Here's the nbviewer link, since github has so much trouble rendering notebooks (including this one!)https://t.co/lSxdNtsQb7
— Jeremy Howard (@jeremyphoward) October 22, 2019
Here's the nbviewer link, since github has so much trouble rendering notebooks (including this one!)https://t.co/lSxdNtsQb7
— Jeremy Howard (@jeremyphoward) October 22, 2019
I made a notebook with examples of cool Python features that either took me a long time to find out or were too intimidating for me to use.
— Chip Huyen (@chipro) October 22, 2019
I especially focus on the features I find useful for machine learning.https://t.co/7LBmu4UuwS
🐍📰 What does Python 3.8 bring to the table? Learn about some of the biggest changes and see you how you can best make use of them https://t.co/8NLDNwNMDW
— PyCoder’s Weekly (@pycoders) October 19, 2019
Ouch - numpy's multithreading made my DataLoader performance 100x slower! Turning it off is not exactly straightforward...
— Jeremy Howard (@jeremyphoward) October 18, 2019
I'm adding this function to fastai so no-one else has to go through this pain! :) pic.twitter.com/QTtuhCOxG5
Mypy 0.740 is out, with type checking of str.format calls and improved checking of unannotated functions. https://t.co/XMzAuE5bBv
— The Mypy Project (@mypyproject) October 17, 2019
Turns out opencv by default uses RLE compression only, and that's the reason for the difference.
— Jeremy Howard (@jeremyphoward) October 16, 2019
PIL doesn't seem to support that (I couldn't see it it in docs, anyway), so you can't get quite the same mix of speed and size as opencv
h/t @eiopa for putting me on the right track https://t.co/WYq9HXKN66
I just discovered that PIL takes 8x longer to save a 16 bit single channel PNG than opencv does. Ouch! Anyone else noticed this before, or know why this is? (I'm only using one opencv thread).
— Jeremy Howard (@jeremyphoward) October 16, 2019
VS Code Adds Native Editing of Jupyter Notebooks https://t.co/L6vsu3s3EW
— PyCoder’s Weekly (@pycoders) October 16, 2019
PyWaffle: Make Waffle Charts in Python https://t.co/8WczIQWuj5
— PyCoder’s Weekly (@pycoders) October 14, 2019
Love R? Watch "Interfacing R and Python" from PyData Berlin 2018 #rstats #pythonhttps://t.co/8XtsOIThEM
— PyData (@PyData) October 14, 2019
Both #rstats and #python used in the GitHub repo for this important analysis #ddj https://t.co/BkwMKebMcc https://t.co/GEzZ9fnWUX
— Sharon Machlis (@sharon000) October 13, 2019
e.g. don't write methods like
— Jake VanderPlas (@jakevdp) October 13, 2019
def load_data(filename):
data = json.load(filename)
# ...
This means *you* are responsible to support every imaginable filetype/database/etc. Instead write methods like
def load_data(data_dict):
# ...
Let your users use the language.