Api

ESM2Config dataclass

基类: ESM2GenericConfig, IOMixinWithGettersSetters

ESM2 模型的配置类。

源代码位于 bionemo/esm2/model/model.py

@dataclass
class ESM2Config(ESM2GenericConfig, iom.IOMixinWithGettersSetters):
    """Configuration class for ESM2 model."""

    model_cls: Type[ESM2Model] = ESM2Model
    num_layers: int = 33  # 650M
    hidden_size: int = 1280  # 650M

ESM2GenericConfig dataclass

基类: BioBertConfig[ESM2ModelT, MegatronLossType]

ESM2 模型的配置类。

属性

名称	类型	描述
num_layers	int	模型中的层数。
hidden_size	int	模型的隐藏层大小。
num_attention_heads	int	模型中的注意力头数。
ffn_hidden_size	int	前馈网络的隐藏层大小。
hidden_dropout	float	隐藏层的 Dropout 率。
attention_dropout	float	注意力层的 Dropout 率。
apply_residual_connection_post_layernorm	bool	是否在层归一化后应用残差连接。
layernorm_epsilon	float	层归一化的 Epsilon 值。
layernorm_zero_centered_gamma	float	是否在层归一化中将 gamma 参数零中心化。
activation_func	Callable	模型中使用的激活函数。
init_method_std	float	权重初始化的标准差。
apply_query_key_layer_scaling	float	是否对查询和键层应用缩放。
masked_softmax_fusion	float	是否使用将注意力 softmax 与其掩码融合的内核。
fp16_lm_cross_entropy	bool	是否将用于语言模型头的交叉熵未缩减损失计算移至 fp16。
share_embeddings_and_output_weights	bool	是否共享嵌入和输出权重。
enable_autocast	bool	是否为混合精度启用自动类型转换。
biobert_spec_option	BiobertSpecOption	模型的 BiobertSpecOption。
position_embedding_type	PositionEmbeddingKinds	模型中使用的位置嵌入类型。
seq_length	int	输入序列的长度。
make_vocab_size_divisible_by	int	使词汇表大小可被此值整除。
token_dropout	bool	是否应用 token dropout。
use_attention_mask	bool	是否使用注意力掩码。
use_esm_attention	bool	是否使用 ESM 注意力。
attention_softmax_in_fp32	bool	是否对注意力 softmax 使用 fp32。
optimizer_fn	Optional[Callable[[MegatronBioBertModel], Optimizer]]	模型的可选优化器函数。
parallel_output	bool	是否使用并行输出。
rotary_base	int	旋转位置编码的基值。
rotary_percent	float	旋转位置编码的百分比。
seq_len_interpolation_factor	Optional[float]	序列长度的插值因子。
get_attention_mask_from_fusion	Optional[float]	是否从融合中获取注意力掩码。
nemo1_ckpt_path	str | None	NEMO1 检查点的路径。
return_only_hidden_states	bool	是否仅返回隐藏状态。
loss_reduction_class	bool	模型的损失缩减类。默认为 BERTMLMLossWithReduction。

源代码位于 bionemo/esm2/model/model.py

@dataclass
class ESM2GenericConfig(BioBertConfig[ESM2ModelT, MegatronLossType]):
    """Configuration class for ESM2 model.

    Attributes:
        num_layers: Number of layers in the model.
        hidden_size: Hidden size of the model.
        num_attention_heads: Number of attention heads in the model.
        ffn_hidden_size: Hidden size of the feed-forward network.
        hidden_dropout: Dropout rate for hidden layers.
        attention_dropout: Dropout rate for attention layers.
        apply_residual_connection_post_layernorm: Whether to apply residual connection after layer normalization.
        layernorm_epsilon: Epsilon value for layer normalization.
        layernorm_zero_centered_gamma: Whether to zero-center the gamma parameter in layer normalization.
        activation_func: Activation function used in the model.
        init_method_std: Standard deviation for weight initialization.
        apply_query_key_layer_scaling: Whether to apply scaling to query and key layers.
        masked_softmax_fusion: Whether to use a kernel that fuses attention softmax with its mask.
        fp16_lm_cross_entropy: Whether to move the cross entropy unreduced loss calculation for lm head to fp16.
        share_embeddings_and_output_weights: Whether to share embeddings and output weights.
        enable_autocast: Whether to enable autocast for mixed precision.
        biobert_spec_option: BiobertSpecOption for the model.
        position_embedding_type: Type of position embedding used in the model.
        seq_length: Length of the input sequence.
        make_vocab_size_divisible_by: Make the vocabulary size divisible by this value.
        token_dropout: Whether to apply token dropout.
        use_attention_mask: Whether to use attention mask.
        use_esm_attention: Whether to use ESM attention.
        attention_softmax_in_fp32: Whether to use fp32 for attention softmax.
        optimizer_fn: Optional optimizer function for the model.
        parallel_output: Whether to use parallel output.
        rotary_base: Base value for rotary positional encoding.
        rotary_percent: Percentage of rotary positional encoding.
        seq_len_interpolation_factor: Interpolation factor for sequence length.
        get_attention_mask_from_fusion: Whether to get attention mask from fusion.
        nemo1_ckpt_path: Path to NEMO1 checkpoint.
        return_only_hidden_states: Whether to return only hidden states.
        loss_reduction_class: Loss reduction class for the model. Default to BERTMLMLossWithReduction.
    """

    # When overriding fields in a dataclass _always_ declare types: https://github.com/python/cpython/issues/123269
    model_cls: Type[ESM2ModelT] = ESM2Model
    num_layers: int = 33  # 650M
    hidden_size: int = 1280  # 650M
    num_attention_heads: int = 20
    ffn_hidden_size: int = 4 * 1280  # Transformer FFN hidden size. Usually 4 * hidden_size.
    hidden_dropout: float = 0  # ESM2 removes dropout from hidden layers and attention
    attention_dropout: float = 0.0  # ESM2 does not use attention dropout
    apply_residual_connection_post_layernorm: bool = False  # TODO: farhadr False is new default, True was BERT pub.
    layernorm_epsilon: float = 1.0e-5
    bias_activation_fusion: bool = True  # True degrades accuracy slightly, but is faster.
    activation_func: Callable = F.gelu  # esm_gelu_func  # ESM2 MLP
    init_method_std: float = 0.02
    softmax_scale: float = 1.0

    # embedding
    token_dropout: bool = True
    use_attention_mask: bool = True

    # core attention
    use_esm_attention: bool = False  # Skip ESM2 custom attention for TE acceleration. Still passes golden value test.
    attention_softmax_in_fp32: bool = False
    normalize_attention_scores: bool = False

    # From megatron.core.models.gpt.bert_model.GPTModel
    fp16_lm_cross_entropy: bool = False  # Move the cross entropy unreduced loss calculation for lm head to fp16
    parallel_output: bool = True
    share_embeddings_and_output_weights: bool = True
    make_vocab_size_divisible_by: int = 128
    position_embedding_type: PositionEmbeddingKinds = "rope"  # ESM2 uses relative positional encoding 'ROPE' to extrapolate to longer sequences unseen during training
    rotary_base: int = 10000
    rotary_percent: float = 1.0
    seq_len_interpolation_factor: Optional[float] = None
    seq_length: int = 1024
    biobert_spec_option: BiobertSpecOption = BiobertSpecOption.esm2_bert_layer_with_transformer_engine_spec

    optimizer_fn: Optional[Callable[[MegatronBioBertModel], Optimizer]] = None
    # TODO (@skothenhill,@georgea) update to use the nemo2 checkpoint mixins
    #  support HF (requires weight interleaving on qkv layer) and nemo1 checkpoints ideally.
    nemo1_ckpt_path: str | None = None
    # The following checkpoint path is for nemo2 checkpoints. Config parameters not present in
    #  self.override_parent_fields will be loaded from the checkpoint and override those values here.
    initial_ckpt_path: str | None = None
    # TODO (@jstjohn) come up with a cleaner way in the biobert module to return user requested
    #  things as part of the workflow for inference and fine-tuning.
    return_embeddings: bool = False
    include_embeddings: bool = False
    include_input_ids: bool = False
    skip_logits: bool = False
    return_only_hidden_states: bool = False  # return logits

    def __post_init__(self):
        # TODO, as a validator?
        """Check configuration compatibility."""
        # reset moe_token_dispatcher_type when variable_seq_lengths is True.
        # must be performed before super().__post_init__()
        if self.variable_seq_lengths and self.moe_token_dispatcher_type in ["allgather", "alltoall_seq"]:
            logging.warning(
                "MoE token dispatcher type 'allgather' and 'alltoall_seq' are not supported with variable sequence lengths. Setting moe_token_dispatcher_type to 'alltoall'."
            )
            self.moe_token_dispatcher_type = "alltoall"

        # reset apply_query_key_layer_scaling based on biobert_spec_option
        super().__post_init__()
        if self.biobert_spec_option == BiobertSpecOption.esm2_bert_layer_with_transformer_engine_spec:
            self.apply_query_key_layer_scaling = False
        elif self.biobert_spec_option == BiobertSpecOption.esm2_bert_layer_local_spec:
            logging.warning(
                "BiobertSpecOption.esm2_bert_layer_local_spec is depreciated. Use BiobertSpecOption.esm2_bert_layer_with_transformer_engine_spec instead."
            )
            self.apply_query_key_layer_scaling = True
        else:
            raise ValueError(f"Unknown biobert_spec_option: {self.biobert_spec_option}")

__post_init__()

检查配置兼容性。

源代码位于 bionemo/esm2/model/model.py

def __post_init__(self):
    # TODO, as a validator?
    """Check configuration compatibility."""
    # reset moe_token_dispatcher_type when variable_seq_lengths is True.
    # must be performed before super().__post_init__()
    if self.variable_seq_lengths and self.moe_token_dispatcher_type in ["allgather", "alltoall_seq"]:
        logging.warning(
            "MoE token dispatcher type 'allgather' and 'alltoall_seq' are not supported with variable sequence lengths. Setting moe_token_dispatcher_type to 'alltoall'."
        )
        self.moe_token_dispatcher_type = "alltoall"

    # reset apply_query_key_layer_scaling based on biobert_spec_option
    super().__post_init__()
    if self.biobert_spec_option == BiobertSpecOption.esm2_bert_layer_with_transformer_engine_spec:
        self.apply_query_key_layer_scaling = False
    elif self.biobert_spec_option == BiobertSpecOption.esm2_bert_layer_local_spec:
        logging.warning(
            "BiobertSpecOption.esm2_bert_layer_local_spec is depreciated. Use BiobertSpecOption.esm2_bert_layer_with_transformer_engine_spec instead."
        )
        self.apply_query_key_layer_scaling = True
    else:
        raise ValueError(f"Unknown biobert_spec_option: {self.biobert_spec_option}")

ESM2Model

基类: MegatronBioBertModel

ESM2 Transformer 语言模型。

源代码位于 bionemo/esm2/model/model.py

class ESM2Model(MegatronBioBertModel):
    """ESM2 Transformer language model."""

    def __init__(
        self,
        config: TransformerConfig,
        num_tokentypes: int,
        transformer_layer_spec: spec_utils.ModuleSpec,
        vocab_size: int,
        max_sequence_length: int,
        tokenizer: Optional[BioNeMoESMTokenizer] = None,
        pre_process: bool = True,
        post_process: bool = True,
        fp16_lm_cross_entropy: bool = False,
        parallel_output: bool = True,
        share_embeddings_and_output_weights: bool = False,
        position_embedding_type: Literal["learned_absolute", "rope"] = "learned_absolute",
        rotary_percent: float = 1.0,
        seq_len_interpolation_factor: Optional[float] = None,
        add_binary_head: bool = True,
        return_embeddings: bool = False,
        include_embeddings: bool = False,
        include_input_ids: bool = False,
        use_full_attention_mask: bool = False,
        include_hiddens: bool = False,
        skip_logits: bool = False,
    ) -> None:
        """Initialize the ESM2 model.

        Args:
            config (TransformerConfig): transformer config
            num_tokentypes (int): Set to 2 when args.bert_binary_head is True, and 0 otherwise. Defaults to 0.
            transformer_layer_spec (ModuleSpec): Specifies module to use for transformer layers
            vocab_size (int): vocabulary size
            max_sequence_length (int): maximum size of sequence. This is used for positional embedding
            tokenizer (AutoTokenizer): optional tokenizer object (currently only used in the constructor of ESM2Model)
            pre_process (bool): Include embedding layer (used with pipeline parallelism)
            post_process (bool): Include an output layer (used with pipeline parallelism)
            fp16_lm_cross_entropy: Whether to move the cross entropy unreduced loss calculation for lm head to fp16.
            parallel_output (bool): Do not gather the outputs, keep them split across tensor parallel ranks
            share_embeddings_and_output_weights (bool): When True, input embeddings and output logit weights are shared. Defaults to False.
            position_embedding_type (string): Position embedding type. Options ['learned_absolute', 'rope'].
                Defaults is 'learned_absolute'.
            rotary_percent (float): Percent of rotary dimension to use for rotary position embeddings.
                Defaults to 1.0 (100%). Ignored unless position_embedding_type is 'rope'.
            seq_len_interpolation_factor (Optional[float]): Interpolation factor for sequence length. Defaults to None.
            add_binary_head (bool): Whether to add a binary head. Defaults to True.
            return_embeddings (bool): Whether to return embeddings. Defaults to False.
            include_embeddings (bool): Whether to include embeddings in the output dictionary. Defaults to False.
            include_input_ids (bool): Whether to include input_ids in the output dictionary. Defaults to False.
            use_full_attention_mask (bool): Whether to use full attention mask. Defaults to False.
            include_hiddens (bool): Whether to include hidden states in the output dictionary. Defaults to False.
            skip_logits (bool): Skip writing the token logits in output dict
        """
        super(MegatronBioBertModel, self).__init__(config=config)
        self.post_process = post_process
        self.add_binary_head = add_binary_head
        if return_embeddings:
            assert self.post_process, "only return embeddings on the last pipeline stage"
        # `b` = batch, `s` = sequence.
        # The old flash attention mechanism apparently wants you to use a b x 1 x s x s attention mask while
        #  the new one wants a b x 1 x 1 x s attention mask. This is a hack to allow us to switch between the two.
        self.use_full_attention_mask = use_full_attention_mask
        self.config: TransformerConfig = config
        self.transformer_layer_spec: spec_utils.ModuleSpec = transformer_layer_spec
        self.vocab_size = vocab_size
        self.max_sequence_length = max_sequence_length
        self.pre_process = pre_process
        self.post_process = post_process
        self.fp16_lm_cross_entropy = fp16_lm_cross_entropy
        self.parallel_output = parallel_output
        self.share_embeddings_and_output_weights = share_embeddings_and_output_weights
        self.position_embedding_type = position_embedding_type
        self.add_binary_head = add_binary_head
        self.return_embeddings = return_embeddings
        self.include_embeddings = include_embeddings
        self.include_hiddens = include_hiddens
        self.include_input_ids = include_input_ids
        self.skip_logits = skip_logits

        # megatron core pipelining currently depends on model type
        self.model_type = ModelType.encoder_or_decoder

        # Embeddings.
        if self.pre_process:
            self.register_buffer(
                "bert_position_id_tensor",
                torch.arange(max_sequence_length, dtype=torch.long, requires_grad=False).unsqueeze(0),
                persistent=False,
            )
            # ESM2 Customization: ESM2Embedding instead of LanguageModelEmbedding
            # TODO: call super, overwrite the self.embedding, and setup_embeddings_and_output_layer in constructor.
            # Note: need to avoid calling setup twice: skip with super (super(skip_setup=True))
            self.embedding = ESM2Embedding(
                config=self.config,
                vocab_size=self.vocab_size,
                max_sequence_length=self.max_sequence_length,
                position_embedding_type=position_embedding_type,
                num_tokentypes=num_tokentypes,
                # ESM2 NEW ARGS
                token_dropout=self.config.token_dropout,
                use_attention_mask=self.config.use_attention_mask,
                mask_token_id=tokenizer.mask_token_id,
            )

        if self.position_embedding_type == "rope":
            self.rotary_pos_emb = RotaryEmbedding(
                kv_channels=self.config.kv_channels,
                rotary_percent=rotary_percent,
                rotary_interleaved=self.config.rotary_interleaved,
                seq_len_interpolation_factor=seq_len_interpolation_factor,
            )

        # Transformer.
        self.encoder = TransformerBlock(
            config=self.config,
            spec=self.transformer_layer_spec,
            pre_process=self.pre_process,
            post_process=self.post_process,
        )

        # Output
        if post_process:
            # TODO: Make sure you are passing in the mpu_vocab_size properly
            self.lm_head = BertLMHead(
                config.hidden_size,
                config,
            )

            self.output_layer = tensor_parallel.ColumnParallelLinear(
                config.hidden_size,
                self.vocab_size,
                config=config,
                init_method=config.init_method,
                bias=True,
                skip_bias_add=False,
                gather_output=not self.parallel_output,
                skip_weight_param_allocation=pre_process and share_embeddings_and_output_weights,
            )

            self.binary_head = None
            if self.add_binary_head:
                # TODO: Shoudl switch this to TE ?
                self.binary_head = get_linear_layer(
                    config.hidden_size, 2, config.init_method, config.perform_initialization
                )

                self.pooler = Pooler(config.hidden_size, config.init_method, config, config.sequence_parallel)
        if self.pre_process or self.post_process:
            self.setup_embeddings_and_output_layer()

    def embedding_forward(
        self, input_ids: Tensor, position_ids: Tensor, tokentype_ids: Tensor = None, attention_mask: Tensor = None
    ):
        """Forward pass of the embedding layer.

        Args:
            input_ids: The input tensor of shape (batch_size, sequence_length) containing the input IDs.
            position_ids: The tensor of shape (batch_size, sequence_length) containing the position IDs.
            tokentype_ids: The tensor of shape (batch_size, sequence_length) containing the token type IDs. Defaults to None.
            attention_mask: The tensor of shape (batch_size, sequence_length) containing the attention mask. Defaults to None.

        Returns:
            Tensor: The output tensor of shape (batch_size, sequence_length, hidden_size) containing the embedded representations.
        """
        # ESM2 Customization: ESM2Embedding forward takes attention_mask
        # in addition to the args required by LanguageModelEmbedding
        return self.embedding(
            input_ids=input_ids, position_ids=position_ids, tokentype_ids=tokentype_ids, attention_mask=attention_mask
        )

__init__(config, num_tokentypes, transformer_layer_spec, vocab_size, max_sequence_length, tokenizer=None, pre_process=True, post_process=True, fp16_lm_cross_entropy=False, parallel_output=True, share_embeddings_and_output_weights=False, position_embedding_type='learned_absolute', rotary_percent=1.0, seq_len_interpolation_factor=None, add_binary_head=True, return_embeddings=False, include_embeddings=False, include_input_ids=False, use_full_attention_mask=False, include_hiddens=False, skip_logits=False)

初始化 ESM2 模型。

参数

名称	类型	描述	默认值
config	TransformerConfig	Transformer 配置	必需
num_tokentypes	int	当 args.bert_binary_head 为 True 时设置为 2，否则设置为 0。默认为 0。	必需
transformer_layer_spec	ModuleSpec	指定用于 Transformer 层的模块	必需
vocab_size	int	词汇表大小	必需
max_sequence_length	int	序列的最大大小。这用于位置嵌入	必需
tokenizer	AutoTokenizer	可选的 tokenizer 对象（目前仅在 ESM2Model 的构造函数中使用）	None
pre_process	bool	包含嵌入层（与流水线并行一起使用）	True
post_process	bool	包含输出层（与流水线并行一起使用）	True
fp16_lm_cross_entropy	bool	是否将用于语言模型头的交叉熵未缩减损失计算移至 fp16。	False
parallel_output	bool	不收集输出，保持它们在张量并行 ranks 中拆分	True
share_embeddings_and_output_weights	bool	当为 True 时，输入嵌入和输出 logits 权重是共享的。默认为 False。	False
position_embedding_type	string	位置嵌入类型。选项 ['learned_absolute', 'rope']。默认值为 'learned_absolute'。	'learned_absolute'
rotary_percent	float	用于旋转位置嵌入的旋转维度的百分比。默认为 1.0 (100%)。除非 position_embedding_type 为 'rope'，否则忽略。	1.0
seq_len_interpolation_factor	Optional[float]	序列长度的插值因子。默认为 None。	None
add_binary_head	bool	是否添加二元头。默认为 True。	True
return_embeddings	bool	是否返回嵌入。默认为 False。	False
include_embeddings	bool	是否在输出字典中包含嵌入。默认为 False。	False
include_input_ids	bool	是否在输出字典中包含 input_ids。默认为 False。	False
use_full_attention_mask	bool	是否使用完整注意力掩码。默认为 False。	False
include_hiddens	bool	是否在输出字典中包含隐藏状态。默认为 False。	False
skip_logits	bool	跳过在输出字典中写入 token logits	False

源代码位于 bionemo/esm2/model/model.py

def __init__(
    self,
    config: TransformerConfig,
    num_tokentypes: int,
    transformer_layer_spec: spec_utils.ModuleSpec,
    vocab_size: int,
    max_sequence_length: int,
    tokenizer: Optional[BioNeMoESMTokenizer] = None,
    pre_process: bool = True,
    post_process: bool = True,
    fp16_lm_cross_entropy: bool = False,
    parallel_output: bool = True,
    share_embeddings_and_output_weights: bool = False,
    position_embedding_type: Literal["learned_absolute", "rope"] = "learned_absolute",
    rotary_percent: float = 1.0,
    seq_len_interpolation_factor: Optional[float] = None,
    add_binary_head: bool = True,
    return_embeddings: bool = False,
    include_embeddings: bool = False,
    include_input_ids: bool = False,
    use_full_attention_mask: bool = False,
    include_hiddens: bool = False,
    skip_logits: bool = False,
) -> None:
    """Initialize the ESM2 model.

    Args:
        config (TransformerConfig): transformer config
        num_tokentypes (int): Set to 2 when args.bert_binary_head is True, and 0 otherwise. Defaults to 0.
        transformer_layer_spec (ModuleSpec): Specifies module to use for transformer layers
        vocab_size (int): vocabulary size
        max_sequence_length (int): maximum size of sequence. This is used for positional embedding
        tokenizer (AutoTokenizer): optional tokenizer object (currently only used in the constructor of ESM2Model)
        pre_process (bool): Include embedding layer (used with pipeline parallelism)
        post_process (bool): Include an output layer (used with pipeline parallelism)
        fp16_lm_cross_entropy: Whether to move the cross entropy unreduced loss calculation for lm head to fp16.
        parallel_output (bool): Do not gather the outputs, keep them split across tensor parallel ranks
        share_embeddings_and_output_weights (bool): When True, input embeddings and output logit weights are shared. Defaults to False.
        position_embedding_type (string): Position embedding type. Options ['learned_absolute', 'rope'].
            Defaults is 'learned_absolute'.
        rotary_percent (float): Percent of rotary dimension to use for rotary position embeddings.
            Defaults to 1.0 (100%). Ignored unless position_embedding_type is 'rope'.
        seq_len_interpolation_factor (Optional[float]): Interpolation factor for sequence length. Defaults to None.
        add_binary_head (bool): Whether to add a binary head. Defaults to True.
        return_embeddings (bool): Whether to return embeddings. Defaults to False.
        include_embeddings (bool): Whether to include embeddings in the output dictionary. Defaults to False.
        include_input_ids (bool): Whether to include input_ids in the output dictionary. Defaults to False.
        use_full_attention_mask (bool): Whether to use full attention mask. Defaults to False.
        include_hiddens (bool): Whether to include hidden states in the output dictionary. Defaults to False.
        skip_logits (bool): Skip writing the token logits in output dict
    """
    super(MegatronBioBertModel, self).__init__(config=config)
    self.post_process = post_process
    self.add_binary_head = add_binary_head
    if return_embeddings:
        assert self.post_process, "only return embeddings on the last pipeline stage"
    # `b` = batch, `s` = sequence.
    # The old flash attention mechanism apparently wants you to use a b x 1 x s x s attention mask while
    #  the new one wants a b x 1 x 1 x s attention mask. This is a hack to allow us to switch between the two.
    self.use_full_attention_mask = use_full_attention_mask
    self.config: TransformerConfig = config
    self.transformer_layer_spec: spec_utils.ModuleSpec = transformer_layer_spec
    self.vocab_size = vocab_size
    self.max_sequence_length = max_sequence_length
    self.pre_process = pre_process
    self.post_process = post_process
    self.fp16_lm_cross_entropy = fp16_lm_cross_entropy
    self.parallel_output = parallel_output
    self.share_embeddings_and_output_weights = share_embeddings_and_output_weights
    self.position_embedding_type = position_embedding_type
    self.add_binary_head = add_binary_head
    self.return_embeddings = return_embeddings
    self.include_embeddings = include_embeddings
    self.include_hiddens = include_hiddens
    self.include_input_ids = include_input_ids
    self.skip_logits = skip_logits

    # megatron core pipelining currently depends on model type
    self.model_type = ModelType.encoder_or_decoder

    # Embeddings.
    if self.pre_process:
        self.register_buffer(
            "bert_position_id_tensor",
            torch.arange(max_sequence_length, dtype=torch.long, requires_grad=False).unsqueeze(0),
            persistent=False,
        )
        # ESM2 Customization: ESM2Embedding instead of LanguageModelEmbedding
        # TODO: call super, overwrite the self.embedding, and setup_embeddings_and_output_layer in constructor.
        # Note: need to avoid calling setup twice: skip with super (super(skip_setup=True))
        self.embedding = ESM2Embedding(
            config=self.config,
            vocab_size=self.vocab_size,
            max_sequence_length=self.max_sequence_length,
            position_embedding_type=position_embedding_type,
            num_tokentypes=num_tokentypes,
            # ESM2 NEW ARGS
            token_dropout=self.config.token_dropout,
            use_attention_mask=self.config.use_attention_mask,
            mask_token_id=tokenizer.mask_token_id,
        )

    if self.position_embedding_type == "rope":
        self.rotary_pos_emb = RotaryEmbedding(
            kv_channels=self.config.kv_channels,
            rotary_percent=rotary_percent,
            rotary_interleaved=self.config.rotary_interleaved,
            seq_len_interpolation_factor=seq_len_interpolation_factor,
        )

    # Transformer.
    self.encoder = TransformerBlock(
        config=self.config,
        spec=self.transformer_layer_spec,
        pre_process=self.pre_process,
        post_process=self.post_process,
    )

    # Output
    if post_process:
        # TODO: Make sure you are passing in the mpu_vocab_size properly
        self.lm_head = BertLMHead(
            config.hidden_size,
            config,
        )

        self.output_layer = tensor_parallel.ColumnParallelLinear(
            config.hidden_size,
            self.vocab_size,
            config=config,
            init_method=config.init_method,
            bias=True,
            skip_bias_add=False,
            gather_output=not self.parallel_output,
            skip_weight_param_allocation=pre_process and share_embeddings_and_output_weights,
        )

        self.binary_head = None
        if self.add_binary_head:
            # TODO: Shoudl switch this to TE ?
            self.binary_head = get_linear_layer(
                config.hidden_size, 2, config.init_method, config.perform_initialization
            )

            self.pooler = Pooler(config.hidden_size, config.init_method, config, config.sequence_parallel)
    if self.pre_process or self.post_process:
        self.setup_embeddings_and_output_layer()

embedding_forward(input_ids, position_ids, tokentype_ids=None, attention_mask=None)

嵌入层的前向传播。

参数

名称	类型	描述	默认值
input_ids	Tensor	形状为 (batch_size, sequence_length) 的输入张量，包含输入 ID。	必需
position_ids	Tensor	形状为 (batch_size, sequence_length) 的张量，包含位置 ID。	必需
tokentype_ids	Tensor	形状为 (batch_size, sequence_length) 的张量，包含 token 类型 ID。默认为 None。	None
attention_mask	Tensor	形状为 (batch_size, sequence_length) 的张量，包含注意力掩码。默认为 None。	None

返回

名称	类型	描述
Tensor		形状为 (batch_size, sequence_length, hidden_size) 的输出张量，包含嵌入表示。

源代码位于 bionemo/esm2/model/model.py

def embedding_forward(
    self, input_ids: Tensor, position_ids: Tensor, tokentype_ids: Tensor = None, attention_mask: Tensor = None
):
    """Forward pass of the embedding layer.

    Args:
        input_ids: The input tensor of shape (batch_size, sequence_length) containing the input IDs.
        position_ids: The tensor of shape (batch_size, sequence_length) containing the position IDs.
        tokentype_ids: The tensor of shape (batch_size, sequence_length) containing the token type IDs. Defaults to None.
        attention_mask: The tensor of shape (batch_size, sequence_length) containing the attention mask. Defaults to None.

    Returns:
        Tensor: The output tensor of shape (batch_size, sequence_length, hidden_size) containing the embedded representations.
    """
    # ESM2 Customization: ESM2Embedding forward takes attention_mask
    # in addition to the args required by LanguageModelEmbedding
    return self.embedding(
        input_ids=input_ids, position_ids=position_ids, tokentype_ids=tokentype_ids, attention_mask=attention_mask
    )