Train a LoRA that generates the video between keyframes — supply first/last (and optional middle) frames at inference and the model fills the in-between motion.
No supported capabilities recorded.
/v1/models/fal/fal-ai/ltx23-trainer-v2/interpolate