Curious to learn more? 📚Check out the Google Gemma Dev Day Playlist in the description to explore deeper insights on Gemma2. ruclips.net/p/PLOU2XLYxmsIKOyXflnuPK-qe32hZLc2HB
The Nvidia Jetson line is super interesting as an idea. The RAM used is a lot cheaper per gigabyte than VRAM so you can get monstrous model sizes for the price, and it’s a lot easier to manage in terms of thermals and power, but the software stack feels quite different to desktop CUDA on x86, and not all projects will work out of the box. It does get a bit weird when you try to evaluate a Jetson module against, say, CPU execution on a desktop platform where you have a lot more freedom, or against a Ryzen iGPU where you can actually outperform the Jetson modules in terms of RAM per dollar. I think it’d be really interesting to see cheaper Jetson modules on a modern architecture that had native support for clustering with fast networking of some description, though.
Guys, the Jetson Nano product itself is just gorgeous. but please see what you are delivering as an operating system. It's just terrible. it was made by an amateur. the disk was split incorrectly. it takes a lot of time to make it at least a little usable. It is better to use this time to study programming.
yup, it’s awful. i can’t imagine anyone using this in production uses much of the default system. though it did get better with jp6 (got rid of those awful “overlays”).
![ATEEZ(에이티즈) - [GOLDEN HOUR : Part.2] Preview](http://i.ytimg.com/vi/X4PfgRPBUVo/mqdefault.jpg)
Curious to learn more? 📚Check out the Google Gemma Dev Day Playlist in the description to explore deeper insights on Gemma2. ruclips.net/p/PLOU2XLYxmsIKOyXflnuPK-qe32hZLc2HB
top Asier! Keep pushing
Jensen, PLEASE if you gonna charge me 2K for the 5090, it BEST HAVE a Jetson built-in already and MORE! I want local LLM NOW!!
The Nvidia Jetson line is super interesting as an idea. The RAM used is a lot cheaper per gigabyte than VRAM so you can get monstrous model sizes for the price, and it’s a lot easier to manage in terms of thermals and power, but the software stack feels quite different to desktop CUDA on x86, and not all projects will work out of the box.
It does get a bit weird when you try to evaluate a Jetson module against, say, CPU execution on a desktop platform where you have a lot more freedom, or against a Ryzen iGPU where you can actually outperform the Jetson modules in terms of RAM per dollar.
I think it’d be really interesting to see cheaper Jetson modules on a modern architecture that had native support for clustering with fast networking of some description, though.
Really amazing
Can we get in India ?
I'm having conflicts with DaVinci resolve recently after the latest update to my driver's
Can you add e.g. an A2000 through that M.2?
Can it reed stock charts?
Guys, the Jetson Nano product itself is just gorgeous. but please see what you are delivering as an operating system. It's just terrible. it was made by an amateur. the disk was split incorrectly. it takes a lot of time to make it at least a little usable. It is better to use this time to study programming.
yup, it’s awful. i can’t imagine anyone using this in production uses much of the default system. though it did get better with jp6 (got rid of those awful “overlays”).
I am also facing the same issue, could you please let me also know or share the resource on how did you solve the issues ? Thanks
@@SuchirKavi jetpack 7 and jetson thor is coming in 2025. Thor is based on blackwell and jp7 on 24.04
@@gamekida4007 you don't need to do this. let the NVIDIA people do it.
ok? “ n+1 is coming out next year”. why do I care?
That conversation 6:15 is all hallucinations :) It is the wrong way to teach them talk.
This is not a step-by-step setup, for beginners. The instructions used in the demo are not listed anywhere.