Project Submission


#1

Hi, I did not know about Hackster contest until some previous weeks. But when I checked out this contest and the 155 ideas submitted, I was shocked. Only 30 people were selected for free devices and yet only 1 person submitted the project. Is there any logical reason behind not submitting the project or they submitted their idea just for the board.
–Community Member–


#2

This is quite normal !

  1. Not every submitted idea gets a device, there is no obligation to provide a hardware by Hackster for every idea submitted. Device quantity and idea selection is at the discretion of Sponsor Hardware Company.

  2. Most projects are submitted during the last few days before the contest is over.

  3. Participants are honor bound but not obliged to submit a project because they couldn’t spare the time or they might not have the technical/financial capability to fulfill their proposed project.

  4. Usually, 10 to 30 % of the people who gets hardware submit a somewhat completed project and less than 10% will submit a well documented useful project. Some participants can’t submit a project because they never receive their awarded hardware ( lost in shipment ).

  5. Don’t be shocked, Hackster is trying to improve the situation by updating rules of the contests.


#3

This was bound to happen from a projects standpoint from the requirements.

  1. Projects needing depth will only be picked.

  2. Project needs to be ready in approximately one month.

Two relatively contradictory requirements. In addition to this, there is the issue of shipping time. The only way to beat the above is if it is an ongoing project just waiting for the board, which is highly unlikely.


#4

Many participants put their ideas in hope to get a free Ultra96 and proceed with the idea development. Since most of them didn’t win a board, they have no hardware to finish the project - it’s sometimes a huge problem to buy Ultrascale-class chips outside US((


#5

To be very honest, this is one of the hardest hardware I’ve worked with. I have spent more than a week on this board and was not able to push further.

  1. Normally the software is extremely easy, integrate with APIs and I can get my project going. But Ultra96 requires you to pretty much make your own OS. The developers are not OS makers, this adds a huge degree of difficulty.

  2. Normally we’d receive help from the community, stackoverflow, github. The Xilinx have almost zero presence, making every roadblock strenous and difficult. It was so difficult to just finding a simple usage of “tensorflow on ultra96” or “caffe on Ultra96”, after days of exploring I finally give up.

  3. Linux system that has less features than Raspberry Pi. I was able to get an independent debian running on Ultra96, but the results aren’t good as I could not get any other libraries running. Things as simple as “/dev/ttyUSB0” will not show up after I clearly plugin my USB hardware for the project. and just simple stuff like that took entire day and I couldn’t get any further. Not to mention that I also have problems with the USB device.

I really want to help… but wasting weeks of not much progress I am about to give up…

Most of us are here for fun, as we are not getting paid to do this. And if we are constantly getting stuck after getting stuck, building the project becomes more strenuous than work. If we are getting paid full time to figure these things out, that’s a different story.

If anyone within U.S wants my board please let me know, I can ship it to you.

Just my 2 cents.


#6

Try out the pynq framework. This should allow you to get something up and running quickly. http://www.pynq.io/

https://docs.avnet.com/amer/smart_channel/ultra96_v2.3.zip


#7

I’ve tried, it is missing tons of stuff, for example, -lxilinxopencl and -llmx6.0

This linux kernel also prevents me to install drivers for other hardwares, as it is not supported. Is it possible to just run Ubuntu Mate or Pi on this?