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

Jump to bottom

[Relay] Handle pad value coming from Tensor instead of scalar #14735

Merged
merged 2 commits into from
Apr 28, 2023
Merged

[Relay] Handle pad value coming from Tensor instead of scalar #14735

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
ad58d81
[Relay] Handle pad value coming from Tensor instead of scalar
Apr 27, 2023
cb3fb1d
Move test to tests/python/relay/test_op_level2.py
Apr 28, 2023
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
2 changes: 1 addition & 1 deletion src/relay/op/nn/pad.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -177,7 +177,7 @@ Array<te::Tensor> PadCompute(const Attrs& attrs, const Array<te::Tensor>& inputs
pad_after.push_back(pad_width[i][1]);
}
te::Tensor cast_pad_value = topi::cast(inputs[1], inputs[0]->dtype);
const PrimExpr& pad_value = cast_pad_value(Array<PrimExpr>());
const PrimExpr& pad_value = cast_pad_value(Array<PrimExpr>(inputs[1]->shape.size(), 0));
return Array<te::Tensor>{topi::pad(inputs[0], pad_before, pad_after, pad_value, "T_pad",
topi::kElementWise, param->pad_mode)};
}
Expand Down
19 changes: 19 additions & 0 deletions tests/python/relay/test_op_level2.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1444,6 +1444,25 @@ def _test_run(dtype):
_test_run("int32")


def test_pad_value_in_array():
A = relay.var("A", shape=(32, 32), dtype="int8")

# Extract pad value from an array
p0 = relay.Constant(tvm.nd.array(np.array([2], dtype="int8")))
p1 = relay.nn.pad(A, pad_value=p0, pad_width=((1, 1), (1, 1)))

func = relay.Function(relay.analysis.free_vars(p1), p1)
mod = tvm.IRModule.from_expr(func)

target = "llvm"
lib = relay.build(
mod,
tvm.target.Target(target, host=target),
runtime=relay.backend.Runtime("cpp"),
executor=relay.backend.Executor("aot", {"unpacked-api": False, "interface-api": "packed"}),
)


@tvm.testing.uses_gpu
@pytest.mark.parametrize("dtype", ["float32", "float16"])
def test_lrn(executor_kind, dtype):
Expand Down