Is really a list comprehension faster than map in #Python? Nope. You just proved that Python is very slow at creating functions (the lambda in the map example which gets created from scratch at every loop) and objects
This is just to let you know that panthera.numpy (https://github.com/alanmarazzi/panthera/blob/master/src/panthera/numpy.clj) is actually a generic DSL for accessing and using EVERY #Python package (yes, even #TensorFlow on GPU) from #Clojure
Don't hesitate to ask me anything about it
I can't help it, after using the #Clojure REPL for so long I can't settle for almost anything else. Anytime I have to touch #Python the only way to have something similar is a Jupyter notebook, but then I hate notebooks...
Lucky enough now we can have the best of both worlds: Clojure expressivity & REPL + Python's boatload of machine learning libs! https://github.com/cnuernber/libpython-clj
I like #Python but lately there are many things going "wrong", pipenv is a very good example https://chriswarrick.com/blog/2018/07/17/pipenv-promises-a-lot-delivers-very-little
I build data products, some call me data scientist, some data engineer, some software developer.
This Mastodon instance is for people interested in technology. Discussions aren't limited to technology, because tech folks shouldn't be limited to technology either!