TensorFlow

What do you like best?

The Keras interface is super cool for beginners. Once you get comfortable, there are ample option to customize the data access during training and export the model.

What do you dislike?

Its lerning curve is a little steep once you cross the beginners threshold. But can say its the worst :).

What problems are you solving with the product? What benefits have you realized?

I worked on using ML in solving classification of signal events from background events in data from Particle Physics experiments. I was able to make a new classifier that had 90%+ purity compared to the existing benchmark of 70%. The model deployment was also pretty quick with GPU back-end available. I am now working on porting things to FPGA based accelerator and the framework I am using supports TF too.. The wide documentation and user support forums are plus points too.