Hugging Face's logo

Transformers documentation

DeepSeek-V2

You are viewing main version, which requires installation from source. If you'd like regular pip install, checkout the latest stable version (v5.3.0).
Hugging Face's logo
Join the Hugging Face community

and get access to the augmented documentation experience

to get started

This model was released on 2024-05-07 and added to Hugging Face Transformers on 2025-07-09.

DeepSeek-V2

Overview

The DeepSeek-V2 model was proposed in DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model by DeepSeek-AI Team.

The abstract from the paper is the following: We present DeepSeek-V2, a strong Mixture-of-Experts (MoE) language model characterized by economical training and efficient inference. It comprises 236B total parameters, of which 21B are activated for each token, and supports a context length of 128K tokens. DeepSeek-V2 adopts innovative architectures including Multi-head Latent Attention (MLA) and DeepSeekMoE. MLA guarantees efficient inference through significantly compressing the Key-Value (KV) cache into a latent vector, while DeepSeekMoE enables training strong models at an economical cost through sparse computation. Compared with DeepSeek 67B, DeepSeek-V2 achieves significantly stronger performance, and meanwhile saves 42.5% of training costs, reduces the KV cache by 93.3%, and boosts the maximum generation throughput to 5.76 times. We pretrain DeepSeek-V2 on a high-quality and multi-source corpus consisting of 8.1T tokens, and further perform Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) to fully unlock its potential. Evaluation results show that, even with only 21B activated parameters, DeepSeek-V2 and its chat versions still achieve top-tier performance among open-source models.

This model was contributed by VladOS95-cyber. The original code can be found here.

Usage tips

The model uses Multi-head Latent Attention (MLA) and DeepSeekMoE architectures for efficient inference and cost-effective training. It employs an auxiliary-loss-free strategy for load balancing and multi-token prediction training objective. The model can be used for various language tasks after being pre-trained on 14.8 trillion tokens and going through Supervised Fine-Tuning and Reinforcement Learning stages.

DeepseekV2Config

class transformers.DeepseekV2Config

< >

( vocab_size: int | None = 32000 hidden_size: int | None = 4096 intermediate_size: int | None = 11008 num_hidden_layers: int | None = 32 num_attention_heads: int | None = 32 num_key_value_heads: int | None = None hidden_act: str | None = 'silu' max_position_embeddings: int | None = 2048 initializer_range: float | None = 0.02 rms_norm_eps: int | None = 1e-06 use_cache: bool | None = True pad_token_id: int | None = None bos_token_id: int | None = 1 eos_token_id: int | None = 2 tie_word_embeddings: bool | None = False rope_parameters: transformers.modeling_rope_utils.RopeParameters | dict[str, transformers.modeling_rope_utils.RopeParameters] | None = None attention_bias: bool | None = False attention_dropout: float | None = 0.0 mlp_bias: bool | None = False first_k_dense_replace: int | None = 0 kv_lora_rank: int | None = 512 q_lora_rank: int | None = 1536 n_group: int | None = None n_routed_experts: int | None = 64 n_shared_experts: int | None = 2 qk_nope_head_dim: int | None = 128 qk_rope_head_dim: int | None = 64 routed_scaling_factor: float | None = 1.0 topk_group: int | None = None topk_method: str | None = 'greedy' norm_topk_prob: bool | None = False v_head_dim: int | None = 128 num_experts_per_tok: int | None = None moe_intermediate_size: int | None = 1407 **kwargs )

Parameters

  • vocab_size (int, optional, defaults to 32000) — Vocabulary size of the model. Defines the number of different tokens that can be represented by the input_ids.
  • hidden_size (int, optional, defaults to 4096) — Dimension of the hidden representations.
  • intermediate_size (int, optional, defaults to 11008) — Dimension of the MLP representations.
  • num_hidden_layers (int, optional, defaults to 32) — Number of hidden layers in the Transformer decoder.
  • num_attention_heads (int, optional, defaults to 32) — Number of attention heads for each attention layer in the Transformer decoder.
  • num_key_value_heads (int, optional) — This is the number of key_value heads that should be used to implement Grouped Query Attention. If num_key_value_heads=num_attention_heads, the model will use Multi Head Attention (MHA), if num_key_value_heads=1 the model will use Multi Query Attention (MQA) otherwise GQA is used. When converting a multi-head checkpoint to a GQA checkpoint, each group key and value head should be constructed by meanpooling all the original heads within that group. For more details, check out this paper. If it is not specified, will default to num_attention_heads.
  • hidden_act (str, optional, defaults to silu) — The non-linear activation function (function or string) in the decoder. For example, "gelu", "relu", "silu", etc.
  • max_position_embeddings (int, optional, defaults to 2048) — The maximum sequence length that this model might ever be used with.
  • initializer_range (float, optional, defaults to 0.02) — The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
  • rms_norm_eps (int, optional, defaults to 1e-06) — The epsilon used by the rms normalization layers.
  • 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.
  • pad_token_id (int, optional) — Token id used for padding in the vocabulary.
  • bos_token_id (int, optional, defaults to 1) — Token id used for beginning-of-stream in the vocabulary.
  • eos_token_id (int, optional, defaults to 2) — Token id used for end-of-stream in the vocabulary.
  • tie_word_embeddings (bool, optional, defaults to False) — Whether to tie weight embeddings according to model’s tied_weights_keys mapping.
  • rope_parameters (Union[~modeling_rope_utils.RopeParameters, dict[str, ~modeling_rope_utils.RopeParameters]], optional) — Dictionary containing the configuration parameters for the RoPE embeddings. The dictionary should contain a value for rope_theta and optionally parameters used for scaling in case you want to use RoPE with longer max_position_embeddings.
  • attention_bias (bool, optional, defaults to False) — Whether to use a bias in the query, key, value and output projection layers during self-attention.
  • attention_dropout (float, optional, defaults to 0.0) — The dropout ratio for the attention probabilities.
  • mlp_bias (bool, optional, defaults to False) — Whether to use a bias in up_proj, down_proj and gate_proj layers in the MLP layers.
  • first_k_dense_replace (int, optional, defaults to 0) — Number of dense layers in the shallow layers before switching to MoE layers.
  • kv_lora_rank (int, optional, defaults to 512) — Rank of the LoRA matrices for key and value projections.
  • q_lora_rank (int, optional, defaults to 1536) — Rank of the LoRA matrices for query projections.
  • n_group (int, optional) — Number of groups for routed experts.
  • n_routed_experts (int, optional, defaults to 64) — Number of routed experts.
  • n_shared_experts (int, optional, defaults to 2) — Number of shared experts.
  • qk_nope_head_dim (int, optional, defaults to 128) — Dimension of the query/key heads that don’t use rotary position embeddings.
  • qk_rope_head_dim (int, optional, defaults to 64) — Dimension of the query/key heads that use rotary position embeddings.
  • routed_scaling_factor (float, optional, defaults to 1.0) — Scaling factor or routed experts.
  • topk_group (int, optional) — Number of selected groups for each token (for each token, ensuring the selected experts is only within topk_group groups).
  • topk_method (str, optional, defaults to "greedy") — The method used for selecting top-k experts in the routed gate mechanism.
  • norm_topk_prob (bool, optional, defaults to False) — Whether to normalize the weights of the routed experts.
  • v_head_dim (int, optional, defaults to 128) — Dimension of the value heads.
  • num_experts_per_tok (int, optional) — Number of experts to route each token to. This is the top-k value for the token-choice routing.
  • moe_intermediate_size (int, optional, defaults to 1407) — Intermediate size of the routed expert MLPs.

This is the configuration class to store the configuration of a DeepseekV2Model. It is used to instantiate a Deepseek V2 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 deepseek-ai/DeepSeek-V2-Lite

Configuration objects inherit from PreTrainedConfig and can be used to control the model outputs. Read the documentation from PreTrainedConfig for more information.

Example:

>>> from transformers import DeepseekV2Model, DeepseekV2Config
>>> # Initializing a DeepSeek-V2 style configuration
>>> configuration = DeepseekV2Config()
>>> # Accessing the model configuration
>>> model = DeepseekV2Model(configuration)
>>> print(model.config)

DeepseekV2Model

class transformers.DeepseekV2Model

< >

( config: DeepseekV2Config )

Parameters

  • config (DeepseekV2Config) — 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() method to load the model weights.

The bare Deepseek V2 Model outputting raw hidden-states without any specific head on top.

This model inherits from 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 subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.

forward

< >

( input_ids: torch.LongTensor | None = None attention_mask: torch.Tensor | None = None position_ids: torch.LongTensor | None = None past_key_values: transformers.cache_utils.Cache | None = None inputs_embeds: torch.FloatTensor | None = None use_cache: bool | None = None **kwargs: typing_extensions.Unpack[transformers.utils.generic.TransformersKwargs] ) → BaseModelOutputWithPast or tuple(torch.FloatTensor)

Parameters

  • 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. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.

    What are 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?

  • 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?

  • past_key_values (~cache_utils.Cache, optional) — Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used to speed up sequential decoding. This typically consists in the past_key_values returned by the model at a previous stage of decoding, when use_cache=True or config.use_cache=True.

    Only Cache instance is allowed as input, see our kv cache guide. If no past_key_values are passed, DynamicCache will be initialized by default.

    The model will output the same cache format that is fed as input.

    If past_key_values are used, the user is expected to input only unprocessed input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, unprocessed_length) instead of all input_ids of shape (batch_size, sequence_length).

  • 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.
  • 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).

Returns

BaseModelOutputWithPast or tuple(torch.FloatTensor)

A BaseModelOutputWithPast 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 (DeepseekV2Config) and inputs.

The DeepseekV2Model 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.

    If past_key_values is used only the last hidden-state of the sequences of shape (batch_size, 1, hidden_size) is output.

  • past_key_values (Cache, optional, returned when use_cache=True is passed or when config.use_cache=True) — It is a Cache instance. For more details, see our kv cache guide.

    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.

  • 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.

DeepseekV2ForCausalLM

class transformers.DeepseekV2ForCausalLM

< >

( config )

Parameters

  • config (DeepseekV2ForCausalLM) — 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() method to load the model weights.

The Deepseek V2 Model for causal language modeling.

This model inherits from 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 subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.

forward

< >

( input_ids: torch.LongTensor | None = None attention_mask: torch.Tensor | None = None position_ids: torch.LongTensor | None = None past_key_values: transformers.cache_utils.Cache | None = None inputs_embeds: torch.FloatTensor | None = None labels: torch.LongTensor | None = None use_cache: bool | None = None logits_to_keep: int | torch.Tensor = 0 **kwargs: typing_extensions.Unpack[transformers.utils.generic.TransformersKwargs] ) → CausalLMOutputWithPast or tuple(torch.FloatTensor)

Parameters

  • 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. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.

    What are 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?

  • 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?

  • past_key_values (~cache_utils.Cache, optional) — Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used to speed up sequential decoding. This typically consists in the past_key_values returned by the model at a previous stage of decoding, when use_cache=True or config.use_cache=True.

    Only Cache instance is allowed as input, see our kv cache guide. If no past_key_values are passed, DynamicCache will be initialized by default.

    The model will output the same cache format that is fed as input.

    If past_key_values are used, the user is expected to input only unprocessed input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, unprocessed_length) instead of all input_ids of shape (batch_size, sequence_length).

  • 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 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).
  • logits_to_keep (Union[int, torch.Tensor], optional, defaults to 0) — If an int, compute logits for the last logits_to_keep tokens. If 0, calculate logits for all input_ids (special case). Only last token logits are needed for generation, and calculating them only for that token can save memory, which becomes pretty significant for long sequences or large vocabulary size. If a torch.Tensor, must be 1D corresponding to the indices to keep in the sequence length dimension. This is useful when using packed tensor format (single dimension for batch and sequence length).

Returns

CausalLMOutputWithPast or tuple(torch.FloatTensor)

A CausalLMOutputWithPast 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 (DeepseekV2Config) and inputs.

The DeepseekV2ForCausalLM 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) — Language modeling loss (for next-token prediction).

  • 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).

  • past_key_values (Cache, optional, returned when use_cache=True is passed or when config.use_cache=True) — It is a Cache instance. For more details, see our kv cache guide.

    Contains pre-computed hidden-states (key and values in the self-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.

  • 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:

>>> from transformers import AutoTokenizer, DeepseekV2ForCausalLM

>>> model = DeepseekV2ForCausalLM.from_pretrained("meta-deepseek_v2/DeepseekV2-2-7b-hf")
>>> tokenizer = AutoTokenizer.from_pretrained("meta-deepseek_v2/DeepseekV2-2-7b-hf")

>>> prompt = "Hey, are you conscious? Can you talk to me?"
>>> inputs = tokenizer(prompt, return_tensors="pt")

>>> # Generate
>>> generate_ids = model.generate(inputs.input_ids, max_length=30)
>>> tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
"Hey, are you conscious? Can you talk to me?\nI'm not conscious, but I can talk to you."

DeepseekV2ForSequenceClassification

class transformers.DeepseekV2ForSequenceClassification

< >

( config )

forward

< >

( input_ids: torch.LongTensor | None = None attention_mask: torch.Tensor | None = None position_ids: torch.LongTensor | None = None past_key_values: transformers.cache_utils.Cache | None = None inputs_embeds: torch.FloatTensor | None = None labels: torch.LongTensor | None = None use_cache: bool | None = None **kwargs: typing_extensions.Unpack[transformers.utils.generic.TransformersKwargs] ) → SequenceClassifierOutputWithPast or tuple(torch.FloatTensor)

Parameters

  • 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. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.

    What are 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?

  • 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?

  • past_key_values (~cache_utils.Cache, optional) — Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used to speed up sequential decoding. This typically consists in the past_key_values returned by the model at a previous stage of decoding, when use_cache=True or config.use_cache=True.

    Only Cache instance is allowed as input, see our kv cache guide. If no past_key_values are passed, DynamicCache will be initialized by default.

    The model will output the same cache format that is fed as input.

    If past_key_values are used, the user is expected to input only unprocessed input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, unprocessed_length) instead of all input_ids of shape (batch_size, sequence_length).

  • 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 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).

Returns

SequenceClassifierOutputWithPast or tuple(torch.FloatTensor)

A SequenceClassifierOutputWithPast 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 (None) and inputs.

The GenericForSequenceClassification 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).

  • past_key_values (Cache, optional, returned when use_cache=True is passed or when config.use_cache=True) — It is a Cache instance. For more details, see our kv cache guide.

    Contains pre-computed hidden-states (key and values in the self-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.

  • 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.

Update on GitHub

←DeBERTa-v2 DeepSeek-V3→