrainedModel):
         >>> samples = output[:, 1:].cpu().detach().numpy()
         >>> samples_img = [
         ...     np.reshape(np.rint(127.5 * (clusters[s] + 1.0)), [n_px, n_px, 3]).astype(np.uint8) for s in samples
-        >>> ]  # convert color cluster tokens back to pixels
+        ... ]  # convert color cluster tokens back to pixels
         >>> f, axes = plt.subplots(1, batch_size, dpi=300)
 
         >>> for img, ax in zip(samples_img, axes):

src/transformers/models/longformer/modeling_longformer.py

@@ -1634,10 +1634,10 @@ class LongformerModel(LongformerPreTrainedModel):
 
         >>> attention_mask = torch.ones(
         ...     input_ids.shape, dtype=torch.long, device=input_ids.device
-        >>> )  # initialize to local attention
+        ... )  # initialize to local attention
         >>> global_attention_mask = torch.zeros(
         ...     input_ids.shape, dtype=torch.long, device=input_ids.device
-        >>> )  # initialize to global attention to be deactivated for all tokens
+        ... )  # initialize to global attention to be deactivated for all tokens
         >>> global_attention_mask[
         ...     :,
         ...     [
@@ -1645,7 +1645,7 @@ class LongformerModel(LongformerPreTrainedModel):
         ...         4,
         ...         21,
         ...     ],
-        >>> ] = 1  # Set global attention to random tokens for the sake of this example
+        ... ] = 1  # Set global attention to random tokens for the sake of this example
         >>> # Usually, set global attention based on the task. For example,
         >>> # classification: the <s> token
         >>> # QA: question tokens
@@ -2025,7 +2025,7 @@ class LongformerForQuestionAnswering(LongformerPreTrainedModel):
         >>> answer_tokens = all_tokens[torch.argmax(start_logits) : torch.argmax(end_logits) + 1]
         >>> answer = tokenizer.decode(
         ...     tokenizer.convert_tokens_to_ids(answer_tokens)
-        >>> )  # remove space prepending space token
+        ... )  # remove space prepending space token
         ```"""
         return_dict = return_dict if return_dict is not None else self.config.use_return_dict
 

src/transformers/models/luke/modeling_luke.py

@@ -953,11 +953,11 @@ class LukeModel(LukePreTrainedModel):
         >>> entities = [
         ...     "Beyoncé",
         ...     "Los Angeles",
-        >>> ]  # Wikipedia entity titles corresponding to the entity mentions "Beyoncé" and "Los Angeles"
+        ... ]  # Wikipedia entity titles corresponding to the entity mentions "Beyoncé" and "Los Angeles"
         >>> entity_spans = [
         ...     (0, 7),
         ...     (17, 28),
-        >>> ]  # character-based entity spans corresponding to "Beyoncé" and "Los Angeles"
+        ... ]  # character-based entity spans corresponding to "Beyoncé" and "Los Angeles"
 
         >>> encoding = tokenizer(
         ...     text, entities=entities, entity_spans=entity_spans, add_prefix_space=True, return_tensors="pt"
@@ -1435,7 +1435,7 @@ class LukeForEntityPairClassification(LukePreTrainedModel):
         >>> entity_spans = [
         ...     (0, 7),
         ...     (17, 28),
-        >>> ]  # character-based entity spans corresponding to "Beyoncé" and "Los Angeles"
+        ... ]  # character-based entity spans corresponding to "Beyoncé" and "Los Angeles"
         >>> inputs = tokenizer(text, entity_spans=entity_spans, return_tensors="pt")
         >>> outputs = model(**inputs)
         >>> logits = outputs.logits

src/transformers/models/openai/modeling_openai.py

@@ -674,7 +674,7 @@ class OpenAIGPTDoubleHeadsModel(OpenAIGPTPreTrainedModel):
         >>> model = OpenAIGPTDoubleHeadsModel.from_pretrained("openai-gpt")
         >>> tokenizer.add_special_tokens(
         ...     {"cls_token": "[CLS]"}
-        >>> )  # Add a [CLS] to the vocabulary (we should train it also!)
+        ... )  # Add a [CLS] to the vocabulary (we should train it also!)
         >>> model.resize_token_embeddings(len(tokenizer))
 
         >>> choices = ["Hello, my dog is cute [CLS]", "Hello, my cat is cute [CLS]"]

src/transformers/models/openai/modeling_tf_openai.py

@@ -693,9 +693,9 @@ class TFOpenAIGPTDoubleHeadsModel(TFOpenAIGPTPreTrainedModel):
         >>> inputs = {k: tf.expand_dims(v, 0) for k, v in encoding.items()}
         >>> inputs["mc_token_ids"] = tf.constant(
         ...     [inputs["input_ids"].shape[-1] - 1, inputs["input_ids"].shape[-1] - 1]
-        >>> )[
+        ... )[
         ...     None, :
-        >>> ]  # Batch size 1
+        ... ]  # Batch size 1
         >>> outputs = model(inputs)
         >>> lm_prediction_scores, mc_prediction_scores = outputs[:2]
         ```"""

src/transformers/models/prophetnet/modeling_prophetnet.py

@@ -1813,7 +1813,7 @@ class ProphetNetModel(ProphetNetPreTrainedModel):
 
         >>> input_ids = tokenizer(
         ...     "Studies have been shown that owning a dog is good for you", return_tensors="pt"
-        >>> ).input_ids  # Batch size 1
+        ... ).input_ids  # Batch size 1
         >>> decoder_input_ids = tokenizer("Studies show that", return_tensors="pt").input_ids  # Batch size 1
         >>> outputs = model(input_ids=input_ids, decoder_input_ids=decoder_input_ids)
 
@@ -1935,7 +1935,7 @@ class ProphetNetForConditionalGeneration(ProphetNetPreTrainedModel):
 
         >>> input_ids = tokenizer(
         ...     "Studies have been shown that owning a dog is good for you", return_tensors="pt"
-        >>> ).input_ids  # Batch size 1
+        ... ).input_ids  # Batch size 1
         >>> decoder_input_ids = tokenizer("Studies show that", return_tensors="pt").input_ids  # Batch size 1
         >>> outputs = model(input_ids=input_ids, decoder_input_ids=decoder_input_ids)
 
@@ -2202,7 +2202,7 @@ class ProphetNetForCausalLM(ProphetNetPreTrainedModel):
         >>> input_ids = tokenizer_enc(ARTICLE, return_tensors="pt").input_ids
         >>> labels = tokenizer_dec(
         ...     "us rejects charges against its ambassador in bolivia", return_tensors="pt"
-        >>> ).input_ids
+        ... ).input_ids
         >>> outputs = model(input_ids=input_ids, decoder_input_ids=labels[:, :-1], labels=labels[:, 1:])
 
         >>> loss = outputs.loss

src/transformers/models/rag/modeling_rag.py

@@ -826,7 +826,7 @@ class RagSequenceForGeneration(RagPreTrainedModel):
         >>> docs_dict = retriever(input_ids.numpy(), question_hidden_states.detach().numpy(), return_tensors="pt")
         >>> doc_scores = torch.bmm(
         ...     question_hidden_states.unsqueeze(1), docs_dict["retrieved_doc_embeds"].float().transpose(1, 2)
-        >>> ).squeeze(1)
+        ... ).squeeze(1)
         >>> # 3. Forward to generator
         >>> outputs = model(
         ...     context_input_ids=docs_dict["context_input_ids"],
@@ -1293,7 +1293,7 @@ class RagTokenForGeneration(RagPreTrainedModel):
         >>> docs_dict = retriever(input_ids.numpy(), question_hidden_states.detach().numpy(), return_tensors="pt")
         >>> doc_scores = torch.bmm(
         ...     question_hidden_states.unsqueeze(1), docs_dict["retrieved_doc_embeds"].float().transpose(1, 2)
-        >>> ).squeeze(1)
+        ... ).squeeze(1)
         >>> # 3. Forward to generator
         >>> outputs = model(
         ...     context_input_ids=docs_dict["context_input_ids"],

src/transformers/models/rag/retrieval_rag.py

@@ -354,7 +354,7 @@ class RagRetriever:
 
     >>> dataset = (
     ...     ...
-    >>> )  # dataset must be a datasets.Datasets object with columns "title", "text" and "embeddings", and it must have a faiss index
+    ... )  # dataset must be a datasets.Datasets object with columns "title", "text" and "embeddings", and it must have a faiss index
     >>> retriever = RagRetriever.from_pretrained("facebook/dpr-ctx_encoder-single-nq-base", indexed_dataset=dataset)
 
     >>> # To load your own indexed dataset built with the datasets library that was saved on disk. More info in examples/rag/use_own_knowledge_dataset.py

src/transformers/models/realm/modeling_realm.py

@@ -1782,7 +1782,7 @@ class RealmForOpenQA(RealmPreTrainedModel):
         ...     add_special_tokens=False,
         ...     return_token_type_ids=False,
         ...     return_attention_mask=False,
-        >>> ).input_ids
+        ... ).input_ids
 
         >>> reader_output, predicted_answer_ids = model(**question_ids, answer_ids=answer_ids, return_dict=False)
         >>> predicted_answer = tokenizer.decode(predicted_answer_ids)

src/transformers/models/speech_to_text/modeling_tf_speech_to_text.py

@@ -1387,7 +1387,7 @@ class TFSpeech2TextForConditionalGeneration(TFSpeech2TextPreTrainedModel, TFCaus
 
         >>> input_features = processor(
         ...     ds["speech"][0], sampling_rate=16000, return_tensors="tf"
-        >>> ).input_features  # Batch size 1
+        ... ).input_features  # Batch size 1
         >>> generated_ids = model.generate(input_features)
 
         >>> transcription = processor.batch_decode(generated_ids)

src/transformers/models/t5/modeling_flax_t5.py

@@ -1344,7 +1344,7 @@ FLAX_T5_MODEL_DOCSTRING = """
 
     >>> input_ids = tokenizer(
     ...     "Studies have been shown that owning a dog is good for you", return_tensors="np"
-    >>> ).input_ids
+    ... ).input_ids
     >>> decoder_input_ids = tokenizer("Studies show that", return_tensors="np").input_ids
 
     >>> # forward pass

src/transformers/models/t5/modeling_t5.py

@@ -1375,7 +1375,7 @@ class T5Model(T5PreTrainedModel):
 
         >>> input_ids = tokenizer(
         ...     "Studies have been shown that owning a dog is good for you", return_tensors="pt"
-        >>> ).input_ids  # Batch size 1
+        ... ).input_ids  # Batch size 1
         >>> decoder_input_ids = tokenizer("Studies show that", return_tensors="pt").input_ids  # Batch size 1
 
         >>> # forward pass
@@ -1583,7 +1583,7 @@ class T5ForConditionalGeneration(T5PreTrainedModel):
         >>> # inference
         >>> input_ids = tokenizer(
         ...     "summarize: studies have shown that owning a dog is good for you", return_tensors="pt"
-        >>> ).input_ids  # Batch size 1
+        ... ).input_ids  # Batch size 1
         >>> outputs = model.generate(input_ids)
         >>> print(tokenizer.decode(outputs[0], skip_special_tokens=True))
         >>> # studies have shown that owning a dog is good for you.
@@ -1831,7 +1831,7 @@ class T5EncoderModel(T5PreTrainedModel):
         >>> model = T5EncoderModel.from_pretrained("t5-small")
         >>> input_ids = tokenizer(
         ...     "Studies have been shown that owning a dog is good for you", return_tensors="pt"
-        >>> ).input_ids  # Batch size 1
+        ... ).input_ids  # Batch size 1
         >>> outputs = model(input_ids=input_ids)
         >>> last_hidden_states = outputs.last_hidden_state
         ```"""

src/transformers/models/t5/modeling_tf_t5.py

@@ -1165,7 +1165,7 @@ class TFT5Model(TFT5PreTrainedModel):
 
         >>> input_ids = tokenizer(
         ...     "Studies have been shown that owning a dog is good for you", return_tensors="tf"
-        >>> ).input_ids  # Batch size 1
+        ... ).input_ids  # Batch size 1
         >>> decoder_input_ids = tokenizer("Studies show that", return_tensors="tf").input_ids  # Batch size 1
 
         >>> # forward pass
@@ -1353,7 +1353,7 @@ class TFT5ForConditionalGeneration(TFT5PreTrainedModel, TFCausalLanguageModeling
         >>> # inference
         >>> inputs = tokenizer(
         ...     "summarize: studies have shown that owning a dog is good for you", return_tensors="tf"
-        >>> ).input_ids  # Batch size 1
+        ... ).input_ids  # Batch size 1
         >>> outputs = model.generate(inputs)
         >>> print(tokenizer.decode(outputs[0], skip_special_tokens=True))
         >>> # studies have shown that owning a dog is good for you
@@ -1642,7 +1642,7 @@ class TFT5EncoderModel(TFT5PreTrainedModel):
 
         >>> input_ids = tokenizer(
         ...     "Studies have been shown that owning a dog is good for you", return_tensors="tf"
-        >>> ).input_ids  # Batch size 1
+        ... ).input_ids  # Batch size 1
         >>> outputs = model(input_ids)
         ```"""
 

src/transformers/models/tapas/modeling_tapas.py

@@ -1068,7 +1068,7 @@ class TapasForMaskedLM(TapasPreTrainedModel):
         ... )
         >>> labels = tokenizer(
         ...     table=table, queries="How many movies has George Clooney played in?", return_tensors="pt"
-        >>> )["input_ids"]
+        ... )["input_ids"]
 
         >>> outputs = model(**inputs, labels=labels)
         >>> logits = outputs.logits

src/transformers/models/tapas/modeling_tf_tapas.py

@@ -1095,7 +1095,7 @@ class TFTapasForMaskedLM(TFTapasPreTrainedModel, TFMaskedLanguageModelingLoss):
         ... )
         >>> labels = tokenizer(
         ...     table=table, queries="How many movies has George Clooney played in?", return_tensors="tf"
-        >>> )["input_ids"]
+        ... )["input_ids"]
 
         >>> outputs = model(**inputs, labels=labels)
         >>> logits = outputs.logits

src/transformers/models/vision_encoder_decoder/modeling_tf_vision_encoder_decoder.py

@@ -326,7 +326,7 @@ class TFVisionEncoderDecoderModel(TFPreTrainedModel, TFCausalLanguageModelingLos
 
         >>> output_ids = model.generate(
         ...     pixel_values, max_length=16, num_beams=4, return_dict_in_generate=True
-        >>> ).sequences
+        ... ).sequences
 
         >>> preds = decoder_tokenizer.batch_decode(output_ids, skip_special_tokens=True)
         >>> preds = [pred.strip() for pred in preds]

src/transformers/models/wav2vec2/modeling_flax_wav2vec2.py

@@ -1081,7 +1081,7 @@ FLAX_WAV2VEC2_MODEL_DOCSTRING = """
 
     >>> input_values = processor(
     ...     ds["speech"][0], sampling_rate=16_000, return_tensors="np"
-    >>> ).input_values  # Batch size 1
+    ... ).input_values  # Batch size 1
     >>> hidden_states = model(input_values).last_hidden_state
     ```
 """
@@ -1200,7 +1200,7 @@ FLAX_WAV2VEC2_FOR_CTC_DOCSTRING = """
 
     >>> input_values = processor(
     ...     ds["speech"][0], sampling_rate=16_000, return_tensors="np"
-    >>> ).input_values  # Batch size 1
+    ... ).input_values  # Batch size 1
     >>> logits = model(input_values).logits
     >>> predicted_ids = jnp.argmax(logits, axis=-1)
 

src/transformers/models/xlm/modeling_xlm.py

@@ -1039,7 +1039,7 @@ class XLMForQuestionAnswering(XLMPreTrainedModel):
 
         >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute", add_special_tokens=True)).unsqueeze(
         ...     0
-        >>> )  # Batch size 1
+        ... )  # Batch size 1
         >>> start_positions = torch.tensor([1])
         >>> end_positions = torch.tensor([3])
 

src/transformers/models/xlm_prophetnet/modeling_xlm_prophetnet.py

@@ -98,7 +98,7 @@ class XLMProphetNetModel(ProphetNetModel):
 
     >>> input_ids = tokenizer(
     ...     "Studies have been shown that owning a dog is good for you", return_tensors="pt"
-    >>> ).input_ids  # Batch size 1
+    ... ).input_ids  # Batch size 1
     >>> decoder_input_ids = tokenizer("Studies show that", return_tensors="pt").input_ids  # Batch size 1
     >>> outputs = model(input_ids=input_ids, decoder_input_ids=decoder_input_ids)
 
@@ -124,7 +124,7 @@ class XLMProphetNetForConditionalGeneration(ProphetNetForConditionalGeneration):
 
     >>> input_ids = tokenizer(
     ...     "Studies have been shown that owning a dog is good for you", return_tensors="pt"
-    >>> ).input_ids  # Batch size 1
+    ... ).input_ids  # Batch size 1
     >>> decoder_input_ids = tokenizer("Studies show that", return_tensors="pt").input_ids  # Batch size 1
     >>> outputs = model(input_ids=input_ids, decoder_input_ids=decoder_input_ids)
 

src/transformers/models/xlnet/modeling_tf_xlnet.py

@@ -1281,17 +1281,17 @@ class TFXLNetLMHeadModel(TFXLNetPreTrainedModel, TFCausalLanguageModelingLoss):
         >>> # We show how to setup inputs to predict a next token using a bi-directional context.
         >>> input_ids = tf.constant(tokenizer.encode("Hello, my dog is very <mask>", add_special_tokens=True))[
         ...     None, :
-        >>> ]  # We will predict the masked token
+        ... ]  # We will predict the masked token
 
         >>> perm_mask = np.zeros((1, input_ids.shape[1], input_ids.shape[1]))
         >>> perm_mask[:, :, -1] = 1.0  # Previous tokens don't see last token
 
         >>> target_mapping = np.zeros(
         ...     (1, 1, input_ids.shape[1])
-        >>> )  # Shape [1, 1, seq_length] => let's predict one token
+        ... )  # Shape [1, 1, seq_length] => let's predict one token
         >>> target_mapping[
         ...     0, 0, -1
-        >>> ] = 1.0  # Our first (and only) prediction will be the last token of the sequence (the masked token)
+        ... ] = 1.0  # Our first (and only) prediction will be the last token of the sequence (the masked token)
 
         >>> outputs = model(
         ...     input_ids,
@@ -1301,7 +1301,7 @@ class TFXLNetLMHeadModel(TFXLNetPreTrainedModel, TFCausalLanguageModelingLoss):
 
         >>> next_token_logits = outputs[
         ...     0
-        >>> ]  # Output has shape [target_mapping.size(0), target_mapping.size(1), config.vocab_size]
+        ... ]  # Output has shape [target_mapping.size(0), target_mapping.size(1), config.vocab_size]
         ```"""
         transformer_outputs = self.transformer(
             input_ids=input_ids,

src/transformers/models/xlnet/modeling_xlnet.py

@@ -1400,47 +1400,47 @@ class XLNetLMHeadModel(XLNetPreTrainedModel):
         >>> # We show how to setup inputs to predict a next token using a bi-directional context.
         >>> input_ids = torch.tensor(
         ...     tokenizer.encode("Hello, my dog is very <mask>", add_special_tokens=False)
-        >>> ).unsqueeze(
+        ... ).unsqueeze(
         ...     0
-        >>> )  # We will predict the masked token
+        ... )  # We will predict the masked token
         >>> perm_mask = torch.zeros((1, input_ids.shape[1], input_ids.shape[1]), dtype=torch.float)
         >>> perm_mask[:, :, -1] = 1.0  # Previous tokens don't see last token
         >>> target_mapping = torch.zeros(
         ...     (1, 1, input_ids.shape[1]), dtype=torch.float
-        >>> )  # Shape [1, 1, seq_length] => let's predict one token
+        ... )  # Shape [1, 1, seq_length] => let's predict one token
         >>> target_mapping[
         ...     0, 0, -1
-        >>> ] = 1.0  # Our first (and only) prediction will be the last token of the sequence (the masked token)
+        ... ] = 1.0  # Our first (and only) prediction will be the last token of the sequence (the masked token)
 
         >>> outputs = model(input_ids, perm_mask=perm_mask, target_mapping=target_mapping)
         >>> next_token_logits = outputs[
         ...     0
-        >>> ]  # Output has shape [target_mapping.size(0), target_mapping.size(1), config.vocab_size]
+        ... ]  # Output has shape [target_mapping.size(0), target_mapping.size(1), config.vocab_size]
 
         >>> # The same way can the XLNetLMHeadModel be used to be trained by standard auto-regressive language modeling.
         >>> input_ids = torch.tensor(
         ...     tokenizer.encode("Hello, my dog is very <mask>", add_special_tokens=False)
-        >>> ).unsqueeze(
+        ... ).unsqueeze(
         ...     0
-        >>> )  # We will predict the masked token
+        ... )  # We will predict the masked token
         >>> labels = torch.tensor(tokenizer.encode("cute", add_special_tokens=False)).unsqueeze(0)
         >>> assert labels.shape[0] == 1, "only one word will be predicted"
         >>> perm_mask = torch.zeros((1, input_ids.shape[1], input_ids.shape[1]), dtype=torch.float)
         >>> perm_mask[
         ...     :, :, -1
-        >>> ] = 1.0  # Previous tokens don't see last token as is done in standard auto-regressive lm training
+        ... ] = 1.0  # Previous tokens don't see last token as is done in standard auto-regressive lm training
         >>> target_mapping = torch.zeros(
         ...     (1, 1, input_ids.shape[1]), dtype=torch.float
-        >>> )  # Shape [1, 1, seq_length] => let's predict one token
+        ... )  # Shape [1, 1, seq_length] => let's predict one token
         >>> target_mapping[
         ...     0, 0, -1
-        >>> ] = 1.0  # Our first (and only) prediction will be the last token of the sequence (the masked token)
+        ... ] = 1.0  # Our first (and only) prediction will be the last token of the sequence (the masked token)
 
         >>> outputs = model(input_ids, perm_mask=perm_mask, target_mapping=target_mapping, labels=labels)
         >>> loss = outputs.loss
         >>> next_token_logits = (
         ...     outputs.logits
-        >>> )  # Logits have shape [target_mapping.size(0), target_mapping.size(1), config.vocab_size]
+        ... )  # Logits have shape [target_mapping.size(0), target_mapping.size(1), config.vocab_size]
         ```"""
         return_dict = return_dict if return_dict is not None else self.config.use_return_dict
 
@@ -1980,7 +1980,7 @@ class XLNetForQuestionAnswering(XLNetPreTrainedModel):
 
         >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute", add_special_tokens=True)).unsqueeze(
         ...     0
-        >>> )  # Batch size 1
+        ... )  # Batch size 1
         >>> start_positions = torch.tensor([1])
         >>> end_positions = torch.tensor([3])
         >>> outputs = model(input_ids, start_positions=start_positions, end_positions=end_positions)