NVidia Benchmarks for specific Artificial Intelligence tasks (Generative Adversarial Networks)

I started to do some benchmarks in order to plan the future needed computers and tools for Generative Adversarial Networks (GAN)

bechmark-adrian-minst-1

3 computers until now (all on Windows 10)
Elapsed Time for CPU = 5 hours, 47 minutes, 26 seconds
Elapsed Time for GPU NVidia 920M = 1 hour, 39 minutes, 15 seconds
Elapsed Time for GPU NVidia 940MX + SSD = 1 hour, 13 minutes, 16 seconds

GPU and SSD are a must so you need to
– install tensorflow-gpu instead of tensorflow
– have python 3.6 (on 3.7 id does not work yet, see tensorflow site https://www.tensorflow.org/install/pip)
– install Cuda 9 (I also installed Cuda 10 but it does not work – you can have both Cuda installed)
https://developer.nvidia.com/cuda-90-download-archive
Cuda 10 is here https://developer.nvidia.com/cuda-downloads
– install cuDNN for Cuda 9 (register at NVidia, download package and copy 3 files as described on site )
https://docs.nvidia.com/deeplearning/sdk/cudnn-install/index.html

Tensorflow news:
TensorFlow 1.13.0-rc2 @aselle aselle released this 5 days ago
TensorFlow GPU binaries are now built against CUDA 10 and TensorRT 5.0.
Support for Python3.7 on all operating systems.

https://github.com/tensorflow/tensorflow/releases

I tested with Tensorflow nightly 1.13.0-rc2, Cuda 10, cuDNN v7.4.2. I had to install a new driver from Nvidia GeForce Experience application to make the system work. I am still on Python 3.6.8. I will come back here after Python update. Now I am benchmarking this new configuration

Bad news – Cuda compute capability minimum requirements (solved on Tensorflow nightly 1.13.0-rc2, Cuda 10, cuDNN v7.4.2)

I upgraded to Tensorflow 1.11.0 and the game ended for NVidia 920M (Cuda compute capability required is 3.7 > cuda compute capability of NVidia 920M). Check your board on this table https://developer.nvidia.com/cuda-gpus

Tests where done on super algorithms described at https://www.pyimagesearch.com/

I will add more NVidia boards and configurations on this article 

Enjoy Artificial Intelligence and Generative Adversarial Networks (GAN) !

Dorin

P.S. Strange errors while using other configurations

File “C:\Program Files\Python36\lib\site-packages\tensorflow\python\keras\backend_init_.py”, line 22, in
from tensorflow.python.keras.impl.keras.backend import abs

File “C:\Users\dorin\AppData\Local\Programs\Python\Python36\lib\ctypes_init
.py”, line 348, in init
self._handle = _dlopen(self._name, mode)
OSError: [WinError 126] The specified module could not be found

File “C:\Users\dorin\AppData\Local\Programs\Python\Python36\lib\site-packages\tensorflow\python\platform\self_check.py”, line 97, in preload_check
% (build_info.cudnn_dll_name, build_info.cudnn_version_number))
ImportError: Could not find ‘cudnn64_7.dll’. TensorFlow requires that this DLL be installed in a directory that is named in your %PATH% environment variable. Note that installing cuDNN is a separate step from installing CUDA, and this DLL is often found in a different directory from the CUDA DLLs. You may install the necessary DLL by downloading cuDNN 7 from this URL: https://developer.nvidia.com/cudnn

……………

 File “C:\Users\dorin\AppData\Local\Programs\Python\Python36\lib\site-packages\tensorflow\python\saved_model\model_utils\export_utils.py”, line 51, in <module>
    ModeKeys.TRAIN: signature_constants.DEFAULT_TRAIN_SIGNATURE_DEF_KEY,
AttributeError: module ‘tensorflow.python.saved_model.signature_constants’ has no attribute ‘DEFAULT_TRAIN_SIGNATURE_DEF_KEY’
……………….
Ignoring visible gpu device (device: 0, name: GeForce 920M, pci bus id: 0000:08:00.0, compute capability: 3.5) with Cuda compute capability 3.5. The minimum required Cuda capability is 3.7.

 

Dog AI

dogs-high-resolution

Dogs are undoubtedly one of the friendliest animals in the world towards their owners. Empathetic and companionable, they’re almost like us in their regular lives. Dogs have hobbies and food preferences, they adore active lifestyle, games, and a good sleep. But most importantly — they make us happy and teach us how to enjoy every day of our lives by their own example.

Dogs are a boon to photographers, who love shooting wild animals and nature, yet, these pets are very prone to get into funny situations too. There are hundreds of viral videos and photos, which were taken by vigilant owners and uploaded on the Internet, where different dog breeds make everyone around laugh because of their inborn cheerfulness (and clumsiness). There’s even more of amusing materials stored on flash drivers or external HDDs all around the world. After a few months of ownership, it’s OK to find your whole phone or camera gallery conquered by the images of your pet in all kinds of positions, situations and moods.

Plenty of similar photos in the camera memory is a very common problem for people who like to imprint the most significant moments of their friendship with pets. Moreover, it’s a big trouble when among dog photos you can’t find some important images. Such significant, yet, memo polluting shots accumulate on your SD card until there’s no empty space for new pictures. Fortunately, modern technologies can help you even in such complicated situation.

To sort out the best photos and get rid of duplicates you’ll need to spend hours and hours of monotone manual scrolling. Are you ready to lose a whole day for such questionable enjoyment? If the answer is “no”, then Brian AI will be the best choice of assistant for you. Using all the latest achievements in artificial intelligence researches, we created an app which can do all photo systematization job without any user’s help.

Based on sophisticated algorithms, Brian AI organizes all available images according to their content. Its potential is almost unlimited — app easily finds your nursling midst any landscape and among other dogs. Application attaches key tags, generated by Google Vision and GoogLeNet Inception services, to the photo, which means you can quickly find some specific image amid thousands of them. Application analyzes dog’s breed, coloration, clothes and surroundings, identifies dog’s snout features and unique marks. Brian AI can manage all your dog photos in seconds despite their big quantity or poor quality.

Apart from filtration and management, our app can systematize images according to their tags into albums. It can analyze focus, blur, angle of the photo, automatically choose the best one from the series of similar shots, and then delete the others with one click of yours. The search amidst photos was never so simple as with Brian AI — you can put in required tag and get results immediately, or use the right-sided menu with all the tags collected in one place. About even more functions of this app you can learn by this link.

A unique feature of our application is, of course, its ability to educate. Artificial intelligence can boost its productivity and accuracy using your photo archive thanks to built-in mechanisms based on deep learning technology. The more you use it, the more exact results you get. By repetitive analysis of your pet’s photos, Brian AI can train to recognize its 10 year old appearance, which would be very convenient if you haven’t put in order your SD’s memory in a while.

If you need to sort your pictures into albums with a need for further exploitation, then, again, Brian AI, which you can find on http://www.brianai.com, is a must-have. Sell a whole bunch of same-thematic photos, make a presentation-like movie demonstrating pictures according to their photographing date from the oldest ones up to the newest, upload images from social networks — do whatever you need with your brand-new systematized archive. Brian AI is suitable for every desktop, computer or mobile phone, so there’s no need to transfer pictures from different devices. Only for $10 you’ll get a whole arsenal of AI instruments and a constant support from our consultants, which is nothing compared with a convenience of Brian AI application.

Modern technologies have made our lives much easier and comfortable by taking away a need of hand-operated functions, but today their sphere of influence is growing day by day. Artificial intelligence-based applications like Brian AI are the examples of how high tech can be used in our everyday lives. Download it and use its full potential to make your life much more systematized!

All details about Brian AI application’s creation and history, as well as detailed usage instructions you can find here.
www.brianai.com

How to sell your images

Here are some sites where Brian AI can help you organize and upload images, for sale. See the examples below with images uploaded by me using Brian AI

dreamstime_portofolioDreamstime

getty_images_istock_portofolio-800Getty Images and iStock

picfair_portofolioPicFair

adobe_fotolia_portofolio-800Adobe & Fotolia

123rf-800123RF

See upload and organize instructions on User Manual of Brian AI. Try the application for free for 30 days !

Enjoy and a lot of sales !

Dorin

One Click #ai Money

Upload 100+ images for sale with Artificial Intelligence from brianai.com plus the One Click Option from dreamstime.com and stock.adobe.com

dreamsTime-all-images-jpg
Make Money Doing What You Love! Tag and upload your photos for sale using Artificial Intelligence

Photography is a very popular hobby, which doesn’t require years of training or highly expensive ammunition.
If you have an endow of seeing beauty in the smallest details and you want to spread captured emotions with other people, then photography will be the best hobby option for you.
Every single photo made with passion is sure to turn up fantastically good, whether you’re a newbie or an experienced master.
As a well-known phrase says, “there’s no better job than the one, where you gain money for doing what you love”.
Making photos isn’t an exception.

The demand for qualitative or simply beautiful shots in the Internet is huge – many blogs or sites owners need unique images to diversify their text content.
Surprisingly, not many people consider photographing as a real money source, which is surely a serious problem both for image buyers and sellers.
Enthusiasm and willing to make something extraordinary often smashes against lack of output or even insolent steal of your brainchild.
That’s why platforms like dreamstime.com are so much important for image selling field: they allow photographers to gain money for their work and motivate them to reach higher productivity and to improve professional skills.

1500-images-uploaded-2b-small
1500 images uploaded with one-click
Yet, sorting, editing, uploading several shots by hand is a thankless job. Doing same operations with all of the photos continually is not only plundering your spare time, but also transforms you from the Creator into a simple task making machine.
So, why not leave all the boring work for the application, created specially for such needs?

Brian AI is an artificial intelligence app, made for serving photographers in their hard work.
Apart from essential editing functions, Brian AI can analyze, systematize and manage your photos in seconds, which allows you to leave dull work behind and concentrate on creative process only.

Brian AI is a perfect choice for photographers wishing to reap the monetary fruits of their efforts.
Artificial intelligence of the app, build upon deep learning mechanism, dissects the composition of the image, and then adds appropriate tags according to what it has found on it.
Images can later be organized into specific albums, while tags generation does all the “dirty work” for the user,
who uploads photos on services like dreamstime.com.
It creates a unique CSV file of the album, which is transferred onto the site’s working area together with the photo original via FTP connection.
More about the process of analysis and upload you can learn in a convenient tuition video via this link.


Playlist with links to videos https://www.youtube.com/playlist?list=PL9rGXY173aFoWr-JiXIlrlTTxw5v82KE5

Say “no” to the dull manual work.
Simply by using one on the multiple Brian AI instruments you save hours of free time – imagine, what you can do using all app’s potential.

Artificial Intelligence of Brian AI opens up a huge amount of capabilities for better photo management for photographers.
Whether you have 10 or 10.000 images for processing, it is no longer a problem with this application, as soon as AI work do all the processing while you sleep.
It can search for the shots duplicates and delete them on your request, or even analyze the focus, blur, exposition of the series of images to find the best shot among regular ones.

App has an interface with Google Vision chargeable platform, which makes available even more labels and landmarks for your photos.

Step by step procedure

The procedure to upload photos to Dreamstime.com or Adobe Stock Photos (stock.adobe.com)
– organize all your photos using AI and Brian AI,
– choose the images that you think will impress your clients, will be useful to be used in other works, add them to an album
– Ask “Brian AI” to prepare the CSV files and the Originals to pe uploaded to Adobe Stock or Dreamstime
– on Dreamstime do not forget to upload the CSV using FTP before you upload the images
– FTP upload all the files of your album (on each of Dreamstime and Adobe Stock)
– Using Adobe Stock interface upload the CSV file created by “Brian AI” for you album
– Start bulk operations (on Adobe or Dreamstime cloud interfaces) that will submit all your photos for review to Adobe or Dreamstime photograph experts
– Repeat the process with more and more images
– Watch the approval process of you images and your sales every time you wake up in the morning :-). Make money while you sleep.
– Go out to work with your super photo camera and super smartphone and create more photos everytime something is inspiring you. If you have a different job go to lunch with your camera and make some photos on the road. Take some printed documents with you for getting models approval for photos when necessary. After job stay one more hour in a place that you love and do some pictures
– Late night at home, move the photos to your Hard Drives and let Brian AI work on finding tags and labels and faces and landmarks ….
– Repeat the process
– Become the supplier for the photos used in this article or other blogs and magazines around the world !
– Watch the photos of your life by using “Brian AI” to search and navigate through years of photos, check and remember how beautiful life is 🙂 – watch the full list of movies generated while using “Brian Ai” on this Youtube playlists
Videos created by Brian AI playlist
How to Sell photos with one click bulk upload – playlist

All details about Brian AI application’s creation and history, as well as detailed usage instructions you can find here.
www.brianai.com

shutterstock_100794931
Subscribe on this link to receive a PDF with all the details of the selling process

Matrix reversed

pisa-faces-andrei-anca-w200 While watching the movie Matrix, we were all revolted that all the people are connected to a Matrix and are living virtual lives
Starting today I noticed that some people will deliberately connect to Artificial Intelligence Matrix in the near future. They will become dependent of it.
They will feel bad, lonely and unproductive outside The Matrix.

ioana_anca_300 It is happening today with mobile phones. We all saw jokes about people staring at smartphones and looked at new jobs like helping an “email reader” and a “facebook poster” to cross the street. Netherlands already implemented an upgrade of traffic lights – the big one for drivers we already know. We even start to get used to the small ones for bicycles. But what about the new one for smartphone readers that is a red bar on the ground level, parallel with the white zebra lines, that goes red at the left and right of your smartphone !
1487090741_bodegraven-lights-to-warn-mobile-phone-users_story

The real Matrix is the one with Artificial Intelligence engines delivered in the clouds of Microsoft, Amazon, Google, Nvidia and many others.
Read about the tested performances of this engines from someone how knows how to ask the right question (read the Terra Questions Set asked by Adrian Rosebrock from PyImageSearch) in this article
tensorflow brain l2

Software Development is different with or without Internet. I remember that one day I had to switch activities during a day without brain connection to Google. By having access only to the books that are already downloaded to your PC and paper books, developing speed go 10 times slower. Debugging goes 10 times slower without the advice of your peers around the world. Developing using reading and using chunks of Internet examples goes back to a slower paste of full book reading. Now even the airplanes start to offer Internet and power supply so you do not feel unproductive on Trans Atlantic flights. 20 years ago the quality of slow paste analysis and development was much better but the deadlines where not so tight. We need The Matrix to deliver at today’s client expectations and competition.

Matrix Reversed existed before 1910 see https://en.wikipedia.org/wiki/History_of_amateur_radio
People where connected to planetary radio-communication matrix and spend long nightly hours to connect with the voice of others (through CQ CQ CQ mesages – Hello I am Bravo Mike Oscar Quebec Hotel India nice to meet you) So Matrix Reversed existed before 1969 (see implementation of Internet https://en.wikipedia.org/wiki/History_of_the_Internet),
now even the no-it-school guys have a device called Facebook (even simpler versions starts competing – WhatsApp, brain sensors, …)

Enjoy your soul until we become the tree from Avatar http://www.imdb.com/title/tt0499549/ – !!! So the supreme network is already described 🙂
Avatar tree of connecting brains and all the past brains uploaded into the tree
Avatar tree used to connect the living brains with all the past brains uploaded into the tree

Bonus ! A TED video of a super scared expert in sociology (scared of super “free” networks working on advertising business models)
We can not live outside the Reversed Matrix because we can not deliver the results and deadlines expected by our clients, today, only with printed books. We can not afford to create or buy private networks nearly 30% as good as the public network. We can not accuse the “free” business models of big companies because they are collecting data on us and use AI advertising models on our personal data. We have each day many choices, but we choose them every day, with our habits, by reading and writing in the spider webs, instead of using just our printed books and Starbucks meetings. A lot of years ago I thought that Second Life application was for me and my freak friends. Now it become our world. I noticed how small dependencies where created then, and huge ones today.

Enjoy !

Dorin

How I used Artificial Intelligence to organize 200.000 photos

“When the Student Is Ready the Teacher Will Appear”

I found a very good book about Artificial Inteligence on this site PyImageSearch.com at this link
I used the algoritms inside the book to compute image blur and other properties
This is a photo of the app using Adrian Rosebrock advices and python code (see the focus info at each image)
adrian_pyimagesearch

See the application at work on this link

Enjoy !

Dorin

How I organized 200.000 of my photos using Artificial Intelligence

pisa-faces-andrei-anca
I organized 200.000 of my photos using Artificial Intelligence. Photos where on 20 external HDD or memory sticks that I filled up during years and years of making photos. I wrote an app to
– Navigate,
– Intelligent Upload (keep track of each photo id and URL for future apps),
– Share,
– Sell,
– Create movies,
– Include in apps
– Find faces,
– Detect landmarks, labels
– Detect duplicates
– Time Travel
see details and movies of my work on this google document

film15_026

Enjoy !

Dorin

Artificial Intelligence for images – Bucharest, Romania

Every year I made some pictures of the marvelous city of Bucharest, Romania
This year I used Artificial Intelligence to label and group them and my AI virtual assistant created an album, see bellow

How I organized 200.000 of my photos using Artificial Intelligence (from 20 local external HDD with images,
working only on my laptop and one empty external HDD) ?
– Navigate,
– Intelligent Upload (keep track of each photo id and URL for future apps),
– Share,
– Sell,
– Create movies,
– Include in apps
– Find faces,
– Detect landmarks, labels

Download the app from here

Revolution Square, car, sky, landmark, metropolitan area

Revolution Square, car, sky, landmark, metropolitan area

Revolution Square, car, sky, landmark, metropolitan area, city, infrastructure, road, town, vehicle, daytime, obelisk, flagpole, flagstaff, triumphal arch, cab, hack, taxi, taxicab, pedestal, plinth, footstall

Romanian Athenaeum, car, land vehicle, transport, luxury vehicle

Romanian Athenaeum, car, land vehicle, transport, luxury vehicle

Romanian Athenaeum, car, land vehicle, transport, luxury vehicle, sky, landmark, road, family car, infrastructure, town, cab, hack, taxi, taxicab, minivan, minibus, police van, police wagon, paddy wagon, patrol wagon, wagon, black Maria, ambulance

Revolution Square, car, metropolitan area, sky, landmark

Revolution Square, car, metropolitan area, sky, landmark

Revolution Square, car, metropolitan area, sky, landmark, town square, infrastructure, town, city, road, urban area, obelisk, triumphal arch, flagpole, flagstaff, palace, bell cote, bell cot

See full article on this link

Download the app from here

Artificial Intelligence for images – Wynn & Encore, Las Vegas

Every year I made some pictures of the marvelous garden of Wynn & Encore in Las Vegas
This year I used Artificial Intelligence to label and group them and my AI virtual assistant created an album, see bellow

How I organized 200.000 of my photos using Artificial Intelligence (from 20 local external HDD with images,
working only on my laptop and one empty external HDD) ?
– Navigate,
– Intelligent Upload (keep track of each photo id and URL for future apps),
– Share,
– Sell,
– Create movies,
– Include in apps
– Find faces,
– Detect landmarks, labels

Download the app from here

red, light, lighting, glass, flower, mid autumn festival, lampshade, lamp shade

red, light, lighting, glass, flower, mid autumn festival, lampshade, lamp shade

red, light, lighting, glass, flower, mid autumn festival, lampshade, lamp shade, table lamp, carousel, carrousel, merry-go-round, roundabout, whirligig, confectionery, confectionary, candy store, wall clock

tree, christmas decoration, flower, branch, christmas tree, christmas lights, christmas

tree, christmas decoration, flower, branch, christmas tree, christmas lights, christmas

tree, christmas decoration, flower, branch, christmas tree, christmas lights, christmas, jackfruit, jak, jack, greenhouse, nursery, glasshouse, spider web, spider’s web, vase, lemon

flora, flower arranging, flower, botany, floristry, plant, leaf

flora, flower arranging, flower, botany, floristry, plant, leaf

flora, flower arranging, flower, botany, floristry, plant, leaf, floral design, branch, produce, jackfruit, jak, jack, greenhouse, nursery, glasshouse, pot, flowerpot, vase, orange

See full article on this link

Download the app from here

Artificial Intelligence used for images – my Boats Album

My Boats Album

I searched flowers and found all the flowers that I played with, on my camera, in the last 10 years.
How I organized 200.000 of my photos using Artificial Intelligence (from 20 local external HDD with images, working only on my laptop and one empty external HDD) ?
– Navigate,
– Intelligent Upload (keep track of each photo id and URL for future apps),
– Share,
– Sell,
– Create movies,
– Include in apps
– Find faces,
– Detect landmarks, labels

See below an example of images that are identified, labeled and uploaded to Google Drive, both as original and preview, and then this page is generated by the app

It you send the images as a HTML file with external links like these
My boats album – html file
Landmarks set 21 – html file
360 images with Landmarks – Html file with links to Google Drive

Or inserted in WordPress, see below
sailboat, dinghy sailing, sailing, sail, boat, vehicle, sailing ship

sailboat, dinghy sailing, sailing, sail, boat, vehicle, sailing ship

sailboat, dinghy sailing, sailing, sail, boat, vehicle, sailing ship, sailboat racing, keelboat, catamaran, trimaran, yawl, lycaenid, lycaenid butterfly, snorkel
dinghy sailing, sailboat, boat, sail, sailing, vehicle, sailing ship

dinghy sailing, sailboat, boat, sail, sailing, vehicle, sailing ship

dinghy sailing, sailboat, boat, sail, sailing, vehicle, sailing ship, sailboat racing, keelboat, catamaran, trimaran, yawl, seashore, coast, seacoast, sea-coast, schooner
dinghy sailing, sailing, sail, sailboat, boat, vehicle, sailing ship

dinghy sailing, sailing, sail, sailboat, boat, vehicle, sailing ship

dinghy sailing, sailing, sail, sailboat, boat, vehicle, sailing ship, sailboat racing, windsurfing, catamaran, trimaran, yawl, seashore, coast, seacoast, sea-coast, snorkel

See full article on this link

Download the app from here

Bitnami