Word2vec and generator in Tensorflow 2 – Train and use your own embedding on large datasets

I found a lot of blog posts about word2vec in tensorflow1 that explain the workings of the method very well and this blog post is not about that. It’s simply about implementing it and training the model in tensorflow2, with a large dataset. Thus here is what I’ll talk about: Creating token encoder in tensorflow […]

Read More Word2vec and generator in Tensorflow 2 – Train and use your own embedding on large datasets

Tensorflow and imbalanced data – Am I the asshole?

During this quarantine I took it upon myself to try not be too bored while I couldn’t travel and was waiting for the next thing. One thing I decided to pick up again was machine learning and specifically getting my hands back on tensorflow. A lot as changed since last time I played with tensorflow […]

Read More Tensorflow and imbalanced data – Am I the asshole?

Gmapping install with ROS

Ok a quick post because everytime I need to install it I get confused about what is what:  openslam_gmapping is the gmapping package for install in ROS. Findable here. slam_gmapping is the ros wrapper around Gmapping ROSified. You can find it here. I find it surprising that Gmapping was ROSified, to then need an additional […]

Read More Gmapping install with ROS

ROS Kinetic on Fedora 26/27

It is time for another ROS on Fedora post 😀 ! The tutorial on ROS on fedora 25 is sadly not enough to get ROS working on Fedora 26/27. So here goes. As in all other tutorials, one needs to start by downloading every ros packages. However, opencv3 compiled with qt5 is in the repo […]

Read More ROS Kinetic on Fedora 26/27