feat: manual dtype casting removal #22700

Madjid-CH · 2023-08-28T16:15:12Z

PR Description

This PR is for this task

I started with the Paddle backend. I need some feedback on my work so I can continue with the other backends.

…elementwise.py.

…activations.py.

…date_types.py.

…general.py.

…linear_algebra.py

…manipulation.py

…random.py

…searching.py

…set.py

…sorting.py

…statistical.py

…utility.py

…module.

github-actions · 2023-08-28T16:15:30Z

Thanks for contributing to Ivy! 😊👏
Here are some of the important points from our Contributing Guidelines 📝:
1. Feel free to ignore the run_tests (1), run_tests (2), … jobs, and only look at the display_test_results job. 👀 It contains the following two sections:
- Combined Test Results: This shows the results of all the ivy tests that ran on the PR. ✔️
- New Failures Introduced: This lists the tests that are passing on main, but fail on the PR Fork. Please try to make sure that there are no such tests. 💪
2. The lint / Check formatting / check-formatting tests check for the formatting of your code. 📜 If it fails, please check the exact error message in the logs and fix the same. ⚠️🔧
3. Finally, the test-docstrings / run-docstring-tests check for the changes made in docstrings of the functions. This may be skipped, as well. 📚
Happy coding! 🎉👨‍💻

ivy-leaves · 2023-08-28T16:30:03Z

If you are working on an open task, please edit the PR description to link to the issue you've created.

For more information, please check ToDo List Issues Guide.

Thank you 🤗

…nctions.

vedpatwardhan

Hey @Madjid-CH, thanks for looking into this, just requested a couple of changes, but they also apply to all the other functions if applicable 😄

vedpatwardhan · 2023-08-29T08:14:00Z

ivy/functional/backends/paddle/activations.py

-@with_unsupported_device_and_dtypes(
-    {"2.5.1 and below": {"cpu": ("bfloat16",)}}, backend_version
+@with_supported_dtypes(
+    {"2.5.1 and below": ("float16", "uint16", "float32", "float64", "complex")},


instead of float16, float32 and float64 you could just use float, as all float dtypes are supported.
Also, I think uint16 isn't supported in paddle
Same applies to other similar instances in the paddle backend

Actually I looked into the functions source code to figure out the supported types, for example: here I think is where the supported types checks is performed and uint16 is mentioned.

Yeah it seems like they might've started supporting uint16 in the later versions, @MahmoudAshraf97 what do you think? Thanks!

ivy/functional/backends/paddle/data_type.py

…es wrapper

…g-removal # Conflicts: # ivy/functional/backends/paddle/experimental/statistical.py # ivy/functional/backends/paddle/general.py

…tions.

…nd functions.

…nctions.

Madjid-CH · 2023-09-18T10:39:10Z

Hi @vedpatwardhan and @Madjid-CH , please note that some of the manual casting found in this PR is not aimed at adding support for unsupported dtypes and hence should be left in place because it has other functionality (ensuring consistency for between backends for example), most of the manual casting to add extra support is only found in paddle backend

these 2 commits should be reverted and reviewed first: 282bffc cfcfbea

understood

Madjid-CH · 2023-09-18T10:46:23Z

Hi @vedpatwardhan and @Madjid-CH , please note that some of the manual casting found in this PR is not aimed at adding support for unsupported dtypes and hence should be left in place because it has other functionality (ensuring consistency for between backends for example), most of the manual casting to add extra support is only found in paddle backend
these 2 commits should be reverted and reviewed first: 282bffc cfcfbea

understood

@MahmoudAshraf97 can you please explain how to differentiate between a manual cast for adding support and a manual cast for backends compatibility because my approach is to test the real backend function against all available dtypes and see which dtypes behave correctly (same input same ouput dtype, not throwing error, ...)

MahmoudAshraf97 · 2023-09-18T12:58:27Z

Hi @vedpatwardhan and @Madjid-CH , please note that some of the manual casting found in this PR is not aimed at adding support for unsupported dtypes and hence should be left in place because it has other functionality (ensuring consistency for between backends for example), most of the manual casting to add extra support is only found in paddle backend
these 2 commits should be reverted and reviewed first: 282bffc cfcfbea

understood

@MahmoudAshraf97 can you please explain how to differentiate between a manual cast for adding support and a manual cast for backends compatibility because my approach is to test the real backend function against all available dtypes and see which dtypes behave correctly (same input same ouput dtype, not throwing error, ...)

there is no solid diffrentiation between the two types, but as I said the casting for compatibility is mostly found in paddle backend

vedpatwardhan · 2023-09-19T04:47:08Z

Hi @vedpatwardhan and @Madjid-CH , please note that some of the manual casting found in this PR is not aimed at adding support for unsupported dtypes and hence should be left in place because it has other functionality (ensuring consistency for between backends for example), most of the manual casting to add extra support is only found in paddle backend

these 2 commits should be reverted and reviewed first: 282bffc cfcfbea

Yeah you're right! There are cases where we apply casting to the output to ensure consistency, could you please share some examples in those commits where this is being done? Thanks @MahmoudAshraf97 😄

vedpatwardhan

Hey @Madjid-CH, could you please check through if there's any unresolved conversations in the previous reviews as well? Thanks 😄

Madjid-CH · 2023-09-19T10:46:09Z

Hey @Madjid-CH, could you please check through if there's any unresolved conversations in the previous reviews as well? Thanks 😄

sure Ved 😃

vedpatwardhan

lgtm! Let's wait for @MahmoudAshraf97 to give it another look, thanks @Madjid-CH 😄

vedpatwardhan · 2023-09-21T04:39:32Z

ivy-gardener
✅ Ivy gardener has formatted your code.
If changes are requested, don't forget to pull your fork.

MahmoudAshraf97

LGTM, you should keep the dtypes decorator as is because cpu and gpu support are different in paddle

github-actions

PR Compliance Checks

Thank you for your Pull Request! We have run several checks on this pull request in order to make sure it's suitable for merging into this project. The results are listed in the following section.

Issue Reference

In order to be considered for merging, the pull request description must refer to a specific issue number. This is described in our contributing guide and our PR template.
This check is looking for a phrase similar to: "Fixes #XYZ" or "Resolves #XYZ" where XYZ is the issue number that this PR is meant to address.

github-actions

PR Compliance Checks

Thank you for your Pull Request! We have run several checks on this pull request in order to make sure it's suitable for merging into this project. The results are listed in the following section.

Issue Reference

In order to be considered for merging, the pull request description must refer to a specific issue number. This is described in our contributing guide and our PR template.
This check is looking for a phrase similar to: "Fixes #XYZ" or "Resolves #XYZ" where XYZ is the issue number that this PR is meant to address.

Madjid-CH · 2023-09-29T16:48:47Z

LGTM, you should keep the dtypes decorator as is because cpu and gpu support are different in paddle

understood

…g-removal # Conflicts: # ivy/functional/backends/paddle/elementwise.py

github-actions

PR Compliance Checks

Thank you for your Pull Request! We have run several checks on this pull request in order to make sure it's suitable for merging into this project. The results are listed in the following section.

Issue Reference

In order to be considered for merging, the pull request description must refer to a specific issue number. This is described in our contributing guide and our PR template.
This check is looking for a phrase similar to: "Fixes #XYZ" or "Resolves #XYZ" where XYZ is the issue number that this PR is meant to address.

github-actions

PR Compliance Checks

Thank you for your Pull Request! We have run several checks on this pull request in order to make sure it's suitable for merging into this project. The results are listed in the following section.

Issue Reference

In order to be considered for merging, the pull request description must refer to a specific issue number. This is described in our contributing guide and our PR template.
This check is looking for a phrase similar to: "Fixes #XYZ" or "Resolves #XYZ" where XYZ is the issue number that this PR is meant to address.

github-actions · 2023-10-09T16:33:22Z

Thank you for this PR, here is the CI results:

This pull request does not result in any additional test failures. Congratulations!

vedpatwardhan · 2023-10-10T06:03:01Z

Hey @MahmoudAshraf97, could you please review this PR so that we can get it merged soon? Thanks 😄

…g-removal # Conflicts: # ivy/functional/backends/paddle/activations.py # ivy/functional/backends/paddle/elementwise.py

…to manual-dtype-casting-removal

github-actions

PR Compliance Checks

Thank you for your Pull Request! We have run several checks on this pull request in order to make sure it's suitable for merging into this project. The results are listed in the following section.

Issue Reference

In order to be considered for merging, the pull request description must refer to a specific issue number. This is described in our contributing guide and our PR template.
This check is looking for a phrase similar to: "Fixes #XYZ" or "Resolves #XYZ" where XYZ is the issue number that this PR is meant to address.

Madjid-CH added 16 commits August 26, 2023 14:36

feat(paddle backend)! removed manual dtype casting from functions in …

30db6a3

…elementwise.py.

feat(paddle backend)! removed manual dtype casting from functions in …

b77ad16

…activations.py.

feat(paddle backend)! removed manual dtype casting from functions in …

f40c70d

…date_types.py.

feat(paddle backend)! removed manual dtype casting from functions in …

2bf152f

…general.py.

feat(paddle backend)! removed manual dtype casting from functions in …

5e19928

…linear_algebra.py

feat(paddle backend)! removed manual dtype casting from functions in …

da88855

…manipulation.py

feat(paddle backend)! removed manual dtype casting from functions in …

aae6444

…random.py

feat(paddle backend)! removed manual dtype casting from functions in …

db791d0

…searching.py

feat(paddle backend)! removed manual dtype casting from functions in …

79da6bf

…set.py

feat(paddle backend)! removed manual dtype casting from functions in …

5dff2ac

…sorting.py

feat(paddle backend)! removed manual dtype casting from functions in …

ec7ee1b

…statistical.py

feat(paddle backend)! removed manual dtype casting from functions in …

116a406

…utility.py

fix(paddle backend): reintroduced the support for complex numbers.

58c2976

fix(paddle backend)! fixed failing pre-commit hook.

049cce3

feat(paddle backend)! removed manual dtype casting from experimental …

3225b77

…module.

merging

c0d7e08

Madjid-CH requested a review from vedpatwardhan August 28, 2023 16:17

ivy-leaves added the PaddlePaddle Backend Developing the Paddle Paddle Backend. label Aug 28, 2023

ivy-seed assigned D0m-inic Aug 28, 2023

feat(numpy backend)! removed manual dtype casting from the backend fu…

992af94

…nctions.

vedpatwardhan suggested changes Aug 29, 2023

View reviewed changes

refactor(jax backend): grouped supported types in with_supported_dtyp…

849c99c

…es wrapper

vedpatwardhan mentioned this pull request Aug 31, 2023

refactor(ivy): wrapper for dtype handling in paddle backend functions #22747

Closed

Madjid-CH added 4 commits August 31, 2023 09:52

Merge remote-tracking branch 'upstream/main' into manual-dtype-castin…

4e636c3

…g-removal # Conflicts: # ivy/functional/backends/paddle/experimental/statistical.py # ivy/functional/backends/paddle/general.py

feat(jax backend)! removed manual dtype casting from the backend func…

0622ef8

…tions.

feat(tensorflow backend)! removed manual dtype casting from the backe…

40ee0dc

…nd functions.

feat(torch backend)! removed manual dtype casting from the backend fu…

dab7274

…nctions.

Madjid-CH requested a review from vedpatwardhan August 31, 2023 10:45

Madjid-CH requested a review from vedpatwardhan September 18, 2023 10:46

vedpatwardhan suggested changes Sep 19, 2023

View reviewed changes

Madjid-CH requested a review from vedpatwardhan September 20, 2023 19:45

vedpatwardhan approved these changes Sep 21, 2023

View reviewed changes

Madjid-CH requested a review from MahmoudAshraf97 September 21, 2023 08:28

MahmoudAshraf97 suggested changes Sep 29, 2023

View reviewed changes

replaced dtypes decorator

68e2039

github-actions bot reviewed Sep 29, 2023

View reviewed changes

🤖 Lint code

88ed132

github-actions bot reviewed Sep 29, 2023

View reviewed changes

Merge remote-tracking branch 'upstream/main' into manual-dtype-castin…

0af0309

…g-removal # Conflicts: # ivy/functional/backends/paddle/elementwise.py

github-actions bot reviewed Sep 29, 2023

View reviewed changes

🤖 Lint code

6e52ff4

github-actions bot reviewed Sep 29, 2023

View reviewed changes

Madjid-CH requested a review from MahmoudAshraf97 September 29, 2023 16:56

ivy-leaves added the PyTorch Frontend Developing the PyTorch Frontend, checklist triggered by commenting add_frontend_checklist label Oct 9, 2023

Madjid Chergui added 2 commits November 3, 2023 10:24

Merge remote-tracking branch 'upstream/main' into manual-dtype-castin…

37fb5c3

…g-removal # Conflicts: # ivy/functional/backends/paddle/activations.py # ivy/functional/backends/paddle/elementwise.py

Merge remote-tracking branch 'origin/manual-dtype-casting-removal' in…

64aa4dc

…to manual-dtype-casting-removal

github-actions bot reviewed Nov 3, 2023

View reviewed changes

Madjid-CH merged commit 2def0c8 into ivy-llc:main Nov 3, 2023
51 of 95 checks passed

Madjid-CH deleted the manual-dtype-casting-removal branch November 3, 2023 09:28

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: manual dtype casting removal #22700

feat: manual dtype casting removal #22700

Madjid-CH commented Aug 28, 2023 •

edited

Loading

github-actions bot commented Aug 28, 2023

ivy-leaves commented Aug 28, 2023

vedpatwardhan left a comment

vedpatwardhan Aug 29, 2023

Madjid-CH Aug 29, 2023 •

edited

Loading

vedpatwardhan Aug 31, 2023

Madjid-CH commented Sep 18, 2023

Madjid-CH commented Sep 18, 2023

MahmoudAshraf97 commented Sep 18, 2023

vedpatwardhan commented Sep 19, 2023

vedpatwardhan left a comment

Madjid-CH commented Sep 19, 2023

vedpatwardhan left a comment

vedpatwardhan commented Sep 21, 2023 •

edited by ivy-branch

Loading

MahmoudAshraf97 left a comment

github-actions bot left a comment

github-actions bot left a comment

Madjid-CH commented Sep 29, 2023

github-actions bot left a comment

github-actions bot left a comment

github-actions bot commented Oct 9, 2023

vedpatwardhan commented Oct 10, 2023

github-actions bot left a comment

feat: manual dtype casting removal #22700

feat: manual dtype casting removal #22700

Conversation

Madjid-CH commented Aug 28, 2023 • edited Loading

PR Description

github-actions bot commented Aug 28, 2023

ivy-leaves commented Aug 28, 2023

vedpatwardhan left a comment

Choose a reason for hiding this comment

vedpatwardhan Aug 29, 2023

Choose a reason for hiding this comment

Madjid-CH Aug 29, 2023 • edited Loading

Choose a reason for hiding this comment

vedpatwardhan Aug 31, 2023

Choose a reason for hiding this comment

Madjid-CH commented Sep 18, 2023

Madjid-CH commented Sep 18, 2023

MahmoudAshraf97 commented Sep 18, 2023

vedpatwardhan commented Sep 19, 2023

vedpatwardhan left a comment

Choose a reason for hiding this comment

Madjid-CH commented Sep 19, 2023

vedpatwardhan left a comment

Choose a reason for hiding this comment

vedpatwardhan commented Sep 21, 2023 • edited by ivy-branch Loading

MahmoudAshraf97 left a comment

Choose a reason for hiding this comment

github-actions bot left a comment

Choose a reason for hiding this comment

PR Compliance Checks

Issue Reference

github-actions bot left a comment

Choose a reason for hiding this comment

PR Compliance Checks

Issue Reference

Madjid-CH commented Sep 29, 2023

github-actions bot left a comment

Choose a reason for hiding this comment

PR Compliance Checks

Issue Reference

github-actions bot left a comment

Choose a reason for hiding this comment

PR Compliance Checks

Issue Reference

github-actions bot commented Oct 9, 2023

vedpatwardhan commented Oct 10, 2023

github-actions bot left a comment

Choose a reason for hiding this comment

PR Compliance Checks

Issue Reference

Madjid-CH commented Aug 28, 2023 •

edited

Loading

Madjid-CH Aug 29, 2023 •

edited

Loading

vedpatwardhan commented Sep 21, 2023 •

edited by ivy-branch

Loading