when I was doing an internship back in 2018, I started looking into object detection techniques, because I needed to solve a visual inspection problem. This problem required the detection of many different objects in a stream of images coming from an industrial camera.
To tackle this challenge, I first tried to use classification in combination with a sliding window! Naturally, the system was very slow and unfit for production.
After this, I started looking into end-to-end deep learning models that perform object detection. I stumbled upon a now famous paper from google research titled:
Speed/accuracy trade-offs for modern convolutional…
Introduction to training deep learning models on Google Cloud AI Platform
Around a year ago I was working on a personal project where I needed to be able to train deep learning models on the cloud. It was my first time training on the cloud so I started googling things and trying to figure out which path is the best for what I was trying to achieve.
A few months ago, a friend of mine contacted me on LinkedIn to ask me whether I had some experience with cloud computing and whether I can guide him to learn how to…
I surf the internet everyday looking for machine learning and computer vision content. I like to stay up-to-date with what’s happening in the field of ML because this is a field that can surprise you almost everyday!
One question that I came across a few times is :
Which is better OpenCV or Tensorflow?
To some, this is not a valid question.
To others, this is a question worth thinking about.
The simplest answer is that Tensorflow is better than OpenCV and OpenCV is better than Tensorflow!
I hope I didn’t confuse you! If I did, please read on!
What is it? Why use it? And how to create your custom learning rate schedulers in Tensorflow 2
Learning rate schedule is simply the process of making your learning rate change during the training of your neural networks.
Some publications show that by changing the learning rate during the training, you can approach the global minimum faster. Moreover, your training process will avoid being stuck in a local minimum or in saddle points.
Several researchers have stressed the need for changing the learning rate during the training. …
Common trends and techniques in the industry from an AI engineer
Deep learning is seeing tremendous adoption in different industries. One specific area where deep learning has shown great potential is Computer Vision.
I personally graduated from a computer vision master’s program and went immediately to work in the industry. So what follows is my take on different trends that I am seeing in companies that are using deep learning to tackle challenging computer vision problems.
So going back to my studies, in the middle of the master’s program, I did an internship in a company in Luxembourg that makes…
When my students started complaining about not being able to train their deep learning models on Google AI Platform, I thought they probably have a small error somewhere. Maybe, they didn’t use the command that I gave them correctly. Maybe, they forgot to setup some things on AI platform. Or simply they made a typo somewhere.
The course that my students were enrolled in was about how to train and evaluate deep learning models for object detection(Faster RCNN, SSD and YOLOv3) using Tensorflow 2. One of the last steps when it comes to running your training on google AI platform…