Struct VaeChainArgs 

pub struct VaeChainArgs<'a> {
    pub key: u64,
    pub conv_in: VaeConvRef<'a>,
    pub mid_res1: VaeResnetRef<'a>,
    pub mid_attn: VaeAttnRef<'a>,
    pub mid_res2: VaeResnetRef<'a>,
    pub ups: Vec<VaeUpRef<'a>>,
    pub norm_out: VaeNormRef<'a>,
    pub conv_out: VaeConvRef<'a>,
    pub latent_channels: usize,
    pub scaling_factor: f32,
    pub shift_factor: f32,
}

The whole decoder as borrowed slices, in execution order: conv_in → mid_res1 → mid_attn → mid_res2 → ups[0..] → norm_out+SiLU → conv_out. Consumed by gpu::vae_decode_chain; key is stable for the life of the VaeDecoder, so a backend caches its uploaded weights by it instead of fingerprinting every conv on every call.

Fields

key: u64

conv_in: VaeConvRef<'a>

mid_res1: VaeResnetRef<'a>

mid_attn: VaeAttnRef<'a>

mid_res2: VaeResnetRef<'a>

ups: Vec<VaeUpRef<'a>>

norm_out: VaeNormRef<'a>

conv_out: VaeConvRef<'a>

latent_channels: usize

scaling_factor: f32

shift_factor: f32