Plures/uarray

Quansight

@quansightai_twitter

[Unknown, Quansight]
*1 new commit pushed to master*
Quansight-Labs/udiff@c8d1660 - Add coerce=True to pass tests.

Hameer Abbasi

@hameerabbasi

@sangyx I've fixed the tests. :) Last part of your work is to somehow add a mechanism so we can control what's the base of the differentiation, and what gets differentiated by what, as it was before.

Then we can publish to PyPI.

esang

@sangyx

@hameerabbasi When will the uarray be refactored?

Quansight

@quansightai_twitter

[Hameer Abbasi, Quansight] uarray doesn’t need refactoring, the uarray backend in udiff does. :)

[Hameer Abbasi, Quansight] I’ll do it this week.

esang

@sangyx

OK.

Quansight

@quansightai_twitter

[Unknown, Quansight]
*1 new commit pushed to master*
Quansight-Labs/unumpy@3a96c7a - Multimethods for functional programming routines (#65)

[Unknown, Quansight]
Pull request merged by hameerabbasi
Quansight-Labs/unumpy#65

Quansight

@quansightai_twitter

[Unknown, Quansight]
Pull request opened by peterbell10
Quansight-Labs/uarray#246
As discussed, this skips the final call directly to the default if the last backend was marked as only or coerce.
Reviewers: @hameerabbasi

Quansight

@quansightai_twitter

[Unknown, Quansight]
*2 new commits pushed to master*
Quansight-Labs/uarray@8353f03 - Skip last default call if the last backend was set as only or coerce
Quansight-Labs/uarray@d09f5fb - Merge pull request #246 from Quansight-Labs/skip-default

[Unknown, Quansight]
Pull request merged by hameerabbasi
Quansight-Labs/uarray#246

Quansight

@quansightai_twitter

[Unknown, Quansight]
*1 new commit pushed to master*
Quansight-Labs/unumpy@6edccfd - Multimethods for mathematical functions (#64)

[Unknown, Quansight]
Pull request merged by hameerabbasi
Quansight-Labs/unumpy#64

Quansight

@quansightai_twitter

[Unknown, Quansight]
Pull request opened by sangyx
Quansight-Labs/udiff#30

[Unknown, Quansight]
Pull request opened by joaosferreira
Quansight-Labs/unumpy#66
Picking up where Quansight-Labs/unumpy#61 left off by extending class overriding to other backends, more specifically, Dask and Sparse.
Reviewers: @hameerabbasi

Quansight

@quansightai_twitter

[Unknown, Quansight]
*1 new commit pushed to master*
Quansight-Labs/unumpy@e9d5d8d - Add overriden_class to other backends (#66)

[Unknown, Quansight]
Pull request merged by hameerabbasi
Quansight-Labs/unumpy#66

[Unknown, Quansight]
Pull request closed by hameerabbasi
Quansight-Labs/unumpy#61

Quansight

@quansightai_twitter

[Unknown, Quansight]
Pull request opened by joaosferreira
Quansight-Labs/unumpy#67
This PR intends to expand unumpy's coverage of indexing routines. The multimethods added so far are the following:

Generating index arrays

• indices
• ix_
• ravel_multi_index
• unravel_index
• diag_indices
• diag_indices_from
• mask_indices
• tril_indices
• tril_indices_from
• triu_indices
• triu_indices_from

Some observations:

The failing tests are in the Sparse backend. Because ones creates a COO with a fill_value of 1.0, the call to the next method fails since a zero fill_value is necessary.
I'm working on default implementations for ravel_multi_index and unravel_index. I'll push them as soon as I have a working prototype but tips on how to implement these are welcomed.
There are three classes that should be added as well since they are part of these routines, they are c_, r_ and s_.

*Note:* I'll be adding more multimethods that target these routines later.

joaosferreira

@joaosferreira

Hey everyone, hope you are doing well. Here is my last blog post. Have a nice week!

Also, not sure if the mentors can see it, but I completed my 2nd evaluation of the organization/mentors for GSoC. :)

joaosferreira

@joaosferreira

@hameerabbasi Can we move this week's meeting to tomorrow? Same time is fine by me.

Hameer Abbasi

@hameerabbasi

@joaosferreira I'm on leave tomorrow. If you're blocked, you can ask on Gitter or on the PR, or we'll have a meeting on Monday. Whichever you prefer.

joaosferreira

@joaosferreira

I'll use those channels.

Hameer Abbasi

@hameerabbasi

Great. You're good at writing up the issues either way!

sidhu1012

@sidhu1012

Hi guys

Hameer Abbasi

@hameerabbasi

Hello!

sidhu1012

@sidhu1012

Hi Hameer

sidhu1012

@sidhu1012

I want to get started with this organization , from where shall I begin

Hameer Abbasi

@hameerabbasi

@sidhu1012 Here's a blogpost that explains the basics. https://labs.quansight.org/blog/2019/07/uarray-update-api-changes-overhead-and-comparison-to-__array_function__/

sidhu1012

@sidhu1012

Okay I'll give it a read ,thanks @hameerabbasi

Quansight

@quansightai_twitter

[Unknown, Quansight]
Pull request ready for review by joaosferreira
Quansight-Labs/unumpy#67

joaosferreira

@joaosferreira

What's the best way to check if an array is of a given dtype in a default implementation, in this case a boolean array?

Quansight

@quansightai_twitter

[Peter Bell, Quansight] a.dtype.kind == 'b'?

joaosferreira

@joaosferreira

Hey everyone! Here is my latest blog post. Have a nice week! :smile:

Quansight

@quansightai_twitter

[Unknown, Quansight]
Pull request opened by hameerabbasi
Quansight-Labs/unumpy#68

Quansight

@quansightai_twitter

[Unknown, Quansight]
*1 new commit pushed to master*
Quansight-Labs/unumpy@8d1ca5c - Fix up backends and metaclass so subclasses work correctly. (#68)

[Unknown, Quansight]
Pull request merged by hameerabbasi
Quansight-Labs/unumpy#68

Quansight

@quansightai_twitter

[Unknown, Quansight]
*1 new commit pushed to master*
Quansight-Labs/udiff@f6f0407 - control the base of the differentiation (#30)

[Unknown, Quansight]
Pull request merged by hameerabbasi
Quansight-Labs/udiff#30

Quansight

@quansightai_twitter

[Unknown, Quansight]
Pull request opened by joaosferreira
Quansight-Labs/unumpy#69
As first discussed in #67 (review), this PR intends to extend class overriding to CuPy.
Reviewers: @hameerabbasi, @peterbell10

Quansight

@quansightai_twitter

[Unknown, Quansight]
*1 new commit pushed to master*
Quansight-Labs/unumpy@cc124c2 - Add overridden_class to CuPy backend (#69)

[Unknown, Quansight]
Pull request merged by hameerabbasi
Quansight-Labs/unumpy#69

Quansight

@quansightai_twitter

[Unknown, Quansight]
Pull request opened by sangyx
Quansight-Labs/udiff#31
I have added the diff function of np.stack.

When we extend the code to higher-order derivatives, the situation becomes a bit more complicated. There are some places to improve:

• In some derivative functions, the program may enter an infinite loop. For example, np.sum is used when registering the gradient of np.sum. I plan to add depth attributes to DiffArray or to() to control the order of gradients.
• The function registered in nograd_functions will register the gradient when it is called in the form of ufunc.__call__.
• Consider a cleaner way to control whether the calculation should derive.

Do you have any idea about the above questions?

Quansight

@quansightai_twitter

[Unknown, Quansight]
Pull request opened by joaosferreira
Quansight-Labs/unumpy#70
This pull request adds multimethods for statistical functions. The multimethods added are the following:

Order statistics

• percentile
• nanpercentile
• quantile
• nanquantile

Averages and variances

• median
• average
• mean
• nanmedian
• nanmean
• nanstd
• nanvar

Correlating

• corrcoef
• correlate
• cov

Histograms

• histogram
• histogram2d
• histogramdd
• bincount
• histogram_bin_edges
• digitize

A few things to discuss:

Why is dtype not being dispatched in std and var? Other multimethods that use _reduce_argreplacer also don't dispatch dtype.
The previously added min and max seem to be equal to amin and amax that this PR intends to add:

In [2]: onp.amin                                                                
Out[2]: <function numpy.amin(a, axis=None, out=None, keepdims=<no value>, initial=<no value>, where=<no value>)>

In [3]: onp.min                                                                 
Out[3]: <function numpy.amin(a, axis=None, out=None, keepdims=<no value>, initial=<no value>, where=<no value>)>

I'm looking into implementing some defaults as well. Multimethods like median that reduce array slices along an axis might be the easier ones for now. As I understand these require a for loop over the given axis, so in terms of complexity it would be O(n) where n is the length of that dimension.

joaosferreira

@joaosferreira

Hey everyone! Here is my latest blog post. Have a nice week! :)

Quansight

@quansightai_twitter

[Unknown, Quansight]
Pull request opened by peterbell10
Quansight-Labs/unumpy#71

Quansight

@quansightai_twitter

[Unknown, Quansight]
*2 new commits pushed to master*
Quansight-Labs/unumpy@76a279f - Misc fixes found by the cupy backend
Quansight-Labs/unumpy@9a42b49 - Merge pull request #71 from Quansight-Labs/cupy-fixes

[Unknown, Quansight]
Pull request merged by hameerabbasi
Quansight-Labs/unumpy#71

Where communities thrive

People

Repo info

Activity