BRN Discussion Ongoing

[DK6161]

Carefull what you say @Cyw wouldnt like you to get roasted, lol
@DK6161 may buy more shares based on that.

Lol typical bullsh1t from the same CEO that predicted explosion in sales in 2022 2023 2024.
The only thing that is imminent is this bloke leaving us in 6 months. Bye.

 

[Gemmax]

Bravo.
Don’t run away!!
Maybe around the block though!

 

[HopalongPetrovski]

Hi Chips,

It‘s getting annoying for me too. I really shouldn’t have to feel the need to defend myself over and over again. There are more important things to focus on in life. So perhaps it’s time for me to sign out for a little while.

Good luck to all genuine holders. We’ve stuck together through thick and thin and I’m very hopeful that our luck will turn shortly.

B x

Hi Bravo.
Sure, take a break if you wish/need to.
It's healthy to step away from social media every once and a while.
Smell the kitten mitts, ride the rainbow, have a run and a spa. Reset.
We'll miss you and what you bring here along with your gorgeous personality.
Frangipanni seems to be an excellent researcher, as are you, but I rarely read what she puts here in such excruciating and voluminous detail.
Frankly I don't have the time or inclination.
But, I'm happy to skim over 99% of it in the expectation that one day it may be important to me.
If he/she/they/it are bothering you with their attacks I suggest you treat them as I do the pests over on the crapper.
Don't engage with them. They exist, but that doesn't mean they can, or have to, define your existence, unless you allow it.
They may say all sorts of horrid things or spin myriad distortions to justify existence, but just don't play their game.
Anyhow, I hope you will return to us soon, as I, and I am sure, many others will miss you. xo

 

[Quiltman]

For all those who mistakenly use the SP to value Brainchip.

 

[Esq.111]

Chippers ,

CRANK IT.



Esq.

 

[Hammertime]

First it was fact finder now it’s Bravo , well dun pongy ,absolute disgrace.

Agree, Bravo's posts are about Brainchip, Frangipani's posts look like they are about Brainchip but in reality they are just about Frangipani.

 

[Diogenese]

View attachment 72206

Brainchip’s MetaTF Software Tools ease the Transition of ML Models to the Edge​

Kurt Manninen, BrainChip Senior Solutions Architect​

Empowering Companies: Seamless AI Migration to the Edge​

In their latest survey report, Edge Impulse highlights that the estimated market value of edge AI will be $143.6B by 2032. Companies that are early adopters of AI on the edge, particularly those that are willing to invest in the latest neural network hardware to give themselves an advantage over their competitors, will be looking to their vendors to help make the transition as painless as possible. This is why BrainChip developed our MetaTF framework: as a way for companies to effortlessly transition their existing machine learning models into a format that takes advantage of Akida’s event-based computing platform.

MetaTF is a suite of tools that can be used to convert, quantize and run existing machine learning models on an Akida event-based processor, which reduces computations by only computing on event data, not the raw sparse data as conventional NPUs must do. It was built for the purpose of providing an easy path for current and prospective BrainChip customers to validate their existing models on our advanced processor. As an experienced software and machine learning engineer who is new to event-based computing, BrainChip’s MetaTF allowed me to use tools that I already know (i.e. Python and Jupyter Notebooks) to experiment with ML on the edge. This is a valuable offering that will help engineers at developer organizations quickly migrate their processes to edge-based learning.

Embedded Devices ≠ Servers​

When it comes to deploying machine learning solutions, embedded devices present unique challenges compared to traditional server environments. Unlike servers with standardized architectures and operating systems, the world of IoT devices is incredibly diverse. From smartphones to smart sensors, each device may have a different architecture, run on a different operating system, or in some cases, operate on bare metal without any OS at all. This design diversity makes software packaging and installation a complex task. Moreover, when scaling to thousands or even millions of devices, manually managing each installation becomes impractical. This is where software build tools become invaluable, offering a way to manage complexity, ensure repeatability, and create consistent device images across a wide range of hardware configurations.

Edge Impulse+MetaTF = The Easy Button for Buildroot and Yocto​

Training a model, and making it small enough for low size, weight and power devices is one thing, but packaging and deploying the application to the edge device is another. BrainChip has extensive experience with major embedded build systems such as Buildroot and Yocto. Our goal is to ensure that our Akida System on Chip design can be seamlessly integrated into our customers unique embedded environments. Furthermore, our partnership with Edge Impulse is helping to revolutionize the ML-to-Edge training and deployment process, making it easier than ever to deploy advanced event-based machine learning solutions on embedded devices.

A prime example of our expertise in action is the Akida™ Edge Box, developed in collaboration with our partners VVDNand Edge Impulse. This innovative product embodies the principles discussed in this blog post, showcasing our ability to tackle complex edge AI challenges. The Edge Box runs on an NXP i.MX 8M application processor, featuring dual Akida 1000 chips on reference boards connected to a single M.2 PCI card. Our demonstration models were trained using Edge Impulse Studio and we bundled our MetaTF libraries into the target image using NXP’s i.MX Yocto Project Board Support Package. This real-world application demonstrates our commitment to integrating cutting-edge event-based computing with established embedded system build tools.

To learn more about how BrainChip’s solutions architecture team can integrate cutting-edge event-based machine learning into your embedded device toolchain, or to explore how the BrainChip Edge AI Box could benefit your AI initiatives, contact us for a demonstration. We’re here to help you take your specific ML processing needs and AI products to the edge.

"A prime example of our expertise in action is the Akida™ Edge Box, developed in collaboration with our partners VVDNand Edge Impulse."

The mention of Edge impulse indictes that we have been working with EI to develop RAG small language models adapted (quantized) for Akida.

 

[Mea culpa]

Hi Bravo.
Sure, take a break if you wish/need to.
It's healthy to step away from social media every once and a while.
Smell the kitten mitts, ride the rainbow, have a run and a spa. Reset.
We'll miss you and what you bring here along with your gorgeous personality.
Frangipanni seems to be an excellent researcher, as are you, but I rarely read what she puts here in such excruciating and voluminous detail.
Frankly I don't have the time or inclination.
But, I'm happy to skim over 99% of it in the expectation that one day it may be important to me.
If he/she/they/it are bothering you with their attacks I suggest you treat them as I do the pests over on the crapper.
Don't engage with them. They exist, but that doesn't mean they can, or have to, define your existence, unless you allow it.
They may say all sorts of horrid things or spin myriad distortions to justify existence, but just don't play their game.
Anyhow, I hope you will return to us soon, as I, and I am sure, many others will miss you. xo

Again Hoppy you have conveyed in words what I have, and it appears from the responses that others have felt also. I concur with your comment on Frangipanni's research; similarly to yourself, I have limited time. @Bravo is a contributor of long-standing. Her posts are informative, respectful and often her humour, her mischievous humour, is delightfully wicked. She has shared much of her personal life with us. I have not met her, yet I feel she is a friend. I too, will miss her, and cherish her return.