@@ -110,7 +110,7 @@ New models are constantly released and if you want to implement a new model, ple
...
@@ -110,7 +110,7 @@ New models are constantly released and if you want to implement a new model, ple
If you are willing to contribute the model yourself, let us know so we can help you add it to 🤗 Transformers!
If you are willing to contribute the model yourself, let us know so we can help you add it to 🤗 Transformers!
We have added a [detailed guide and templates](https://github.com/huggingface/transformers/tree/main/templates) to help you get started with adding a new model, and we also have a more technical guide for [how to add a model to 🤗 Transformers](https://huggingface.co/docs/transformers/add_new_model).
We have a technical guide for [how to add a model to 🤗 Transformers](https://huggingface.co/docs/transformers/add_new_model).
@@ -398,12 +398,14 @@ In dem speziellen Fall, dass Sie ein Modell hinzufügen, dessen Architektur gena
...
@@ -398,12 +398,14 @@ In dem speziellen Fall, dass Sie ein Modell hinzufügen, dessen Architektur gena
Modells übereinstimmt, müssen Sie nur ein Konvertierungsskript hinzufügen, wie in [diesem Abschnitt](#write-a-conversion-script) beschrieben.
Modells übereinstimmt, müssen Sie nur ein Konvertierungsskript hinzufügen, wie in [diesem Abschnitt](#write-a-conversion-script) beschrieben.
In diesem Fall können Sie einfach die gesamte Modellarchitektur des bereits vorhandenen Modells wiederverwenden.
In diesem Fall können Sie einfach die gesamte Modellarchitektur des bereits vorhandenen Modells wiederverwenden.
Andernfalls beginnen wir mit der Erstellung eines neuen Modells. Sie haben hier zwei Möglichkeiten:
Andernfalls beginnen wir mit der Erstellung eines neuen Modells. Wir empfehlen die Verwendung des folgenden Skripts, um ein Modell hinzuzufügen
ein bestehendes Modell:
-`transformers-cli add-new-model-like`, um ein neues Modell wie ein bestehendes hinzuzufügen
```bash
-`transformers-cli add-new-model`, um ein neues Modell aus unserer Vorlage hinzuzufügen (sieht dann aus wie BERT oder Bart, je nachdem, welche Art von Modell Sie wählen)
transformers-cli add-new-model-like
```
In beiden Fällen werden Sie mit einem Fragebogen aufgefordert, die grundlegenden Informationen zu Ihrem Modell auszufüllen. Für den zweiten Befehl müssen Sie `cookiecutter` installieren, weitere Informationen dazu finden Sie [hier](https://github.com/huggingface/transformers/tree/main/templates/adding_a_new_model).
Sie werden mit einem Fragebogen aufgefordert, die grundlegenden Informationen Ihres Modells einzugeben.
**Eröffnen Sie einen Pull Request auf dem Haupt-Repositorium huggingface/transformers**
**Eröffnen Sie einen Pull Request auf dem Haupt-Repositorium huggingface/transformers**
@@ -98,7 +98,7 @@ Es werden ständig neue Modelle veröffentlicht. Wenn Sie ein neues Modell imple
...
@@ -98,7 +98,7 @@ Es werden ständig neue Modelle veröffentlicht. Wenn Sie ein neues Modell imple
Lassen Sie es uns wissen, wenn Sie bereit sind, das Modell selbst beizutragen. Dann können wir Ihnen helfen, es zu 🤗 Transformers hinzuzufügen!
Lassen Sie es uns wissen, wenn Sie bereit sind, das Modell selbst beizutragen. Dann können wir Ihnen helfen, es zu 🤗 Transformers hinzuzufügen!
Wir haben eine [detaillierte Anleitung und Vorlagen](https://github.com/huggingface/transformers/tree/main/templates) hinzugefügt, um Ihnen das Hinzufügen eines neuen Modells zu erleichtern, und wir haben auch einen technischen Leitfaden dazu, [wie man ein Modell zu 🤗 Transformers hinzufügt](https://huggingface.co/docs/transformers/add_new_model).
Wir haben auch einen technischen Leitfaden dazu, [wie man ein Modell zu 🤗 Transformers hinzufügt](https://huggingface.co/docs/transformers/add_new_model).
@@ -398,12 +398,14 @@ In the special case that you are adding a model whose architecture exactly match
...
@@ -398,12 +398,14 @@ In the special case that you are adding a model whose architecture exactly match
existing model you only have to add a conversion script as described in [this section](#write-a-conversion-script).
existing model you only have to add a conversion script as described in [this section](#write-a-conversion-script).
In this case, you can just re-use the whole model architecture of the already existing model.
In this case, you can just re-use the whole model architecture of the already existing model.
Otherwise, let's start generating a new model. You have two choices here:
Otherwise, let's start generating a new model. We recommend using the following script to add a model starting from
an existing model:
-`transformers-cli add-new-model-like` to add a new model like an existing one
```bash
-`transformers-cli add-new-model` to add a new model from our template (will look like BERT or Bart depending on the type of model you select)
transformers-cli add-new-model-like
```
In both cases, you will be prompted with a questionnaire to fill in the basic information of your model. The second command requires to install `cookiecutter`, you can find more information on it [here](https://github.com/huggingface/transformers/tree/main/templates/adding_a_new_model).
You will be prompted with a questionnaire to fill in the basic information of your model.
**Open a Pull Request on the main huggingface/transformers repo**
**Open a Pull Request on the main huggingface/transformers repo**
@@ -351,13 +351,14 @@ Nel caso speciale in cui stiate aggiungendo un modello, la cui architettura sia
...
@@ -351,13 +351,14 @@ Nel caso speciale in cui stiate aggiungendo un modello, la cui architettura sia
dovrete solo aggiugnere uno script di conversione, come descritto [qui](#write-a-conversion-script).
dovrete solo aggiugnere uno script di conversione, come descritto [qui](#write-a-conversion-script).
In questo caso, potete riutilizzare l'intera architettura del modello gia esistente.
In questo caso, potete riutilizzare l'intera architettura del modello gia esistente.
Se questo non é il caso, cominciamo con il generare un nuovo modello. Avrete due opzioni:
Se questo non é il caso, cominciamo con il generare un nuovo modello. Ti consigliamo di utilizzare il seguente script per aggiungere un modello a partire da
un modello esistente:
-`transformers-cli add-new-model-like` per aggiungere un nuovo modello come uno che gia esiste
```bash
-`transformers-cli add-new-model` per aggiungere un nuovo modello da un nostro template (questo assomigliera a BERT o Bart, in base al modello che selezionerete)
transformers-cli add-new-model-like
```
In entrambi i casi, l'output vi darà un questionario da riempire con informazioni basi sul modello. Il secondo comando richiede di installare
Ti verrà richiesto con un questionario di compilare le informazioni di base del tuo modello.
un `cookiecutter` - maggiori informazioni [qui](https://github.com/huggingface/transformers/tree/main/templates/adding_a_new_model).
**Aprire una Pull Request in main huggingface/transformers repo**
**Aprire una Pull Request in main huggingface/transformers repo**
다음과 같이 이미 존재하는 모델의 모델 아키텍처와 정확히 일치하는 모델을 추가하는 특별한 경우에는 [이 섹션](#write-a-conversion-script)에 설명된대로 변환 스크립트만 추가하면 됩니다. 이 경우에는 이미 존재하는 모델의 전체 모델 아키텍처를 그대로 재사용할 수 있습니다.
다음과 같이 이미 존재하는 모델의 모델 아키텍처와 정확히 일치하는 모델을 추가하는 특별한 경우에는 [이 섹션](#write-a-conversion-script)에 설명된대로 변환 스크립트만 추가하면 됩니다. 이 경우에는 이미 존재하는 모델의 전체 모델 아키텍처를 그대로 재사용할 수 있습니다.
그렇지 않으면 새로운 모델 생성을 시작합시다. 여기에서 두 가지 선택지가 있습니다:
그렇지 않으면 새 모델 생성을 시작하겠습니다. 다음 스크립트를 사용하여 다음에서 시작하는 모델을 추가하는 것이 좋습니다.
기존 모델:
-`transformers-cli add-new-model-like`를 사용하여 기존 모델과 유사한 새로운 모델 추가하기
```bash
-`transformers-cli add-new-model`을 사용하여 템플릿을 기반으로 한 새로운 모델 추가하기 (선택한 모델 유형에 따라 BERT 또는 Bart와 유사한 모습일 것입니다)
transformers-cli add-new-model-like
```
두 경우 모두, 모델의 기본 정보를 입력하는 설문조사가 제시됩니다. 두 번째 명령어는 `cookiecutter`를 설치해야 합니다. 자세한 정보는 [여기](https://github.com/huggingface/transformers/tree/main/templates/adding_a_new_model)에서 확인할 수 있습니다.
모델의 기본 정보를 입력하는 설문지가 표시됩니다.
**huggingface/transformers 메인 저장소에 Pull Request 열기**
**huggingface/transformers 메인 저장소에 Pull Request 열기**
만약 모델을 직접 기여하고 싶으시다면, 알려주세요. 🤗 Transformers에 추가할 수 있도록 도와드리겠습니다!
만약 모델을 직접 기여하고 싶으시다면, 알려주세요. 🤗 Transformers에 추가할 수 있도록 도와드리겠습니다!
새로운 모델을 추가하는 방법에 대한 [상세 안내서와 템플릿](https://github.com/huggingface/transformers/tree/main/templates)을 제공하고 있으며, [🤗 Transformers에 새로운 모델을 추가하는 방법](https://huggingface.co/docs/transformers/add_new_model)에 대한 기술적인 안내서도 있습니다.
[🤗 Transformers에 새로운 모델을 추가하는 방법](https://huggingface.co/docs/transformers/add_new_model)에 대한 기술적인 안내서도 있습니다.
## 문서를 추가하고 싶으신가요? [[do-you-want-to-add-documentation]]
## 문서를 추가하고 싶으신가요? [[do-you-want-to-add-documentation]]
This model inherits from [`FlaxPreTrainedModel`]. Check the superclass documentation for the
generic methods the library implements for all its model (such as downloading, saving and converting weights from
PyTorch models)
This model is also a Flax Linen [flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html) subclass. Use it as a regular Flax linen Module
and refer to the Flax documentation for all matter related to general usage and behavior.
Finally, this model supports inherent JAX features such as:
# causal mask for cached decoder self-attention: our single query position should only attend to those key positions that have already been generated and cached, not the remaining zero elements.
@add_start_docstrings("""{{cookiecutter.camelcase_modelname}} Model with a `language modeling` head on top for MLM training. """,{{cookiecutter.uppercase_modelname}}_START_DOCSTRING)
@add_start_docstrings("""{{cookiecutter.camelcase_modelname}} Model with a `language modeling` head on top for CLM training. """,{{cookiecutter.uppercase_modelname}}_START_DOCSTRING)
This model inherits from [`FlaxPreTrainedModel`]. Check the superclass documentation for the
generic methods the library implements for all its model (such as downloading or saving, resizing the input
embeddings, pruning heads etc.)
This model is also a Flax Linen [flax.nn.Module](https://flax.readthedocs.io/en/latest/_autosummary/flax.nn.module.html) subclass. Use it as a regular Flax
Module and refer to the Flax documentation for all matter related to general usage and behavior.
Finally, this model supports inherent JAX features such as:
input_ids (`jnp.ndarray` of shape `(batch_size, sequence_length)`):
Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide
it.
Indices can be obtained using [`~{{cookiecutter.camelcase_modelname}}Tokenizer`]. See
[`PreTrainedTokenizer.encode`] and [`PreTrainedTokenizer.__call__`] for
details.
[What are input IDs?](../glossary#input-ids)
attention_mask (`jnp.ndarray` of shape `(batch_size, sequence_length)`, *optional*):
Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
- 1 for tokens that are **not masked**,
- 0 for tokens that are **masked**.
[What are attention masks?](../glossary#attention-mask)
decoder_input_ids (`jnp.ndarray` of shape `(batch_size, target_sequence_length)`, *optional*):
Indices of decoder input sequence tokens in the vocabulary.
Indices can be obtained using [`~{{cookiecutter.camelcase_modelname}}Tokenizer`]. See
[`PreTrainedTokenizer.encode`] and [`PreTrainedTokenizer.__call__`] for
details.
[What are decoder input IDs?](../glossary#decoder-input-ids)
For translation and summarization training, `decoder_input_ids` should be provided. If no
`decoder_input_ids` is provided, the model will create this tensor by shifting the `input_ids` to
the right for denoising pre-training following the paper.
decoder_attention_mask (`jnp.ndarray` of shape `(batch_size, target_sequence_length)`, *optional*):
Default behavior: generate a tensor that ignores pad tokens in `decoder_input_ids`. Causal mask will
also be used by default.
If you want to change padding behavior, you should modify to your needs. See diagram 1 in [the paper](https://arxiv.org/abs/1910.13461) for more information on the default strategy.
position_ids (`numpy.ndarray` of shape `(batch_size, sequence_length)`, *optional*):
Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0, config.max_position_embeddings - 1]`.
decoder_position_ids (`numpy.ndarray` of shape `(batch_size, sequence_length)`, *optional*):
Indices of positions of each decoder input sequence tokens in the position embeddings. Selected in the
range `[0, config.max_position_embeddings - 1]`.
output_attentions (`bool`, *optional*):
Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
tensors for more detail.
output_hidden_states (`bool`, *optional*):
Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
more detail.
return_dict (`bool`, *optional*):
Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
decoder_input_ids (`jnp.ndarray` of shape `(batch_size, target_sequence_length)`):
Indices of decoder input sequence tokens in the vocabulary.
Indices can be obtained using [`~{{cookiecutter.camelcase_modelname}}Tokenizer`]. See
[`PreTrainedTokenizer.encode`] and [`PreTrainedTokenizer.__call__`] for
details.
[What are decoder input IDs?](../glossary#decoder-input-ids)
For translation and summarization training, `decoder_input_ids` should be provided. If no
`decoder_input_ids` is provided, the model will create this tensor by shifting the `input_ids` to
the right for denoising pre-training following the paper.
encoder_outputs (`tuple(tuple(jnp.ndarray)`):
Tuple consists of (`last_hidden_state`, *optional*: `hidden_states`, *optional*:
`attentions`) `last_hidden_state` of shape `(batch_size, sequence_length, hidden_size)`,
*optional*) is a sequence of hidden-states at the output of the last layer of the encoder. Used in the
cross-attention of the decoder.
encoder_attention_mask (`jnp.ndarray` of shape `(batch_size, sequence_length)`, *optional*):
Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
- 1 for tokens that are **not masked**,
- 0 for tokens that are **masked**.
[What are attention masks?](../glossary#attention-mask)
decoder_attention_mask (`jnp.ndarray` of shape `(batch_size, target_sequence_length)`, *optional*):
Default behavior: generate a tensor that ignores pad tokens in `decoder_input_ids`. Causal mask will
also be used by default.
If you want to change padding behavior, you should modify to your needs. See diagram 1 in [the paper](https://arxiv.org/abs/1910.13461) for more information on the default strategy.
decoder_position_ids (`numpy.ndarray` of shape `(batch_size, sequence_length)`, *optional*):
Indices of positions of each decoder input sequence tokens in the position embeddings. Selected in the
range `[0, config.max_position_embeddings - 1]`.
past_key_values (`Dict[str, np.ndarray]`, *optional*, returned by `init_cache` or when passing previous `past_key_values`):
Dictionary of pre-computed hidden-states (key and values in the attention blocks) that can be used for fast
auto-regressive decoding. Pre-computed key and value hidden-states are of shape *[batch_size, max_length]*.
output_attentions (`bool`, *optional*):
Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
tensors for more detail.
output_hidden_states (`bool`, *optional*):
Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
more detail.
return_dict (`bool`, *optional*):
Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
# causal mask for cached decoder self-attention: our single query position should only attend to those key positions that have already been generated and cached, not the remaining zero elements.
`encoder_outputs` consists of (`last_hidden_state`, *optional*: `hidden_states`,
*optional*: `attentions`). `last_hidden_state` of shape `(batch_size, sequence_length, hidden_size)`, *optional*) is a sequence of hidden-states at the output of the last layer of the
encoder. Used in the cross-attention of the decoder.
"The {{cookiecutter.uppercase_modelname}} Model with a language modeling head. Can be used for summarization.",{{cookiecutter.uppercase_modelname}}_START_DOCSTRING
# Transpose the tensor from [batch_size, seq_length, num_attention_heads, attention_head_size] to [batch_size, num_attention_heads, seq_length, attention_head_size]
This model inherits from [`TFPreTrainedModel`]. Check the superclass documentation for the
generic methods the library implements for all its model (such as downloading or saving, resizing the input
embeddings, pruning heads etc.)
This model is also a [keras.Model](https://www.tensorflow.org/api_docs/python/tf/keras/Model) subclass.
Use it as a regular TF 2.0 Keras Model and refer to the TF 2.0 documentation for all matter related to general
usage and behavior.
<Tip>
TensorFlow models and layers in `transformers` accept two formats as input:
- having all inputs as keyword arguments (like PyTorch models), or
- having all inputs as a list, tuple or dict in the first positional argument.
The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second format outside of Keras methods like `fit()` and `predict()`, such as when creating
your own layers or models with the Keras `Functional` API, there are three possibilities you
can use to gather all the input Tensors in the first positional argument:
- a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
- a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
`model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
- a dictionary with one or several input Tensors associated to the input names given in the docstring:
input_ids (`np.ndarray`, `tf.Tensor`, `List[tf.Tensor]`, `Dict[str, tf.Tensor]` or `Dict[str, np.ndarray]` and each example must have the shape `({0})`):
Indices of input sequence tokens in the vocabulary.
Indices can be obtained using [`AutoTokenizer`]. See
[`PreTrainedTokenizer.__call__`] and [`PreTrainedTokenizer.encode`] for
details.
[What are input IDs?](../glossary#input-ids)
attention_mask (`np.ndarray` or `tf.Tensor` of shape `({0})`, *optional*):
Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
- 1 for tokens that are **not masked**,
- 0 for tokens that are **masked**.
[What are attention masks?](../glossary#attention-mask)
token_type_ids (`np.ndarray` or `tf.Tensor` of shape `({0})`, *optional*):
Segment token indices to indicate first and second portions of the inputs. Indices are selected in `[0, 1]`:
- 0 corresponds to a *sentence A* token,
- 1 corresponds to a *sentence B* token.
[What are token type IDs?](../glossary#token-type-ids)
position_ids (`np.ndarray` or `tf.Tensor` of shape `({0})`, *optional*):
Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0, config.max_position_embeddings - 1]`.
[What are position IDs?](../glossary#position-ids)
head_mask (`np.ndarray` or `tf.Tensor` of shape `(num_heads,)` or `(num_layers, num_heads)`, *optional*):
Mask to nullify selected heads of the self-attention modules. Mask values selected in `[0, 1]`:
- 1 indicates the head is **not masked**,
- 0 indicates the head is **masked**.
inputs_embeds (`np.ndarray` or `tf.Tensor` of shape `({0}, hidden_size)`, *optional*):
Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation.
This is useful if you want more control over how to convert `input_ids` indices into associated
vectors than the model's internal embedding lookup matrix.
output_attentions (`bool`, *optional*):
Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
tensors for more detail. This argument can be used only in eager mode, in graph mode the value in the
config will be used instead.
output_hidden_states (`bool`, *optional*):
Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
more detail. This argument can be used only in eager mode, in graph mode the value in the config will be
used instead.
return_dict (`bool`, *optional*):
Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. This
argument can be used in eager mode, in graph mode the value will always be set to True.
training (`bool`, *optional*, defaults to `False`):
Whether or not to use the model in training mode (some modules like dropout modules have different
behaviors between training and evaluation).
"""
@add_start_docstrings(
"The bare {{cookiecutter.modelname}} Model transformer outputing raw hidden-states without any specific head on top.",
@add_start_docstrings("""{{cookiecutter.modelname}} Model with a `language modeling` head on top. """,{{cookiecutter.uppercase_modelname}}_START_DOCSTRING)
labels (`tf.Tensor` or `np.ndarray` of shape `(batch_size, sequence_length)`, *optional*):
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]`
"""{{cookiecutter.modelname}} Model with a `language modeling` head on top for CLM fine-tuning. """,{{cookiecutter.uppercase_modelname}}_START_DOCSTRING
labels (`tf.Tensor` or `np.ndarray` of shape `(batch_size,)`, *optional*):
Labels for computing the sequence classification/regression loss. Indices should be in `[0, ..., config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss),
If `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
labels (`tf.Tensor` or `np.ndarray` of shape `(batch_size,)`, *optional*):
Labels for computing the multiple choice classification loss. Indices should be in `[0, ..., num_choices]` where `num_choices` is the size of the second dimension of the input tensors. (See
This model inherits from [`TFPreTrainedModel`]. Check the superclass documentation for the
generic methods the library implements for all its model (such as downloading or saving, resizing the input
embeddings, pruning heads etc.)
This model is also a [keras.Model](https://www.tensorflow.org/api_docs/python/tf/keras/Model) subclass. Use
it as a regular TF 2.0 Keras Model and refer to the TF 2.0 documentation for all matter related to general usage
and behavior.
<Tip>
TensorFlow models and layers in `transformers` accept two formats as input:
- having all inputs as keyword arguments (like PyTorch models), or
- having all inputs as a list, tuple or dict in the first positional argument.
The reason the second format is supported is that Keras methods prefer this format when passing inputs to models
and layers. Because of this support, when using methods like `model.fit()` things should "just work" for you - just
pass your inputs and labels in any format that `model.fit()` supports! If, however, you want to use the second format outside of Keras methods like `fit()` and `predict()`, such as when creating
your own layers or models with the Keras `Functional` API, there are three possibilities you
can use to gather all the input Tensors in the first positional argument:
- a single Tensor with `input_ids` only and nothing else: `model(input_ids)`
- a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
`model([input_ids, attention_mask])` or `model([input_ids, attention_mask, token_type_ids])`
- a dictionary with one or several input Tensors associated to the input names given in the docstring:
input_ids (`tf.Tensor` of shape `(batch_size, sequence_length)`):
Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you
provide it.
Indices can be obtained using [`~{{cookiecutter.camelcase_modelname}}Tokenizer`]. See
[`PreTrainedTokenizer.encode`] and [`PreTrainedTokenizer.__call__`]
for details.
[What are input IDs?](../glossary#input-ids)
attention_mask (`tf.Tensor` of shape `(batch_size, sequence_length)`, *optional*):
Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
- 1 for tokens that are **not masked**,
- 0 for tokens that are **masked**.
[What are attention masks?](../glossary#attention-mask)
head_mask (`tf.Tensor` of shape `(encoder_layers, encoder_attention_heads)`, `optional): Mask to nullify selected heads of the attention modules. Mask values selected in `[0, 1]`:
- 1 indicates the head is **not masked**,
- 0 indicates the head is **masked**.
inputs_embeds (`tf.Tensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*):
Optionally, instead of passing `input_ids` you can choose to directly pass an embedded
representation. This is useful if you want more control over how to convert `input_ids` indices
into associated vectors than the model's internal embedding lookup matrix.
output_attentions (`bool`, *optional*):
Whether or not to return the attentions tensors of all attention layers. See `attentions` under
returned tensors for more detail. This argument can be used only in eager mode, in graph mode the value
in the config will be used instead.
output_hidden_states (`bool`, *optional*):
Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors
for more detail. This argument can be used only in eager mode, in graph mode the value in the config
will be used instead.
return_dict (`bool`, *optional*):
Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. This
argument can be used in eager mode, in graph mode the value will always be set to True.
training (`bool`, *optional*, defaults to `False`):
Whether or not to use the model in training mode (some modules like dropout modules have different
behaviors between training and evaluation).
"""
ifinput_idsisnotNoneandinputs_embedsisnotNone:
raiseValueError("You cannot specify both input_ids and inputs_embeds at the same time")
elifinput_idsisnotNone:
input_shape=shape_list(input_ids)
elifinputs_embedsisnotNone:
input_shape=shape_list(inputs_embeds)[:-1]
else:
raiseValueError("You have to specify either input_ids or inputs_embeds")
input_ids (`tf.Tensor` of shape `(batch_size, sequence_length)`):
Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you
provide it.
Indices can be obtained using [`~{{cookiecutter.camelcase_modelname}}Tokenizer`]. See
[`PreTrainedTokenizer.encode`] and [`PreTrainedTokenizer.__call__`]
for details.
[What are input IDs?](../glossary#input-ids)
attention_mask (`tf.Tensor` of shape `(batch_size, sequence_length)`, *optional*):
Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
- 1 for tokens that are **not masked**,
- 0 for tokens that are **masked**.
[What are attention masks?](../glossary#attention-mask)
encoder_hidden_states (`tf.Tensor` of shape `(batch_size, encoder_sequence_length, hidden_size)`, *optional*):
Sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention
of the decoder.
encoder_attention_mask (`tf.Tensor` of shape `(batch_size, encoder_sequence_length)`, *optional*):
Mask to avoid performing cross-attention on padding tokens indices of encoder input_ids. Mask values
selected in `[0, 1]`:
- 1 for tokens that are **not masked**,
- 0 for tokens that are **masked**.
[What are attention masks?](../glossary#attention-mask)
head_mask (`tf.Tensor` of shape `(decoder_layers, decoder_attention_heads)`, *optional*):
Mask to nullify selected heads of the attention modules. Mask values selected in `[0, 1]`:
- 1 indicates the head is **not masked**,
- 0 indicates the head is **masked**.
cross_attn_head_mask (`tf.Tensor` of shape `(decoder_layers, decoder_attention_heads)`, *optional*):
Mask to nullify selected heads of the cross-attention modules. Mask values selected in `[0, 1]`:
- 1 indicates the head is **not masked**,
- 0 indicates the head is **masked**.
past_key_values (`Tuple[Tuple[tf.Tensor]]` of length `config.n_layers` with each tuple having 2 tuples each of which has 2 tensors of shape `(batch_size, num_heads, sequence_length - 1, embed_size_per_head)`):
Contains precomputed key and value hidden-states of the attention blocks. Can be used to speed up
decoding.
If `past_key_values` are used, the user can optionally input only the last
`decoder_input_ids` (those that don't have their past key value states given to this model) of
shape `(batch_size, 1)` instead of all `decoder_input_ids` of shape `(batch_size,
sequence_length)`.
inputs_embeds (`tf.Tensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*):
Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation.
This is useful if you want more control over how to convert `input_ids` indices
into associated vectors than the model's internal embedding lookup matrix.
output_attentions (`bool`, *optional*):
Whether or not to return the attentions tensors of all attention layers. See `attentions` under
returned tensors for more detail. This argument can be used only in eager mode, in graph mode the value
in the config will be used instead.
output_hidden_states (`bool`, *optional*):
Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors
for more detail. This argument can be used only in eager mode, in graph mode the value in the config
will be used instead.
return_dict (`bool`, *optional*):
Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple. This
argument can be used in eager mode, in graph mode the value will always be set to True.
training (`bool`, *optional*, defaults to `False`):
Whether or not to use the model in training mode (some modules like dropout modules have different
behaviors between training and evaluation).
"""
ifinput_idsisnotNoneandinputs_embedsisnotNone:
raiseValueError("You cannot specify both decoder_input_ids and decoder_inputs_embeds at the same time")
elifinput_idsisnotNone:
input_shape=shape_list(input_ids)
elifinputs_embedsisnotNone:
input_shape=shape_list(inputs_embeds)[:-1]
else:
raiseValueError("You have to specify either decoder_input_ids or decoder_inputs_embeds")
encoder_hidden_states (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*):
Sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention
if the model is configured as a decoder.
encoder_attention_mask (`torch.FloatTensor` of shape `(batch_size, sequence_length)`, *optional*):
Mask to avoid performing attention on the padding token indices of the encoder input. This mask
is used in the cross-attention if the model is configured as a decoder.
Mask values selected in `[0, 1]`:
- 1 for tokens that are **not masked**,
- 0 for tokens that are **masked**.
past_key_values (`tuple(tuple(torch.FloatTensor))` of length `config.n_layers` with each tuple having 4 tensors of shape `(batch_size, num_heads, sequence_length - 1, embed_size_per_head)`):
Contains precomputed key and value hidden states of the attention blocks. Can be used to speed up decoding.
If `past_key_values` are used, the user can optionally input only the last `decoder_input_ids`
(those that don't have their past key value states given to this model) of shape `(batch_size, 1)`
instead of all `decoder_input_ids` of shape `(batch_size, sequence_length)`.
use_cache (`bool`, *optional*):
If set to `True`, `past_key_values` key value states are returned and can be used to speed up
@add_start_docstrings("""{{cookiecutter.modelname}} Model with a `language modeling` head on top. """,{{cookiecutter.uppercase_modelname}}_START_DOCSTRING)
"""{{cookiecutter.modelname}} Model with a `language modeling` head on top for CLM fine-tuning. """,{{cookiecutter.uppercase_modelname}}_START_DOCSTRING
"The {{cookiecutter.modelname}} Model with a language modeling head. Can be used for summarization.",{{cookiecutter.uppercase_modelname}}_START_DOCSTRING
labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
Labels for computing the masked language modeling loss. Indices should either be in `[0, ..., config.vocab_size]` or -100 (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]`.
Returns:
Conditional generation example:
```python
>>> from transformers import {{cookiecutter.camelcase_modelname}}Tokenizer, {{cookiecutter.camelcase_modelname}}ForConditionalGeneration
labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
Labels for computing the sequence classification/regression loss. Indices should be in `[0, ..., config.num_labels - 1]`. If `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you
provide it.
Indices can be obtained using [`~{{cookiecutter.camelcase_modelname}}Tokenizer`]. See
[`PreTrainedTokenizer.encode`] and [`PreTrainedTokenizer.__call__`]
for details.
[What are input IDs?](../glossary#input-ids)
attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*):
Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
- 1 for tokens that are **not masked**,
- 0 for tokens that are **masked**.
[What are attention masks?](../glossary#attention-mask)
encoder_hidden_states (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*):
Sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention
if the model is configured as a decoder.
encoder_attention_mask (`torch.FloatTensor` of shape `(batch_size, sequence_length)`, *optional*):
Mask to avoid performing attention on the padding token indices of the encoder input. This mask is used
in the cross-attention if the model is configured as a decoder. Mask values selected in `[0, 1]`:
head_mask (`torch.Tensor` of shape `(decoder_layers, decoder_attention_heads)`, *optional*):
Mask to nullify selected heads of the attention modules. Mask values selected in `[0, 1]`:
- 1 indicates the head is **not masked**,
- 0 indicates the head is **masked**.
cross_attn_head_mask (`torch.Tensor` of shape `(decoder_layers, decoder_attention_heads)`, *optional*):
Mask to nullify selected heads of the cross-attention modules. Mask values selected in `[0, 1]`:
- 1 indicates the head is **not masked**,
- 0 indicates the head is **masked**.
past_key_values (`tuple(tuple(torch.FloatTensor))` of length `config.n_layers` with each tuple having 4 tensors of shape `(batch_size, num_heads, sequence_length - 1, embed_size_per_head)`):
Contains precomputed key and value hidden-states of the attention blocks. Can be used to speed up
decoding.
If `past_key_values` are used, the user can optionally input only the last `decoder_input_ids`
(those that don't have their past key value states given to this model) of shape `(batch_size, 1)`
instead of all `decoder_input_ids` of shape `(batch_size, sequence_length)`.
labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
Labels for computing the masked language modeling loss. Indices should either be in `[0, ..., config.vocab_size]` or -100 (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]`.
use_cache (`bool`, *optional*):
If set to `True`, `past_key_values` key value states are returned and can be used to speed up
decoding (see `past_key_values`).
- 1 for tokens that are **not masked**,
- 0 for tokens that are **masked**.
output_attentions (`bool`, *optional*):
Whether or not to return the attentions tensors of all attention layers. See `attentions` under
returned tensors for more detail.
output_hidden_states (`bool`, *optional*):
Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors
for more detail.
return_dict (`bool`, *optional*):
Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
Returns:
Example:
```python
>>> from transformers import {{cookiecutter.camelcase_modelname}}Tokenizer, {{cookiecutter.camelcase_modelname}}ForCausalLM
