Class IPAdapterAttnProcessor2_0

models_diffusers/attention_processor.py:2100–2231

r""" Attention processor for IP-Adapater for PyTorch 2.0. Args: hidden_size (`int`): The hidden size of the attention layer. cross_attention_dim (`int`): The number of channels in the `encoder_hidden_states`. num_tokens (`int`, defaults to 4):

class IPAdapterAttnProcessor2_0(torch.nn.Module):
    r"""
    Attention processor for IP-Adapater for PyTorch 2.0.

    Args:
        hidden_size (`int`):
            The hidden size of the attention layer.
        cross_attention_dim (`int`):
            The number of channels in the `encoder_hidden_states`.
        num_tokens (`int`, defaults to 4):
            The context length of the image features.
        scale (`float`, defaults to 1.0):
            the weight scale of image prompt.
    """

    def __init__(self, hidden_size, cross_attention_dim=None, num_tokens=4, scale=1.0):
        super().__init__()

        if not hasattr(F, "scaled_dot_product_attention"):
            raise ImportError(
                f"{self.__class__.__name__} requires PyTorch 2.0, to use it, please upgrade PyTorch to 2.0."
            )

        self.hidden_size = hidden_size
        self.cross_attention_dim = cross_attention_dim
        self.num_tokens = num_tokens
        self.scale = scale

        self.to_k_ip = nn.Linear(cross_attention_dim or hidden_size, hidden_size, bias=False)
        self.to_v_ip = nn.Linear(cross_attention_dim or hidden_size, hidden_size, bias=False)

    def __call__(
        self,
        attn,
        hidden_states,
        encoder_hidden_states=None,
        attention_mask=None,
        temb=None,
        scale=1.0,
    ):
        if scale != 1.0:
            logger.warning("`scale` of IPAttnProcessor should be set by `set_ip_adapter_scale`.")
        residual = hidden_states

        if attn.spatial_norm is not None:
            hidden_states = attn.spatial_norm(hidden_states, temb)

        input_ndim = hidden_states.ndim

        if input_ndim == 4:
            batch_size, channel, height, width = hidden_states.shape
            hidden_states = hidden_states.view(batch_size, channel, height * width).transpose(1, 2)

        batch_size, sequence_length, _ = (
            hidden_states.shape if encoder_hidden_states is None else encoder_hidden_states.shape
        )

        if attention_mask is not None: