Hmm. If there's no copyright notice, it might be okay to copy the initial, non-gibberish lines here. I expect those first lines work some quick magic to determine some characteristics of the Linux system on which the script is running, and based upon that, the script then decodes the gibberish lines into one or more binaries that are placed on the system in appropriate places. Most likely, if the script even works that far on the tablets, the binaries will fail because they'll be for x86.

That's just a guess though. I can't say for sure until I see the script. (And maybe not even then. I've never done much work with shell scripts.)