What's different between these two BC policies? It's the same architecture, training budget, and data collection setup — the only difference is the controller gains!
Controller gains are an understudied design parameter in robot learning. In our new work (w/ @BronarsToni*, @pulkitology), we show how they act as an inductive bias across BC, RL, and Sim2Real transfer, with real consequences on performance. Here's what we found 🧵
* Equal Contribution
📄arxiv: https://t.co/SMYgh7i8cA
🔗website: https://t.co/cLCd1FYCdJ
Although here is no direct support for things like mesh skinning in newton-physics, it's a nice experience to implement it with their straight-forward API. I struggled a bunch trying to achieve the same thing with the Isaac Sim APIs.
Strawberry plant simulated as a tree of rods. I've been trying to find a way to get this working for long time now. The initial trials with @NewtonPhysEng are very promising 🤩🌱
We just open-sourced our OpenUSD -> @rerundotio logger. While still a work-in-progress, the tool is already powering several of our @nvidiaomniverse projects.
https://t.co/pJZg2LnPUS
The new graphs feature in @rerundotio already helped to catch some otherwise invisible bugs in my plant-simulation project🎉 Thanks for the release!!✨❤️
🎉 I'm proud to announce Gatto 🐈. It's a #github notification manager app built with #React ⚛️ and #Electron ⚡️.
👀 Let's you peek into notifications
💻 Get native Slack like notifications
⚙️ Filter your notifications by organizations
🌙 Dark mode
🔗 https://t.co/nb75p5Wyw4