feat: gradient numpy frontend #23104

sineha17 · 2023-09-05T22:20:00Z

added gradient function to numpy frontend

out of 5 test cases, 4 pass successfully. paddle backend function expects a tensor attribute but the numpy gradient function uses an array. i'm unable to convert the array to a tensor element for the paddle test case without affecting the input datatype of the other test cases.

please provide some insight on how to fix this error.

close #23053

Checklist

Did you add a function?
Did you add the tests?
Did you follow the steps we provided?

Socials:

github-actions · 2023-09-05T22:20:18Z

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 · 2023-09-05T22:30:02Z

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 🤗

reformatting

MahmoudAshraf97 · 2023-09-08T21:45:52Z

ivy_tests/test_ivy/test_frontends/test_numpy/test_mathematical_functions/test_miscellaneous.py

+@handle_frontend_test(
+    fn_tree="numpy.gradient",
+    dtype_input_axis=helpers.dtype_values_axis(
+        available_dtypes=("float32", "float16", "float64"),


the available dtypes should be a strategy that gets all valid dtypes, you should then set the supported dtypes for the function using one of the decorators specified here https://unify.ai/docs/ivy/overview/deep_dive/data_types.html#supported-and-unsupported-data-types

MahmoudAshraf97 · 2023-09-08T21:46:22Z

ivy_tests/test_ivy/test_frontends/test_numpy/test_mathematical_functions/test_miscellaneous.py

please add support for edge_order argument in testing

MahmoudAshraf97 · 2023-09-08T21:48:50Z

ivy_tests/test_ivy/test_frontends/test_numpy/test_mathematical_functions/test_miscellaneous.py

+    varargs=helpers.ints(
+        min_value=-3,
+        max_value=3,
+    ),


the varargs argument can be one of the following as per the function documentation:

single scalar to specify a sample distance for all dimensions.

N scalars to specify a constant sample distance for each dimension.
i.e. dx, dy, dz, ...

N arrays to specify the coordinates of the values along each
dimension of F. The length of the array must match the size of
the corresponding dimension

Any combination of N scalars/arrays with the meaning of 2. and 3.

so please extend the testing strategy to accommodate these cases

okay, i'm working on it. thank you

sineha17 · 2023-09-14T14:09:09Z

On adding edge_order argument in testing, an exception is raised stating edge_order should not be greater than axis.shape. For example, if axis=0 and edge_order=1 (default value), there is no error produced but in case of axis=0 and edge_order=2, the said exception is raised. For multidimentional arrays the shape of axis will be greater than 1 and in this case edge_order=2 does not produce any error.
I do not know how to restrict the value of edge_order to be less than axis size in the frontend. So far, I've noticed that not adding the edge_order argument in testing and passing only the default value (edge_order=1) is not throwing any error.

sineha17 · 2023-09-14T14:22:58Z

in case of paddle backend, the following error occurs.

ivy.utils.exceptions.IvyValueError: paddle: gradient: paddle: prod: (InvalidArgument) prod(): argument 'x' (position 0) must be Tensor, but got Array (at /paddle/paddle/fluid/pybind/eager_utils.cc:1005)

I have tried adding function decorators which convert numpy array to tensor, also tried importing ivy.paddle/ivy.torch in test file and calling the funtion paddle.tensor/torch.to_tensor inside an if block when backend=='paddle' and still could not resolve this error.

MahmoudAshraf97 · 2023-09-15T08:43:04Z

On adding edge_order argument in testing, an exception is raised stating edge_order should not be greater than axis.shape. For example, if axis=0 and edge_order=1 (default value), there is no error produced but in case of axis=0 and edge_order=2, the said exception is raised. For multidimentional arrays the shape of axis will be greater than 1 and in this case edge_order=2 does not produce any error. I do not know how to restrict the value of edge_order to be less than axis size in the frontend. So far, I've noticed that not adding the edge_order argument in testing and passing only the default value (edge_order=1) is not throwing any error.

to fix this, you might need to write a more complex testing strategy that samples edge_order based on the shape and the axis argument.

in case of paddle backend, the following error occurs.

ivy.utils.exceptions.IvyValueError: paddle: gradient: paddle: prod: (InvalidArgument) prod(): argument 'x' (position 0) must be Tensor, but got Array (at /paddle/paddle/fluid/pybind/eager_utils.cc:1005)

I have tried adding function decorators which convert numpy array to tensor, also tried importing ivy.paddle/ivy.torch in test file and calling the funtion paddle.tensor/torch.to_tensor inside an if block when backend=='paddle' and still could not resolve this error.

can you please share the entire error trace?

sineha17 · 2023-09-19T15:26:05Z

ivy_tests/test_ivy/test_frontends/test_numpy/test_mathematical_functions/test_miscellaneous.py:375: in test_numpy_gradient
    fn_tree="numpy.gradient",
ivy_tests/test_ivy/test_frontends/test_numpy/test_mathematical_functions/test_miscellaneous.py:404: in test_numpy_gradient
    helpers.test_frontend_function(
ivy_tests/test_ivy/helpers/function_testing.py:747: in test_frontend_function
    ret = get_frontend_ret(
ivy_tests/test_ivy/helpers/function_testing.py:2409: in get_frontend_ret
    with ivy_backend.PreciseMode(precision_mode):
ivy/functional/ivy/general.py:86: in __exit__
    raise exc_val
ivy_tests/test_ivy/helpers/function_testing.py:2410: in get_frontend_ret
    ret = frontend_fn(*args, **kwargs)
ivy/functional/frontends/numpy/func_wrapper.py:502: in _outputs_to_frontend_arrays
    ret = fn(*args, **kwargs)
ivy/functional/frontends/numpy/func_wrapper.py:454: in _inputs_to_ivy_arrays_np
    return fn(*ivy_args, **ivy_kwargs)
ivy/functional/frontends/numpy/mathematical_functions/miscellaneous.py:340: in gradient
    return ivy.gradient(f, spacing=varargs, axis=axis, edge_order=edge_order)
ivy/utils/exceptions.py:362: in _handle_exceptions
    _handle_exceptions_helper(e, type(e))
ivy/utils/exceptions.py:371: in _handle_exceptions_helper
    raise cls(fn.__name__, str(e), include_backend=True)
E   ivy.utils.exceptions.IvyValueError: paddle: gradient: paddle: prod: (InvalidArgument) prod(): argument 'x' (position 0) must be Tensor, but got Array (at /paddle/paddle/fluid/pybind/eager_utils.cc:1005)
E   
E   Falsifying example: test_numpy_gradient(
E       frontend='numpy',
E       backend_fw='paddle',
E       on_device='cpu',
E       dtype_input_axis=(['float64'], [array([-1., -1.])], 0),
E       varargs=0,
E       fn_tree='ivy.functional.frontends.numpy.gradient',
E       test_flags=FrontendFunctionTestFlags(
E           num_positional_args=0,
E           with_out=False,
E           inplace=False,
E           as_variable=[False],
E           native_arrays=[False],
E           test_compile=False,
E           generate_frontend_arrays=False,
E           transpile=False,
E           precision_mode=False,
E       ),
E   )```

MahmoudAshraf97 · 2023-10-01T21:27:00Z

The problem is solved now, please sync your branch with main and try again

sineha17 · 2023-10-04T15:41:06Z

The problem is solved now, please sync your branch with main and try again

Will do, thank you for the update.

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.

Protected Branch

In order to be considered for merging, the pull request changes must not be implemented on the "main" branch. This is described in our Contributing Guide. We are closing this pull request and we would suggest that you implement your changes as described in our Contributing Guide and open a new pull request.

github-actions · 2023-10-04T16:43:01Z

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

Failed tests:

This PR introduces the following new failing tests:
ivy_tests/test_ivy/test_functional/test_experimental/test_core/test_elementwise.py::test_diff,paddle
ivy_tests/test_ivy/test_frontends/test_tensorflow/test_math.py::test_tensorflow_scalar_mul,numpy
ivy_tests/test_ivy/test_frontends/test_tensorflow/test_math.py::test_tensorflow_scalar_mul,jax
ivy_tests/test_ivy/test_frontends/test_tensorflow/test_math.py::test_tensorflow_scalar_mul,tensorflow
ivy_tests/test_ivy/test_frontends/test_tensorflow/test_math.py::test_tensorflow_scalar_mul,torch
ivy_tests/test_ivy/test_frontends/test_torch/test_tensor.py::test_torch_tensor_neg,numpy
ivy_tests/test_ivy/test_frontends/test_torch/test_tensor.py::test_torch_tensor_neg,jax
ivy_tests/test_ivy/test_frontends/test_torch/test_tensor.py::test_torch_tensor_neg,torch
ivy_tests/test_ivy/test_frontends/test_torch/test_tensor.py::test_torch_tensor_neg,paddle
ivy_tests/test_ivy/test_frontends/test_torch/test_tensor.py::test_torch_tensor_neg,tensorflow

MahmoudAshraf97 · 2023-10-30T10:00:22Z

Hi, are you still working on this?

sineha17 · 2023-10-31T13:54:55Z

Hi, are you still working on this?

in my local, all the test cases are passing. I've committed all the changes here.

Ishticode

@sineha17
Thanks for the PR and following through with the instruction above. I've tested locally and tests are passing. Apologies for the delays in merging on our end. Thank you very much.

sineha17 added 2 commits September 6, 2023 03:01

added gradient function to numpy

d2c25bc

added gradient function to numpy

1b6eae3

ivy-leaves added the NumPy Frontend Developing the NumPy Frontend, checklist triggered by commenting add_frontend_checklist label Sep 5, 2023

sineha17 added 7 commits September 6, 2023 04:02

Update miscellaneous.py

db642e6

Update test_miscellaneous.py

59e3052

Update miscellaneous.py

37de128

Update test_miscellaneous.py

5315a5b

Update miscellaneous.py

e838e71

Update test_miscellaneous.py

14e8ec8

reformatting

Update test_miscellaneous.py

4680bdd

ivy-seed assigned MahmoudAshraf97 Sep 5, 2023

MahmoudAshraf97 suggested changes Sep 8, 2023

View reviewed changes

sineha17 added 2 commits September 10, 2023 19:00

Merge branch 'unifyai:main' into main

c774b5e

added unsupported dtypes

daa8949

sineha17 requested a review from MahmoudAshraf97 September 14, 2023 13:52

Update available dtypes

574bb07

MahmoudAshraf97 removed their request for review October 1, 2023 21:26

Merge branch 'unifyai:main' into main

97a7123

github-actions bot reviewed Oct 4, 2023

View reviewed changes

sineha17 changed the title ~~gradient numpy frontend~~ feat: gradient numpy frontend Oct 4, 2023

MahmoudAshraf97 closed this Oct 30, 2023

MahmoudAshraf97 reopened this Oct 30, 2023

ivy-leaves unassigned MahmoudAshraf97 Dec 23, 2023

ivy-seed assigned niixxaaa Dec 23, 2023

Ishticode assigned Ishticode and unassigned niixxaaa Mar 3, 2024

Ishticode approved these changes Mar 3, 2024

View reviewed changes

Ishticode merged commit 3d8c5e7 into ivy-llc:main Mar 3, 2024
240 of 282 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: gradient numpy frontend #23104

feat: gradient numpy frontend #23104

sineha17 commented Sep 5, 2023 •

edited

Loading

github-actions bot commented Sep 5, 2023

ivy-leaves commented Sep 5, 2023

MahmoudAshraf97 Sep 8, 2023

MahmoudAshraf97 Sep 8, 2023

MahmoudAshraf97 Sep 8, 2023

sineha17 Sep 10, 2023

sineha17 commented Sep 14, 2023

sineha17 commented Sep 14, 2023

MahmoudAshraf97 commented Sep 15, 2023

sineha17 commented Sep 19, 2023 •

edited by MahmoudAshraf97

Loading

MahmoudAshraf97 commented Oct 1, 2023

sineha17 commented Oct 4, 2023

github-actions bot left a comment •

edited

Loading

github-actions bot commented Oct 4, 2023

MahmoudAshraf97 commented Oct 30, 2023

sineha17 commented Oct 31, 2023

Ishticode left a comment

feat: gradient numpy frontend #23104

feat: gradient numpy frontend #23104

Conversation

sineha17 commented Sep 5, 2023 • edited Loading

added gradient function to numpy frontend

out of 5 test cases, 4 pass successfully. paddle backend function expects a tensor attribute but the numpy gradient function uses an array. i'm unable to convert the array to a tensor element for the paddle test case without affecting the input datatype of the other test cases.

Checklist

Socials:

github-actions bot commented Sep 5, 2023

ivy-leaves commented Sep 5, 2023

MahmoudAshraf97 Sep 8, 2023

Choose a reason for hiding this comment

MahmoudAshraf97 Sep 8, 2023

Choose a reason for hiding this comment

MahmoudAshraf97 Sep 8, 2023

Choose a reason for hiding this comment

sineha17 Sep 10, 2023

Choose a reason for hiding this comment

sineha17 commented Sep 14, 2023

sineha17 commented Sep 14, 2023

MahmoudAshraf97 commented Sep 15, 2023

sineha17 commented Sep 19, 2023 • edited by MahmoudAshraf97 Loading

MahmoudAshraf97 commented Oct 1, 2023

sineha17 commented Oct 4, 2023

github-actions bot left a comment • edited Loading

Choose a reason for hiding this comment

PR Compliance Checks

Protected Branch

github-actions bot commented Oct 4, 2023

Failed tests:

MahmoudAshraf97 commented Oct 30, 2023

sineha17 commented Oct 31, 2023

Ishticode left a comment

Choose a reason for hiding this comment

sineha17 commented Sep 5, 2023 •

edited

Loading

sineha17 commented Sep 19, 2023 •

edited by MahmoudAshraf97

Loading

github-actions bot left a comment •

edited

Loading