Please make sure that this is a bug. As per our
GitHub Policy,
we only address code/doc bugs, performance issues, feature requests and
build/installation issues on GitHub. tag:bug_template

System information

• Have I written custom code (as opposed to using a stock
  example script provided in TensorFlow): Yes

(e.g. for links and images), because some of these examples are now being rendered in the docs.

Added by @fchollet in requests for contributions.

Describe the issue linked to the documentation

https://scikit-learn.org/stable/developers/contributing.html#guidelines-for-writing-documentation specifies how to write parameters with a shape attribute in a docstring, e.g.

array_parameter : {array-like, sparse matrix, dataframe} of shape (n_samples, n_features) or (n_samples,)
    This parameter accepts data in either of the mentio

WARNING: /data/users/ezyang/pytorch-tmp/BUILD.bazel:580:1: in hdrs attribute of cc_library rule //:aten_headers: Artifact 
'bazel-out/k8-fastbuild/bin/aten/src/ATen/Config.h' is duplicated (through '//:aten/src/ATen/Config.h' and '//:aten_src_AT
en_config'). Since this rule was created by the macro 'cc_library', the error might have been caused by the macro implemen
tation                 

I just noticed the Random stdlib has an UnsafeView type for filling arrays. With the new layout optimizations, this should no longer be necessary. We should try using the normal SubArray type for this. If it's still too slow/complex, then at least the type can be made safe by holding the underlying array directly instead of a Ptr.