[doc] :obj: hunt (#14954)

* redo sans examples

* style

docs/source/testing.mdx

@@ -738,7 +738,7 @@ leave any data in there.
 <Tip>
 
 In order to run the equivalent of `rm -r` safely, only subdirs of the project repository checkout are allowed if
-an explicit obj:*tmp_dir* is used, so that by mistake no `/tmp` or similar important part of the filesystem will
+an explicit `tmp_dir` is used, so that by mistake no `/tmp` or similar important part of the filesystem will
 get nuked. i.e. please always pass paths that start with `./`.
 
 </Tip>

src/transformers/generation_utils.py

@@ -1320,7 +1320,7 @@ class GenerationMixin:
 
         Return:
             [`~generation_utils.GreedySearchDecoderOnlyOutput`], [`~generation_utils.GreedySearchEncoderDecoderOutput`]
-            or obj:*torch.LongTensor*: A `torch.LongTensor` containing the generated tokens (default behaviour) or a
+            or `torch.LongTensor`: A `torch.LongTensor` containing the generated tokens (default behaviour) or a
             [`~generation_utils.GreedySearchDecoderOnlyOutput`] if `model.config.is_encoder_decoder=False` and
             `return_dict_in_generate=True` or a [`~generation_utils.GreedySearchEncoderDecoderOutput`] if
             `model.config.is_encoder_decoder=True`.
@@ -1547,7 +1547,7 @@ class GenerationMixin:
 
         Return:
             [`~generation_utils.SampleDecoderOnlyOutput`], [`~generation_utils.SampleEncoderDecoderOutput`] or
-            obj:*torch.LongTensor*: A `torch.LongTensor` containing the generated tokens (default behaviour) or a
+            `torch.LongTensor`: A `torch.LongTensor` containing the generated tokens (default behaviour) or a
             [`~generation_utils.SampleDecoderOnlyOutput`] if `model.config.is_encoder_decoder=False` and
             `return_dict_in_generate=True` or a [`~generation_utils.SampleEncoderDecoderOutput`] if
             `model.config.is_encoder_decoder=True`.
@@ -1785,7 +1785,7 @@ class GenerationMixin:
 
         Return:
             [`generation_utilsBeamSearchDecoderOnlyOutput`], [`~generation_utils.BeamSearchEncoderDecoderOutput`] or
-            obj:*torch.LongTensor*: A `torch.LongTensor` containing the generated tokens (default behaviour) or a
+            `torch.LongTensor`: A `torch.LongTensor` containing the generated tokens (default behaviour) or a
             [`~generation_utils.BeamSearchDecoderOnlyOutput`] if `model.config.is_encoder_decoder=False` and
             `return_dict_in_generate=True` or a [`~generation_utils.BeamSearchEncoderDecoderOutput`] if
             `model.config.is_encoder_decoder=True`.
@@ -2079,7 +2079,7 @@ class GenerationMixin:
 
         Return:
             [`~generation_utils.BeamSampleDecoderOnlyOutput`], [`~generation_utils.BeamSampleEncoderDecoderOutput`] or
-            obj:*torch.LongTensor*: A `torch.LongTensor` containing the generated tokens (default behaviour) or a
+            `torch.LongTensor`: A `torch.LongTensor` containing the generated tokens (default behaviour) or a
             [`~generation_utils.BeamSampleDecoderOnlyOutput`] if `model.config.is_encoder_decoder=False` and
             `return_dict_in_generate=True` or a [`~generation_utils.BeamSampleEncoderDecoderOutput`] if
             `model.config.is_encoder_decoder=True`.
@@ -2375,7 +2375,7 @@ class GenerationMixin:
 
         Return:
             [`~generation_utils.BeamSearchDecoderOnlyOutput`], [`~generation_utils.BeamSearchEncoderDecoderOutput`] or
-            obj:*torch.LongTensor*: A `torch.LongTensor` containing the generated tokens (default behaviour) or a
+            `torch.LongTensor`: A `torch.LongTensor` containing the generated tokens (default behaviour) or a
             [`~generation_utils.BeamSearchDecoderOnlyOutput`] if [`~generation_utils.BeamSearchDecoderOnlyOutput`] if
             `model.config.is_encoder_decoder=False` and `return_dict_in_generate=True` or a
             [`~generation_utils.BeamSearchEncoderDecoderOutput`] if `model.config.is_encoder_decoder=True`.

src/transformers/modeling_utils.py

@@ -1840,8 +1840,8 @@ class PoolerEndLogits(nn.Module):
 
         <Tip>
 
-        One of `start_states` or `start_positions` should be not obj:`None`. If both are set, `start_positions`
-        overrides `start_states`.
+        One of `start_states` or `start_positions` should be not `None`. If both are set, `start_positions` overrides
+        `start_states`.
 
         </Tip>
 
@@ -1906,8 +1906,8 @@ class PoolerAnswerClass(nn.Module):
 
         <Tip>
 
-        One of `start_states` or `start_positions` should be not obj:`None`. If both are set, `start_positions`
-        overrides `start_states`.
+        One of `start_states` or `start_positions` should be not `None`. If both are set, `start_positions` overrides
+        `start_states`.
 
         </Tip>
 

src/transformers/models/encoder_decoder/modeling_encoder_decoder.py

@@ -293,7 +293,7 @@ class EncoderDecoderModel(PreTrainedModel):
         the model, you need to first set it back in training mode with `model.train()`.
 
         Params:
-            encoder_pretrained_model_name_or_path (:obj: *str*, *optional*):
+            encoder_pretrained_model_name_or_path (`str`, *optional*):
                 Information necessary to initiate the encoder. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co.
@@ -306,7 +306,7 @@ class EncoderDecoderModel(PreTrainedModel):
                       `config` argument. This loading path is slower than converting the TensorFlow checkpoint in a
                       PyTorch model using the provided conversion scripts and loading the PyTorch model afterwards.
 
-            decoder_pretrained_model_name_or_path (:obj: *str*, *optional*, defaults to `None`):
+            decoder_pretrained_model_name_or_path (`str`, *optional*, defaults to `None`):
                 Information necessary to initiate the decoder. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co.

src/transformers/models/encoder_decoder/modeling_flax_encoder_decoder.py

@@ -746,7 +746,7 @@ class FlaxEncoderDecoderModel(FlaxPreTrainedModel):
         checkpoints.
 
         Params:
-            encoder_pretrained_model_name_or_path (:obj: *Union[str, os.PathLike]*, *optional*):
+            encoder_pretrained_model_name_or_path (`Union[str, os.PathLike]`, *optional*):
                 Information necessary to initiate the encoder. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co.
@@ -755,7 +755,7 @@ class FlaxEncoderDecoderModel(FlaxPreTrainedModel):
                     - A path to a *directory* containing model weights saved using
                       [`~FlaxPreTrainedModel.save_pretrained`], e.g., `./my_model_directory/`.
 
-            decoder_pretrained_model_name_or_path (:obj: *Union[str, os.PathLike]*, *optional*, defaults to `None`):
+            decoder_pretrained_model_name_or_path (`Union[str, os.PathLike]`, *optional*, defaults to `None`):
                 Information necessary to initiate the decoder. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co.

src/transformers/models/encoder_decoder/modeling_tf_encoder_decoder.py

@@ -308,7 +308,7 @@ class TFEncoderDecoderModel(TFPreTrainedModel):
 
 
         Params:
-            encoder_pretrained_model_name_or_path (:obj: *str*, *optional*):
+            encoder_pretrained_model_name_or_path (`str`, *optional*):
                 Information necessary to initiate the encoder. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co.
@@ -319,7 +319,7 @@ class TFEncoderDecoderModel(TFPreTrainedModel):
                     - A path or url to a *pytorch index checkpoint file* (e.g, `./pt_model/`). In this case,
                       `encoder_from_pt` should be set to `True`.
 
-            decoder_pretrained_model_name_or_path (:obj: *str*, *optional*, defaults to `None`):
+            decoder_pretrained_model_name_or_path (`str`, *optional*, defaults to `None`):
                 Information necessary to initiate the decoder. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co.

src/transformers/models/ibert/quant_modules.py

@@ -713,7 +713,7 @@ def batch_frexp(inputs, max_bit=31):
             Target scaling factor to decompose.
 
     Returns:
-        :obj:``Tuple(torch.Tensor, torch.Tensor)`: mantisa and exponent
+        ``Tuple(torch.Tensor, torch.Tensor)`: mantisa and exponent
     """
 
     shape_of_input = inputs.size()

src/transformers/models/layoutlm/modeling_layoutlm.py

@@ -108,7 +108,7 @@ class LayoutLMEmbeddings(nn.Module):
             right_position_embeddings = self.x_position_embeddings(bbox[:, :, 2])
             lower_position_embeddings = self.y_position_embeddings(bbox[:, :, 3])
         except IndexError as e:
-            raise IndexError("The :obj:`bbox`coordinate values should be within 0-1000 range.") from e
+            raise IndexError("The `bbox`coordinate values should be within 0-1000 range.") from e
 
         h_position_embeddings = self.h_position_embeddings(bbox[:, :, 3] - bbox[:, :, 1])
         w_position_embeddings = self.w_position_embeddings(bbox[:, :, 2] - bbox[:, :, 0])

src/transformers/models/layoutlm/modeling_tf_layoutlm.py

@@ -162,7 +162,7 @@ class TFLayoutLMEmbeddings(tf.keras.layers.Layer):
             right_position_embeddings = tf.gather(self.x_position_embeddings, bbox[:, :, 2])
             lower_position_embeddings = tf.gather(self.y_position_embeddings, bbox[:, :, 3])
         except IndexError as e:
-            raise IndexError("The :obj:`bbox`coordinate values should be within 0-1000 range.") from e
+            raise IndexError("The `bbox`coordinate values should be within 0-1000 range.") from e
         h_position_embeddings = tf.gather(self.h_position_embeddings, bbox[:, :, 3] - bbox[:, :, 1])
         w_position_embeddings = tf.gather(self.w_position_embeddings, bbox[:, :, 2] - bbox[:, :, 0])
 

src/transformers/models/layoutlmv2/modeling_layoutlmv2.py

@@ -86,7 +86,7 @@ class LayoutLMv2Embeddings(nn.Module):
             right_position_embeddings = self.x_position_embeddings(bbox[:, :, 2])
             lower_position_embeddings = self.y_position_embeddings(bbox[:, :, 3])
         except IndexError as e:
-            raise IndexError("The :obj:`bbox` coordinate values should be within 0-1000 range.") from e
+            raise IndexError("The `bbox` coordinate values should be within 0-1000 range.") from e
 
         h_position_embeddings = self.h_position_embeddings(bbox[:, :, 3] - bbox[:, :, 1])
         w_position_embeddings = self.w_position_embeddings(bbox[:, :, 2] - bbox[:, :, 0])

src/transformers/models/lxmert/modeling_tf_lxmert.py

@@ -1324,7 +1324,7 @@ class TFLxmertForPreTraining(TFLxmertPreTrainedModel):
             Labels for computing the masked language modeling loss. Indices should be in `[-100, 0, ...,
             config.vocab_size]` (see `input_ids` docstring) Tokens with indices set to `-100` are ignored (masked), the
             loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`
-        obj_labels: (`Dict[Str: Tuple[tf.Tensor, tf.Tensor]]`, *optional*, defaults to :obj: `None`):
+        obj_labels: (`Dict[Str: Tuple[tf.Tensor, tf.Tensor]]`, *optional*, defaults to `None`):
             each key is named after each one of the visual losses and each element of the tuple is of the shape
             `(batch_size, num_features)` and `(batch_size, num_features, visual_feature_dim)` for each the label id and
             the label score respectively
@@ -1334,7 +1334,7 @@ class TFLxmertForPreTraining(TFLxmertPreTrainedModel):
 
             - 0 indicates that the sentence does not match the image,
             - 1 indicates that the sentence does match the image.
-        ans (`Torch.Tensor` of shape `(batch_size)`, *optional*, defaults to :obj: `None`):
+        ans (`Torch.Tensor` of shape `(batch_size)`, *optional*, defaults to `None`):
             a one hot representation hof the correct answer *optional*
 
         Returns:

src/transformers/models/rag/modeling_rag.py

@@ -258,7 +258,7 @@ class RagPreTrainedModel(PreTrainedModel):
         the model, you need to first set it back in training mode with `model.train()`.
 
         Params:
-            question_encoder_pretrained_model_name_or_path (:obj: *str*, *optional*, defaults to `None`):
+            question_encoder_pretrained_model_name_or_path (`str`, *optional*, defaults to `None`):
                 Information necessary to initiate the question encoder. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co.
@@ -271,7 +271,7 @@ class RagPreTrainedModel(PreTrainedModel):
                       `config` argument. This loading path is slower than converting the TensorFlow checkpoint in a
                       PyTorch model using the provided conversion scripts and loading the PyTorch model afterwards.
 
-            generator_pretrained_model_name_or_path (:obj: *str*, *optional*, defaults to `None`):
+            generator_pretrained_model_name_or_path (`str`, *optional*, defaults to `None`):
                 Information necessary to initiate the generator. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co.

src/transformers/models/rag/modeling_tf_rag.py

@@ -233,7 +233,7 @@ class TFRagPreTrainedModel(TFPreTrainedModel):
         model checkpoints.
 
         Params:
-            question_encoder_pretrained_model_name_or_path (:obj: *str*, *optional*):
+            question_encoder_pretrained_model_name_or_path (`str`, *optional*):
                 Information necessary to initiate the question encoder. Can be either:
 
                     - A string with the *shortcut name* of a pretrained model to load from cache or download, e.g.,
@@ -245,7 +245,7 @@ class TFRagPreTrainedModel(TFPreTrainedModel):
                     - A path or url to a *pytorch index checkpoint file* (e.g, `./pt_model/`). In this case,
                       `question_encoder_from_pt` should be set to `True`.
 
-            generator_pretrained_model_name_or_path (:obj: *str*, *optional*, defaults to `None`):
+            generator_pretrained_model_name_or_path (`str`, *optional*, defaults to `None`):
                 Information necessary to initiate the generator. Can be either:
 
                     - A string with the *shortcut name* of a pretrained model to load from cache or download, e.g.,

src/transformers/models/speech_encoder_decoder/modeling_speech_encoder_decoder.py

@@ -287,7 +287,7 @@ class SpeechEncoderDecoderModel(PreTrainedModel):
         the model, you need to first set it back in training mode with `model.train()`.
 
         Params:
-            encoder_pretrained_model_name_or_path (:obj: *str*, *optional*):
+            encoder_pretrained_model_name_or_path (`str`, *optional*):
                 Information necessary to initiate the encoder. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co.
@@ -300,7 +300,7 @@ class SpeechEncoderDecoderModel(PreTrainedModel):
                       `config` argument. This loading path is slower than converting the TensorFlow checkpoint in a
                       PyTorch model using the provided conversion scripts and loading the PyTorch model afterwards.
 
-            decoder_pretrained_model_name_or_path (:obj: *str*, *optional*, defaults to `None`):
+            decoder_pretrained_model_name_or_path (`str`, *optional*, defaults to `None`):
                 Information necessary to initiate the decoder. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co.

src/transformers/models/t5/modeling_t5.py

@@ -915,7 +915,7 @@ class T5Stack(T5PreTrainedModel):
         mask_seq_length = past_key_values[0][0].shape[2] + seq_length if past_key_values is not None else seq_length
 
         if use_cache is True:
-            assert self.is_decoder, f":obj:`use_cache` can only be set to `True` if {self} is used as a decoder"
+            assert self.is_decoder, f"`use_cache` can only be set to `True` if {self} is used as a decoder"
 
         if attention_mask is None:
             attention_mask = torch.ones(batch_size, mask_seq_length).to(inputs_embeds.device)

src/transformers/models/tapas/modeling_tapas.py

@@ -2277,7 +2277,7 @@ def _calculate_expected_result(
             Numeric values of every token. Nan for tokens which are not numeric values.
         numeric_values_scale (`torch.FloatTensor` of shape `(batch_size, seq_length)`):
             Scale of the numeric values of every token.
-        input_mask_float (:obj: *torch.FloatTensor* of shape `(batch_size, seq_length)`):
+        input_mask_float (`torch.FloatTensor` of shape `(batch_size, seq_length)`):
             Mask for the table, without question tokens and table headers.
         logits_aggregation (`torch.FloatTensor` of shape `(batch_size, num_aggregation_labels)`):
             Logits per aggregation operation.
@@ -2371,9 +2371,9 @@ def _calculate_regression_loss(
     Calculates the regression loss per example.
 
     Args:
-        answer (:obj: *torch.FloatTensor* of shape `(batch_size,)`):
+        answer (`torch.FloatTensor` of shape `(batch_size,)`):
             Answer for every example in the batch. Nan if there is no scalar answer.
-        aggregate_mask (:obj: *torch.FloatTensor* of shape `(batch_size,)`):
+        aggregate_mask (`torch.FloatTensor` of shape `(batch_size,)`):
             A mask set to 1 for examples that should use aggregation functions.
         dist_per_cell (`torch.distributions.Bernoulli`):
             Cell selection distribution for each cell.
@@ -2381,9 +2381,9 @@ def _calculate_regression_loss(
             Numeric values of every token. Nan for tokens which are not numeric values.
         numeric_values_scale (`torch.FloatTensor` of shape `(batch_size, seq_length)`):
             Scale of the numeric values of every token.
-        input_mask_float (:obj: *torch.FloatTensor* of shape `(batch_size, seq_length)`):
+        input_mask_float (`torch.FloatTensor` of shape `(batch_size, seq_length)`):
             Mask for the table, without question tokens and table headers.
-        logits_aggregation (:obj: *torch.FloatTensor* of shape `(batch_size, num_aggregation_labels)`):
+        logits_aggregation (`torch.FloatTensor` of shape `(batch_size, num_aggregation_labels)`):
             Logits per aggregation operation.
         config ([`TapasConfig`]):
             Model configuration class with all the parameters of the model

src/transformers/models/tapas/modeling_tf_tapas.py

@@ -2241,7 +2241,7 @@ def _calculate_expected_result(
             Numeric values of every token. Nan for tokens which are not numeric values.
         numeric_values_scale (`tf.Tensor` of shape `(batch_size, seq_length)`):
             Scale of the numeric values of every token.
-        input_mask_float (:obj: *tf.Tensor* of shape `(batch_size, seq_length)`):
+        input_mask_float (`tf.Tensor` of shape `(batch_size, seq_length)`):
             Mask for the table, without question tokens and table headers.
         logits_aggregation (`tf.Tensor` of shape `(batch_size, num_aggregation_labels)`):
             Logits per aggregation operation.
@@ -2321,9 +2321,9 @@ def _calculate_regression_loss(
     Calculates the regression loss per example.
 
     Args:
-        answer (:obj: *tf.Tensor* of shape `(batch_size,)`):
+        answer (`tf.Tensor` of shape `(batch_size,)`):
             Answer for every example in the batch. Nan if there is no scalar answer.
-        aggregate_mask (:obj: *tf.Tensor* of shape `(batch_size,)`):
+        aggregate_mask (`tf.Tensor` of shape `(batch_size,)`):
             A mask set to 1 for examples that should use aggregation functions.
         dist_per_cell (`torch.distributions.Bernoulli`):
             Cell selection distribution for each cell.
@@ -2331,9 +2331,9 @@ def _calculate_regression_loss(
             Numeric values of every token. Nan for tokens which are not numeric values.
         numeric_values_scale (`tf.Tensor` of shape `(batch_size, seq_length)`):
             Scale of the numeric values of every token.
-        input_mask_float (:obj: *tf.Tensor* of shape `(batch_size, seq_length)`):
+        input_mask_float (`tf.Tensor` of shape `(batch_size, seq_length)`):
             Mask for the table, without question tokens and table headers.
-        logits_aggregation (:obj: *tf.Tensor* of shape `(batch_size, num_aggregation_labels)`):
+        logits_aggregation (`tf.Tensor` of shape `(batch_size, num_aggregation_labels)`):
             Logits per aggregation operation.
         config ([`TapasConfig`]):
             Model configuration class with all the parameters of the model

src/transformers/models/unispeech/configuration_unispeech.py

@@ -73,7 +73,7 @@ class UniSpeechConfig(PretrainedConfig):
         feat_extract_activation (`str, `optional`, defaults to `"gelu"`):
             The non-linear activation function (function or string) in the 1D convolutional layers of the feature
             extractor. If string, `"gelu"`, `"relu"`, `"selu"` and `"gelu_new"` are supported.
-        feat_quantizer_dropout (obj:*float*, *optional*, defaults to 0.0):
+        feat_quantizer_dropout (`float`, *optional*, defaults to 0.0):
             The dropout probabilitiy for quantized feature extractor states.
         conv_dim (`Tuple[int]`, *optional*, defaults to `(512, 512, 512, 512, 512, 512, 512)`):
             A tuple of integers defining the number of input and output channels of each 1D convolutional layer in the

src/transformers/models/unispeech_sat/configuration_unispeech_sat.py

@@ -73,7 +73,7 @@ class UniSpeechSatConfig(PretrainedConfig):
         feat_extract_activation (`str, `optional`, defaults to `"gelu"`):
             The non-linear activation function (function or string) in the 1D convolutional layers of the feature
             extractor. If string, `"gelu"`, `"relu"`, `"selu"` and `"gelu_new"` are supported.
-        feat_quantizer_dropout (obj:*float*, *optional*, defaults to 0.0):
+        feat_quantizer_dropout (`float`, *optional*, defaults to 0.0):
             The dropout probabilitiy for quantized feature extractor states.
         conv_dim (`Tuple[int]`, *optional*, defaults to `(512, 512, 512, 512, 512, 512, 512)`):
             A tuple of integers defining the number of input and output channels of each 1D convolutional layer in the

src/transformers/models/vision_encoder_decoder/modeling_flax_vision_encoder_decoder.py

@@ -712,7 +712,7 @@ class FlaxVisionEncoderDecoderModel(FlaxPreTrainedModel):
         checkpoints.
 
         Params:
-            encoder_pretrained_model_name_or_path (:obj: *Union[str, os.PathLike]*, *optional*):
+            encoder_pretrained_model_name_or_path (`Union[str, os.PathLike]`, *optional*):
                 Information necessary to initiate the encoder. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co. An
@@ -720,7 +720,7 @@ class FlaxVisionEncoderDecoderModel(FlaxPreTrainedModel):
                     - A path to a *directory* containing model weights saved using
                       [`~FlaxPreTrainedModel.save_pretrained`], e.g., `./my_model_directory/`.
 
-            decoder_pretrained_model_name_or_path (:obj: *Union[str, os.PathLike]*, *optional*, defaults to `None`):
+            decoder_pretrained_model_name_or_path (`Union[str, os.PathLike]`, *optional*, defaults to `None`):
                 Information necessary to initiate the decoder. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co.

src/transformers/models/vision_encoder_decoder/modeling_vision_encoder_decoder.py

@@ -260,7 +260,7 @@ class VisionEncoderDecoderModel(PreTrainedModel):
         the model, you need to first set it back in training mode with `model.train()`.
 
         Params:
-            encoder_pretrained_model_name_or_path (:obj: *str*, *optional*):
+            encoder_pretrained_model_name_or_path (`str`, *optional*):
                 Information necessary to initiate the image encoder. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co. An
@@ -272,7 +272,7 @@ class VisionEncoderDecoderModel(PreTrainedModel):
                       `config` argument. This loading path is slower than converting the TensorFlow checkpoint in a
                       PyTorch model using the provided conversion scripts and loading the PyTorch model afterwards.
 
-            decoder_pretrained_model_name_or_path (:obj: *str*, *optional*, defaults to `None`):
+            decoder_pretrained_model_name_or_path (`str`, *optional*, defaults to `None`):
                 Information necessary to initiate the text decoder. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co.

src/transformers/models/vision_text_dual_encoder/modeling_flax_vision_text_dual_encoder.py

@@ -403,7 +403,7 @@ class FlaxVisionTextDualEncoderModel(FlaxPreTrainedModel):
     ) -> FlaxPreTrainedModel:
         """
         Params:
-            vision_model_name_or_path (:obj: *str*, *optional*, defaults to `None`):
+            vision_model_name_or_path (`str`, *optional*, defaults to `None`):
                 Information necessary to initiate the vision model. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co.
@@ -416,7 +416,7 @@ class FlaxVisionTextDualEncoderModel(FlaxPreTrainedModel):
                       loading path is slower than converting the PyTorch checkpoint in a Flax model using the provided
                       conversion scripts and loading the Flax model afterwards.
 
-            text_model_name_or_path (:obj: *str*, *optional*):
+            text_model_name_or_path (`str`, *optional*):
                 Information necessary to initiate the text model. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co.

src/transformers/models/vision_text_dual_encoder/modeling_vision_text_dual_encoder.py

@@ -404,7 +404,7 @@ class VisionTextDualEncoderModel(PreTrainedModel):
     ) -> PreTrainedModel:
         """
         Params:
-            vision_model_name_or_path (:obj: *str*, *optional*, defaults to `None`):
+            vision_model_name_or_path (`str`, *optional*, defaults to `None`):
                 Information necessary to initiate the vision model. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co.
@@ -417,7 +417,7 @@ class VisionTextDualEncoderModel(PreTrainedModel):
                       loading path is slower than converting the PyTorch checkpoint in a Flax model using the provided
                       conversion scripts and loading the Flax model afterwards.
 
-            text_model_name_or_path (:obj: *str*, *optional*):
+            text_model_name_or_path (`str`, *optional*):
                 Information necessary to initiate the text model. Can be either:
 
                     - A string, the *model id* of a pretrained model hosted inside a model repo on huggingface.co.

src/transformers/models/wav2vec2/configuration_wav2vec2.py

@@ -73,7 +73,7 @@ class Wav2Vec2Config(PretrainedConfig):
         feat_extract_activation (`str, `optional`, defaults to `"gelu"`):
             The non-linear activation function (function or string) in the 1D convolutional layers of the feature
             extractor. If string, `"gelu"`, `"relu"`, `"selu"` and `"gelu_new"` are supported.
-        feat_quantizer_dropout (obj:*float*, *optional*, defaults to 0.0):
+        feat_quantizer_dropout (`float`, *optional*, defaults to 0.0):
             The dropout probabilitiy for quantized feature extractor states.
         conv_dim (`Tuple[int]`, *optional*, defaults to `(512, 512, 512, 512, 512, 512, 512)`):
             A tuple of integers defining the number of input and output channels of each 1D convolutional layer in the

src/transformers/models/wavlm/configuration_wavlm.py

@@ -72,7 +72,7 @@ class WavLMConfig(PretrainedConfig):
         feat_extract_activation (`str, `optional`, defaults to `"gelu"`):
             The non-linear activation function (function or string) in the 1D convolutional layers of the feature
             extractor. If string, `"gelu"`, `"relu"`, `"selu"` and `"gelu_new"` are supported.
-        feat_quantizer_dropout (obj:*float*, *optional*, defaults to 0.0):
+        feat_quantizer_dropout (`float`, *optional*, defaults to 0.0):
             The dropout probabilitiy for quantized feature extractor states.
         conv_dim (`Tuple[int]`, *optional*, defaults to `(512, 512, 512, 512, 512, 512, 512)`):
             A tuple of integers defining the number of input and output channels of each 1D convolutional layer in the

src/transformers/models/xlnet/modeling_tf_xlnet.py

@@ -512,15 +512,15 @@ class TFXLNetMainLayer(tf.keras.layers.Layer):
             curr_out = curr_out[: self.reuse_len]
 
         if self.mem_len is None or self.mem_len == 0:
-            # If :obj:`use_mems` is active but no `mem_len` is defined, the model behaves like GPT-2 at inference time
+            # If `use_mems` is active but no `mem_len` is defined, the model behaves like GPT-2 at inference time
             # and returns all of the past and current hidden states.
             cutoff = 0
         else:
-            # If :obj:`use_mems` is active and `mem_len` is defined, the model returns the last `mem_len` hidden
+            # If `use_mems` is active and `mem_len` is defined, the model returns the last `mem_len` hidden
             # states. This is the preferred setting for training and long-form generation.
             cutoff = -self.mem_len
         if prev_mem is None:
-            # if :obj:`use_mems` is active and `mem_len` is defined, the model
+            # if `use_mems` is active and `mem_len` is defined, the model
             new_mem = curr_out[cutoff:]
         else:
             new_mem = tf.concat([prev_mem, curr_out], 0)[cutoff:]

src/transformers/models/xlnet/modeling_xlnet.py

@@ -1000,15 +1000,15 @@ class XLNetModel(XLNetPreTrainedModel):
             curr_out = curr_out[: self.reuse_len]
 
         if self.mem_len is None or self.mem_len == 0:
-            # If :obj:`use_mems` is active but no `mem_len` is defined, the model behaves like GPT-2 at inference time
+            # If `use_mems` is active but no `mem_len` is defined, the model behaves like GPT-2 at inference time
             # and returns all of the past and current hidden states.
             cutoff = 0
         else:
-            # If :obj:`use_mems` is active and `mem_len` is defined, the model returns the last `mem_len` hidden
+            # If `use_mems` is active and `mem_len` is defined, the model returns the last `mem_len` hidden
             # states. This is the preferred setting for training and long-form generation.
             cutoff = -self.mem_len
         if prev_mem is None:
-            # if :obj:`use_mems` is active and `mem_len` is defined, the model
+            # if `use_mems` is active and `mem_len` is defined, the model
             new_mem = curr_out[cutoff:]
         else:
             new_mem = torch.cat([prev_mem, curr_out], dim=0)[cutoff:]

src/transformers/trainer.py

@@ -2466,7 +2466,7 @@ class Trainer:
         ignore_keys: Optional[List[str]] = None,
     ) -> Tuple[Optional[torch.Tensor], Optional[torch.Tensor], Optional[torch.Tensor]]:
         """
-        Perform an evaluation step on `model` using obj:*inputs*.
+        Perform an evaluation step on `model` using `inputs`.
 
         Subclass and override to inject custom behavior.
 

src/transformers/trainer_pt_utils.py

@@ -226,8 +226,8 @@ def torch_distributed_zero_first(local_rank: int):
 
 class DistributedSamplerWithLoop(DistributedSampler):
     """
-    Like a :obj:torch.utils.data.distributed.DistributedSampler` but loops at the end back to the beginning of the
-    shuffled samples to make each process have a round multiple of batch_size samples.
+    Like a torch.utils.data.distributed.DistributedSampler` but loops at the end back to the beginning of the shuffled
+    samples to make each process have a round multiple of batch_size samples.
 
     Args:
         dataset (`torch.utils.data.Dataset`):

src/transformers/trainer_seq2seq.py

@@ -126,7 +126,7 @@ class Seq2SeqTrainer(Trainer):
         ignore_keys: Optional[List[str]] = None,
     ) -> Tuple[Optional[float], Optional[torch.Tensor], Optional[torch.Tensor]]:
         """
-        Perform an evaluation step on `model` using obj:*inputs*.
+        Perform an evaluation step on `model` using `inputs`.
 
         Subclass and override to inject custom behavior.
 

src/transformers/training_args.py

@@ -175,8 +175,8 @@ class TrainingArguments:
         logging_steps (`int`, *optional*, defaults to 500):
             Number of update steps between two logs if `logging_strategy="steps"`.
         logging_nan_inf_filter (`bool`, *optional*, defaults to `True`):
-            Whether to filter `nan` and `inf` losses for logging. If set to obj:`True` the loss of every step that is
-            `nan` or `inf` is filtered and the average loss of the current logging window is taken instead.
+            Whether to filter `nan` and `inf` losses for logging. If set to `True` the loss of every step that is `nan`
+            or `inf` is filtered and the average loss of the current logging window is taken instead.
 
             <Tip>
 

tests/test_doc_samples.py

@@ -45,11 +45,11 @@ class TestCodeExamples(unittest.TestCase):
         the doctests in those files
 
         Args:
-            directory (:obj:`Path`): Directory containing the files
-            identifier (:obj:`str`): Will parse files containing this
-            ignore_files (:obj:`List[str]`): List of files to skip
-            n_identifier (:obj:`str` or :obj:`List[str]`): Will not parse files containing this/these identifiers.
-            only_modules (:obj:`bool`): Whether to only analyze modules
+            directory (`Path`): Directory containing the files
+            identifier (`str`): Will parse files containing this
+            ignore_files (`List[str]`): List of files to skip
+            n_identifier (`str` or `List[str]`): Will not parse files containing this/these identifiers.
+            only_modules (`bool`): Whether to only analyze modules
         """
         files = [file for file in os.listdir(directory) if os.path.isfile(os.path.join(directory, file))]
 

tests/test_modeling_xlnet.py

@@ -556,7 +556,7 @@ class XLNetModelTest(ModelTesterMixin, GenerationTesterMixin, unittest.TestCase)
         self.model_tester.create_and_check_xlnet_base_model(*config_and_inputs)
 
     def test_xlnet_base_model_use_mems(self):
-        # checking that in auto-regressive mode, :obj:`use_mems` gives the same results
+        # checking that in auto-regressive mode, `use_mems` gives the same results
         self.model_tester.set_seed()
         config_and_inputs = self.model_tester.prepare_config_and_inputs()
         self.model_tester.create_and_check_xlnet_model_use_mems(*config_and_inputs)