Deepstream Jetson Nano

Dear all,

Do you have a schedule for the launch of Deepstream based on Jetson Nano ?

Thanks.

Regards,

Hi Christophe, please refer to this recent post - thanks!

[url]https://devtalk.nvidia.com/default/topic/1055628/jetson-nano/new-jetson-software-modules-and-pricing/post/5358866/#5358866[/url]

Hi Dusty,

Thanks a lot, I missed the link ;-)

Regards,

Hi all, DeepStream 4.0 was released yesterday (including support for Nano), see here for more info: [url]https://developer.nvidia.com/deepstream-sdk[/url]

Will the JetPack image be updated to include it, or is that scheduled for a new release?

Hi mdegans, DeepStream isn’t actually planned to be included in the base Nano image in order to keep the size down. However you can install it on top of the image from the downloadable tarball or Debian package, or the SDK Manager can install it.

Would it be possible to provide a link to a direct download so I can curl/wget it from the nano? I understand you have a license agreement, but that’s in the package itself as well isn’t it? I mean there is a LICENSE file.

I can’t decompress the tarball on windows, which means I need to move a 200mb tarball over scp to a linux machine, extract it, and move those many, files back, which is slow no matter the protocol.

Please can you provide a zip, xz, or recompress it tar.bz2 somehow so it actually works. I downloaded a boost tarball just last night. 7zip liked it. Python liked it. I really think there may be something wrong with the way tarballs are generated at nvidia.

I believe you that it may be possible that bzip is at fault, but I looked at the only documented issue I can find and many of the bug reports are from nvidia tarballs.

On top of that, I can recompress those same files on a Linux machine using a current version of rar and libbzip2
with “tar cjf …” and it extracts fine on everywhere, including on Windows. I mean I can’t replicate a tarball that breaks like nvidia tarballs do. IDGI. Anyway, please provide an alternative for those of us who sometimes work on a Windows laptop.

I re-compressed the files in the bzip2 on nano with it’s version of bzip and copied it back to my windows laptop. It extracts fine.
What version of libbzip2 are you using to compress?

From this issue on the Gentoo list it seems it might be when BZ_MAX_SELECTORS is exceeded, but I’m not sure of the version. I know that for future releases we are looking into changing it - sorry for the nuisance.

No, I feel like the nuisance, really. Thanks for the reference. According to the discussion, however, it should segfault (or maybe something worse) if BZ_MAX_SELECTORS exceeds ~18k becuase of a lack of a bounds check.

Assuming whatever version of libbzip2 you are using is generating tarballs with > 18k selectors, and whatever current version I have of libbzip2 on Windows is correctly checking for it (and correctly reporting a bad header), wouldn’t upgrading libbzip2 on your end fix the issue? That’s why I was hoping to reproduce the issue with an older version. If I can, then i can file the problem away in my head as 100% certain and it can stop bothering my OCD.