Hugging Face's logo

Transformers documentation

MiMo-V2-Flash

You are viewing main version, which requires installation from source. If you'd like regular pip install, checkout the latest stable version (v5.12.0).
Hugging Face's logo
Join the Hugging Face community

and get access to the augmented documentation experience

to get started

This model was contributed to Hugging Face Transformers on 2026-06-30.

PyTorch

MiMo-V2-Flash

Overview

MiMo-V2-Flash is a Mixture-of-Experts (MoE) language model developed by the Xiaomi MiMo team. Designed to establish a new balance between long-context modeling capabilities and inference efficiency, the model is built for strong performance in complex reasoning and agentic tasks. Trained on 27T tokens with native 32k sequence lengths, MiMo-V2-Flash seamlessly supports an extended 256K context window while significantly reducing KV-cache storage compared to standard global attention models.

Key Features

  • Hybrid Attention Architecture: Interleaves Sliding Window Attention (SWA) and Global Attention (GA) at a 5:1 ratio, using an aggressive 128-token window. This approach reduces KV-cache storage by nearly 6x while utilizing a learnable attention sink bias to preserve excellent performance on long contexts.
  • Agentic Capabilities: Enhanced through Multi-Teacher On-Policy Distillation (MOPD) and large-scale agentic RL during post-training, the model demonstrates superior tool-use capabilities and exceptional performance on benchmarks like SWE-Bench.
  • Inference Efficiency: Pre-trained using FP8 mixed precision, making it highly optimized for practical deployments and modern accelerators.

For more details, please refer to the technical report, and the official repository.
This model was contributed by casinca.

Usage examples

Text generation

The example below demonstrates how to generate text with Pipeline or the AutoModelForCausalLM class.

Pipeline
AutoModelForCausalLM 
import torch
from transformers import pipeline

pipe = pipeline(
    task="text-generation",
    model="XiaomiMiMo/MiMo-V2-Flash",
)
pipe("Explain why sparse MoE models can be efficient at inference.")

Chat template generation

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "XiaomiMiMo/MiMo-V2-Flash"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    device_map="auto",
)

messages = [
    {"role": "system", "content": "You are MiMo, a helpful assistant."},
    {"role": "user", "content": "Write a short summary of MiMo-V2-Flash."},
]

input_ids = tokenizer.apply_chat_template(
    messages,
    add_generation_prompt=True,
    return_tensors="pt",
).to(model.device)

generated_ids = model.generate(input_ids, max_new_tokens=128)
response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
print(response)

MiMoV2FlashConfig

class transformers.MiMoV2FlashConfig

< >

( transformers_version: str | None = None architectures: list[str] | None = None output_hidden_states: bool | None = False return_dict: bool | None = True dtype: typing.Union[str, ForwardRef('torch.dtype'), NoneType] = None chunk_size_feed_forward: int = 0 is_encoder_decoder: bool = False id2label: dict[int, str] | dict[str, str] | None = None label2id: dict[str, int] | dict[str, str] | None = None problem_type: typing.Optional[typing.Literal['regression', 'single_label_classification', 'multi_label_classification']] = None vocab_size: int = 152576 hidden_size: int = 4096 intermediate_size: int = 16384 num_hidden_layers: int = 48 num_attention_heads: int = 64 num_key_value_heads: int = 4 hidden_act: str = 'silu' max_position_embeddings: int = 131072 initializer_range: float = 0.02 rms_norm_eps: float = 1e-05 use_cache: bool = True tie_word_embeddings: bool = False rope_parameters: transformers.modeling_rope_utils.RopeParameters | dict | None = None attention_bias: bool = False attention_dropout: float | int = 0.0 moe_intermediate_size: int = 2048 num_experts_per_tok: int = 8 n_routed_experts: int = 256 routed_scaling_factor: float | None = 1.0 n_group: int = 1 topk_group: int = 1 norm_topk_prob: bool = True bos_token_id: int | None = 1 eos_token_id: int | list[int] | None = None pad_token_id: int | None = None head_dim: int = 192 v_head_dim: int = 128 sliding_window: int = 128 layer_types: list[str] | None = None mlp_layer_types: list[str] | None = None attention_value_scale: float | None = 0.707 )

Parameters

  • vocab_size (int, optional, defaults to 152576) — 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 16384) — Dimension of the MLP representations.
  • num_hidden_layers (int, optional, defaults to 48) — Number of hidden layers in the Transformer decoder.
  • num_attention_heads (int, optional, defaults to 64) — Number of attention heads for each attention layer in the Transformer decoder.
  • num_key_value_heads (int, optional, defaults to 4) — 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 131072) — 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 (float, optional, defaults to 1e-05) — 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.
  • 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], 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 (Union[float, int], optional, defaults to 0.0) — The dropout ratio for the attention probabilities.
  • moe_intermediate_size (int, optional, defaults to 2048) — Intermediate size of the routed expert MLPs.
  • num_experts_per_tok (int, optional, defaults to 8) — Number of experts to route each token to. This is the top-k value for the token-choice routing.
  • n_routed_experts (int, optional, defaults to 256) — Number of routed experts.
  • routed_scaling_factor (float, optional, defaults to 1.0) — Scaling factor or routed experts.
  • n_group (int, optional, defaults to 1) — Number of expert groups for group-based top-k routing.
  • topk_group (int, optional, defaults to 1) — Number of groups selected per token in group-based top-k routing.
  • norm_topk_prob (bool, optional, defaults to True) — Whether to normalize the weights of the routed experts.
  • bos_token_id (int, optional, defaults to 1) — Token id used for beginning-of-stream in the vocabulary.
  • eos_token_id (Union[int, list[int]], optional) — Token id used for end-of-stream in the vocabulary.
  • pad_token_id (int, optional) — Token id used for padding in the vocabulary.
  • head_dim (int, optional, defaults to 192) — Dimension of query and key heads.
  • v_head_dim (int, optional, defaults to 128) — Dimension of value heads (special case because MiMo uses a smaller v head dim than (qk) head dim )
  • sliding_window (int, optional, defaults to 128) — Sliding window attention window size. If None, no sliding window is applied.
  • layer_types (list[str], optional) — A list that explicitly maps each layer index with its layer type. If not provided, it will be automatically generated based on config values.
  • mlp_layer_types (list, optional) — MLP pattern for each layer ("dense" or "sparse"). Defaults to 1 dense + rest sparse.
  • attention_value_scale (float, optional, defaults to 0.707 (which is the decimal approximation — of sqrt(hidden_size / (num_attention_heads * v_head_dim))): Constant multiplier applied to rescale the attention values.

This is the configuration class to store the configuration of a MiMoV2FlashModel. It is used to instantiate a Mimo V2 Flash 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 XiaomiMiMo/MiMo-V2-Flash

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

MiMoV2FlashModel

class transformers.MiMoV2FlashModel

< >

( config: MiMoV2FlashConfig )

Parameters

  • config (MiMoV2FlashConfig) — 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 Mimo V2 Flash 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] ) → MoeModelOutputWithPast 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

MoeModelOutputWithPast or tuple(torch.FloatTensor)

A MoeModelOutputWithPast 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 (MiMoV2FlashConfig) and inputs.

The MiMoV2FlashModel 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.

  • 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.

  • router_logits (tuple(torch.FloatTensor), optional, returned when output_router_probs=True and config.add_router_probs=True is passed or when config.output_router_probs=True) — Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, sequence_length, num_experts).

    Raw router logtis (post-softmax) that are computed by MoE routers, these terms are used to compute the auxiliary loss for Mixture of Experts models.

MiMoV2FlashForCausalLM

class transformers.MiMoV2FlashForCausalLM

< >

( config )

Parameters

  • config (MiMoV2FlashForCausalLM) — 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 Mimo V2 Flash 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 (MiMoV2FlashConfig) and inputs.

The MiMoV2FlashForCausalLM 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, MiMoV2FlashForCausalLM

>>> model = MiMoV2FlashForCausalLM.from_pretrained("meta-mimo_v2_flash/MiMoV2Flash-2-7b-hf")
>>> tokenizer = AutoTokenizer.from_pretrained("meta-mimo_v2_flash/MiMoV2Flash-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."
Update on GitHub

←Mellum MiniCPM3→