Skip to content

Updated padding validation in max_pool functions to account for dilation #154395

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

Open
wants to merge 4 commits into
base: main
Choose a base branch
from
+53 −16
Open

Updated padding validation in max_pool functions to account for dilation #154395

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
db48618
Updated max_pool2d and max_pool3d padding validation to account for d…
narekmalk May 26, 2025
ff5c35f
Minor fix
narekmalk May 27, 2025
6787f6a
Fixed linter error
narekmalk May 28, 2025
caca0c6
Updated max_pool1d padding validation
narekmalk May 28, 2025
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
10 changes: 6 additions & 4 deletions aten/src/ATen/native/MaxPooling.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -40,6 +40,8 @@ inline void check_max_pool1d(
stride = kernel_size;
}

int effective_kernel_size = (kernel_size[0] - 1) * dilation[0] + 1;

TORCH_CHECK(
kernel_size[0] > 0,
"max_pool1d() kernel_size must be greater than zero, but got ",
Expand All @@ -49,11 +51,11 @@ inline void check_max_pool1d(
TORCH_CHECK(
padding[0] >= 0, "max_pool1d() padding must be non-negative, but got ", padding[0]);
TORCH_CHECK(
padding[0] <= kernel_size[0] / 2,
"max_pool1d() padding should be at most half of kernel size, but got padding=",
padding[0] <= effective_kernel_size / 2,
"max_pool1d() padding should be at most half of effective kernel size, but got padding=",
padding[0],
" and kernel_size=",
kernel_size[0]);
" and effective_kernel_size=",
effective_kernel_size);
TORCH_CHECK(
dilation[0] > 0, "max_pool1d() dilation must be greater than zero, but got ", dilation[0]);

Expand Down
39 changes: 33 additions & 6 deletions aten/src/ATen/native/Pool.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -153,9 +153,20 @@ pool2d_shape_check(
input.sizes());
}

TORCH_CHECK(kW/2 >= padW && kH/2 >= padH,
"pad should be smaller than or equal to half of kernel size, but got ",
"padW = ", padW, ", padH = ", padH, ", kW = ", kW, ", kH = ", kH);
int effectiveKW = (kW - 1) * dilationW + 1;
int effectiveKH = (kH - 1) * dilationH + 1;

TORCH_CHECK(
effectiveKW / 2 >= padW && effectiveKH / 2 >= padH,
"pad should be smaller than or equal to half of effective kernel size, but got ",
"padW = ",
padW,
", padH = ",
padH,
", effectiveKW = ",
effectiveKW,
", effectiveKH = ",
effectiveKH);

TORCH_CHECK(outputWidth >= 1 && outputHeight >= 1,
"Given input size: (",
Expand Down Expand Up @@ -276,9 +287,25 @@ pool3d_shape_check(
"kernel size ", "(kT: ", kT, " kH: ", kH, " kW: ", kW, ")");
}

TORCH_CHECK(kT/2 >= pT && kW/2 >= pW && kH/2 >= pH,
"pad should be smaller than or equal to half of kernel size, but got "
"kT: ", kT, " kW: ", kW, " kH: ", kH, " padT: ", pT, " padW: ", pW, " padH: ", pH);
int effectiveKT = (kT - 1) * dilationT + 1;
int effectiveKW = (kW - 1) * dilationW + 1;
int effectiveKH = (kH - 1) * dilationH + 1;

TORCH_CHECK(
effectiveKT / 2 >= pT && effectiveKW / 2 >= pW && effectiveKH / 2 >= pH,
"pad should be smaller than or equal to half of effective kernel size, but got "
"padT: ",
pT,
" padW: ",
pW,
" padH: ",
pH,
" effectiveKT: ",
effectiveKT,
" effectiveKW: ",
effectiveKW,
" effectiveKH: ",
effectiveKH);

TORCH_CHECK(otime >= 1 && owidth >= 1 && oheight >= 1,
"Given input size: (",
Expand Down
20 changes: 14 additions & 6 deletions torch/_meta_registrations.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4451,10 +4451,13 @@ def pool2d_shape_check(
lambda: f"Expected 3D or 4D (batch mode) tensor with optional 0 dim batch size for input, but got: {input.size()}",
)

effectiveKW = (kW - 1) * dilationW + 1
effectiveKH = (kH - 1) * dilationH + 1

torch._check(
kW // 2 >= padW and kH // 2 >= padH,
lambda: "pad should be smaller than or equal to half of kernel size, but got "
f"padW = {padW}, padH = {padH}, kW = {kW}, kH = {kH}",
effectiveKW // 2 >= padW and effectiveKH // 2 >= padH,
lambda: "pad should be smaller than or equal to half of effective kernel size, but got "
f"padW = {padW}, padH = {padH}, effectiveKW = {effectiveKW}, effectiveKH = {effectiveKH}",
)

torch._check(
Expand Down Expand Up @@ -4539,11 +4542,16 @@ def pool3d_shape_check(
),
)

effectiveKT = (kT - 1) * dilationT + 1
effectiveKW = (kW - 1) * dilationW + 1
effectiveKH = (kH - 1) * dilationH + 1

torch._check(
kT / 2 >= pT and kW / 2 >= pW and kH / 2 >= pH,
effectiveKT / 2 >= pT and effectiveKW / 2 >= pW and effectiveKH / 2 >= pH,
lambda: (
f"pad should be smaller than or equal to half of kernel size, but got "
f"kT: {kT} kW: {kW} kH: {kH} padT: {pT} padW: {pW} padH: {pH}"
f"pad should be smaller than or equal to half of effective kernel size, but got "
f"padT: {pT}, padW: {pW}, padH: {pH}, "
f"effectiveKT: {effectiveKT}, effectiveKW: {effectiveKW}, effectiveKH: {effectiveKH}"
),
)

Expand Down