Search papers, labs, and topics across Lattice.
LiveAnimate introduces a groundbreaking approach to real-time human animation by leveraging a 14B-parameter video Diffusion Transformer (DiT) that enables stable long-form streaming. The system employs a two-stage training pipeline that adapts a pretrained DiT into an autoregressive generator and incorporates a novel Pose-Retrieval Sink Attention mechanism to efficiently manage memory and appearance context. As a result, LiveAnimate achieves 19.63 FPS streaming inference while maintaining consistent perceptual quality across extended durations, significantly outperforming existing methods that struggle with quality and latency.
Achieving real-time human animation at nearly 20 FPS while preserving identity and quality over three-minute streams sets a new standard for interactive applications.
Pose-driven human animation synthesizes a video of a target person from a single reference image and a driving pose stream. Real-time generation is essential for interactive applications such as live streaming, telepresence, and virtual avatars, yet diffusion-based systems require minutes to hours per clip, precluding responsive interaction. We present LiveAnimate, to our knowledge the first animation system to combine real-time streaming with stable long-form generation at billion scale, built on a 14B-parameter video Diffusion Transformer (DiT). A two-stage training pipeline first adapts a pretrained bidirectional DiT into a block-causal autoregressive generator through Reference-Anchored Teacher-Forcing Adaptation, and then reduces the sampling budget to three steps through Block-wise Self-Forcing Distillation. To preserve appearance over extended streams, we introduce Pose-Retrieval Sink Attention (PR-Sink), a bounded KV-cache mechanism combining a Static Sink that permanently anchors the first generated block, a Dynamic Sink that holds a pose-retrieved historical block, and a three-slot Rolling Window. When a pose recurs, PR-Sink restores the relevant appearance context without retaining the entire sequence, so memory and per-block latency remain constant regardless of stream duration. Together with Ulysses sequence parallelism and operator fusion, these designs enable 19.63\,FPS streaming inference on two NVIDIA H100 GPUs. On a three-minute benchmark, LiveAnimate maintains nearly constant perceptual quality and identity from the first 30 seconds to the final minute, while prior systems degrade substantially or require hours of offline computation for the same rollout. These results establish a new operating point in quality, latency, and duration for interactive full-body animation.
