🐛 Bug

To Reproduce

Run following from jupyter lab console

import torch
foo = torch.arange(5)
foo.as_strided((5,), (-1,), storage_offset=4)

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/daniil/.local/lib/python3.6/site-packages/torch/tensor.py", line 159, i

I was having a very hard time figuring out

fill = A.stack().mean()
A.add(B, fill_value=fill)

fill = 4.5. However I computed a value of 3.2 because I was taking the mean from the column of A not the DataFrame A.
This coming after the Indexing chapter where "explicit is better than implicit." I was thinking that this should be a little more explicit.

"Bokeh is a Python interactive visualization library that targets modern web browsers for presentation. Its goal is to provide elegant, concise construction of novel graphics in the style of D3.js, but also deliver this capability with high-performance interactivity over very large or streaming datasets. Bokeh can help anyone who would like to quickly and easi

This might be less of an implementation question, and more of a "philosophy" question, but shouldn't the Frobenius Norm work on Vectors? Source: Wolfram

Currently, the Frobenius Norm in numpy does not accept vectors:

import numpy as np
a = np.random.rand(10, 1)
b = np.squeeze(a)
print(np.linalg.norm(a, 'fro'))
print(np.linalg.

Hi, I'm trying to visualise the graph involved when xarray writes to a netcdf file, but no visualisation is appearing.

import dask
import xarray

ds = xarray.Dataset({'a': 0})
task = ds.to_netcdf('dat.nc',compute=False)
task.compute()
print(task)
task.visualise(filename='netcdf.svg')  # This file is never produced


def inc(x):
    return x + 1

z = dask.delayed(inc)(1)

Sphinx (2.2.1 or master) produces the following two kinds of warnings in my environment.

duplicate object description

I think cross refs to such object is ambiguous.

autosummary: stub file not found

There are

• chainer.dataset.Converter base class and
• chainer.dataset.converter decorator.

Therefore the filesystem has to allow to store `chainer.dataset.Conver

Hyperlink of the edit on Github button on the top right corner of this page https://docs-cupy.chainer.org/en/stable/reference/generated/cupy.ndarray.html#cupy.ndarray.nonzero seems to be broken.

right now it is linked to https://github.com/cupy/cupy/blob/107c7d2899fa34af8ea3a80f79edaf950709b764/docs/source/reference/generated/cupy.ndarray.rst

Since Trax is a successor of tensor2tensor (according to the release notes of tensor2tensor v1.15.0), it would be helpful if you could provide examples for more advanced machine learning tasks. An outstanding feature of tensor2tensor are the numerous (and useful) examples which Trax is currently lacking. Such examples would especi

This is working and should be documented in the install instructions.

mkdir build
cd build
cmake -G "MinGW Makefiles" -DCMAKE_INSTALL_PREFIX=path_to_prefix ..
mingw32-make
mingw32-make install

As listed in https://docs.scipy.org/doc/scipy/reference/special.html#bessel-functions, we can implement the universal functions in the list first.

MCVE Code Sample

# Your code here
import numpy as np
import xarray as xr

data = np.zeros((10, 4))
example_xr = xr.DataArray(data, coords=[range(10), ["

The random_walk function is not a direct equivalent to the RandomWalker class method. A strict equivalent would be this:

def random_walk(n):
    position = 0
    for i in range(n):
        yield position
        position += 2*random.randint(0,1)-1

It is still not much faster, but it's a more fair comparison.

Change the default value for quantiles in this function to quantiles=None. When using bins instead of quantiles with the utils.get_clean_factor_and_forward_returns() function, there is an error in the utils.quantize_factor() inferring that no bin size or quantile sizes were passed but it is actually because utils.quantize_factor() passes quantiles = 5 by default, regardless of if bins in not None.