mvp-engine
Collection
mvp-engine • 4 items • Updated
MVP Lab
community
AI & ML interests
multi-modal foundation models
Recent Activity
Papers
View all Papers
geoffreychen777
updated a
collection 5 days ago
geoffreychen777
updated a
dataset 5 days ago
geoffreychen777
published a
dataset 5 days ago
geoffreychen777
updated a
dataset 5 days ago
xiangan
authored 2
papers 6 days ago
geoffreychen777
updated a
collection 6 days ago
geoffreychen777
published a
dataset 6 days ago
xiangan
submitted a
paper to Daily Papers 7 days ago
geoffreychen777
updated a
collection 13 days ago
geoffreychen777
updated a
model 13 days ago
geoffreychen777
published a
model 13 days ago
geoffreychen777
updated a
collection 13 days ago
geoffreychen777
updated a
dataset 13 days ago
geoffreychen777
published a
dataset 13 days ago
xiangan
updated a
collection 14 days ago
oliveryanzuolu
authored a
paper 16 days ago
oliveryanzuolu
posted an update 20 days ago
Post
119
Excited to share RAVEN, my first PhD project. Paper, code, and models are all released.
RAVEN is for real-time autoregressive video generation. Instead of simply appending future chunks, we train the model to better remember and use its own generated history, leading to more realistic and natural long-horizon videos.
Technically, RAVEN repacks self-rollouts into interleaved clean historical endpoints and noisy denoising states, aligning training-time attention with inference-time extrapolation.
We also introduce CM-GRPO: by reformulating consistency-model sampling as a conditional Gaussian transition kernel, online RL can directly optimize the sampler transition used at inference.
Project Page: https://yanzuo.lu/raven
Paper: https://arxiv.org/abs/2605.15190
Code: https://github.com/mvp-ai-lab/RAVEN
Model: mvp-lab/RAVEN
RAVEN is for real-time autoregressive video generation. Instead of simply appending future chunks, we train the model to better remember and use its own generated history, leading to more realistic and natural long-horizon videos.
Technically, RAVEN repacks self-rollouts into interleaved clean historical endpoints and noisy denoising states, aligning training-time attention with inference-time extrapolation.
We also introduce CM-GRPO: by reformulating consistency-model sampling as a conditional Gaussian transition kernel, online RL can directly optimize the sampler transition used at inference.
Project Page: https://yanzuo.lu/raven
Paper: https://arxiv.org/abs/2605.15190
Code: https://github.com/mvp-ai-lab/RAVEN
Model: mvp-lab/RAVEN
oliveryanzuolu
updated a
collection 20 days ago
oliveryanzuolu
submitted a
paper to Daily Papers 20 days ago