[przemoch]

Hello,

What is your experience about Nvidia Jetson Nano performance? Do you think it is correct with official data in nvidia datasheet?

We are trying to detect people in video recordings and we can't reach those parameters. We are able to analize video with YOLO Tiny algorithm with only 1 fps.  
We've also tried the same algorithm with Jetson TX2 and it's almost the same result. 

What do you thik about it? Are we not able to set up our script correctly or it can be the Jetson Nano issue?

Do you have any similar expierience in working with those microcomputers? Or mayby do you have some ideas how to improve our results.?

Best regards,
Przemyslaw

[Klojum]

Our forum only covers a single home theater software application, Kodi, which plays videos and music files.
I think I can safely say that the number of people on this forum working with Nvidia Jetson boards is very, very close to zero.

There is also entirely no science or A.I. involved in our software for detecting people, animals or other Mother Earth occupants.
I think you took a wrong turn by mistake, and ended up here. Unfortunately, we are not the target audience you are looking for.

[honcho]

Hello,

What is your experience about Nvidia Jetson Nano performance? Do you think it is correct with official data in nvidia datasheet?

We are trying to detect people in video recordings and we can't reach those parameters. We are able to analize video with YOLO Tiny algorithm with only 1 fps.  
We've also tried the same algorithm with Jetson TX2 and it's almost the same result. 

What do you thik about it? Are we not able to set up our script correctly or it can be the Jetson Nano issue?

Do you have any similar expierience in working with those microcomputers? Or mayby do you have some ideas how to improve our results.?

Best regards,
Przemyslaw

It's very old, and very slow. Why wouldn't you just buy a google TPU.

[RockerC]

It's very old, and very slow. Why wouldn't you just buy a google TPU. 

I believe that you must be thinking of the old Nvidia Jetson, and not the new Nvidia Jetson Nano

Nvidia Jetson Nano is a new development board from Nvidia that only cost $99 and at least on paper has the potential to be a great platform for Kodi

https://www.cnx-software.com/2019/03/19/...loper-kit/

Jetson Nano developer kit specifications:

• Jetson Nano CPU Module
  • 128-core Maxwell GPU
    
  • Quad-core Arm A57 processor @ 1.43 GHz
    
  • System Memory  – 4GB 64-bit LPDDR4 @ 25.6 GB/s
    
  • Storage  – microSD card slot (devkit) or 16GB eMMC flash (production)
    
  • Video Encode – 4K @ 30 | 4x 1080p @ 30 | 9x 720p @ 30 (H.264/H.265)
    
  • Video Decode – 4K @ 60 | 2x 4K @ 30 | 8x 1080p @ 30 | 18x 720p @ 30 (H.264/H.265)
    
  • Dimensions – 70 x 45 mm
    
• Baseboard
  • 260-pin SO-DIMM connector for Jetson Nano module.
    
  • Video Output – HDMI 2.0 and eDP 1.4 (video only)
    
  • Connectivity – Gigabit Ethernet (RJ45) + 4-pin PoE header
    
  • USB – 4x USB 3.0 ports, 1x USB 2.0 Micro-B port for power or device mode
    
  • Camera I/F – 1x MIPI CSI-2 DPHY lanes compatible with Leopard Imaging LI-IMX219-MIPI-FF-NANO camera module and Raspberry Pi Camera Module V2
    
• Expansion
  • M.2 Key E socket (PCIe x1, USB 2.0, UART, I2S, and I2C) for wireless networking cards
    
  • 40-pin expansion header with GPIO, I2C, I2S, SPI, UART signals
    
  • 8-pin button header with system power, reset, and force recovery related signals
    
• Misc – Power LED, 4-pin fan header
  
• Power Supply – 5V/4A via power barrel or 5V/2A via micro USB port; optional PoE support
  
• Dimensions – 100 x 80 x 29 mm
  

[mpottinger]

Well, yes, it is a new product, but based on the same old technology that the tx1/tx2 are. He is saying the tx1 is old and slow, and sure enough it is. The nano is just a trimmed down version of that. Half the gpu cores.

Don't get me wrong, it is good for what it is. A tiny, relatively cheap computer based on the same platform as the 2015 NVIDIA shield tv box.

I just got one and I am playing around with it. I have mixed feelings.

On the one hand it is the fastest cheap little computer I have ever used, and on the other hand it still petforms like a cheap computer, unable to do serious work.

For streaming etc, I like the Android OS on the Shield better. Just seems smoother than Linux4tegra even though it's the same hardware.

I'm finding it not quite or just barely up to the tasks I had in mind for it.

It's very old, and very slow. Why wouldn't you just buy a google TPU. 

I believe that you must be thinking of the old Nvidia Jetson, and not the new Nvidia Jetson Nano

Nvidia Jetson Nano is a new development board from Nvidia that only cost $99 and at least on paper has the potential to be a great platform for Kodi

https://www.cnx-software.com/2019/03/19/...loper-kit/

Jetson Nano developer kit specifications:

• Jetson Nano CPU Module
  • 128-core Maxwell GPU
    
  • Quad-core Arm A57 processor @ 1.43 GHz
    
  • System Memory  – 4GB 64-bit LPDDR4 @ 25.6 GB/s
    
  • Storage  – microSD card slot (devkit) or 16GB eMMC flash (production)
    
  • Video Encode – 4K @ 30 | 4x 1080p @ 30 | 9x 720p @ 30 (H.264/H.265)
    
  • Video Decode – 4K @ 60 | 2x 4K @ 30 | 8x 1080p @ 30 | 18x 720p @ 30 (H.264/H.265)
    
  • Dimensions – 70 x 45 mm
    
• Baseboard
  • 260-pin SO-DIMM connector for Jetson Nano module.
    
  • Video Output – HDMI 2.0 and eDP 1.4 (video only)
    
  • Connectivity – Gigabit Ethernet (RJ45) + 4-pin PoE header
    
  • USB – 4x USB 3.0 ports, 1x USB 2.0 Micro-B port for power or device mode
    
  • Camera I/F – 1x MIPI CSI-2 DPHY lanes compatible with Leopard Imaging LI-IMX219-MIPI-FF-NANO camera module and Raspberry Pi Camera Module V2
    
• Expansion
  • M.2 Key E socket (PCIe x1, USB 2.0, UART, I2S, and I2C) for wireless networking cards
    
  • 40-pin expansion header with GPIO, I2C, I2S, SPI, UART signals
    
  • 8-pin button header with system power, reset, and force recovery related signals
    
• Misc – Power LED, 4-pin fan header
  
• Power Supply – 5V/4A via power barrel or 5V/2A via micro USB port; optional PoE support
  
• Dimensions – 100 x 80 x 29 mm
  

[RockerC]

Well, yes, it is a new product, but based on the same old technology that the tx1/tx2 are. He is saying the tx1 is old and slow, and sure enough it is. The nano is just a trimmed down version of that. Half the gpu cores.

...Don't get me wrong, it is good for what it is. A tiny, relatively cheap computer based on the same platform as the 2015 NVIDIA shield tv box.

Worth noting then that the Nvidia Shield TV is widely considered the best Android-based box for Kodi and is also among the best platforms for Kodi if you have a budget under $200(US).

So now that the Nvidia Jetson Nano dev-board only cost $100(US), half of a what an Nvidia Shield TV cost, I don't think it's fair to shoot it down as fast for Kodi, even if it does not come with a remote and case.

For use with only Kodi, the Nvidia Jetson Nano dev-board could be a steal at $100(US) to be used as a 4K HDR capable video player platform, only time will tell.

This is the official Kodi forum, and the only use case we should really discuss for it is as a Kodi media player platform.

[cestrague]

Hello,

What is your experience about Nvidia Jetson Nano performance? Do you think it is correct with official data in nvidia datasheet?

We are trying to detect people in video recordings and we can't reach those parameters. We are able to analize video with YOLO Tiny algorithm with only 1 fps.  
We've also tried the same algorithm with Jetson TX2 and it's almost the same result. 

What do you thik about it? Are we not able to set up our script correctly or it can be the Jetson Nano issue?

Do you have any similar expierience in working with those microcomputers? Or mayby do you have some ideas how to improve our results.?

Best regards,
Przemyslaw

It's very old, and very slow. Why wouldn't you just buy a google TPU. 

It's not very slow it's very powerful than Google TPU 

Also forget having an opengl driver es 3.2, Opengl 4.6 and vulkan 100% implemented leaves much to be desired in that regard 
so Manufacturers outside nvidia do not mind implementing a driver extensions on the latest current APIS in addition to ignoring OpenGL desktop or partially implemented and not fully compilant extensions that is They do not respect the standards when they leave.

The main problem is Kodi does not support Gstreamer .

[cestrague]

 

 

It's not very slow it's very powerful than Google TPU 

Also forget having an opengl driver es 3.2, Opengl 4.6 and vulkan 100% implemented leaves much to be desired in that regard 
so Manufacturers outside nvidia do not mind implementing a driver extensions on the latest current APIS in addition to ignoring OpenGL desktop or partially implemented and not fully compilant extensions that is They do not respect the standards when they leave.

The main problem is Kodi does not support Gstreamer . 

Also  tegra soc will be working if V4L2 are supported see https://docs.nvidia.com/jetson/archives/...L2Dec.html