Object detection is useful for understanding what’s in an image, describing both what is in an image and where those objects are found. In general, there are two different approaches for this task –
Managing multiple research experiments at a time can be overwhelming. The same applies to deep learning research as well. Beyond the usual challenges in software development, machine learning developers face new challenges - experiment management (tracking which parameters, code, and data went into a result) and reproducibility (running the same code and environment later)!
As a researcher, I have to keep myself up-to-date with latest research in my field. Given the pace with which deep learning research is moving currently, it has become quite a gargantuan task lately. Large quantity also brings a lot of noise with it.
Lately, a lot of my friends have been asking about my deep learning workstation setup. In this post I am going to describe my hardware, OS, and different packages that I use. In particular, based on the question, I found that the most of the interest have been around managing different python versions, and modules like pytorch/tensorflow libraries etc.
Usually in a conventional neural network, one tries to predict a target vector y from input vectors x. In an auto-encoder network, one tries to predict x from x. It is trivial to learn a mapping from x to x if the network has no constraints, but if the network is constrained the learning process becomes more interesting.