Skip to content
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

Implement Cast in PyTorch backend #1023

Merged
merged 1 commit into from
Oct 10, 2024
Merged

Implement Cast in PyTorch backend #1023

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
11 changes: 11 additions & 0 deletions pytensor/link/pytorch/dispatch/scalar.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,7 @@

from pytensor.link.pytorch.dispatch.basic import pytorch_funcify
from pytensor.scalar.basic import (
Cast,
ScalarOp,
)

Expand Down Expand Up @@ -38,3 +39,13 @@ def pytorch_func(*args):
)

return pytorch_func


@pytorch_funcify.register(Cast)
def pytorch_funcify_Cast(op: Cast, node, **kwargs):
dtype = getattr(torch, op.o_type.dtype)

def cast(x):
return x.to(dtype=dtype)

return cast
13 changes: 13 additions & 0 deletions tests/link/pytorch/test_elemwise.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,9 @@
from tests.link.pytorch.test_basic import compare_pytorch_and_py


torch = pytest.importorskip("torch")


def test_pytorch_Dimshuffle():
a_pt = matrix("a")

Expand Down Expand Up @@ -143,3 +146,13 @@ def test_softmax_grad(axis):
out = SoftmaxGrad(axis=axis)(dy, sm)
fgraph = FunctionGraph([dy, sm], [out])
compare_pytorch_and_py(fgraph, [dy_value, sm_value])


def test_cast():
x = matrix("x", dtype="float32")
out = pt.cast(x, "int32")
fgraph = FunctionGraph([x], [out])
_, [res] = compare_pytorch_and_py(
fgraph, [np.arange(6, dtype="float32").reshape(2, 3)]
)
assert res.dtype == torch.int32
Loading