Allows explicit control over the causal_fix flag passed to
latent_to_pixel_coords. Defaults to frame_idx == 0 when not
specified, fixing the previous heuristic.
Implements per-guide attention attenuation via log-space additive bias
in self-attention. Each guide reference tracks its own strength and
optional spatial mask in conditioning metadata (guide_attention_entries).
Tavi HalperinandGitHub