With the Jetson Nano, NVIDIA has done a fantastic job of bringing GPU-accelerated machine learning to the masses. For less than the cost of a used graphics card, you get a turn-key Linux computer that’s ready and able to handle whatever AI code you throw at it. But if you’re trying to set up a lab for 30 students, the cost of even relatively affordable development boards can really add up.
Which is why [Tea Vui Huang] has developed jetson-emulator. This Python library provides a work-alike environment to NVIDIA’s own “Hello AI World” tutorials designed for the Jetson family of devices, with one big difference: you don’t need the actual hardware. In fact, it doesn’t matter what kind of computer you’ve got; with this library, anything that can run Python 3.7.9 or better can take you through NVIDIA’s getting started tutorial.
So what’s the trick? Well, if you haven’t guessed already, it’s all fake. Obviously it can’t actually run GPU-accelerated code without a GPU, so the library [Tea] has developed simply pretends. It provides virtual images and even “live” camera feeds to which randomly generated objects have been assigned.
The original NVIDIA functions have been rewritten to work with these feeds, so when you call something like net.Classify(img) against one of them you’ll get a report of what faux objects were detected. The output will look just like it would if you were running on a real Jetson, down to providing fictitious dimensions and positions for the bounding boxes.
If you’re a hacker looking to dive into machine learning and computer vision, you’d be better off getting a $59 Jetson Nano and a webcam. But if you’re putting together a workshop that shows a dozen people the basics of NVIDIA’s AI workflow, jetson-emulator will allow everyone in attendance to run code and get results back regardless of what they’ve got under the hood.
Over the past decade, we’ve seen great strides made in the area of AI and neural networks. When trained appropriately, they can be coaxed into generating impressive output, whether it be in text, images, or simply in classifying objects. There’s also much fun to be had in pushing them outside their prescribed operating region, as [Jon Warlick] attempted recently.
[Jon]’s work began using NVIDIA’s GauGAN tool. It’s capable of generating pseudo-photorealistic images of landscapes from segmentation maps, where different colors of a 2D image represent things such as trees, dirt, or mountains, or water. After spending much time toying with the software, [Jon] decided to see if it could be pressed into service to generate video instead.
The GauGAN tool is only capable of taking in a single segmentation map, and outputting a single image, so [Jon] had to get creative. Experiments were undertaken wherein a video was generated and exported as individual frames, with these frames fed to GauGAN as individual segmentation maps. The output frames from GauGAN were then reassembled into a video again.
The results are somewhat psychedelic, as one would expect. GauGAN’s single image workflow means there is only coincidental relevance between consecutive frames, creating a wild, shifting visage. While it’s not a technique we expect to see used for serious purposes anytime soon, it’s a great experiment at seeing how far the technology can be pushed. It’s not the first time we’ve seen such technology used to create full motion video, either. Video after the break.
NVIDIA kicked off their line of GPU-accelerated single board computers back in 2014 with the Jetson TK1, a $200 USD development system for those looking to get involved with the burgeoning world of so-called “edge computing”. It was designed to put high performance computing in a small and energy efficient enough package that it could be integrated directly into products, rather than connecting to a data center half-way across the world.
The TK1 was an impressive piece of hardware, but not something the hacker and maker community was necessarily interested in. For one thing, it was fairly expensive. But perhaps more importantly, it was clearly geared more towards industry types than consumers. We did see the occasional project using the TK1 and the subsequent TX1 and TX2 boards, but they were few and far between.
Then came the Jetson Nano. Its 128 core Maxwell CPU still packed plenty of power and was fully compatible with NVIDIA’s CUDA architecture, but its smaller size and $99 price tag made it far more attractive for hobbyists. According to the company’s own figures, the number of active Jetson developers has more than tripled since the Nano’s introduction in March of 2019. With the platform accessible to a larger and more diverse group of users, new and innovative applications for machine learning started pouring in.
Cutting the price of the entry level Jetson hardware in half was clearly a step in the right direction, but NVIDIA wanted to bring even more developers into the fray. So why not see if lightning can strike twice? Today they’ve officially announced that the new Jetson Nano 2GB will go on sale later this month for just $59. Let’s take a close look at this new iteration of the Nano to see what’s changed (and what hasn’t) from last year’s model.
In this age of headlines that use the b-word in place of nine zeros it’s easy to lose track, so you may be wondering, didn’t SoftBank just buy Arm? That was all the way back in July of 2016 to the tune of $32 billion. SoftBank is a holding company, so that deal didn’t ruffle any feathers, but this week’s move by Nvidia might.
Arm Limited is the company behind the ARM architecture, but they don’t actually produce the chips themselves, instead licensing them to other companies who pay a fee to use the core design and build their own chip around it. Nvidia licenses the ARM core for some of their chips, and with this deal they will be in a position to set terms for how their competitors may license the ARM core. The deal still needs regulatory approval so time will tell if this becomes a kink in the acquisition plan.
Chances are you know the Nvidia name for their role as purveyors of fine graphics cards. They got a major boost as the world ramped up Bitcoin and other cryptocurrency mining hardware which early on was mainly based on the heavy lifting of graphics processors. But the company also has their eye on the ongoing wave of hardware targeting AI applications like computer vision. Nvidia’s line of Jetson boards, marketed for “next-generation autonomous machines”, all feature ARM cores.
Assuming the deal goes through without a hitch, what will be the fallout? Your guess is as good ours. There is certainly a conflict of interest in a company who competes in the ARM market owning the Arm. But it’s impossible to say what efforts they will make to firewall those parts of the business. Some might predict a mass exodus from the ARM ecosystem in favor of an open standard like RISC-V, but that is unlikely in the near-term. Momentum is difficult to overcome — look at how long it took ARM to climb that mountain and it was primarily the advent of a new mobile ecosystem lacking an established dominant player that let ARM thrive.
The frame is built out of the same brackets and aluminum tubing used to add handrails to stairwells on buildings. Not only is this a fast way to do it, the set-up can be guaranteed to be sturdy since hand rails are often literally standing between life and death. The high ground clearance allows for all sorts of sensors and devices to be mounted while still being able to clear the plants below.
For motion hub motors driven by an ODrive were re-purposed for the task. He explored turning the wheels as well, but it seems like differential steer and casters works well for this set-up. ROS on an Nividia Jetson runs the show and deals with the various sensors such as a stereoscopic camera and IMU.
We’re excited to see what hacks people come up with as research in this area grows. (Tee-hee!) For example, [Junglist] wants to see the effect of simply running a UV light over a field rather than spraying with pesticides or fungicides would have.
Every beginning standard needs a test case, and OSK’s is a simple one. A bowl that tracks what you eat. While a simple concept, the way in which the data is shared, tracked, logged, and communicated is the real goal.
The current demo uses a Nvidia Jetson Nano as its processing center. This $100 US board packs a bit of a punch in its weight class. It processes the video from a camera held above the bowl of fruit, suspended by a scale in a squirrel shaped hangar, determining the calories in and calories out.
It’s an interesting idea. One wonders how the IoT boom might have played out if there had been a widespread standard ready to go before people started walling their gardens.
Found yourself with a shiny new NVIDIA Jetson Nano but tired of having it slide around your desk whenever cables get yanked? You need a stand! If only there was a convenient repository of options that anyone could print out to attach this hefty single-board computer to nearly anything. But wait, there is! [Madeline Gannon]’s accurately named jetson-nano-accessories repository supports a wider range of mounting options that you might expect, with modular interconnect-ability to boot!
A device like the Jetson Nano is a pretty incredible little System On Module (SOM), more so when you consider that it can be powered by a boring USB battery. Mounted to NVIDIA’s default carrier board the entire assembly is quite a bit bigger than something like a Raspberry Pi. With a huge amount of computing power and an obvious proclivity for real-time computer vision, the Nano is a device that wants to go out into the world! Enter these accessories.
At their core is an easily printable slot-and-tab modular interlock system which facilitates a wide range of attachments. Some bolt the carrier board to a backplate (like the gardening spike). Others incorporate clips to hold everything together and hang onto a battery and bicycle. And yes, there are boring mounts for desks, tripods, and more. Have we mentioned we love good documentation? Click into any of the mount types to find more detailed descriptions, assembly directions, and even dimensioned drawings. This is a seriously professional collection of useful kit.