# ESM [[esm]]

## 개요 [[overview]]

이 페이지는 Meta AI의 Fundamental AI Research 팀에서 제공하는 Transformer 단백질 언어 모델에 대한 코드와 사전 훈련된 가중치를 제공합니다. 여기에는 최첨단인 ESMFold와 ESM-2, 그리고 이전에 공개된 ESM-1b와 ESM-1v가 포함됩니다. Transformer 단백질 언어 모델은 Alexander Rives, Joshua Meier, Tom Sercu, Siddharth Goyal, Zeming Lin, Jason Liu, Demi Guo, Myle Ott, C. Lawrence Zitnick, Jerry Ma, Rob Fergus의 논문 [Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences](https://www.pnas.org/content/118/15/e2016239118)에서 소개되었습니다. 이 논문의 첫 번째 버전은 2019년에 [출판 전 논문](https://www.biorxiv.org/content/10.1101/622803v1?versioned=true) 형태로 공개되었습니다.

ESM-2는 다양한 구조 예측 작업에서 테스트된 모든 단일 시퀀스 단백질 언어 모델을 능가하며, 원자 수준의 구조 예측을 가능하게 합니다. 이 모델은 Zeming Lin, Halil Akin, Roshan Rao, Brian Hie, Zhongkai Zhu, Wenting Lu, Allan dos Santos Costa, Maryam Fazel-Zarandi, Tom Sercu, Sal Candido, Alexander Rives의 논문 [Language models of protein sequences at the scale of evolution enable accurate structure prediction](https://doi.org/10.1101/2022.07.20.500902)에서 공개되었습니다.

이 논문에서 함께 소개된 ESMFold는 ESM-2 스템을 사용하며, 최첨단의 정확도로 단백질 접힘 구조를 예측할 수 있는 헤드를 갖추고 있습니다. [AlphaFold2](https://www.nature.com/articles/s41586-021-03819-2)와 달리, 이는 대형 사전 훈련된 단백질 언어 모델 스템의 토큰 임베딩에 의존하며, 추론 시 다중 시퀀스 정렬(MSA) 단계를 수행하지 않습니다. 이는 ESMFold 체크포인트가 완전히 "독립적"이며, 예측을 위해 알려진 단백질 시퀀스와 구조의 데이터베이스, 그리고 그와 관련 외부 쿼리 도구를 필요로 하지 않는다는 것을 의미합니다. 그리고 그 결과, 훨씬 빠릅니다.

"Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences"의 초록은 다음과 같습니다:

*인공지능 분야에서는 대규모의 데이터와 모델 용량을 갖춘 비지도 학습의 조합이 표현 학습과 통계적 생성에서 주요한 발전을 이끌어냈습니다. 생명 과학에서는 시퀀싱 기술의 성장이 예상되며, 자연 시퀀스 다양성에 대한 전례 없는 데이터가 나올 것으로 기대됩니다. 진화적 단계에서 볼 때, 단백질 언어 모델링은 생물학을 위한 예측 및 생성 인공지능을 향한 논리적인 단계에 있습니다. 이를 위해 우리는 진화적 다양성을 아우르는 2억 5천만 개의 단백질 시퀀스에서 추출한 860억 개의 아미노산에 대해 심층 컨텍스트 언어 모델을 비지도 학습으로 훈련합니다. 그 결과 모델은 그 표현에서 생물학적 속성에 대한 정보를 포함합니다. 이 표현은 시퀀스 데이터만으로 학습됩니다. 학습된 표현 공간은 아미노산의 생화학적 특성 수준에서부터 단백질의 원거리 상동성까지 구조를 반영하는 다중 규모의 조직을 가지고 있습니다. 이 표현에는 2차 및 3차 구조에 대한 정보가 인코딩되어 있으며, 선형 전사에 의해 식별 될 수 있습니다. 표현 학습은 돌연변이에 의한 효과와 2차 구조의 최첨단 지도 예측을 가능하게 하고, 넓은 범위의 접촉 부위 예측을 위한 최첨단 특징을 향상시킵니다.*

"Language models of protein sequences at the scale of evolution enable accurate structure prediction"의 초록은 다음과 같습니다:

*대형 언어 모델은 최근 규모가 커짐에 따라 긴급한 기능을 개발하여 단순한 패턴 매칭을 넘어 더 높은 수준의 추론을 수행하고 생생한 이미지와 텍스트를 생성하는 것으로 나타났습니다. 더 작은 규모에서 훈련된 단백질 시퀀스의 언어 모델이 연구되었지만, 그들이 규모가 커짐에 따라 생물학에 대해 무엇을 배우는지는 거의 알려져 있지 않습니다. 이 연구에서 우리는 현재까지 평가된 가장 큰 150억 개의 매개변수를 가진 모델을 훈련합니다. 우리는 모델이 규모가 커짐에 따라 단일 아미노산의 해상도로 단백질의 3차원 구조를 예측할 수 있는 정보를 학습한다는 것을 발견했습니다. 우리는 개별 단백질 시퀀스로부터 직접 고정밀 원자 수준의 엔드-투-엔드 구조 예측을 하기 위한 ESMFold를 제시합니다. ESMFold는 언어 모델에 잘 이해되는 낮은 퍼플렉서티를 가진 시퀀스에 대해 AlphaFold2와 RoseTTAFold와 유사한 정확도를 가지고 있습니다. ESMFold의 추론은 AlphaFold2보다 한 자릿수 빠르며, 메타게놈 단백질의 구조적 공간을 실용적인 시간 내에 탐색할 수 있게 합니다.*

원본 코드는 [여기](https://github.com/facebookresearch/esm)에서 찾을 수 있으며, Meta AI의 Fundamental AI Research 팀에서 개발되었습니다. ESM-1b, ESM-1v, ESM-2는 [jasonliu](https://huggingface.co/jasonliu)와 [Matt](https://huggingface.co/Rocketknight1)에 의해 HuggingFace에 기여되었습니다.

ESMFold는 [Matt](https://huggingface.co/Rocketknight1)와 [Sylvain](https://huggingface.co/sgugger)에 의해 HuggingFace에 기여되었으며, 이 과정에서 많은 도움을 준 Nikita Smetanin, Roshan Rao, Tom Sercu에게 큰 감사를 드립니다!

## 사용 팁 [[usage-tips]]

- ESM 모델은 마스크드 언어 모델링(MLM) 목표로 훈련되었습니다.
- HuggingFace의 ESMFold 포트는 [openfold](https://github.com/aqlaboratory/openfold) 라이브러리의 일부를 사용합니다. `openfold` 라이브러리는 Apache License 2.0에 따라 라이선스가 부여됩니다.

## 리소스 [[resources]]

- [텍스트 분류 작업 가이드](../tasks/sequence_classification)
- [토큰 분류 작업 가이드](../tasks/token_classification)
- [마스킹드 언어 모델링 작업 가이드](../tasks/masked_language_modeling)

## EsmConfig [[transformers.EsmConfig]][[transformers.EsmConfig]]

#### transformers.EsmConfig[[transformers.EsmConfig]]

[Source](https://github.com/huggingface/transformers/blob/v5.6.0/src/transformers/models/esm/configuration_esm.py#L157)

This is the configuration class to store the configuration of a EsmModel. It is used to instantiate a Esm
model according to the specified arguments, defining the model architecture. Instantiating a configuration with the
defaults will yield a similar configuration to that of the [facebook/esm-1b](https://huggingface.co/facebook/esm-1b)

Configuration objects inherit from [PreTrainedConfig](/docs/transformers/v5.6.0/ko/main_classes/configuration#transformers.PreTrainedConfig) and can be used to control the model outputs. Read the
documentation from [PreTrainedConfig](/docs/transformers/v5.6.0/ko/main_classes/configuration#transformers.PreTrainedConfig) for more information.

Examples:

```python
>>> from transformers import EsmModel, EsmConfig

>>> # Initializing a ESM facebook/esm-1b style configuration
>>> configuration = EsmConfig(vocab_size=33)

>>> # Initializing a model from the configuration
>>> model = EsmModel(configuration)

>>> # Accessing the model configuration
>>> configuration = model.config
```

**Parameters:**

vocab_size (`int`, *optional*) : Vocabulary size of the model. Defines the number of different tokens that can be represented by the `input_ids`.

mask_token_id (`int`, *optional*) : The index of the mask token in the vocabulary. This must be included in the config because of the "mask-dropout" scaling trick, which will scale the inputs depending on the number of masked tokens.

pad_token_id (`int`, *optional*) : Token id used for padding in the vocabulary.

hidden_size (`int`, *optional*, defaults to `768`) : Dimension of the hidden representations.

num_hidden_layers (`int`, *optional*, defaults to `12`) : Number of hidden layers in the Transformer decoder.

num_attention_heads (`int`, *optional*, defaults to `12`) : Number of attention heads for each attention layer in the Transformer decoder.

intermediate_size (`int`, *optional*, defaults to `3072`) : Dimension of the MLP representations.

hidden_dropout_prob (`float`, *optional*, defaults to `0.1`) : The dropout probability for all fully connected layers in the embeddings, encoder, and pooler.

attention_probs_dropout_prob (`float`, *optional*, defaults to `0.1`) : The dropout ratio for the attention probabilities.

max_position_embeddings (`int`, *optional*, defaults to `1026`) : The maximum sequence length that this model might ever be used with.

rope_theta (`float`, defaults to 10000.0) : The base period of the RoPE embeddings. Only used when `position_embedding_type` is set to `"rotary"`.

initializer_range (`float`, *optional*, defaults to `0.02`) : The standard deviation of the truncated_normal_initializer for initializing all weight matrices.

layer_norm_eps (`float`, *optional*, defaults to `1e-12`) : The epsilon used by the layer normalization layers.

position_embedding_type (`str`, *optional*, defaults to `"absolute"`) : Type of position embedding. Choose either `"absolute"` or "rotary"`.

use_cache (`bool`, *optional*, defaults to `True`) : Whether or not the model should return the last key/values attentions (not used by all models). Only relevant if `config.is_decoder=True` or when the model is a decoder-only generative model.

emb_layer_norm_before (`bool`, *optional*) : Whether to apply layer normalization after embeddings but before the main stem of the network.

token_dropout (`bool`, defaults to `False`) : When this is enabled, masked tokens are treated as if they had been dropped out by input dropout.

is_folding_model (`bool`, defaults to `False`) : When this is enabled, ESMFold model will be initialized.

esmfold_config (`dict`, *optional*) : Configuration to initiate the ESMFold module.

vocab_list (`list`, *optional*) : List of the vocabulary items.

is_decoder (`bool`, *optional*, defaults to `False`) : Whether the model is used as a decoder or not. If `False`, the model is used as an encoder.

add_cross_attention (`bool`, *optional*, defaults to `False`) : Whether cross-attention layers should be added to the model.

tie_word_embeddings (`bool`, *optional*, defaults to `True`) : Whether to tie weight embeddings according to model's `tied_weights_keys` mapping.

bos_token_id (`int`, *optional*) : Token id used for beginning-of-stream in the vocabulary.

eos_token_id (`Union[int, list[int]]`, *optional*, defaults to `2`) : Token id used for end-of-stream in the vocabulary.

## EsmTokenizer [[transformers.EsmTokenizer]][[transformers.EsmTokenizer]]

#### transformers.EsmTokenizer[[transformers.EsmTokenizer]]

[Source](https://github.com/huggingface/transformers/blob/v5.6.0/src/transformers/models/esm/tokenization_esm.py#L33)

Constructs an ESM tokenizer.

build_inputs_with_special_tokenstransformers.EsmTokenizer.build_inputs_with_special_tokenshttps://github.com/huggingface/transformers/blob/v5.6.0/src/transformers/models/esm/tokenization_esm.py#L89[{"name": "token_ids_0", "val": ": list"}, {"name": "token_ids_1", "val": ": list[int] | None = None"}]
#### get_special_tokens_mask[[transformers.EsmTokenizer.get_special_tokens_mask]]

[Source](https://github.com/huggingface/transformers/blob/v5.6.0/src/transformers/models/esm/tokenization_esm.py#L103)

Retrieves sequence ids from a token list that has no special tokens added. This method is called when adding
special tokens using the tokenizer `prepare_for_model` or `encode_plus` methods.

**Parameters:**

token_ids_0 (`list[int]`) : List of ids of the first sequence.

token_ids_1 (`list[int]`, *optional*) : List of ids of the second sequence.

already_has_special_tokens (`bool`, *optional*, defaults to `False`) : Whether or not the token list is already formatted with special tokens for the model.

**Returns:**

`A list of integers in the range [0, 1]`

1 for a special token, 0 for a sequence token.
#### create_token_type_ids_from_sequences[[transformers.EsmTokenizer.create_token_type_ids_from_sequences]]

[Source](https://github.com/huggingface/transformers/blob/v5.6.0/src/transformers/tokenization_python.py#L1292)

Create a mask from the two sequences passed to be used in a sequence-pair classification task.

This method dynamically builds the token type IDs based on the tokenizer's configuration attributes:
- `token_type_ids_pattern`: Pattern to use ("all_zeros" or "bert_style")
- `token_type_ids_include_special_tokens`: Whether to account for special tokens in length calculation

Examples:
```python
# All zeros pattern (default, used by RoBERTa, BART, etc.)
tokenizer.token_type_ids_pattern = "all_zeros"
# Returns: [0, 0, 0, ...] for both sequences

# BERT-style pattern (first sequence gets 0s, second gets 1s)
tokenizer.token_type_ids_pattern = "bert_style"
# Returns: [0, 0, 0, ..., 1, 1, 1, ...] for sequence pairs
```

**Parameters:**

token_ids_0 (`list[int]`) : List of IDs.

token_ids_1 (`list[int]`, *optional*) : Optional second list of IDs for sequence pairs.

**Returns:**

``list[int]``

Token type IDs according to the configured pattern.
#### save_vocabulary[[transformers.EsmTokenizer.save_vocabulary]]

[Source](https://github.com/huggingface/transformers/blob/v5.6.0/src/transformers/models/esm/tokenization_esm.py#L134)

## EsmModel [[transformers.EsmModel]][[transformers.EsmModel]]

#### transformers.EsmModel[[transformers.EsmModel]]

[Source](https://github.com/huggingface/transformers/blob/v5.6.0/src/transformers/models/esm/modeling_esm.py#L627)

The bare Esm Model outputting raw hidden-states without any specific head on top.

This model inherits from [PreTrainedModel](/docs/transformers/v5.6.0/ko/main_classes/model#transformers.PreTrainedModel). 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 PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass.
Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage
and behavior.

forwardtransformers.EsmModel.forwardhttps://github.com/huggingface/transformers/blob/v5.6.0/src/transformers/models/esm/modeling_esm.py#L695[{"name": "input_ids", "val": ": torch.Tensor | None = None"}, {"name": "attention_mask", "val": ": torch.Tensor | None = None"}, {"name": "position_ids", "val": ": torch.Tensor | None = None"}, {"name": "inputs_embeds", "val": ": torch.Tensor | None = None"}, {"name": "encoder_hidden_states", "val": ": torch.Tensor | None = None"}, {"name": "encoder_attention_mask", "val": ": torch.Tensor | None = None"}, {"name": "**kwargs", "val": ": typing_extensions.Unpack[transformers.utils.generic.TransformersKwargs]"}]- **input_ids** (`torch.LongTensor` of shape `((batch_size, sequence_length))`) --
  Indices of input sequence tokens in the vocabulary.

  Indices can be obtained using [AutoTokenizer](/docs/transformers/v5.6.0/ko/model_doc/auto#transformers.AutoTokenizer). See [PreTrainedTokenizer.encode()](/docs/transformers/v5.6.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.encode) and
  [PreTrainedTokenizer.__call__()](/docs/transformers/v5.6.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.__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)
- **position_ids** (`torch.LongTensor` 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]`.

  [What are position IDs?](../glossary#position-ids)
- **inputs_embeds** (`torch.FloatTensor` 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.
- **encoder_hidden_states** (`torch.Tensor` 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.Tensor` 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**.0[BaseModelOutputWithPoolingAndCrossAttentions](/docs/transformers/v5.6.0/ko/main_classes/output#transformers.modeling_outputs.BaseModelOutputWithPoolingAndCrossAttentions) or `tuple(torch.FloatTensor)`A [BaseModelOutputWithPoolingAndCrossAttentions](/docs/transformers/v5.6.0/ko/main_classes/output#transformers.modeling_outputs.BaseModelOutputWithPoolingAndCrossAttentions) or a tuple of
`torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various
elements depending on the configuration ([EsmConfig](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmConfig)) and inputs.
The [EsmModel](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmModel) forward method, overrides the `__call__` special method.

Although the recipe for forward pass needs to be defined within this function, one should call the `Module`
instance afterwards instead of this since the former takes care of running the pre and post processing steps while
the latter silently ignores them.

- **last_hidden_state** (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`) -- Sequence of hidden-states at the output of the last layer of the model.
- **pooler_output** (`torch.FloatTensor` of shape `(batch_size, hidden_size)`) -- Last layer hidden-state of the first token of the sequence (classification token) after further processing
  through the layers used for the auxiliary pretraining task. E.g. for BERT-family of models, this returns
  the classification token after processing through a linear layer and a tanh activation function. The linear
  layer weights are trained from the next sentence prediction (classification) objective during pretraining.
- **hidden_states** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`) -- Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
  one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`.

  Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
- **attentions** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`) -- Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
  sequence_length)`.

  Attentions weights after the attention softmax, used to compute the weighted average in the self-attention
  heads.
- **cross_attentions** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` and `config.add_cross_attention=True` is passed or when `config.output_attentions=True`) -- Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
  sequence_length)`.

  Attentions weights of the decoder's cross-attention layer, after the attention softmax, used to compute the
  weighted average in the cross-attention heads.
- **past_key_values** (`Cache`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`) -- It is a [Cache](/docs/transformers/v5.6.0/ko/internal/generation_utils#transformers.Cache) instance. For more details, see our [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache).

  Contains pre-computed hidden-states (key and values in the self-attention blocks and optionally if
  `config.is_encoder_decoder=True` in the cross-attention blocks) that can be used (see `past_key_values`
  input) to speed up sequential decoding.

**Parameters:**

config ([EsmModel](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmModel)) : Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the [from_pretrained()](/docs/transformers/v5.6.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained) method to load the model weights.

add_pooling_layer (`bool`, *optional*, defaults to `True`) : Whether to add a pooling layer

**Returns:**

`[BaseModelOutputWithPoolingAndCrossAttentions](/docs/transformers/v5.6.0/ko/main_classes/output#transformers.modeling_outputs.BaseModelOutputWithPoolingAndCrossAttentions) or `tuple(torch.FloatTensor)``

A [BaseModelOutputWithPoolingAndCrossAttentions](/docs/transformers/v5.6.0/ko/main_classes/output#transformers.modeling_outputs.BaseModelOutputWithPoolingAndCrossAttentions) or a tuple of
`torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various
elements depending on the configuration ([EsmConfig](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmConfig)) and inputs.

## EsmForMaskedLM [[transformers.EsmForMaskedLM]][[transformers.EsmForMaskedLM]]

#### transformers.EsmForMaskedLM[[transformers.EsmForMaskedLM]]

[Source](https://github.com/huggingface/transformers/blob/v5.6.0/src/transformers/models/esm/modeling_esm.py#L816)

The Esm Model with a `language modeling` head on top."

This model inherits from [PreTrainedModel](/docs/transformers/v5.6.0/ko/main_classes/model#transformers.PreTrainedModel). 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 PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass.
Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage
and behavior.

forwardtransformers.EsmForMaskedLM.forwardhttps://github.com/huggingface/transformers/blob/v5.6.0/src/transformers/models/esm/modeling_esm.py#L839[{"name": "input_ids", "val": ": torch.LongTensor | None = None"}, {"name": "attention_mask", "val": ": torch.Tensor | None = None"}, {"name": "position_ids", "val": ": torch.LongTensor | None = None"}, {"name": "inputs_embeds", "val": ": torch.FloatTensor | None = None"}, {"name": "encoder_hidden_states", "val": ": torch.FloatTensor | None = None"}, {"name": "encoder_attention_mask", "val": ": torch.Tensor | None = None"}, {"name": "labels", "val": ": torch.LongTensor | None = None"}, {"name": "**kwargs", "val": ": typing_extensions.Unpack[transformers.utils.generic.TransformersKwargs]"}]- **input_ids** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) --
  Indices of input sequence tokens in the vocabulary. Padding will be ignored by default.

  Indices can be obtained using [AutoTokenizer](/docs/transformers/v5.6.0/ko/model_doc/auto#transformers.AutoTokenizer). See [PreTrainedTokenizer.encode()](/docs/transformers/v5.6.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.encode) and
  [PreTrainedTokenizer.__call__()](/docs/transformers/v5.6.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.__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)
- **position_ids** (`torch.LongTensor` 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.n_positions - 1]`.

  [What are position IDs?](../glossary#position-ids)
- **inputs_embeds** (`torch.FloatTensor` 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.
- **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.Tensor` 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**.
- **labels** (`torch.LongTensor` 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]`0[MaskedLMOutput](/docs/transformers/v5.6.0/ko/main_classes/output#transformers.modeling_outputs.MaskedLMOutput) or `tuple(torch.FloatTensor)`A [MaskedLMOutput](/docs/transformers/v5.6.0/ko/main_classes/output#transformers.modeling_outputs.MaskedLMOutput) or a tuple of
`torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various
elements depending on the configuration ([EsmConfig](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmConfig)) and inputs.
The [EsmForMaskedLM](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmForMaskedLM) forward method, overrides the `__call__` special method.

Although the recipe for forward pass needs to be defined within this function, one should call the `Module`
instance afterwards instead of this since the former takes care of running the pre and post processing steps while
the latter silently ignores them.

- **loss** (`torch.FloatTensor` of shape `(1,)`, *optional*, returned when `labels` is provided) -- Masked language modeling (MLM) loss.
- **logits** (`torch.FloatTensor` of shape `(batch_size, sequence_length, config.vocab_size)`) -- Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
- **hidden_states** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`) -- Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
  one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`.

  Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
- **attentions** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`) -- Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
  sequence_length)`.

  Attentions weights after the attention softmax, used to compute the weighted average in the self-attention
  heads.

Example:

```python
>>> from transformers import AutoTokenizer, EsmForMaskedLM
>>> import torch

>>> tokenizer = AutoTokenizer.from_pretrained("facebook/esm-1b")
>>> model = EsmForMaskedLM.from_pretrained("facebook/esm-1b")

>>> inputs = tokenizer("The capital of France is .", return_tensors="pt")

>>> with torch.no_grad():
...     logits = model(**inputs).logits

>>> # retrieve index of 
>>> mask_token_index = (inputs.input_ids == tokenizer.mask_token_id)[0].nonzero(as_tuple=True)[0]

>>> predicted_token_id = logits[0, mask_token_index].argmax(axis=-1)
>>> tokenizer.decode(predicted_token_id)
...

>>> labels = tokenizer("The capital of France is Paris.", return_tensors="pt")["input_ids"]
>>> # mask labels of non- tokens
>>> labels = torch.where(inputs.input_ids == tokenizer.mask_token_id, labels, -100)

>>> outputs = model(**inputs, labels=labels)
>>> round(outputs.loss.item(), 2)
...
```

**Parameters:**

config ([EsmForMaskedLM](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmForMaskedLM)) : Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the [from_pretrained()](/docs/transformers/v5.6.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained) method to load the model weights.

**Returns:**

`[MaskedLMOutput](/docs/transformers/v5.6.0/ko/main_classes/output#transformers.modeling_outputs.MaskedLMOutput) or `tuple(torch.FloatTensor)``

A [MaskedLMOutput](/docs/transformers/v5.6.0/ko/main_classes/output#transformers.modeling_outputs.MaskedLMOutput) or a tuple of
`torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various
elements depending on the configuration ([EsmConfig](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmConfig)) and inputs.

## EsmForSequenceClassification [[transformers.EsmForSequenceClassification]][[transformers.EsmForSequenceClassification]]

#### transformers.EsmForSequenceClassification[[transformers.EsmForSequenceClassification]]

[Source](https://github.com/huggingface/transformers/blob/v5.6.0/src/transformers/models/esm/modeling_esm.py#L916)

ESM Model transformer with a sequence classification/regression head on top (a linear layer on top of the pooled
output) e.g. for GLUE tasks.

This model inherits from [PreTrainedModel](/docs/transformers/v5.6.0/ko/main_classes/model#transformers.PreTrainedModel). 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 PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass.
Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage
and behavior.

forwardtransformers.EsmForSequenceClassification.forwardhttps://github.com/huggingface/transformers/blob/v5.6.0/src/transformers/models/esm/modeling_esm.py#L927[{"name": "input_ids", "val": ": torch.LongTensor | None = None"}, {"name": "attention_mask", "val": ": torch.Tensor | None = None"}, {"name": "position_ids", "val": ": torch.LongTensor | None = None"}, {"name": "inputs_embeds", "val": ": torch.FloatTensor | None = None"}, {"name": "labels", "val": ": torch.LongTensor | None = None"}, {"name": "**kwargs", "val": ": typing_extensions.Unpack[transformers.utils.generic.TransformersKwargs]"}]- **input_ids** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) --
  Indices of input sequence tokens in the vocabulary. Padding will be ignored by default.

  Indices can be obtained using [AutoTokenizer](/docs/transformers/v5.6.0/ko/model_doc/auto#transformers.AutoTokenizer). See [PreTrainedTokenizer.encode()](/docs/transformers/v5.6.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.encode) and
  [PreTrainedTokenizer.__call__()](/docs/transformers/v5.6.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.__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)
- **position_ids** (`torch.LongTensor` 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.n_positions - 1]`.

  [What are position IDs?](../glossary#position-ids)
- **inputs_embeds** (`torch.FloatTensor` 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.
- **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 regression loss is computed (Mean-Square loss), If
  `config.num_labels > 1` a classification loss is computed (Cross-Entropy).0[SequenceClassifierOutput](/docs/transformers/v5.6.0/ko/main_classes/output#transformers.modeling_outputs.SequenceClassifierOutput) or `tuple(torch.FloatTensor)`A [SequenceClassifierOutput](/docs/transformers/v5.6.0/ko/main_classes/output#transformers.modeling_outputs.SequenceClassifierOutput) or a tuple of
`torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various
elements depending on the configuration ([EsmConfig](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmConfig)) and inputs.
The [EsmForSequenceClassification](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmForSequenceClassification) forward method, overrides the `__call__` special method.

Although the recipe for forward pass needs to be defined within this function, one should call the `Module`
instance afterwards instead of this since the former takes care of running the pre and post processing steps while
the latter silently ignores them.

- **loss** (`torch.FloatTensor` of shape `(1,)`, *optional*, returned when `labels` is provided) -- Classification (or regression if config.num_labels==1) loss.
- **logits** (`torch.FloatTensor` of shape `(batch_size, config.num_labels)`) -- Classification (or regression if config.num_labels==1) scores (before SoftMax).
- **hidden_states** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`) -- Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
  one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`.

  Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
- **attentions** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`) -- Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
  sequence_length)`.

  Attentions weights after the attention softmax, used to compute the weighted average in the self-attention
  heads.

Example of single-label classification:

```python
>>> import torch
>>> from transformers import AutoTokenizer, EsmForSequenceClassification

>>> tokenizer = AutoTokenizer.from_pretrained("facebook/esm-1b")
>>> model = EsmForSequenceClassification.from_pretrained("facebook/esm-1b")

>>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")

>>> with torch.no_grad():
...     logits = model(**inputs).logits

>>> predicted_class_id = logits.argmax().item()
>>> model.config.id2label[predicted_class_id]
...

>>> # To train a model on `num_labels` classes, you can pass `num_labels=num_labels` to `.from_pretrained(...)`
>>> num_labels = len(model.config.id2label)
>>> model = EsmForSequenceClassification.from_pretrained("facebook/esm-1b", num_labels=num_labels)

>>> labels = torch.tensor([1])
>>> loss = model(**inputs, labels=labels).loss
>>> round(loss.item(), 2)
...
```

Example of multi-label classification:

```python
>>> import torch
>>> from transformers import AutoTokenizer, EsmForSequenceClassification

>>> tokenizer = AutoTokenizer.from_pretrained("facebook/esm-1b")
>>> model = EsmForSequenceClassification.from_pretrained("facebook/esm-1b", problem_type="multi_label_classification")

>>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")

>>> with torch.no_grad():
...     logits = model(**inputs).logits

>>> predicted_class_ids = torch.arange(0, logits.shape[-1])[torch.sigmoid(logits).squeeze(dim=0) > 0.5]

>>> # To train a model on `num_labels` classes, you can pass `num_labels=num_labels` to `.from_pretrained(...)`
>>> num_labels = len(model.config.id2label)
>>> model = EsmForSequenceClassification.from_pretrained(
...     "facebook/esm-1b", num_labels=num_labels, problem_type="multi_label_classification"
... )

>>> labels = torch.sum(
...     torch.nn.functional.one_hot(predicted_class_ids[None, :].clone(), num_classes=num_labels), dim=1
... ).to(torch.float)
>>> loss = model(**inputs, labels=labels).loss
```

**Parameters:**

config ([EsmForSequenceClassification](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmForSequenceClassification)) : Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the [from_pretrained()](/docs/transformers/v5.6.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained) method to load the model weights.

**Returns:**

`[SequenceClassifierOutput](/docs/transformers/v5.6.0/ko/main_classes/output#transformers.modeling_outputs.SequenceClassifierOutput) or `tuple(torch.FloatTensor)``

A [SequenceClassifierOutput](/docs/transformers/v5.6.0/ko/main_classes/output#transformers.modeling_outputs.SequenceClassifierOutput) or a tuple of
`torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various
elements depending on the configuration ([EsmConfig](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmConfig)) and inputs.

## EsmForTokenClassification [[transformers.EsmForTokenClassification]][[transformers.EsmForTokenClassification]]

#### transformers.EsmForTokenClassification[[transformers.EsmForTokenClassification]]

[Source](https://github.com/huggingface/transformers/blob/v5.6.0/src/transformers/models/esm/modeling_esm.py#L989)

The Esm transformer with a token classification head on top (a linear layer on top of the hidden-states
output) e.g. for Named-Entity-Recognition (NER) tasks.

This model inherits from [PreTrainedModel](/docs/transformers/v5.6.0/ko/main_classes/model#transformers.PreTrainedModel). 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 PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass.
Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage
and behavior.

forwardtransformers.EsmForTokenClassification.forwardhttps://github.com/huggingface/transformers/blob/v5.6.0/src/transformers/models/esm/modeling_esm.py#L1000[{"name": "input_ids", "val": ": torch.LongTensor | None = None"}, {"name": "attention_mask", "val": ": torch.Tensor | None = None"}, {"name": "position_ids", "val": ": torch.LongTensor | None = None"}, {"name": "inputs_embeds", "val": ": torch.FloatTensor | None = None"}, {"name": "labels", "val": ": torch.LongTensor | None = None"}, {"name": "**kwargs", "val": ": typing_extensions.Unpack[transformers.utils.generic.TransformersKwargs]"}]- **input_ids** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) --
  Indices of input sequence tokens in the vocabulary. Padding will be ignored by default.

  Indices can be obtained using [AutoTokenizer](/docs/transformers/v5.6.0/ko/model_doc/auto#transformers.AutoTokenizer). See [PreTrainedTokenizer.encode()](/docs/transformers/v5.6.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.encode) and
  [PreTrainedTokenizer.__call__()](/docs/transformers/v5.6.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.__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)
- **position_ids** (`torch.LongTensor` 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.n_positions - 1]`.

  [What are position IDs?](../glossary#position-ids)
- **inputs_embeds** (`torch.FloatTensor` 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.
- **labels** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) --
  Labels for computing the token classification loss. Indices should be in `[0, ..., config.num_labels - 1]`.0[TokenClassifierOutput](/docs/transformers/v5.6.0/ko/main_classes/output#transformers.modeling_outputs.TokenClassifierOutput) or `tuple(torch.FloatTensor)`A [TokenClassifierOutput](/docs/transformers/v5.6.0/ko/main_classes/output#transformers.modeling_outputs.TokenClassifierOutput) or a tuple of
`torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various
elements depending on the configuration ([EsmConfig](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmConfig)) and inputs.
The [EsmForTokenClassification](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmForTokenClassification) forward method, overrides the `__call__` special method.

Although the recipe for forward pass needs to be defined within this function, one should call the `Module`
instance afterwards instead of this since the former takes care of running the pre and post processing steps while
the latter silently ignores them.

- **loss** (`torch.FloatTensor` of shape `(1,)`, *optional*, returned when `labels` is provided) -- Classification loss.
- **logits** (`torch.FloatTensor` of shape `(batch_size, sequence_length, config.num_labels)`) -- Classification scores (before SoftMax).
- **hidden_states** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`) -- Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
  one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`.

  Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
- **attentions** (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`) -- Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
  sequence_length)`.

  Attentions weights after the attention softmax, used to compute the weighted average in the self-attention
  heads.

Example:

```python
>>> from transformers import AutoTokenizer, EsmForTokenClassification
>>> import torch

>>> tokenizer = AutoTokenizer.from_pretrained("facebook/esm-1b")
>>> model = EsmForTokenClassification.from_pretrained("facebook/esm-1b")

>>> inputs = tokenizer(
...     "HuggingFace is a company based in Paris and New York", add_special_tokens=False, return_tensors="pt"
... )

>>> with torch.no_grad():
...     logits = model(**inputs).logits

>>> predicted_token_class_ids = logits.argmax(-1)

>>> # Note that tokens are classified rather then input words which means that
>>> # there might be more predicted token classes than words.
>>> # Multiple token classes might account for the same word
>>> predicted_tokens_classes = [model.config.id2label[t.item()] for t in predicted_token_class_ids[0]]
>>> predicted_tokens_classes
...

>>> labels = predicted_token_class_ids
>>> loss = model(**inputs, labels=labels).loss
>>> round(loss.item(), 2)
...
```

**Parameters:**

config ([EsmForTokenClassification](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmForTokenClassification)) : Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the [from_pretrained()](/docs/transformers/v5.6.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained) method to load the model weights.

**Returns:**

`[TokenClassifierOutput](/docs/transformers/v5.6.0/ko/main_classes/output#transformers.modeling_outputs.TokenClassifierOutput) or `tuple(torch.FloatTensor)``

A [TokenClassifierOutput](/docs/transformers/v5.6.0/ko/main_classes/output#transformers.modeling_outputs.TokenClassifierOutput) or a tuple of
`torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various
elements depending on the configuration ([EsmConfig](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmConfig)) and inputs.

## EsmForProteinFolding [[transformers.EsmForProteinFolding]][[transformers.EsmForProteinFolding]]

#### transformers.EsmForProteinFolding[[transformers.EsmForProteinFolding]]

[Source](https://github.com/huggingface/transformers/blob/v5.6.0/src/transformers/models/esm/modeling_esmfold.py#L1967)

ESMForProteinFolding is the HuggingFace port of the original ESMFold model. It consists of an ESM-2 "stem" followed
by a protein folding "head", although unlike most other output heads, this "head" is similar in size and runtime to
the rest of the model combined! It outputs a dictionary containing predicted structural information about the input
protein(s).

This model inherits from [PreTrainedModel](/docs/transformers/v5.6.0/ko/main_classes/model#transformers.PreTrainedModel). 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 PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass.
Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage
and behavior.

forwardtransformers.EsmForProteinFolding.forwardhttps://github.com/huggingface/transformers/blob/v5.6.0/src/transformers/models/esm/modeling_esmfold.py#L2043[{"name": "input_ids", "val": ": Tensor"}, {"name": "attention_mask", "val": ": torch.Tensor | None = None"}, {"name": "position_ids", "val": ": torch.Tensor | None = None"}, {"name": "masking_pattern", "val": ": torch.Tensor | None = None"}, {"name": "num_recycles", "val": ": int | None = None"}, {"name": "output_hidden_states", "val": ": bool | None = False"}, {"name": "**kwargs", "val": ""}]- **input_ids** (`torch.Tensor` of shape `(batch_size, sequence_length)`) --
  Indices of input sequence tokens in the vocabulary. Padding will be ignored by default.

  Indices can be obtained using [AutoTokenizer](/docs/transformers/v5.6.0/ko/model_doc/auto#transformers.AutoTokenizer). See [PreTrainedTokenizer.encode()](/docs/transformers/v5.6.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.encode) and
  [PreTrainedTokenizer.__call__()](/docs/transformers/v5.6.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.__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)
- **position_ids** (`torch.Tensor` 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.n_positions - 1]`.

  [What are position IDs?](../glossary#position-ids)
- **masking_pattern** (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*) --
  Locations of tokens to mask during training as a form of regularization. Mask values selected in `[0, 1]`.
- **num_recycles** (`int`, *optional*, defaults to `None`) --
  Number of times to recycle the input sequence. If `None`, defaults to `config.num_recycles`. "Recycling"
  consists of passing the output of the folding trunk back in as input to the trunk. During training, the
  number of recycles should vary with each batch, to ensure that the model learns to output valid predictions
  after each recycle. During inference, num_recycles should be set to the highest value that the model was
  trained with for maximum accuracy. Accordingly, when this value is set to `None`, config.max_recycles is
  used.
- **output_hidden_states** (`bool`, *optional*, defaults to `False`) --
  Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
  more detail.0`EsmForProteinFoldingOutput` or `tuple(torch.FloatTensor)`A `EsmForProteinFoldingOutput` or a tuple of
`torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various
elements depending on the configuration ([EsmConfig](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmConfig)) and inputs.
The [EsmForProteinFolding](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmForProteinFolding) forward method, overrides the `__call__` special method.

Although the recipe for forward pass needs to be defined within this function, one should call the `Module`
instance afterwards instead of this since the former takes care of running the pre and post processing steps while
the latter silently ignores them.

- **frames** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Output frames.
- **sidechain_frames** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Output sidechain frames.
- **unnormalized_angles** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Predicted unnormalized backbone and side chain torsion angles.
- **angles** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Predicted backbone and side chain torsion angles.
- **positions** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Predicted positions of the backbone and side chain atoms.
- **states** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Hidden states from the protein folding trunk.
- **s_s** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Per-residue embeddings derived by concatenating the hidden states of each layer of the ESM-2 LM stem.
- **s_z** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Pairwise residue embeddings.
- **distogram_logits** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Input logits to the distogram used to compute residue distances.
- **lm_logits** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Logits output by the ESM-2 protein language model stem.
- **aatype** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Input amino acids (AlphaFold2 indices).
- **atom14_atom_exists** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Whether each atom exists in the atom14 representation.
- **residx_atom14_to_atom37** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Mapping between atoms in the atom14 and atom37 representations.
- **residx_atom37_to_atom14** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Mapping between atoms in the atom37 and atom14 representations.
- **atom37_atom_exists** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Whether each atom exists in the atom37 representation.
- **residue_index** (`torch.FloatTensor`, *optional*, defaults to `None`) -- The index of each residue in the protein chain. Unless internal padding tokens are used, this will just be
  a sequence of integers from 0 to `sequence_length`.
- **lddt_head** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Raw outputs from the lddt head used to compute plddt.
- **plddt** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Per-residue confidence scores. Regions of low confidence may indicate areas where the model's prediction is
  uncertain, or where the protein structure is disordered.
- **ptm_logits** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Raw logits used for computing ptm.
- **ptm** (`torch.FloatTensor`, *optional*, defaults to `None`) -- TM-score output representing the model's high-level confidence in the overall structure.
- **aligned_confidence_probs** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Per-residue confidence scores for the aligned structure.
- **predicted_aligned_error** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Predicted error between the model's prediction and the ground truth.
- **max_predicted_aligned_error** (`torch.FloatTensor`, *optional*, defaults to `None`) -- Per-sample maximum predicted error.

Example:

```python
>>> from transformers import AutoTokenizer, EsmForProteinFolding

>>> model = EsmForProteinFolding.from_pretrained("facebook/esmfold_v1")
>>> tokenizer = AutoTokenizer.from_pretrained("facebook/esmfold_v1")
>>> inputs = tokenizer(["MLKNVQVQLV"], return_tensors="pt", add_special_tokens=False)  # A tiny random peptide
>>> outputs = model(**inputs)
>>> folded_positions = outputs.positions
```

**Parameters:**

config ([EsmForProteinFolding](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmForProteinFolding)) : Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the [from_pretrained()](/docs/transformers/v5.6.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained) method to load the model weights.

**Returns:**

``EsmForProteinFoldingOutput` or `tuple(torch.FloatTensor)``

A `EsmForProteinFoldingOutput` or a tuple of
`torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various
elements depending on the configuration ([EsmConfig](/docs/transformers/v5.6.0/ko/model_doc/esm#transformers.EsmConfig)) and inputs.