We're sorry but this page doesn't work properly without JavaScript enabled. Please enable it to continue.
Feedback

Classify things in Go: the easy way.

00:00

Formal Metadata

Title
Classify things in Go: the easy way.
Subtitle
Building classifiers quickly with the community contributions.
Title of Series
Number of Parts
490
Author
License
CC Attribution 2.0 Belgium:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor.
Identifiers
Publisher
Release Date
Language

Content Metadata

Subject Area
Genre
Abstract
Go and public training models can provide great potential: A fast way to build "eyes around the world", also known as classifiers. And with great powers, come great opportunities, such as building fantastic applications to turn our world a better place to live through the technology with few steps. Go and public training models can provide great potential: A fast way to build "eyes around the world", also known as classifiers. And with great powers, come great opportunities, such as building fantastic applications to turn our world a better place to live through the technology. The GO language, have GoCV package, and it provides the most modern and advanced Computational Vision libraries that exist like OpenCV. In this talk, I'll demonstrate how to use public models from TensorFlow Hub and OpenCV library to easily build classifiers for APIs, taking a super leap from draft to a working classifier, in a few steps! The idea is to demystify the concepts behind classifiers and show how to build one in a few steps and make rankers accessible to the business, showing how GO does this in a unique, scalable and self-performing way, and of course, encouraging the community to contribute and sharing more training models to they can turn more and more accurate!
BuildingAverageDifferent (Kate Ryan album)PlanningProgramming languageComputer hardwareSoftwareInternet der DingeComputer animationLecture/Conference
Open setModel theoryContent (media)Computer-generated imageryVideoconferencingMachine visionComputerEndliche ModelltheorieObject (grammar)Wave packetClient (computing)Product (business)Proof theoryTerm (mathematics)Execution unitEndliche ModelltheorieTerm (mathematics)Multiplication signFormal languageMedical imagingComputational visualisticsVirtual machineProduct (business)Client (computing)System callBasis <Mathematik>Marginal distributionReading (process)Chaos (cosmogony)DampingLattice (order)Scheduling (computing)Independence (probability theory)Real numberInternet der DingeComputer animation
Open sourceIntelMachine visionLibrary (computing)CodeOpen setComputer fileMereologySystem callLibrary (computing)Computational visualisticsRule of inferenceLie groupWindowEndliche ModelltheorieSource codeWave packetBlock (periodic table)Machine visionSoftware engineeringRevision controlTunisOpen sourceComputer animation
Presentation of a groupComputer animation
Machine visionDirected graphVideo gameTheoryComputer programmingCodeComputer animation
VotingWindowCodeSource codeComputer animationMeeting/Interview
Meeting/Interview
Computer programmingClosed setLaptopPresentation of a groupTheorySystem callComputational visualisticsCodeTouch typingLecture/Conference
Chi-squared distributionRankingSoftware testingSource code
Chi-squared distributionSource codeComputer animationLecture/Conference
Web pageMaxima and minimaComputer programmingSource code
Execution unitLecture/Conference
Chi-squared distributionLogic gateWebsiteDataflowLecture/ConferenceSource code
Formal languageModel theoryModulo (jargon)Modul <Datentyp>Library (computing)Virtual machineHeat transferTask (computing)Process (computing)Endliche ModelltheorieComputer animation
Image warpingSource codeLecture/Conference
Computer animationLecture/Conference
Virtual machineRippingGraph coloringComputer animationLecture/Conference
Mach's principleField (computer science)Computer animationLecture/Conference
Computer animation
Computer animationLecture/Conference
Video gameWater vaporComputer animationLecture/Conference
Computer animationLecture/Conference
Computer animationLecture/Conference
Goodness of fitComputer animationLecture/Conference
MaizeLocal ringMaxima and minimaInclusion mapCore dumpView (database)Endliche ModelltheorieProcess (computing)Lecture/ConferenceSource code
QR codePresentation of a groupPresentation of a groupMachine visionSoftware frameworkLink (knot theory)Office suiteComputational visualisticsLibrary (computing)Open sourceSoftware repositoryLecture/ConferenceComputer animation
Process (computing)Ring (mathematics)Water vaporBlock (periodic table)BuildingReading (process)Machine visionType theoryAreaEndliche ModelltheorieFormal languageDiscrete element methodGoogolInheritance (object-oriented programming)Computational visualisticsComputer-assisted translationSoftwareGraph coloringIdentifiabilityModel theoryFreewareComputer animation
Open sourcePoint cloudFacebookLecture/Conference
Transcript: English(auto-generated)
It's okay? Okay. Hi. I'm just, okay. Who I am. My name is a kind of different name,
and that give show how to speak my name. Just a little, so close. It's Cheyney. I'm
working with AI at a company called Novell, and me and my team use a lot of Go program language to build things, and yes, I book a lot of software things and codes, and we
give, we built the software, and then we give some artificial intelligence to the hardwares. We have hardwares to new hardwares. Like, I don't know if I can tell about it, but yes, imagine IoT new things. This is our schedule today. I will talk about introduction
to CV classifiers, a challenge for you, and models from community, and other libraries.
Here is why use Go CV and not Python, for example, and Go first, but. So, what is computer vision? I think, I don't know if she, everybody knows what it is. It's basically a field
inside of the artificial intelligence that trains computer to, like, computer has eyes and brain, like us, and can see, so that way, the computer can see our, or the human world, and then, if you have luck, then they can react to the, to what the machines
see. This is my challenge, a scenario. IoT company, it's not my company, I promise. I received a proposal for a new client who asked for 10 distinct image classifiers for
a different product. Some classifiers will be wrapped into API, and put in connecting a Kafka SKU, and other real world independence. They want 10 classifiers. You need to do
a pack to show to your companies if it's viable to do it in the time that was proposed. How you do it? Anybody? You're the imaginary classifier from the scratch? No. Wrong answer.
Because, I imagine it, you don't, you don't throw money away, and I'm correct, because it's hard. You need to knows math, knows how to do models, and know basically, basically
some language, and it's a huge team to do it. And how they say time is money?
The other is through the community, and open source things, like GoCV. We can use the thing the other people just built before, like TensorFlow hub, it's a source,
Kafka is another source, and GoCV. Basically, the TensorFlow hub is a library, open library.
You can imagine a lot of pieces of Lego blocks, and that pieces are trained models, and you can just download that, and use that. And then you can do what I did, is a minimal modifications in the files, in the .pb files, and the text files, and like fine tuning.
And why GoCV? And why use GoCV? GoCV gives to the gofers,
access to the OpenCV computer vision library. GoCV packages the last release of Go, and the last version is for Linux, Mac OS, and Windows. Why I tell about Windows? Because I
bring a computer with Windows, because I want to show and prove that makes GoCV and computer vision accessible for everyone. If you are a software engineer or not,
and GoCV supports the Intel OpenVINO toolkit, too. It's a very nice thing. Okay, show me the code, because blah, blah, blah. Just a second.
Big square on top. Oh, a mirror.
This is my sheeting. So, first start, I will run a program that
our code we can found on the GoCV platform, and in my last slide, I put all the material. So, just, I just, we will try to run, and yeah, that, that, that, that, that, that
error, it's because Windows, but how, how, how it works. It's just, just to, just to
show that I can open a window and communicate my code.
Better? Yeah. I blame you, Windows, for this. I try, I try, but it's my camera.
I think it might be, no. Okay. Hello? Yeah, it's that thing. Boo. Okay. Okay.
Close. I, I, I, how I show my code if I don't, can't touch my notebook. But it's literally like, but it's, the computer is muted. It's for me.
Thank you, Microsoft. Okay.
Some problems like that, the rest of my presentation. It's just to test my camera,
and the camera works, and now, okay, I will do another test with my camera,
because I try to do, I try to see if that camera will recognize that it, I think, or recognize me. Margie, come here. A picture of me. I can, if you, anyone can, if you,
we can, the, the program will count how crazy golfers there are. Can I try to,
okay. Sorry. Now, I will, what I told you, told you before, I just entering,
like, I, I, you show before. In that site, the TensorFlow web, you can search here for,
for models. Then, I download that model, and then I do a finite, finite turning, and I try to record the, the, the, the, the, the, the, the, the, the, the, to, to, to the new, new classifier can recor, recor, recor, recor, recor, recor, recore the, the golfer.
And we will see if he works. This is the, this is the, the, what the program,
Another color of gopher?
I can, I can, I can do it on other things,
like water, water.
I have mug, and I borrow it from my hotel.
This is that baguole. Okay, we can do with the models, the pre-building models. I obviously just download the model,
just have the classification for the water, the cough mug, and I put another layer,
another two layers, indeed, the gopher plushie, and the gopher, but back to my presentation.
Go cv.io, it's the platter farm. We can learn everything about Go cv there, and there is a link to the GitHub repo. TensorFlow rub, it's the library I told before.
Caffe, it's the framework from computer vision. It's open source too, and my presentation are in the scare code, and it's just available through, you can access for the speaker deck.
This, that Giphy, I made it, and if you access my GitHub, you can download this, that's for free too. And you can put it in your Slack team,
and some water you can do. There is another version, it's a gopher, but just do it, and with the parrot party too. They change colors and looks like very gay, I love it. I don't know if you understand me,
sorry about my English, not my native language. I try to pass the idea here, anyone can build computer vision softwares and models and use in your academic life or your job for fun.
You can download, for example, in my job, we download the model of Google to identify sheets,
blank sheets, and sheets with some text, and we did the final turning about it, and now you use it in my job. It's another example. Of course, it's a barrentine, so I prefer
bring gophers, because show the blank sheet, it's bleh. That's it, there is a lot of models, car models, pet models, you can identify
a lot of type of cats, for example, cars, counting cars, it's insane. How much things you can do with the pre-building models? It's like Lego blocks, you can build everything.
That's it.