Hugging Face's logo

Transformers documentation

Dots 3 Note Preview

You are viewing main version, which requires installation from source. If you'd like regular pip install, checkout the latest stable version (v5.22.2).
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 2026-11-12.

Dots 3 Note Preview

Dots 3 Note Preview is a mixture-of-experts causal language model with native text, image, video, and audio inputs. It uses a shared vision encoder for images and videos and a Whisper-style audio encoder. Both encoders project their outputs into the language model’s hidden space before autoregressive text generation.

BF16 and FP8 inference are supported.

The checkpoint needs a standard processor_config.json with Qwen image/video components and Dots preprocessing settings. Saved configs use the native Transformers schema; older serving frameworks may need a compatibility adapter.

Usage

from transformers import AutoModelForMultimodalLM, AutoProcessor


model_id = "dots-studio/dots3-note-prev"
processor = AutoProcessor.from_pretrained(model_id)
model = AutoModelForMultimodalLM.from_pretrained(model_id, device_map="auto")

conversation = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/cats.png"},
            {"type": "text", "text": "Describe this image."},
        ],
    }
]
inputs = processor.apply_chat_template(
    conversation,
    add_generation_prompt=True,
    tokenize=True,
    return_dict=True,
    return_tensors="pt",
).to(model.device, dtype=model.dtype)

output_ids = model.generate(**inputs, max_new_tokens=128)
print(processor.batch_decode(output_ids, skip_special_tokens=True)[0])

Use the same processing and generation calls with any of these user messages:

# Text
conversation = [{"role": "user", "content": [{"type": "text", "text": "Briefly introduce yourself."}]}]

# Audio: replace the path with a local recording.
conversation = [{"role": "user", "content": [
    {"type": "audio", "path": "speech.wav"},
    {"type": "text", "text": "Transcribe this recording."},
]}]

# Video frames. Pass audio separately to include the soundtrack.
conversation = [{"role": "user", "content": [
    {"type": "video", "path": "concert.mp4"},
    {"type": "text", "text": "Describe what you see."},
]}]

Dots3NoteModel combines the text, vision and audio encoders. Dots3NoteForConditionalGeneration adds the language-model head and generation interface. Dots3NoteForCausalLM is the text-only variant.

Dots3NoteConfig

class transformers.Dots3NoteConfig

< >

( transformers_version: str | None = Nonearchitectures: list[str] | None = Noneoutput_hidden_states: bool | None = Falsereturn_dict: bool | None = Truedtype: str | torch.dtype | None = Nonechunk_size_feed_forward: int = 0is_encoder_decoder: bool = Falseid2label: dict[int, str] | dict[str, str] | None = Nonelabel2id: dict[str, int] | dict[str, str] | None = Noneproblem_type: Literal['regression', 'single_label_classification', 'multi_label_classification'] | None = Nonetext_config: dict | transformers.models.dots3_note.configuration_dots3_note.Dots3NoteTextConfig | None = Nonevision_config: dict | transformers.models.dots3_note.configuration_dots3_note.Dots3NoteVisionConfig | None = Noneaudio_config: dict | transformers.models.dots3_note.configuration_dots3_note.Dots3NoteAudioConfig | None = Noneimage_token_id: int = 151660video_token_id: int = 151680audio_token_id: int = 151720tie_word_embeddings: bool = False )

Parameters

  • text_config (Union[dict, ~models.dots3_note.configuration_dots3_note.Dots3NoteTextConfig], optional) — The config object or dictionary of the text backbone.
  • vision_config (Union[dict, ~models.dots3_note.configuration_dots3_note.Dots3NoteVisionConfig], optional) — The config object or dictionary of the vision backbone.
  • audio_config (Union[dict, ~models.dots3_note.configuration_dots3_note.Dots3NoteAudioConfig], optional) — The config object or dictionary of the audio backbone.
  • image_token_id (int, optional, defaults to 151660) — The image token index used as a placeholder for input images.
  • video_token_id (int, optional, defaults to 151680) — The video token index used as a placeholder for input videos.
  • audio_token_id (int, optional, defaults to 151720) — The audio token index used as a placeholder for input audio.
  • tie_word_embeddings (bool, optional, defaults to False) — Whether to tie weight embeddings according to model’s tied_weights_keys mapping.

This is the configuration class to store the configuration of a Dots3NoteModel. It is used to instantiate a Dots3 Note 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 dots-studio/dots3-note-prev

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

Dots3NoteTextConfig

class transformers.Dots3NoteTextConfig

< >

( transformers_version: str | None = Nonearchitectures: list[str] | None = Noneoutput_hidden_states: bool | None = Falsereturn_dict: bool | None = Truedtype: str | torch.dtype | None = Nonechunk_size_feed_forward: int = 0is_encoder_decoder: bool = Falseid2label: dict[int, str] | dict[str, str] | None = Nonelabel2id: dict[str, int] | dict[str, str] | None = Noneproblem_type: Literal['regression', 'single_label_classification', 'multi_label_classification'] | None = Nonevocab_size: int = 152064hidden_size: int = 5120intermediate_size: int = 13824moe_intermediate_size: int = 1536num_hidden_layers: int = 46num_attention_heads: int = 128num_key_value_heads: int = 128n_shared_experts: int = 1n_routed_experts: int = 256output_router_logits: bool = Falserouted_scaling_factor: float = 1.0kv_lora_rank: int = 512q_lora_rank: int | None = 1024qk_rope_head_dim: int = 64v_head_dim: int = 128qk_nope_head_dim: int = 128n_group: int = 1topk_group: int = 1num_experts_per_tok: int = 8norm_topk_prob: bool = Truehidden_act: str = 'silu'max_position_embeddings: int = 524288initializer_range: float = 0.02rms_norm_eps: float = 1e-05use_cache: bool = Truepad_token_id: int | None = 151659bos_token_id: int | None = 151643eos_token_id: int | list[int] | None = 151668tie_word_embeddings: bool = Falserope_parameters: dict | None = Nonemlp_layer_types: list[str] | None = Noneattention_bias: bool = Falseattention_dropout: float | int = 0.0index_topk: int = 2048index_head_dim: int = 128index_n_heads: int = 64mlp_bias: bool = Falsehead_dim: int = 256layer_types: list[str] | None = Nonesliding_window: int | None = Noneshared_experts_intermediate_size: int = 1536 )

Parameters

  • vocab_size (int, optional, defaults to 152064) — 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 5120) — Dimension of the hidden representations.
  • intermediate_size (int, optional, defaults to 13824) — Dimension of the MLP representations.
  • moe_intermediate_size (int, optional, defaults to 1536) — Intermediate size of the routed expert MLPs.
  • num_hidden_layers (int, optional, defaults to 46) — Number of hidden layers in the Transformer decoder.
  • num_attention_heads (int, optional, defaults to 128) — Number of attention heads for each attention layer in the Transformer decoder.
  • num_key_value_heads (int, optional, defaults to 128) — 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.
  • n_shared_experts (int, optional, defaults to 1) — Number of shared experts.
  • n_routed_experts (int, optional, defaults to 256) — Number of routed experts.
  • output_router_logits (bool, optional, defaults to False) — Whether or not the router logits should be returned by the model. Enabling this will also allow the model to output the auxiliary loss, including load balancing loss and router z-loss.
  • routed_scaling_factor (float, optional, defaults to 1.0) — Scaling factor or routed experts.
  • 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 1024) — Rank of the LoRA matrices for query projections.
  • qk_rope_head_dim (int, optional, defaults to 64) — Dimension of the query/key heads that use rotary position embeddings.
  • v_head_dim (int, optional, defaults to 128) — Dimension of the value heads.
  • qk_nope_head_dim (int, optional, defaults to 128) — Dimension of the query/key heads that don’t use rotary position embeddings.
  • n_group (int, optional, defaults to 1) — Number of groups into which routed experts are partitioned during selection.
  • topk_group (int, optional, defaults to 1) — Number of selected groups for each token (for each token, ensuring the selected experts is only within topk_group groups).
  • 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.
  • norm_topk_prob (bool, optional, defaults to True) — Whether to normalize the weights of the routed experts.
  • 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 524288) — 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.
  • pad_token_id (int, optional, defaults to 151659) — Token id used for padding in the vocabulary.
  • bos_token_id (int, optional, defaults to 151643) — Token id used for beginning-of-stream in the vocabulary.
  • eos_token_id (Union[int, list[int]], optional, defaults to 151668) — 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 (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.
  • mlp_layer_types (list[str], optional) — Per-layer "dense" or "sparse" MLP types, derived from the released MoE schedule when omitted.
  • 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.
  • index_topk (int, optional, defaults to 2048) — Number of key positions selected by dynamic sparse attention.
  • index_head_dim (int, optional, defaults to 128) — Per-head width of the dynamic sparse attention indexer.
  • index_n_heads (int, optional, defaults to 64) — Number of heads in the dynamic sparse attention indexer.
  • 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.
  • head_dim (int, optional, defaults to 256) — The attention head dimension. If None, it will default to hidden_size // num_attention_heads
  • 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.
  • sliding_window (int, optional) — Sliding window attention window size. If None, no sliding window is applied.
  • shared_experts_intermediate_size (int, optional, defaults to 1536) — Intermediate width of each shared expert.

This is the configuration class to store the configuration of a Dots3NoteModel. It is used to instantiate a Dots3 Note 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 dots-studio/dots3-note-prev

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

convert_rope_params_to_dict

< >

( **kwargs )

Convert legacy RoPE settings to separate DSA/SWA entries, using each layer type’s rotary head fraction.

Dots3NoteVisionConfig

class transformers.Dots3NoteVisionConfig

< >

( transformers_version: str | None = Nonearchitectures: list[str] | None = Noneoutput_hidden_states: bool | None = Falsereturn_dict: bool | None = Truedtype: str | torch.dtype | None = Nonechunk_size_feed_forward: int = 0is_encoder_decoder: bool = Falseid2label: dict[int, str] | dict[str, str] | None = Nonelabel2id: dict[str, int] | dict[str, str] | None = Noneproblem_type: Literal['regression', 'single_label_classification', 'multi_label_classification'] | None = Nonehidden_size: int = 1536intermediate_size: int = 4224num_hidden_layers: int = 42num_attention_heads: int = 24num_channels: int = 3patch_size: int = 14hidden_act: str = 'silu'attention_dropout: float | int = 0.0initializer_range: float = 0.02rope_parameters: dict | None = Nonemoe_intermediate_size: int = 2112spatial_merge_size: int = 2rms_norm_eps: float = 1e-05attention_bias: bool = Falsemlp_bias: bool = Falsemlp_layer_types: list[str] | None = Nonenum_local_experts: int = 64num_experts_per_tok: int | None = Nonerouter_scaling_factor: float | None = Noneadapter_in_dim: int = 1536adapter_out_dim: int = 5120adapter_merge_size: int = 2 )

Parameters

  • hidden_size (int, optional, defaults to 1536) — Dimension of the hidden representations.
  • intermediate_size (int, optional, defaults to 4224) — Dimension of the MLP representations.
  • num_hidden_layers (int, optional, defaults to 42) — Number of hidden layers in the Transformer decoder.
  • num_attention_heads (int, optional, defaults to 24) — Number of attention heads for each attention layer in the Transformer decoder.
  • num_channels (int, optional, defaults to 3) — The number of input channels.
  • patch_size (int, optional, defaults to 14) — The size (resolution) of each patch.
  • hidden_act (str, optional, defaults to silu) — The non-linear activation function (function or string) in the decoder. For example, "gelu", "relu", "silu", etc.
  • attention_dropout (Union[float, int], optional, defaults to 0.0) — The dropout ratio for the attention probabilities.
  • initializer_range (float, optional, defaults to 0.02) — The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
  • rope_parameters (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.
  • moe_intermediate_size (int, optional, defaults to 2112) — Intermediate size of the routed expert MLPs.
  • spatial_merge_size (int, optional, defaults to 2) — The size of the spatial merge window used to reduce the number of visual tokens by merging neighboring patches.
  • rms_norm_eps (float, optional, defaults to 1e-05) — The epsilon used by the rms normalization layers.
  • attention_bias (bool, optional, defaults to False) — Whether to use a bias in the query, key, value and output projection layers during self-attention.
  • 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.
  • mlp_layer_types (list[str], optional) — Dense or sparse MLP for each layer. Expert counts are specified through per_layer_config.
  • num_local_experts (int, optional, defaults to 64) — Number of local experts on each device. num_experts should be divisible by num_local_experts.
  • 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.
  • router_scaling_factor (float, optional, defaults to 1.0) — Scale applied to the vision router scores.
  • adapter_in_dim (int, optional, defaults to 1536) — Input width of the vision adapter. It must match hidden_size.
  • adapter_out_dim (int, optional, defaults to 5120) — Output width of the vision adapter. It must match the text hidden size.
  • adapter_merge_size (int, optional, defaults to 2) — Spatial merge factor used by the vision adapter.

This is the configuration class to store the configuration of a Dots3NoteModel. It is used to instantiate a Dots3 Note 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 dots-studio/dots3-note-prev

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

Dots3NoteAudioConfig

class transformers.Dots3NoteAudioConfig

< >

( transformers_version: str | None = Nonearchitectures: list[str] | None = Noneoutput_hidden_states: bool | None = Falsereturn_dict: bool | None = Truedtype: str | torch.dtype | None = Nonechunk_size_feed_forward: int = 0is_encoder_decoder: bool = Falseid2label: dict[int, str] | dict[str, str] | None = Nonelabel2id: dict[str, int] | dict[str, str] | None = Noneproblem_type: Literal['regression', 'single_label_classification', 'multi_label_classification'] | None = Nonehidden_size: int = 1280intermediate_size: int = 5120num_hidden_layers: int = 32num_attention_heads: int = 20num_key_value_heads: int | None = Nonehidden_act: str = 'silu'max_position_embeddings: int = 6000initializer_range: float = 0.02rope_parameters: dict | None = Noneattention_dropout: float | int = 0.0dropout: float = 0.0sampling_rate: int = 16000feature_size: int = 128hop_length: int = 160downsample_hidden_size: int = 480head_dim: int | None = Noneattention_bias: bool = Trueadapter_input_size: int = 1280adapter_out_dim: int = 5120 )

Parameters

  • hidden_size (int, optional, defaults to 1280) — Dimension of the hidden representations.
  • intermediate_size (int, optional, defaults to 5120) — 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 20) — 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 6000) — 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.
  • rope_parameters (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_dropout (Union[float, int], optional, defaults to 0.0) — The dropout ratio for the attention probabilities.
  • dropout (float, optional, defaults to 0.0) — The ratio for all dropout layers.
  • sampling_rate (int, optional, defaults to 16000) — The sampling rate at which the audio files should be digitalized expressed in hertz (Hz).
  • feature_size (int, optional, defaults to 128) — Number of log-mel filter-bank channels.
  • hop_length (int, optional, defaults to 160) — Number of waveform samples between adjacent log-mel frames.
  • downsample_hidden_size (int, optional, defaults to 480) — Hidden width of the temporal downsampling projection.
  • head_dim (int, optional) — The attention head dimension. If None, it will default to hidden_size // num_attention_heads
  • attention_bias (bool, optional, defaults to True) — Whether to use a bias in the query, key, value and output projection layers during self-attention.
  • adapter_input_size (int, optional, defaults to 1280) — Input width of the audio-to-text adapter.
  • adapter_out_dim (int, optional, defaults to 5120) — Output width of the audio-to-text adapter. It must match the text hidden size.

This is the configuration class to store the configuration of a Dots3NoteModel. It is used to instantiate a Dots3 Note 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 dots-studio/dots3-note-prev

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

Dots3NoteForCausalLM

class transformers.Dots3NoteForCausalLM

< >

( config: Dots3NoteTextConfig )

Parameters

  • config (Dots3NoteTextConfig) — 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 Dots3 Note 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: typing.Optional[torch.LongTensor] = Noneattention_mask: typing.Optional[torch.Tensor] = Noneposition_ids: typing.Optional[torch.LongTensor] = Nonepast_key_values: transformers.cache_utils.Cache | None = Noneinputs_embeds: typing.Optional[torch.Tensor] = Nonelabels: typing.Optional[torch.LongTensor] = Noneuse_cache: bool | None = Nonelogits_to_keep: typing.Union[int, torch.Tensor] = 0**kwargs: *<class 'transformers.utils.generic.TransformersKwargs'> ) → MoeCausalLMOutputWithPast or tuple(torch.Tensor)

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.Tensor of shape (batch_size, sequence_length, hidden_size), optional) — Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.
  • labels (torch.LongTensor of shape (batch_size, sequence_length), optional) — Labels for computing the causal 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].

    The labels are shifted internally so that labels[..., 1:] are used to predict logits[..., :-1, :]. Pass unshifted labels matching input_ids (i.e. labels = input_ids). Do not shift the labels yourself before passing them.

    If the labels are already shifted (for example with sequence parallelism), pass them as the shift_labels keyword argument and they are used as is. labels must still be passed for the loss to be computed.

  • 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

MoeCausalLMOutputWithPast or tuple(torch.Tensor)

A MoeCausalLMOutputWithPast or a tuple of torch.Tensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (Dots3NoteConfig) and inputs.

The Dots3NoteForCausalLM 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.Tensor of shape (1,), optional, returned when labels is provided) — Language modeling loss (for next-token prediction).

  • logits (torch.Tensor of shape (batch_size, sequence_length, config.vocab_size)) — Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).

  • aux_loss (torch.Tensor, optional, returned when output_router_logits=True is passed or when config.output_router_logits=True, and the model trains its router with a load-balancing loss) — Load-balancing auxiliary loss for the sparse modules. Models that balance their experts with a router bias instead (DeepSeek-V3 and the architectures derived from it) return None here while still returning router_logits.

  • router_logits (tuple(torch.Tensor), optional, returned when output_router_logits=True is passed or when config.output_router_logits=True) — Tuple of torch.Tensor (one for each sparse layer) of shape (batch_size * sequence_length, num_experts).

    Raw router logits computed by the MoE routers. They can be used to compute a load-balancing loss or to inspect how tokens are dispatched to experts.

  • 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.Tensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) — Tuple of torch.Tensor (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.Tensor), optional, returned when output_attentions=True is passed or when config.output_attentions=True) — Tuple of torch.Tensor (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 AutoModelForCausalLM, AutoTokenizer

>>> model_id = "dots-studio/dots3-note-prev"
>>> tokenizer = AutoTokenizer.from_pretrained(model_id)
>>> model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto")
>>> inputs = tokenizer("Briefly introduce yourself.", return_tensors="pt").to(model.device)
>>> output = model.generate(**inputs, max_new_tokens=32)
>>> text = tokenizer.batch_decode(output, skip_special_tokens=True)

Dots3NoteModel

class transformers.Dots3NoteModel

< >

( config: Dots3NoteConfig )

Parameters

  • config (Dots3NoteConfig) — 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 Dots3 Note 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: typing.Optional[torch.LongTensor] = Noneattention_mask: typing.Optional[torch.Tensor] = Noneposition_ids: typing.Optional[torch.LongTensor] = Nonepast_key_values: transformers.cache_utils.Cache | None = Noneinputs_embeds: typing.Optional[torch.FloatTensor] = Nonepixel_values: typing.Optional[torch.Tensor] = Noneimage_grid_thw: typing.Optional[torch.Tensor] = Nonepixel_values_videos: typing.Optional[torch.Tensor] = Nonevideo_grid_thw: typing.Optional[torch.Tensor] = Noneinput_features: typing.Optional[torch.Tensor] = Nonechunk_sample_lengths: typing.Optional[torch.Tensor] = Nonemm_encoder_outputs: dict[str, transformers.modeling_outputs.BaseModelOutputWithPooling] | None = None**kwargs ) → MoeModelOutputWithPast or tuple(torch.Tensor)

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.
  • pixel_values (torch.Tensor of shape (batch_size, num_channels, image_size, image_size), optional) — The tensors corresponding to the input images. Pixel values can be obtained using Qwen2VLImageProcessor. See Qwen2VLImageProcessor.call() for details (Dots3NoteProcessor uses Qwen2VLImageProcessor for processing images).
  • image_grid_thw (torch.Tensor of shape (num_images, 3), optional) — The temporal, height and width of feature shape of each image in LLM.
  • pixel_values_videos (torch.Tensor of shape (batch_size, num_frames, num_channels, frame_size, frame_size), optional) — The tensors corresponding to the input video. Pixel values for videos can be obtained using Qwen3VLVideoProcessor. See Qwen3VLVideoProcessor.call() for details (Dots3NoteProcessor uses Qwen3VLVideoProcessor for processing videos).
  • video_grid_thw (torch.Tensor of shape (num_videos, 3), optional) — The temporal, height and width of feature shape of each video in LLM.
  • input_features (torch.Tensor of shape (batch_size, sequence_length, feature_dim), optional) — The tensors corresponding to the input audio features. Audio features can be obtained using Dots3NoteFeatureExtractor. See Dots3NoteFeatureExtractor.__call__() for details (Dots3NoteProcessor uses Dots3NoteFeatureExtractor for processing audios).
  • chunk_sample_lengths (torch.Tensor, optional) — Waveform sample count for each audio feature chunk.
  • mm_encoder_outputs (dict[str, ~modeling_outputs.BaseModelOutputWithPooling], optional) — Dict where keys are supported modalities and values are encoded outputs for that modality. Each encoded output is a tuple that consists of (pooler_output, optional: last_hidden_states, optional: hidden_states, optional: attentions) pooler_output of shape (batch_size, sequence_length, hidden_size), optional) is a sequence of multimmodal features of the encoder merged into text embeddings.

Returns

MoeModelOutputWithPast or tuple(torch.Tensor)

A MoeModelOutputWithPast or a tuple of torch.Tensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (Dots3NoteConfig) and inputs.

The Dots3NoteModel 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.Tensor 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.Tensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) — Tuple of torch.Tensor (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.Tensor), optional, returned when output_attentions=True is passed or when config.output_attentions=True) — Tuple of torch.Tensor (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.Tensor), 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.Tensor (one for each layer) of shape (batch_size, sequence_length, num_experts).

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

Dots3NoteForConditionalGeneration

class transformers.Dots3NoteForConditionalGeneration

< >

( config: Dots3NoteConfig )

Parameters

  • config (Dots3NoteConfig) — 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 Dots3 Note Model for token generation conditioned on other modalities (e.g. image-text-to-text generation).

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: typing.Optional[torch.LongTensor] = Noneinput_features: typing.Optional[torch.FloatTensor] = Noneattention_mask: typing.Optional[torch.Tensor] = Noneposition_ids: typing.Optional[torch.LongTensor] = Nonepast_key_values: transformers.cache_utils.Cache | None = Noneinputs_embeds: typing.Optional[torch.FloatTensor] = Nonelabels: typing.Optional[torch.LongTensor] = Noneuse_cache: bool | None = Nonelogits_to_keep: typing.Union[int, torch.Tensor] = 0pixel_values: typing.Optional[torch.Tensor] = Nonepixel_values_videos: typing.Optional[torch.Tensor] = Noneimage_grid_thw: typing.Optional[torch.LongTensor] = Nonevideo_grid_thw: typing.Optional[torch.LongTensor] = Nonechunk_sample_lengths: typing.Optional[torch.Tensor] = Nonemm_encoder_outputs: dict[str, transformers.modeling_outputs.BaseModelOutputWithPooling] | None = None**kwargs: *<class 'transformers.utils.generic.TransformersKwargs'> ) → MoeCausalLMOutputWithPast or tuple(torch.Tensor)

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?

  • input_features (torch.FloatTensor of shape (batch_size, sequence_length, feature_dim), optional) — The tensors corresponding to the input audio features. Audio features can be obtained using Dots3NoteFeatureExtractor. See Dots3NoteFeatureExtractor.__call__() for details (Dots3NoteProcessor uses Dots3NoteFeatureExtractor for processing audios).
  • 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).
  • pixel_values (torch.Tensor of shape (batch_size, num_channels, image_size, image_size), optional) — The tensors corresponding to the input images. Pixel values can be obtained using Qwen2VLImageProcessor. See Qwen2VLImageProcessor.call() for details (Dots3NoteProcessor uses Qwen2VLImageProcessor for processing images).
  • pixel_values_videos (torch.Tensor of shape (batch_size, num_frames, num_channels, frame_size, frame_size), optional) — The tensors corresponding to the input video. Pixel values for videos can be obtained using Qwen3VLVideoProcessor. See Qwen3VLVideoProcessor.call() for details (Dots3NoteProcessor uses Qwen3VLVideoProcessor for processing videos).
  • image_grid_thw (torch.LongTensor of shape (num_images, 3), optional) — The temporal, height and width of feature shape of each image in LLM.
  • video_grid_thw (torch.LongTensor of shape (num_videos, 3), optional) — The temporal, height and width of feature shape of each video in LLM.
  • chunk_sample_lengths (torch.Tensor, optional) — Number of waveform samples in each audio chunk.
  • mm_encoder_outputs (dict[str, ~modeling_outputs.BaseModelOutputWithPooling], optional) — Dict where keys are supported modalities and values are encoded outputs for that modality. Each encoded output is a tuple that consists of (pooler_output, optional: last_hidden_states, optional: hidden_states, optional: attentions) pooler_output of shape (batch_size, sequence_length, hidden_size), optional) is a sequence of multimmodal features of the encoder merged into text embeddings.

Returns

MoeCausalLMOutputWithPast or tuple(torch.Tensor)

A MoeCausalLMOutputWithPast or a tuple of torch.Tensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (Dots3NoteConfig) and inputs.

The Dots3NoteForConditionalGeneration 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.Tensor of shape (1,), optional, returned when labels is provided) — Language modeling loss (for next-token prediction).

  • logits (torch.Tensor of shape (batch_size, sequence_length, config.vocab_size)) — Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).

  • aux_loss (torch.Tensor, optional, returned when output_router_logits=True is passed or when config.output_router_logits=True, and the model trains its router with a load-balancing loss) — Load-balancing auxiliary loss for the sparse modules. Models that balance their experts with a router bias instead (DeepSeek-V3 and the architectures derived from it) return None here while still returning router_logits.

  • router_logits (tuple(torch.Tensor), optional, returned when output_router_logits=True is passed or when config.output_router_logits=True) — Tuple of torch.Tensor (one for each sparse layer) of shape (batch_size * sequence_length, num_experts).

    Raw router logits computed by the MoE routers. They can be used to compute a load-balancing loss or to inspect how tokens are dispatched to experts.

  • 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.Tensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) — Tuple of torch.Tensor (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.Tensor), optional, returned when output_attentions=True is passed or when config.output_attentions=True) — Tuple of torch.Tensor (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 PIL import Image
>>> from transformers import AutoProcessor, Dots3NoteForConditionalGeneration

>>> model = Dots3NoteForConditionalGeneration.from_pretrained("dots-studio/dots3-note-prev")
>>> processor = AutoProcessor.from_pretrained("dots-studio/dots3-note-prev")

>>> messages = [
...     {
...         "role": "user", "content": [
...             {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/pipeline-cat-chonk.jpeg"},
...             {"type": "text", "text": "Where is the cat standing?"},
...         ]
...     },
... ]

>>> inputs = processor.apply_chat_template(
...     messages,
...     tokenize=True,
...     return_dict=True,
...     return_tensors="pt",
...     add_generation_prompt=True
... )
>>> # Generate
>>> generate_ids = model.generate(**inputs)
>>> processor.batch_decode(generate_ids, skip_special_tokens=True)[0]

Dots3NoteTextModel

class transformers.Dots3NoteTextModel

< >

( config: Dots3NoteTextConfig )

Parameters

  • config (Dots3NoteTextConfig) — 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 Dots3 Note Text 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: typing.Optional[torch.LongTensor] = Noneattention_mask: typing.Optional[torch.Tensor] = Noneposition_ids: typing.Optional[torch.LongTensor] = Nonepast_key_values: transformers.cache_utils.Cache | None = Noneinputs_embeds: typing.Optional[torch.FloatTensor] = Noneuse_cache: bool | None = None**kwargs: *<class 'transformers.utils.generic.TransformersKwargs'> ) → MoeModelOutputWithPast or tuple(torch.Tensor)

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.Tensor)

A MoeModelOutputWithPast or a tuple of torch.Tensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (Dots3NoteConfig) and inputs.

The Dots3NoteTextModel 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.Tensor 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.Tensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) — Tuple of torch.Tensor (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.Tensor), optional, returned when output_attentions=True is passed or when config.output_attentions=True) — Tuple of torch.Tensor (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.Tensor), 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.Tensor (one for each layer) of shape (batch_size, sequence_length, num_experts).

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

Dots3NoteVisionModel

class transformers.Dots3NoteVisionModel

< >

( config: Dots3NoteVisionConfig )

Parameters

  • config (Dots3NoteVisionConfig) — 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 Dots3 Note 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

< >

( pixel_values: Tensorgrid_thw: Tensor**kwargs ) → BaseModelOutputWithPooling or tuple(torch.Tensor)

Parameters

Returns

BaseModelOutputWithPooling or tuple(torch.Tensor)

A BaseModelOutputWithPooling or a tuple of torch.Tensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (Dots3NoteConfig) and inputs.

The Dots3NoteVisionModel 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.Tensor 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.Tensor 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.Tensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) — Tuple of torch.Tensor (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.Tensor), optional, returned when output_attentions=True is passed or when config.output_attentions=True) — Tuple of torch.Tensor (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.

Dots3NoteAudioModel

class transformers.Dots3NoteAudioModel

< >

( config: Dots3NoteAudioConfig )

Parameters

  • config (Dots3NoteAudioConfig) — 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 Dots3 Note 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_features: Tensorchunk_sample_lengths: Tensor**kwargs: *<class 'transformers.utils.generic.TransformersKwargs'> ) → BaseModelOutputWithPooling or tuple(torch.Tensor)

Parameters

  • input_features (torch.Tensor of shape (batch_size, sequence_length, feature_dim)) — The tensors corresponding to the input audio features. Audio features can be obtained using Dots3NoteFeatureExtractor. See Dots3NoteFeatureExtractor.__call__() for details (Dots3NoteProcessor uses Dots3NoteFeatureExtractor for processing audios).
  • chunk_sample_lengths (torch.LongTensor of shape (num_chunks,)) — Number of waveform samples in each audio chunk.

Returns

BaseModelOutputWithPooling or tuple(torch.Tensor)

A BaseModelOutputWithPooling or a tuple of torch.Tensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (Dots3NoteConfig) and inputs.

The Dots3NoteAudioModel 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.Tensor 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.Tensor 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.Tensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) — Tuple of torch.Tensor (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.Tensor), optional, returned when output_attentions=True is passed or when config.output_attentions=True) — Tuple of torch.Tensor (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.

Dots3NoteProcessor

class transformers.Dots3NoteProcessor

< >

( image_processor = Nonetokenizer = Nonevideo_processor = Nonefeature_extractor = Nonechat_template = None )

Parameters

  • image_processor (Qwen2VLImageProcessor) — The image processor is a required input.
  • tokenizer (Qwen2Tokenizer) — The tokenizer is a required input.
  • video_processor (Qwen3VLVideoProcessor) — The video processor is a required input.
  • feature_extractor (Dots3NoteFeatureExtractor) — The feature extractor is a required input.
  • chat_template (str) — A Jinja template to convert lists of messages in a chat into a tokenizable string.

Constructs a Dots3NoteProcessor which wraps a image processor, a tokenizer, a video processor, and a feature extractor into a single processor.

Dots3NoteProcessor offers all the functionalities of Qwen2VLImageProcessor, Qwen2Tokenizer, Qwen3VLVideoProcessor, and Dots3NoteFeatureExtractor. See the ~Qwen2VLImageProcessor, ~Qwen2Tokenizer, ~Qwen3VLVideoProcessor, and ~Dots3NoteFeatureExtractor for more information.

__call__

< >

( images: typing.Union[ForwardRef('PIL.Image.Image'), numpy.ndarray, ForwardRef('torch.Tensor'), list['PIL.Image.Image'], list[numpy.ndarray], list['torch.Tensor'], NoneType] = Nonetext: str | list[str] | list[list[str]] | None = Nonevideos: typing.Union[list['PIL.Image.Image'], numpy.ndarray, ForwardRef('torch.Tensor'), list[numpy.ndarray], list['torch.Tensor'], list[list['PIL.Image.Image']], list[list[numpy.ndarray]], list[list['torch.Tensor']], transformers.video_utils.URL, list[transformers.video_utils.URL], list[list[transformers.video_utils.URL]], transformers.video_utils.Path, list[transformers.video_utils.Path], list[list[transformers.video_utils.Path]], NoneType] = Noneaudio: typing.Union[numpy.ndarray, ForwardRef('torch.Tensor'), collections.abc.Sequence[numpy.ndarray], collections.abc.Sequence['torch.Tensor'], NoneType] = None**kwargs: *<class 'transformers.processing_utils.ProcessingKwargs'> )

Parameters

  • images (Union[PIL.Image.Image, numpy.ndarray, torch.Tensor, list[PIL.Image.Image], list[numpy.ndarray], list[torch.Tensor]], optional) — Image to preprocess. Expects a single or batch of images with pixel values ranging from 0 to 255. If passing in images with pixel values between 0 and 1, set do_rescale=False.
  • text (Union[str, list[str], list[list[str]]], optional) — The sequence or batch of sequences to be encoded. Each sequence can be a string or a list of strings (pretokenized string). If you pass a pretokenized input, set is_split_into_words=True to avoid ambiguity with batched inputs.
  • videos (Union[list[PIL.Image.Image], numpy.ndarray, torch.Tensor, list[numpy.ndarray], list[torch.Tensor], list[list[PIL.Image.Image]], list[list[numpy.ndarray]], list[list[torch.Tensor]], ~video_utils.URL, list[~video_utils.URL], list[list[~video_utils.URL]], ~video_utils.Path, list[~video_utils.Path], list[list[~video_utils.Path]]], optional) — Video to preprocess. Expects a single or batch of videos with pixel values ranging from 0 to 255. If passing in videos with pixel values between 0 and 1, set do_rescale=False.
  • audio (Union[numpy.ndarray, torch.Tensor, collections.abc.Sequence[numpy.ndarray], collections.abc.Sequence[torch.Tensor]], optional) — The audio or batch of audios to be prepared. Each audio can be a NumPy array or PyTorch tensor. In case of a NumPy array/PyTorch tensor, each audio should be of shape (C, T), where C is a number of channels, and T is the sample length of the audio.
  • return_mm_token_type_ids (bool, kwargs) — Whether to return multimodal token type ids indicating multimodal placeholder token positions.
  • return_text_replacement_offsets (bool, kwargs) — Whether to return character offsets for each multimodal placeholder and its replacement.
  • return_tensors (str or TensorType, optional) — If set, will return tensors of a particular framework. Acceptable values are:
    • 'pt': Return PyTorch torch.Tensor objects.
    • 'np': Return NumPy np.ndarray objects.
  • **kwargs (ProcessingKwargs, optional) — Additional processing options for each modality (text, images, videos, audio). Model-specific parameters are listed above; see the TypedDict class for the complete list of supported arguments.

Dots3NoteFeatureExtractor

class transformers.Dots3NoteFeatureExtractor

< >

( feature_size: int = 128sampling_rate: int = 16000padding_value: float = 0.0n_fft: int = 400hop_length: int = 160chunk_length: int = 60chunk_seconds: int | None = Nonereturn_attention_mask: bool = Falsedither: float = 0.0**kwargs )

Convert 16 kHz mono waveforms into Dots 3 Note Preview log-mel chunks.

Update on GitHub

←Donut EdgeTAM→