Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Highlighted
New Contributor I
1,172 Views

Any Information about the General Vision 128 Node Hardware Neural Network

Jump to solution

Hi,

I've managed to piece together some info on the 101's hardware neural network provided by General Vision - see my blog here

http://www.electricalengineeringschools.org/2016/computer-engineering-and-the-hidden-neural-network-... http://www.electricalengineeringschools.org/2016/computer-engineering-and-the-hidden-neural-network-...http://www.electricalengineeringschools.org/2016/computer-engineering-and-the-hidden-neural-network-...

I know for instance, node 0 cannot be used for training a pattern, but rather for un-training a pattern, and I know there are 128 max nodes etc. I also have the professional arduino libs from general vision (cost $19). But where is all the other info ?

For example

1. How do I chain 2 together ?

2. What is max number of inputs in a training/classification vector ?

There doesn't seem to be an Intel officiated document for their hardware NN on their own board. It's rather bizarre, don't you want people to use this amazing piece of tech ?

Marcus


Accepted Solutions
Highlighted
Community Manager
21 Views

Hi Marcus,

 

 

We have some information that can be helpful for you:

 

 

Just to clarify the Intel® Curie™ contains a pattern matching accelerator that is unique to Intel. It may be similar to other devices, but it is unique to the Curie.

 

 

About your questions, there is no hardware facility with Curie to chain multiple pattern matching engines together. However, it may be possible to accomplish such a function through software, and the maximum vector size for the Intel Curie pattern matching accelerator it 128 unsigned bytes.

 

 

Additionally, you would have to check with the vendor of the library for specific support information.

 

 

Intel offers a completely free open source library for Arduino and embedded C++ at:

 

https://github.com/01org/Intel-Pattern-Matching-Technology

 

 

I hope you find this information helpful.

 

 

Regards,

 

-Leonardo

View solution in original post

0 Kudos
5 Replies
Highlighted
Community Manager
21 Views

Hi Marcus,

 

 

Thank you for the information provided.

 

 

About your doubts, we're currently working on this, since you created this other thread: https://communities.intel.com/thread/107839 https://communities.intel.com/thread/107839.

 

 

We will let you know in this thread when we find something helpful.

 

 

I appreciate your patience.

 

 

Regards,

 

-Leonardo
0 Kudos
Highlighted
Community Manager
22 Views

Hi Marcus,

 

 

We have some information that can be helpful for you:

 

 

Just to clarify the Intel® Curie™ contains a pattern matching accelerator that is unique to Intel. It may be similar to other devices, but it is unique to the Curie.

 

 

About your questions, there is no hardware facility with Curie to chain multiple pattern matching engines together. However, it may be possible to accomplish such a function through software, and the maximum vector size for the Intel Curie pattern matching accelerator it 128 unsigned bytes.

 

 

Additionally, you would have to check with the vendor of the library for specific support information.

 

 

Intel offers a completely free open source library for Arduino and embedded C++ at:

 

https://github.com/01org/Intel-Pattern-Matching-Technology

 

 

I hope you find this information helpful.

 

 

Regards,

 

-Leonardo

View solution in original post

0 Kudos
Highlighted
Community Manager
21 Views

Hi Marcus,

 

 

Do you have any updates on this? Was it helpful?

 

 

Regards,

 

-Leonardo
Highlighted
New Contributor I
21 Views

Hi Leonardo,

Yes the help was very useful, I'm pretty disappointed that I spent $19 on the pro library from General Vision, I wish it had been published somewhere that the Intel library is free. But anyway, the info was great, so thanks very much.

Marcus

0 Kudos
Highlighted
Community Manager
21 Views

Hi Marcus,

 

 

I'm glad that we could help you.

 

 

Just let us know when you have doubts, we will do our best to assist you.

 

 

Have a nice day.

 

 

Regards,

 

-Leonardo
0 Kudos