I have been more than usually quiet here. This is not because I have been writing little, but because I have been writing much. I'm very thrilled to work with Eli Stevens and Luca Antiga on our book, Deep Learning with PyTorch.
We are rapidly approaching completion, so I am very confident that the book will be appear this spring. One part that I'm particularly exited about is Chapter 15, on deployment. With the desire to go beyond the "put it behind flask" tutorials you can find on the web, we present a simple webserver that
- Uses Sanic based on Python AsyncIO for parallelism,
- and actually implements request batching (and limiting).
Sadly, writing a "safe decoder" (say for images) has been out of scope, but do take a look at the server if you want a simple but somewhat efficient server for your PyTorch models.
We also take a tour of the JIT, C++ (with JIT and the torch::nn api) and PyTorch mobile.
I hope you'll enjoy that chapter and the entire book! The code is also linked from Manning's page for the book.
In addition to the pleasure of working with Eli and Luca who are awesome co-authors, I also enjoy working with the great team at Manning (Frances, Brian, Tiffany being the ones I work with the most, but there are quite a few people covering all aspects of going from a decent manuscript to a great book). They do an awesome job in closing the gap between the things I write and the things I would like to have written. Thanks!