Skip to content

Revert uninit'd allocs in Numpy.reduce_max introduced in #632. #636

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
svlandeg merged 4 commits into from
Apr 21, 2022
Merged

Revert uninit'd allocs in Numpy.reduce_max introduced in #632. #636

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
bd75c9e
Merge pull request #1 from explosion/develop
shadeMe Apr 8, 2022
25f3771
Merge branch 'develop' of github.com:explosion/thinc into develop
shadeMe Apr 14, 2022
2434fb1
`NumpyOps`: Revert uninit'd alloc in `reduce_max`
shadeMe Apr 19, 2022
11906bc
Test `which` in `test_reduce_max`
shadeMe Apr 19, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion thinc/backends/numpy_ops.pyx
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -396,7 +396,8 @@ class NumpyOps(Ops):
assert O != 0

cdef np.ndarray maxes
cdef np.ndarray which = self.alloc(shape=(B, O), dtype="i", zeros=False)
# Needs to be zero-initialized as we start by assuming that the first element is the max value.
cdef np.ndarray which = self.alloc(shape=(B, O), dtype="i", zeros=True)
if reals2d_ft is float2d_t:
maxes = self.alloc(shape=(B, O), dtype="float32", zeros=False)
cpu_reduce_max(<float*>maxes.data, <int*>which.data, &X[0, 0], &lengths[0], B, T, O)
Expand Down
6 changes: 6 additions & 0 deletions thinc/tests/backends/test_ops.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -763,6 +763,9 @@ def test_reduce_max_sm(ops, dtype):
lengths = ops.xp.array([2, 2, 2], dtype="i")
maxes, which = ops.reduce_max(X, lengths)
assert maxes.dtype == dtype
assert ops.xp.all(which >= 0)
assert ops.xp.all(which < X.shape[0])

start = 0
for i, length in enumerate(lengths):
truth = X[start : start + length].max(axis=0)
Expand All @@ -781,6 +784,9 @@ def test_reduce_max(ops, dtype):
# m[1, 3] = 3
maxes, which = ops.reduce_max(m, lengths)
assert maxes.dtype == dtype
assert ops.xp.all(which >= 0)
assert ops.xp.all(which < m.shape[0])

start = 0
for i, length in enumerate(lengths):
truth = m[start : start + length].max(axis=0)
Expand Down