I’m trying to work on the Jetson TX1 to run models and make use of NVIDIA’s GazeNet, however, by the time I’m done flashing and installing the DeepStreamSDK I have no more room.
I’ve seen a few videos on Youtube of people run their TX1 from SD cards or even SSDs.
Is it still possible to this day ?
If so which JetPack and DeepStream version should I go with ?
Will I encounter any issues ?
Btw, if you have other storage, e.g., an eMMC model which also has SD card, or a USB hard drive, or m.2, or SATA drive…anything basically…then it is easy to copy the content from “/usr/local” to that drive, and then mount the drive in place of “/usr/local”. This is where the CUDA and most AI software is from. Testing is safe as well since the original content is still there as soon as you unmount the external storage (but if it tests well you could permanently delete the “/usr/local” content of the eMMC). Actually booting all of rootfs from other media is a bit of a pain.