Host machine setup instruction?

Hi there,

I’ve setup my Jetson nano JetBot (https://github.com/NVIDIA-AI-IOT/jetbot/wiki), running the first Jupyter Notebook successfully. Now I’m moving into the Collision Avoidance, and completed the first part, but in the end of the notebook, it assumes I’ve setup the host machine, which I didn’t do. So now I’m looking around the Internet everywhere for the instructions to setup the host machine but looks like the documents are in a chaos and not enlightening at all…
I installed the Ubuntu 16.04 in my PC as well as the SDK manager. I used the SDK manager to download and install the Host components successfully (I didn’t select the target component since I setup the JetBot already), but when I tried http://<host_pc_ip>:8888, no luck.

Can anyone share a complete instruction to setup the host machine?
And also, I’m having a lot of troubles following https://github.com/NVIDIA-AI-IOT/jetbot/wiki mostly because of a lot of software versions not matched, looks like NV updated the SD image but not the other parts. Maybe it’s time for a complete update?
Thanks.

Up

First, I think the current preferred version for the SDK manager is Ubuntu 18.04. At some point, I expect them to roll forward to 20.04, but I haven’t seen that happen yet.
Second, if you’re having “version problems,” then being more specific about which versions you’re seeing, and which specific error messages you’re getting, will enable people on the forum to help you better.

Thanks for the reply.

So do you mean if I install Ubuntu 18.04 and follow the SDK manager there, it should be all right? I’ll give it a try tonight.

Hi yufan.lu.esp,

Thanks for reaching out.

You should be able to run the train_model.ipynb directly on the Jetson Nano if you do not have another machine.

Apologies, the initial assumption was that a separate machine would be needed for training. We’ve since discovered that Jetson Nano is capable of training relatively small datasets for this task on board in a moderate time frame.

I can see how this is confusing, we should probably clarify in the notebook. That said, if you do want to use a more powerful machine for training, I’m happy to provide some guidance. This assumes you have access to a GPU enabled machine.

Please let me know if this helps or you run into issues.

Best,
John

1 Like