Tag: python

Do you run long, time-demanding calculations in Jupyter Notebook?

With `jupyter-notify` extension, you will get a browser notification when your calculation is done (=when the cell is completed)https://t.co/2xjocGl2S4

Example: 👇 pic.twitter.com/uugYC26qCk

— Daily Python Tip (@python_tip) December 1, 2019

jaw. drop. https://t.co/wK97u9PqoA

— Delip Rao (@deliprao) November 29, 2019

autohooks: Git Hooks With Python https://t.co/kUCGQZcQDQ

— PyCoder’s Weekly (@pycoders) November 26, 2019

Have a huuuuuge dataset and want to read in a random fraction of the data with pandas saving enormous amounts of memory?

Set skiprows to a function that tests a random value against some constant. The below function skips 99% of rows and keeps the first for the column header. pic.twitter.com/tKKR2dyMqy

— Ted Petrou (@TedPetrou) November 22, 2019

The scikit-learn TransformedTargetRegressor can make a fantastic addition to your workflow. Instead of transforming your target before fitting and then after prediction, it automates the process for you.

Here's an example from the Kaggle Housing competition pic.twitter.com/GoHHDth2Tg

— Ted Petrou (@TedPetrou) November 21, 2019

You can visualize multiple dimensions thanks to the MultiIndex, but need to use pivot_table as pivot isn't able to handle multiple columns.

Here is all of the code and 4 hyperparameters pic.twitter.com/IgueKEWC7k

— Ted Petrou (@TedPetrou) November 17, 2019

`numpy.split(ary, indices_or_sections, axis=0)` vs. `torch.split(tensor, split_size_or_sections, dim=0)`; indices or sizes for each chunk, fight! (more seriously though still finding it hard to keep track of which var is tensor/array and remembering the random api differences 🧐)

— Andrej Karpathy (@karpathy) November 6, 2019

TensorTrade - An open source reinforcement learning framework for training, evaluating, and deploying robust trading agents. https://t.co/40Xpd4OC7X #python pic.twitter.com/Bufkg119NS

— Python Weekly (@PythonWeekly) October 29, 2019

pyhttptest - A command-line tool for HTTP tests over RESTful APIs. https://t.co/g9I6JXTbAl #python #rest #api

— Python Weekly (@PythonWeekly) October 28, 2019

Number 5, `import *`, is only true if the package author hasn't defined `__all__`. Packages designed for use in a REPL should endeavor to always define this carefully - e.g. every fastai module does.

If you're using such a package in a REPL, `import *` can save a lot of bother!

— Jeremy Howard (@jeremyphoward) October 22, 2019

Another trick:
```
def flatten (x): return list(itertools.chain.from_iterable(x))
```
NB: the `sum` approach to this is *very* slow on long lists! :O

— Jeremy Howard (@jeremyphoward) October 22, 2019

Here's a little trick:
```
import operator
product = reduce(operator.mul, nums)
```

— Jeremy Howard (@jeremyphoward) October 22, 2019