deploying model

[rubenftech]

Hi,
Do you have any news about the deploying models for unitree go1?
Thank you for your amazing work!

[YandongJi]

Hi,

The Jetson Nano board typically faced limitations in deployment frequency. To address this, we replaced all Nano boards with NX boards, which were not originally available in the manufactured GO1 units.

Additionally, the previous version had some less-than-ideal tuning for combining detection results from the front and chin cameras. We’re still not entirely satisfied with this approach and may refine it further, potentially releasing updated code based on future work. However, there is no specific timeline for this yet.

Issue #3 proposes a working version that uses the front camera for deployment, closely resembling our original pipeline. Could you give it a try?

Thanks

[rubenftech]

Hi,

Thank you for your detailed response and for taking the time to explain the updates!

I understand the challenges with the Jetson Nano and the tuning issues for combining detection results. It's great to hear about the improvements with the NX boards and the possibility of refining the approach in the future.

I’ll definitely take a look at Issue #3 and try the suggested version that uses the front camera for deployment. I’ll let you know if I encounter any issues or have feedback.

Thanks again for your amazing work and support!