Class FunAudioChatConfig

funaudiochat/configuration_funaudiochat.py:157–252

r""" This is the configuration class to store the configuration of a [`FunAudioChatForConditionalGeneration`]. It is used to instantiate an Qwen2-Audio model according to the specified arguments, defining the model architecture. Instantiating a configuration with the defaults will yield

        self.pad_token_id = pad_token_id

class FunAudioChatConfig(PretrainedConfig):
    r"""
    This is the configuration class to store the configuration of a [`FunAudioChatForConditionalGeneration`]. It is used to instantiate an
    Qwen2-Audio 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 Qwen2-Audio.

    e.g. [FunAudioLLM/Fun-Audio-Chat-8B](https://huggingface.co/FunAudioLLM/Fun-Audio-Chat-8B)

    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
    documentation from [`PretrainedConfig`] for more information.

    Args:
        audio_config (`Union[AutoConfig, dict]`,  *optional*, defaults to `CLIPVisionConfig`):
            The config object or dictionary of the audio backbone.
        text_config (`Union[AutoConfig, dict]`, *optional*, defaults to `LlamaConfig`):
            The config object or dictionary of the text backbone.
        audio_token_index (`int`, *optional*, defaults to 151646):
            The audio token index to encode the audio prompt.
        ignore_index (`int`, *optional*, defaults to -100):
            The index to ignore in loss calculation.
        hidden_size (`int`, *optional*):
            Hidden size of the model. If not specified, will use text_config.hidden_size.

    Example:

    ```python
    >>> from transformers import FunAudioChatForConditionalGeneration, FunAudioChatConfig, FunAudioChatEncoderConfig, Qwen2Config

    >>> # Initializing a FunAudioChatEncoder config
    >>> audio_config = FunAudioChatEncoderConfig()

    >>> # Initializing a Qwen2 config
    >>> text_config = Qwen2Config()

    >>> # Initializing a FunAudioChat configuration
    >>> configuration = FunAudioChatConfig(audio_config, text_config)

    >>> # Initializing a model from the qwen2-audio style configuration
    >>> model = FunAudioChatForConditionalGeneration(configuration)

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

    model_type = "funaudiochat"
    attribute_map = {
        "audio_token_id": "audio_token_index",
    }
    sub_configs = {"text_config": AutoConfig, "audio_config": AutoConfig}

    def __init__(
        self,
        audio_config=None,
        text_config=None,
        audio_token_index=151646,
        ignore_index=-100,
        hidden_size=None,
        **kwargs,