Runway Characters: Serving a 57B Video Diffusion Transformer for Real-Time Conversation
cs.DC
Submitted: 2026-11-03
Updated: 2026-11-03
Code: https://github.com/Dao-AILab/flash-attention
Terminology
Sources
- FLUX: Fast Software-based Communication Overlap On GPUs Through Kernel Fusion
- LongLive-2.0: An NVFP4 Parallel Infrastructure for Long Video Generation
- LiveTalk: Real-Time Multimodal Interactive Video Diffusion via Improved On-Policy Distillation
- Moshi: a speech-text foundation model for real-time dialogue
- USP: A Unified Sequence Parallelism Approach for Long Context Generative AI
- xDiT: an Inference Engine for Diffusion Transformers (DiTs) with Massive Parallelism
- Wan-Streamer v0.2: Higher Resolution, Same Latency
- Live Avatar: Streaming Real-time Audio-Driven Avatar Generation with Infinite Length
- Forcing-KV: Hybrid KV Cache Compression for Efficient Autoregressive Video Diffusion Models
- AVTR-1: Open Stack for Real-Time Interactive Avatars
- Avatar-Forever: Decoupled Parallel Training for High-Quality Real-Time Infinite Avatars
- Sol-H3: Recursive Self-Improvement for MiniMax-H3 Inference Acceleration on Sol-Engine across Cloud and Edge
- TalkingMachines: Real-Time Audio-Driven FaceTime-Style Video via Autoregressive Diffusion Models
- mKernel: Fast Multi-GPU, Multi-Node Fused Kernels
- MAGI-1: Autoregressive Video Generation at Scale
- SoulX-FlashTalk: Real-Time Infinite Streaming of Audio-Driven Avatars via Self-Correcting Bidirectional Distillation
- Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism
- Matrix-Game 3.0: Real-Time and Streaming Interactive World Model with Long-Horizon Memory
- Adaptive Resource Management and Quality Control for Streaming Video Generation
- SoulX-LiveAct: Towards Hour-Scale Real-Time Human Animation with Neighbor Forcing and ConvKV Memory
Related papers
- Programmable Probabilistic Computer with 1,000,000 p-bits
- BatchGen: An Architecture for Scalable and Efficient Batch Inference
- Seamless Reconfiguration for DAG-BFT (Technical Report)
- ELDR: Expert-Locality-Aware Decode Routing for PD-Disaggregated MoE Serving
- iScheduler: Reinforcement Learning-Driven Continual Optimization for Large-Scale Resource Investment Problems
- TIRx: A Unified Tile Primitive ML Compiler Across Heterogeneous Backends