Method __call__

models_diffusers/attention_processor.py:2016–2097

(
        self,
        attn,
        hidden_states,
        encoder_hidden_states=None,
        attention_mask=None,
        temb=None,
        scale=1.0,
    )

        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 with `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
        )
        attention_mask = attn.prepare_attention_mask(attention_mask, sequence_length, batch_size)

        if attn.group_norm is not None:
            hidden_states = attn.group_norm(hidden_states.transpose(1, 2)).transpose(1, 2)

        query = attn.to_q(hidden_states)

        if encoder_hidden_states is None:
            encoder_hidden_states = hidden_states
        elif attn.norm_cross:
            encoder_hidden_states = attn.norm_encoder_hidden_states(encoder_hidden_states)

        # split hidden states
        end_pos = encoder_hidden_states.shape[1] - self.num_tokens
        encoder_hidden_states, ip_hidden_states = (
            encoder_hidden_states[:, :end_pos, :],
            encoder_hidden_states[:, end_pos:, :],
        )

        key = attn.to_k(encoder_hidden_states)
        value = attn.to_v(encoder_hidden_states)

        query = attn.head_to_batch_dim(query)
        key = attn.head_to_batch_dim(key)
        value = attn.head_to_batch_dim(value)

        attention_probs = attn.get_attention_scores(query, key, attention_mask)
        hidden_states = torch.bmm(attention_probs, value)
        hidden_states = attn.batch_to_head_dim(hidden_states)

        # for ip-adapter
        ip_key = self.to_k_ip(ip_hidden_states)
        ip_value = self.to_v_ip(ip_hidden_states)