Transformers documentation

Apertus 1.5

You are viewing main version, which requires installation from source. If you'd like regular pip install, checkout the latest stable version (v5.23.0).
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 2026-11-16.

Apertus 1.5

Overview

Apertus 1.5 is a multimodal model from the Swiss AI Initiative that accepts text, images, and audio and generates text. It extends the Apertus language model with the EMU3.5 vision tokenizer and WavTokenizer for audio. The EMU3.5 vision tokenizer is an image-only VQGAN with the same encoder as Chameleon’s, unlike the Emu3 tokenizer, which adds temporal layers for video. It assigns codes by the largest inner product with the codebook instead of the smallest L2 distance.

The media tokenizers convert images and audio into discrete tokens, which the language backbone processes alongside text in a shared input vocabulary. The output vocabulary contains only text tokens. The integration includes the media encoders and quantizers; it does not support image or audio generation, or training the vision tokenizer.

Usage example

Use the processor’s chat template to prepare a conversation with images and audio. It loads the referenced media, resamples audio, and constructs the model inputs automatically.

from transformers import AutoModelForMultimodalLM, AutoProcessor

model = AutoModelForMultimodalLM.from_pretrained(
    "swiss-ai/Apertus-v1.5-8B", device_map="auto"
)
processor = AutoProcessor.from_pretrained("swiss-ai/Apertus-v1.5-8B")

messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/bee.jpg"},
            {"type": "audio", "url": "https://huggingface.co/datasets/hf-internal-testing/dummy-audio-samples/resolve/main/belinda.wav"},
            {"type": "text", "text": "Describe the image and transcribe the audio."},
        ],
    }
]
inputs = processor.apply_chat_template(
    messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt"
).to(model.device)

generated = model.generate(**inputs, max_new_tokens=128)
print(processor.decode(generated[0, inputs["input_ids"].shape[1] :], skip_special_tokens=True))

Usage notes

  • Tokenizer precision: Both media tokenizers require float32 for stable code assignments. from_pretrained preserves this precision when loading the language backbone in half precision. Avoid casting the entire model to a lower precision or running the tokenizers under mixed-precision autocast. If quantizing the model, ensure the backend excludes both media tokenizers.
  • Output vocabulary: Returned logits cover only text_config.output_vocab_size when set, otherwise text_config.vocab_size. Input-only multimodal tokens cannot be generated. Mask these tokens with -100 in training labels, and restrict generation constraints to tokens within the output vocabulary.

Apertus1p5Config

class transformers.Apertus1p5Config

< >

( 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.apertus1p5.configuration_apertus1p5.Apertus1p5TextConfig | None = Nonevision_config: dict | transformers.models.apertus1p5.configuration_apertus1p5.Apertus1p5VisionTokenizerConfig | None = Noneaudio_config: dict | transformers.configuration_utils.PreTrainedConfig | None = Noneimage_token_id: int = 131079audio_token_id: int = 131085image_token_offset: int = 131272audio_token_offset: int = 262344tie_word_embeddings: bool = False )

Parameters

  • text_config (Union[dict, ~models.apertus1p5.configuration_apertus1p5.Apertus1p5TextConfig], optional) — The config object or dictionary of the text backbone.
  • vision_config (Union[dict, ~models.apertus1p5.configuration_apertus1p5.Apertus1p5VisionTokenizerConfig], optional) — The config object or dictionary of the vision backbone.
  • audio_config (Union[dict, ~configuration_utils.PreTrainedConfig], optional) — The config object or dictionary of the audio backbone.
  • image_token_id (int, optional, defaults to 131079) — The image token index used as a placeholder for input images.
  • audio_token_id (int, optional, defaults to 131085) — The audio token index used as a placeholder for input audio.
  • image_token_offset (int, optional, defaults to 131272) — Offset mapping a visual codebook index i to the vocabulary id of <|visual token i|>.
  • audio_token_offset (int, optional, defaults to 262344) — Offset mapping an audio codebook index i to the vocabulary id of <|audio token i|>.
  • 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 Apertus1p5Model. It is used to instantiate a Apertus1P5 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 swiss-ai/Apertus-v1.5-8B

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 Apertus1p5Config, Apertus1p5ForConditionalGeneration

>>> # Initializing a configuration (Apertus 1.5 backbone + vision/audio tokenizer defaults)
>>> configuration = Apertus1p5Config()

>>> # Initializing a model (with random weights) from the configuration
>>> model = Apertus1p5ForConditionalGeneration(configuration)

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

Apertus1p5TextConfig

class transformers.Apertus1p5TextConfig

< >

( 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 = 266752hidden_size: int = 4096intermediate_size: int = 14336num_hidden_layers: int = 32num_attention_heads: int = 32num_key_value_heads: int | None = Nonehidden_act: str = 'xielu'max_position_embeddings: int = 65536initializer_range: float = 0.02rms_norm_eps: float = 1e-05use_cache: bool = Truepad_token_id: int | None = 3bos_token_id: int | None = 1eos_token_id: int | list[int] | None = 2tie_word_embeddings: bool = Falserope_parameters: transformers.modeling_rope_utils.RopeParameters | dict | None = Noneattention_bias: bool = Falseattention_dropout: float | int = 0.0output_vocab_size: int | None = None )

Parameters

  • vocab_size (int, optional, defaults to 266752) — 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 14336) — 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 xielu) — The non-linear activation function (function or string) in the decoder. For example, "gelu", "relu", "silu", etc.
  • max_position_embeddings (int, optional, defaults to 65536) — 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 3) — 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 (Union[int, list[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], 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.
  • output_vocab_size (int, optional) — Number of LM-head rows kept after pruning the multimodal token rows from the output projection; the retained output ids are 0..output_vocab_size - 1. None means the head is unpruned (vocab_size rows). Input embeddings always use vocab_size, while returned logits use the LM-head width, with or without labels. Released checkpoints use tie_word_embeddings=False to keep the pruned head separate from the input embeddings.

This is the configuration class to store the configuration of a Apertus1p5Model. It is used to instantiate a Apertus1P5 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 swiss-ai/Apertus-v1.5-8B

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 Apertus1p5TextConfig, Apertus1p5TextForCausalLM

>>> # Initializing an Apertus 1.5 text configuration (extended vocabulary, optionally pruned output head)
>>> configuration = Apertus1p5TextConfig()

>>> # Initializing a model (with random weights) from the configuration
>>> model = Apertus1p5TextForCausalLM(configuration)

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

Apertus1p5VisionTokenizerConfig

class transformers.Apertus1p5VisionTokenizerConfig

< >

( 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 = Nonecodebook_size: int = 131072embed_dim: int = 256latent_channels: int = 256in_channels: int = 3base_channels: int = 256channel_multiplier: list[int] | tuple[int, ...] = (1, 1, 2, 2, 4)num_res_blocks: int = 4attn_resolutions: list[int] | tuple[int, ...] = (16,)resolution: int = 256dropout: float = 0.0 )

Parameters

  • codebook_size (int, optional, defaults to 131072) — Number of entries in the IBQ codebook.
  • embed_dim (int, optional, defaults to 256) — Dimension of the codebook vectors (the quantizer input after quant_conv).
  • latent_channels (int, optional, defaults to 256) — Number of channels output by the encoder (before quant_conv).
  • in_channels (int, optional, defaults to 3) — The number of input channels.
  • base_channels (int, optional, defaults to 256) — Base channel count of the encoder; per-stage widths are base_channels * channel_multiplier[stage].
  • channel_multiplier (list[int], optional, defaults to (1, 1, 2, 2, 4)) — Channel scaling factor per encoder stage. The number of stages is len(channel_multiplier), giving a spatial downsampling factor of 2**(len(channel_multiplier) - 1) (16 by default).
  • num_res_blocks (int, optional, defaults to 4) — Number of residual blocks per encoder stage.
  • attn_resolutions (list[int], optional, defaults to (16,)) — Feature-map resolutions (relative to a resolution-sized input) at which attention blocks are inserted.
  • resolution (int, optional, defaults to 256) — Reference input resolution used (statically, together with attn_resolutions) to decide which encoder stages carry attention blocks. Does not restrict the actual input size.
  • dropout (float, optional, defaults to 0.0) — Dropout inside the residual blocks.

This is the configuration class to store the configuration of a Apertus1p5Model. It is used to instantiate a Apertus1P5 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 swiss-ai/Apertus-v1.5-8B

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 Apertus1p5VisionTokenizerConfig, Apertus1p5VisionTokenizerModel

>>> # Initializing a vision tokenizer configuration (Apertus 1.5 defaults)
>>> configuration = Apertus1p5VisionTokenizerConfig()

>>> # Initializing the encode-only tokenizer (with random weights) from the configuration
>>> model = Apertus1p5VisionTokenizerModel(configuration)

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

Apertus1p5Processor

class transformers.Apertus1p5Processor

< >

( image_processor = Nonefeature_extractor = Nonetokenizer = Nonechat_template = None**kwargs )

Parameters

  • image_processor (Apertus1p5ImageProcessor) — The image processor is a required input.
  • feature_extractor (WavTokenizerFeatureExtractor) — The feature extractor is a required input.
  • tokenizer (tokenizer_class) — The tokenizer is a required input.
  • chat_template (str) — A Jinja template to convert lists of messages in a chat into a tokenizable string.

Constructs an Apertus 1.5 processor for text, images, and audio.

__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.

Apertus1p5ImageProcessor

class transformers.Apertus1p5ImageProcessor

< >

( **kwargs: *<class 'transformers.models.apertus1p5.image_processing_apertus1p5.Apertus1p5ImageProcessorKwargs'> )

Parameters

  • do_convert_rgb (bool, kwargs, optional, defaults to True) — Whether to convert the image to RGB.
  • do_resize (bool, kwargs, optional, defaults to True) — Whether to resize the image.
  • size (dict, kwargs, optional, defaults to {"min_pixels" -- 128 * 128, "max_pixels": 1400 * 1400}): Minimum and maximum pixel areas. Images are resized to approximately this area range.
  • default_to_square (bool, kwargs, optional, defaults to True) — Whether to default to a square image when resizing, if size is an int.
  • crop_size (Annotated[int | list[int] | tuple[int, ...] | dict[str, int] | None, None], kwargs) — Size of the output image after applying center_crop.
  • resample (Annotated[Union[int, PILImageResampling, NoneType], None], kwargs, defaults to 3) — Resampling filter to use if resizing the image. This can be one of the enum PILImageResampling. Only has an effect if do_resize is set to True.
  • do_rescale (bool, kwargs, optional, defaults to True) — Whether to rescale the image.
  • rescale_factor (float, kwargs, optional, defaults to 0.00392156862745098) — Rescale factor to rescale the image by if do_rescale is set to True.
  • do_normalize (bool, kwargs, optional, defaults to True) — Whether to normalize the image.
  • image_mean (Union[float, list[float], tuple[float, ...]], kwargs, optional, defaults to [0.5, 0.5, 0.5]) — Image mean to use for normalization. Only has an effect if do_normalize is set to True.
  • image_std (Union[float, list[float], tuple[float, ...]], kwargs, optional, defaults to [0.5, 0.5, 0.5]) — Image standard deviation to use for normalization. Only has an effect if do_normalize is set to True.
  • do_pad (bool, kwargs, optional, defaults to True) — Whether to pad the image. Padding is done either to the largest size in the batch or to a fixed square size per image. The exact padding strategy depends on the model.
  • pad_size (Annotated[int | list[int] | tuple[int, ...] | dict[str, int] | None, None], kwargs) — The size in {"height": int, "width" int} to pad the images to. Must be larger than any image size provided for preprocessing. If pad_size is not provided, images will be padded to the largest height and width in the batch. Applied only when do_pad=True.
  • do_center_crop (bool, kwargs, optional) — Whether to center crop the image.
  • data_format (Union[str, ~image_utils.ChannelDimension], kwargs, optional) — Only ChannelDimension.FIRST is supported. Added for compatibility with slow processors.
  • input_data_format (Union[str, ~image_utils.ChannelDimension], kwargs, optional) — The channel dimension format for the input image. If unset, the channel dimension format is inferred from the input image. Can be one of:
    • "channels_first" or ChannelDimension.FIRST: image in (num_channels, height, width) format.
    • "channels_last" or ChannelDimension.LAST: image in (height, width, num_channels) format.
    • "none" or ChannelDimension.NONE: image in (height, width) format.
  • device (Annotated[Union[str, torch.device, NoneType], None], kwargs) — The device to process the videos on. If unset, the device is inferred from the input videos.
  • return_tensors (Annotated[str | ~utils.generic.TensorType | None, None], kwargs) — Returns stacked tensors if set to 'pt', otherwise returns a list of tensors.
  • disable_grouping (bool, kwargs, optional) — Whether to disable grouping of images by size to process them individually and not in batches. If None, will be set to True if the images are on CPU, and False otherwise. This choice is based on empirical observations, as detailed here: https://github.com/huggingface/transformers/pull/38157
  • image_seq_length (int, kwargs, optional) — The number of image tokens to be used for each image in the input. Added for backward compatibility but this should be set as a processor attribute in future models.
  • spatial_factor (int, kwargs, optional, defaults to 16) — Both output sides are rounded to multiples of this factor (the vision tokenizer’s downsampling factor); each spatial_factor x spatial_factor patch becomes one discrete code.

Constructs an Apertus 1.5 image processor with the training pipeline’s resize policy.

preprocess

< >

( images: typing.Union[ForwardRef('PIL.Image.Image'), numpy.ndarray, ForwardRef('torch.Tensor'), list['PIL.Image.Image'], list[numpy.ndarray], list['torch.Tensor']]**kwargs: *<class 'transformers.models.apertus1p5.image_processing_apertus1p5.Apertus1p5ImageProcessorKwargs'> ) → ~image_processing_base.BatchFeature

Parameters

  • images (Union[PIL.Image.Image, numpy.ndarray, torch.Tensor, list[PIL.Image.Image], list[numpy.ndarray], list[torch.Tensor]]) — 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.
  • do_convert_rgb (bool, kwargs, optional) — Whether to convert the image to RGB.
  • do_resize (bool, kwargs, optional) — Whether to resize the image.
  • size (dict, kwargs, optional, defaults to {"min_pixels" -- 128 * 128, "max_pixels": 1400 * 1400}): Minimum and maximum pixel areas. Images are resized to approximately this area range.
  • default_to_square (bool, kwargs, optional) — Whether to default to a square image when resizing, if size is an int.
  • crop_size (Annotated[int | list[int] | tuple[int, ...] | dict[str, int] | None, None], kwargs) — Size of the output image after applying center_crop.
  • resample (Annotated[Union[int, PILImageResampling, NoneType], None], kwargs) — Resampling filter to use if resizing the image. This can be one of the enum PILImageResampling. Only has an effect if do_resize is set to True.
  • do_rescale (bool, kwargs, optional) — Whether to rescale the image.
  • rescale_factor (float, kwargs, optional) — Rescale factor to rescale the image by if do_rescale is set to True.
  • do_normalize (bool, kwargs, optional) — Whether to normalize the image.
  • image_mean (Union[float, list[float], tuple[float, ...]], kwargs, optional) — Image mean to use for normalization. Only has an effect if do_normalize is set to True.
  • image_std (Union[float, list[float], tuple[float, ...]], kwargs, optional) — Image standard deviation to use for normalization. Only has an effect if do_normalize is set to True.
  • do_pad (bool, kwargs, optional) — Whether to pad the image. Padding is done either to the largest size in the batch or to a fixed square size per image. The exact padding strategy depends on the model.
  • pad_size (Annotated[int | list[int] | tuple[int, ...] | dict[str, int] | None, None], kwargs) — The size in {"height": int, "width" int} to pad the images to. Must be larger than any image size provided for preprocessing. If pad_size is not provided, images will be padded to the largest height and width in the batch. Applied only when do_pad=True.
  • do_center_crop (bool, kwargs, optional) — Whether to center crop the image.
  • data_format (Union[str, ~image_utils.ChannelDimension], kwargs, optional) — Only ChannelDimension.FIRST is supported. Added for compatibility with slow processors.
  • input_data_format (Union[str, ~image_utils.ChannelDimension], kwargs, optional) — The channel dimension format for the input image. If unset, the channel dimension format is inferred from the input image. Can be one of:
    • "channels_first" or ChannelDimension.FIRST: image in (num_channels, height, width) format.
    • "channels_last" or ChannelDimension.LAST: image in (height, width, num_channels) format.
    • "none" or ChannelDimension.NONE: image in (height, width) format.
  • device (Annotated[Union[str, torch.device, NoneType], None], kwargs) — The device to process the videos on. If unset, the device is inferred from the input videos.
  • return_tensors (Annotated[str | ~utils.generic.TensorType | None, None], kwargs) — Returns stacked tensors if set to 'pt', otherwise returns a list of tensors.
  • disable_grouping (bool, kwargs, optional) — Whether to disable grouping of images by size to process them individually and not in batches. If None, will be set to True if the images are on CPU, and False otherwise. This choice is based on empirical observations, as detailed here: https://github.com/huggingface/transformers/pull/38157
  • image_seq_length (int, kwargs, optional) — The number of image tokens to be used for each image in the input. Added for backward compatibility but this should be set as a processor attribute in future models.
  • spatial_factor (int, kwargs, optional, defaults to 16) — Both output sides are rounded to multiples of this factor (the vision tokenizer’s downsampling factor); each spatial_factor x spatial_factor patch becomes one discrete code.

Returns

~image_processing_base.BatchFeature

  • data (dict) — Dictionary of lists/arrays/tensors returned by the call method (‘pixel_values’, etc.).
  • tensor_type (Union[None, str, TensorType], optional) — You can give a tensor_type here to convert the lists of integers in PyTorch/Numpy Tensors at initialization.

Apertus1p5VisionTokenizerModel

class transformers.Apertus1p5VisionTokenizerModel

< >

( config: Apertus1p5VisionTokenizerConfig )

Parameters

  • config (Apertus1p5VisionTokenizerConfig) — 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.

Encode-only EMU3.5 vision tokenizer used by Apertus 1.5: a 2D VQGAN encoder as in Chameleon, followed by a quantizer that picks the codebook entry with the largest inner product. It returns hard codebook indices; run it in float32 for stable code assignment (Apertus 1.5 does this automatically).

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.

encode

< >

( pixel_values: Tensor ) → Apertus1p5VisionTokenizerEncoderOutput

Parameters

  • pixel_values (torch.Tensor of shape (batch_size, channels, height, width)) — RGB images normalized to [-1, 1].

Returns

Apertus1p5VisionTokenizerEncoderOutput

Discrete image-code grids in image_tokens.

Tokenizes images into a grid of discrete codebook indices.

Apertus1p5TextModel

class transformers.Apertus1p5TextModel

< >

( config: Apertus1p5TextConfig )

Parameters

  • config (Apertus1p5TextConfig) — 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 Apertus1P5 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] = Nonecache: transformers.cache_utils.Cache | None = Noneinputs_embeds: typing.Optional[torch.Tensor] = Noneuse_cache: bool | None = None**kwargs: *<class 'transformers.utils.generic.TransformersKwargs'> ) → BaseModelOutputWithPast 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?

  • cache (~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 cache 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 cache are passed, DynamicCache will be initialized by default.

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

    If cache 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.
  • use_cache (bool, optional) — If set to True, cache key value states are returned and can be used to speed up decoding (see cache).

Returns

BaseModelOutputWithPast or tuple(torch.Tensor)

A BaseModelOutputWithPast 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 (Apertus1p5Config) and inputs.

The Apertus1p5TextModel 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.

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

  • cache (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 cache 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.

Apertus1p5TextForCausalLM

class transformers.Apertus1p5TextForCausalLM

< >

( config: Apertus1p5TextConfig )

Parameters

  • config (Apertus1p5TextConfig) — 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 Apertus1P5 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] = Nonecache: 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'> ) → CausalLMOutputWithPast 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?

  • cache (~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 cache 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 cache are passed, DynamicCache will be initialized by default.

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

    If cache 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) — Target token ids. -100 ignores a position; other values must be within the physical LM-head range. With a pruned head, mask multimodal/input-only ids with -100.
  • use_cache (bool, optional) — If set to True, cache key value states are returned and can be used to speed up decoding (see cache).
  • 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.Tensor)

Logits use the physical LM-head width (config.output_vocab_size when set, otherwise config.vocab_size), with or without labels.

The Apertus1p5TextForCausalLM 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.

Apertus1p5Model

class transformers.Apertus1p5Model

< >

( config: Apertus1p5Config )

Parameters

  • config (Apertus1p5Config) — 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 Apertus1P5 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] = Nonepixel_values: typing.Optional[torch.Tensor] = Noneimage_sizes: typing.Optional[torch.Tensor] = Noneinput_values: typing.Optional[torch.Tensor] = Nonepadding_mask: typing.Optional[torch.Tensor] = Noneattention_mask: typing.Optional[torch.Tensor] = Noneposition_ids: typing.Optional[torch.LongTensor] = Nonecache: transformers.cache_utils.Cache | None = Noneinputs_embeds: typing.Optional[torch.Tensor] = Noneuse_cache: bool | None = Nonemm_encoder_outputs: dict[str, transformers.modeling_outputs.BaseModelOutputWithPooling] | None = None**kwargs: *<class 'transformers.utils.generic.TransformersKwargs'> ) → BaseModelOutputWithPast 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?

  • 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 image_processor_class. See image_processor_class.__call__ for details (processor_class uses image_processor_class for processing images).
  • image_sizes (torch.Tensor of shape (batch_size, 2), optional) — The sizes of the images in the batch, being (height, width) for each image.
  • input_values (torch.Tensor of shape (batch_size, sequence_length), optional) — Float values of input raw speech waveform. Values can be obtained by loading a .flac or .wav audio file into an array of type list[float], a numpy.ndarray or a torch.Tensor, e.g. via the torchcodec library (pip install torchcodec) or the soundfile library (pip install soundfile). To prepare the array into input_values, the AutoProcessor should be used for padding and conversion into a tensor of type torch.Tensor. See processor_class.__call__ for details.
  • padding_mask (torch.Tensor of shape (num_clips, max_length), optional) — Right-padded mask with 1 for valid samples and 0 for padding, required with input_values.
  • 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?

  • cache (~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 cache 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 cache are passed, DynamicCache will be initialized by default.

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

    If cache 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.
  • use_cache (bool, optional) — If set to True, cache key value states are returned and can be used to speed up decoding (see cache).
  • 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

BaseModelOutputWithPast or tuple(torch.Tensor)

A BaseModelOutputWithPast 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 (None) and inputs.

The Apertus1p5Model 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.

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

  • cache (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 cache 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.

Apertus1p5ForConditionalGeneration

class transformers.Apertus1p5ForConditionalGeneration

< >

( config: Apertus1p5Config )

Parameters

  • config (Apertus1p5Config) — 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 Apertus1P5 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] = Nonepixel_values: typing.Optional[torch.Tensor] = Noneimage_sizes: typing.Optional[torch.Tensor] = Noneinput_values: typing.Optional[torch.Tensor] = Nonepadding_mask: typing.Optional[torch.Tensor] = Noneattention_mask: typing.Optional[torch.Tensor] = Noneposition_ids: typing.Optional[torch.LongTensor] = Nonecache: transformers.cache_utils.Cache | None = Noneinputs_embeds: typing.Optional[torch.Tensor] = Noneuse_cache: bool | None = Nonelabels: typing.Optional[torch.LongTensor] = Nonelogits_to_keep: typing.Union[int, torch.Tensor] = 0mm_encoder_outputs: dict[str, transformers.modeling_outputs.BaseModelOutputWithPooling] | None = None**kwargs: *<class 'transformers.utils.generic.TransformersKwargs'> ) → CausalLMOutputWithPast 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?

  • 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 image_processor_class. See image_processor_class.__call__ for details (processor_class uses image_processor_class for processing images).
  • image_sizes (torch.Tensor of shape (batch_size, 2), optional) — The sizes of the images in the batch, being (height, width) for each image.
  • input_values (torch.Tensor of shape (batch_size, sequence_length), optional) — Float values of input raw speech waveform. Values can be obtained by loading a .flac or .wav audio file into an array of type list[float], a numpy.ndarray or a torch.Tensor, e.g. via the torchcodec library (pip install torchcodec) or the soundfile library (pip install soundfile). To prepare the array into input_values, the AutoProcessor should be used for padding and conversion into a tensor of type torch.Tensor. See processor_class.__call__ for details.
  • padding_mask (torch.Tensor of shape (num_clips, max_length), optional) — Audio sample mask: 1 for valid samples, 0 for right padding. Required with input_values.
  • 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?

  • cache (~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 cache 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 cache are passed, DynamicCache will be initialized by default.

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

    If cache 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.
  • use_cache (bool, optional) — If set to True, cache key value states are returned and can be used to speed up decoding (see cache).
  • labels (torch.LongTensor of shape (batch_size, sequence_length), optional) — Next-token labels in [0, output_vocab_size) (or [0, text_config.vocab_size) when unset). Use -100 for input-only multimodal tokens and other ignored positions.
  • 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).
  • 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

CausalLMOutputWithPast or tuple(torch.Tensor)

Logits have width text_config.output_vocab_size, or text_config.vocab_size when unset, with or without labels.

The Apertus1p5ForConditionalGeneration 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.

get_image_features

< >

( pixel_values: Tensorimage_sizes: LongTensor**kwargs: *<class 'transformers.utils.generic.TransformersKwargs'> ) → Apertus1p5VisionTokenizerModelOutput or tuple(torch.Tensor)

Parameters

Returns

Apertus1p5VisionTokenizerModelOutput or tuple(torch.Tensor)

A Apertus1p5VisionTokenizerModelOutput 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 (Apertus1p5Config) and inputs.

  • 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 (tuple[torch.Tensor], optional) — One (num_tokens, hidden_size) tensor of embedded image tokens per input image.

  • 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.

  • image_tokens (list[torch.LongTensor], optional) — One flat tensor of shared-vocabulary token ids per input image: its (height // factor) * (width // factor) codes in row-major order, offset by config.image_token_offset, where factor is the vision tokenizer’s spatial_scale_factor (16 by default).

Example:

>>> from PIL import Image
>>> from transformers import AutoProcessor, Apertus1p5ForConditionalGeneration

>>> model = Apertus1p5ForConditionalGeneration.from_pretrained("swiss-ai/Apertus-v1.5-8B")
>>> processor = AutoProcessor.from_pretrained("swiss-ai/Apertus-v1.5-8B")

>>> 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]

get_audio_features

< >

( input_values: Tensorpadding_mask: Tensor**kwargs: *<class 'transformers.utils.generic.TransformersKwargs'> ) → Apertus1p5AudioTokenizerModelOutput or tuple(torch.Tensor)

Parameters

  • input_values (torch.Tensor of shape (batch_size, sequence_length)) — Float values of input raw speech waveform. Values can be obtained by loading a .flac or .wav audio file into an array of type list[float], a numpy.ndarray or a torch.Tensor, e.g. via the torchcodec library (pip install torchcodec) or the soundfile library (pip install soundfile). To prepare the array into input_values, the AutoProcessor should be used for padding and conversion into a tensor of type torch.Tensor. See Apertus1p5Processor.call() for details.
  • padding_mask (torch.Tensor of shape (num_clips, max_length)) — Mask with 1 for valid samples and 0 for padding, as returned by the feature extractor.

Returns

Apertus1p5AudioTokenizerModelOutput or tuple(torch.Tensor)

A Apertus1p5AudioTokenizerModelOutput 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 (Apertus1p5Config) and inputs.

  • 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 (tuple[torch.Tensor], optional) — One (num_tokens, hidden_size) tensor of embedded audio tokens per input clip.

  • 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.

  • audio_tokens (list[torch.LongTensor], optional) — One tensor of ceil(num_valid_samples / hop_length) shared-vocabulary token ids per input clip, offset by config.audio_token_offset.

Example:

>>> from PIL import Image
>>> from transformers import AutoProcessor, Apertus1p5ForConditionalGeneration

>>> model = Apertus1p5ForConditionalGeneration.from_pretrained("swiss-ai/Apertus-v1.5-8B")
>>> processor = AutoProcessor.from_pretrained("swiss-ai/Apertus-v1.5-8B")

>>> 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]
Update on GitHub