A few decisions I'm least settled on, and would love some pushback/feedback on:
- single arm vs. bimanual (I went single for cost/space, knowing it rules out things like folding cloth)
- not calibrating camera extrinsics/intrinsics for now
- RGB vs. RGB-D for from-scratch policies (ACT / Diffusion Policy)
And one I'm more confident about but expect disagreement on: not building on ROS 2 / LeRobot, and writing my own stack instead. Happy to get into the reasoning.
wxw•20m ago
- I've heard the advantage of ROS besides the architecture is the ecosystem (driver integrations, etc). Is that not an issue because the arm supports a Python SDK OOTB?
- Any issues you've been running into with this setup?
- How do you determine if a session recording is good enough for training? Is 50/100 samples really all you need?