Log inSign up
Jason Peng
208 posts
Jason Peng profile banner
@xbpeng4

Jason Peng

@xbpeng4
Assistant Prof at @SFU and Research Scientist at @NVIDIA
Vancouver, Canada
xbpeng.github.io
Joined April 2018
48
Following
6,330
Followers
RepliesRepliesRepostsRepostsMediaMedia

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
  • Pinned
    @xbpeng4
    Jason Peng
    @xbpeng4
    Oct 8, 2025
    Implementing motion imitation methods involves lots of nuisances. Not many codebases get all the details right. So, we're excited to release MimicKit! github.com/xbpeng/MimicKit A framework with high quality implementations of our methods: DeepMimic, AMP, ASE, ADD, and more to come!
    Image
    00:00
    9
  • @xbpeng4
    Jason Peng
    @xbpeng4
    Apr 27
    Getting SDS to work for motion imitation has been tricky. But we finally got it to work! With SMP, you can train a diffusion prior on a dataset, freeze it, and reuse it over and over again as a style reward to train new controllers. Try it out: github.com/xbpeng/MimicKit
    @YuxuanMu16173
    Yuxuan Mu
    @YuxuanMu16173
    Apr 27
    Can we build a standalone, modular, and reusable naturalness reward for training motor controllers? #SMP is a step toward that vision. Once SMP has been trained on a motion dataset, the priors can be reused to train new controllers to perform diverse tasks while adhering to the
    Image
    00:00
  • @xbpeng4
    Jason Peng
    @xbpeng4
    Mar 17
    It's always challenging to find the right motion data for training humanoid controllers. With Kimodo you can now generate exactly the data you want through an intuitive and expressive interface. Like having a mocap actor on-demand!
    @davrempe
    Davis Rempe
    @davrempe
    Mar 17
    Need high-quality motion for humanoid robots or digital humans? Meet Kimodo: our new diffusion model trained on 700 hours of optical mocap data for easy, controllable, and high-fidelity motion generation. @NVIDIAAI research.nvidia.com/labs/sil/proje…
    Image
    00:00
  • @xbpeng4
    Jason Peng
    @xbpeng4
    Mar 4
    A nice little quality-of-life update, MimicKit now supports video logging. You can monitor the agent's behaviors during training on WandB and Tensorboard: github.com/xbpeng/MimicKi… We also added an implementation of Lipschitz-Constrained Policies for training smooth controllers.
    Image
    00:00
    1
  • @xbpeng4
    Jason Peng
    @xbpeng4
    Jan 12
    For all the MuJoCo fans: MimicKit now supports Newton / MuJoCo Warp! Fast, lightweight MuJoCo physics with Warp acceleration. github.com/xbpeng/MimicKi… Coming up: God's physics engine
    Image
    00:00
    11
Advertisement
Advertisement