autograd · Issues· 186 open
Open on GitHubLocally synced open issues (discussions stay on GitHub)
- #555
copy.deepcopy leads to incorrect gradient
Updated Aug 31, 2026 - #630
Implement better testing and against multiple NumPy versions
Updated Aug 31, 2026 - #802
Gradient of where() is sometimes erroneously NaN when the inactive case is NaN
Updated Aug 30, 2026 - #788
Gaps in NumPy API coverage: unwrapped subpackages and missing gradients
Updated Jul 8, 2026 - #644
Add Support for `numpy.convolve()`
PR welcomeUpdated Jun 16, 2026 - #623
Support numpy.mgrid
PR welcomeUpdated Jun 16, 2026 - #551
np.issubdtype(np.int32, np.integer) fails
Updated Apr 25, 2026 - #752
Convenience: cbrt and atan2.
Updated Jan 15, 2026 - #47
Numba support
enhancementUpdated Dec 15, 2025 - #743
support numpy.take
PR welcomegood first issueUpdated Nov 25, 2025 - #706
Add utility for debugging intermediate graph nodes values
enhancementPR welcomeUpdated Jun 30, 2025 - #662
Do away with `combo_check` and switch to test parametrization
enhancementPR welcomegood first issueUpdated Jun 30, 2025 - #681
Primitive called with ArrayBox object on forward pass when argument passed by keyword
Updated Feb 27, 2025 - #482
TypeError: Can't differentiate w.r.t. type <class 'int'>
bugUpdated Jan 8, 2025 - #552
np.array([object()]) fails
bugUpdated Dec 2, 2024