Skip to content

fix: get_padded_shape_tensors can now handle dynamic pads #3123

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
merged 2 commits into from
Sep 3, 2024
Merged

fix: get_padded_shape_tensors can now handle dynamic pads #3123

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
06afd81
fix: get_padded_shape_tensors can now handle dynamic pads
Aug 28, 2024
17960d0
fix: use `padded_shape_tensor.dtype` instead of hard-coded `np.int32`
Aug 30, 2024
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
33 changes: 22 additions & 11 deletions py/torch_tensorrt/dynamo/conversion/impl/pad.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,7 +26,7 @@ def get_padded_shape_tensors(
source_ir: Optional[SourceIR],
name: str,
input: TRTTensor,
pad: Sequence[int],
pad: Sequence[Union[int, TRTTensor]],
) -> TRTTensor:
rank = len(input.shape)
if len(pad) // 2 > rank:
Expand All @@ -47,11 +47,11 @@ def get_padded_shape_tensors(
start_list = [0] * rank
for i in range(len(pad) // 2):
dim_index = rank - (i + 1)
pad_before = pad[i * 2]
pad_after = pad[i * 2 + 1]
pad_before = get_trt_tensor(ctx, pad[i * 2], f"{name}_pad_before_{i}")
pad_after = get_trt_tensor(ctx, pad[i * 2 + 1], f"{name}_pad_after_{i}")

pad_sum = get_trt_tensor(
ctx, pad_before + pad_after, f"{name}_pad_sum_{i}", dtype=np.int32
pad_sum = impl.elementwise.add(
ctx, target, source_ir, f"{name}_pad_sum_{i}", pad_before, pad_after
)
dim_shape = ctx.net.add_slice(
input_shape_tensor,
Expand All @@ -63,7 +63,9 @@ def get_padded_shape_tensors(
new_dim_shape = impl.elementwise.add(
ctx, target, source_ir, f"{name}_shape_dim_{i}", dim_shape, pad_sum
)
start_list[dim_index] = -pad_before
start_list[dim_index] = impl.elementwise.sub(
ctx, target, source_ir, f"{name}_pad_before_neg_{i}", 0, pad_before
)

slices = []
for j in range(rank):
Expand All @@ -79,14 +81,23 @@ def get_padded_shape_tensors(
).get_output(0)
)
padded_shape_tensor = impl.cat.cat(
ctx, target, source_ir, f"{name}_cat_dim_{i}", slices, 0
ctx,
target,
source_ir,
f"{name}_cat_dim_{i}",
slices,
0,
cast_dtype=padded_shape_tensor.dtype,
)

start_indices_tensor = get_trt_tensor(
start_indices_tensor = impl.cat.cat(
ctx,
np.array(start_list, dtype=np.int32),
target,
source_ir,
f"{name}_start_indices_tensor",
dtype=np.int32,
start_list,
0,
cast_dtype=padded_shape_tensor.dtype,
)

return start_indices_tensor, padded_shape_tensor
Expand All @@ -98,7 +109,7 @@ def constant_padNd(
source_ir: Optional[SourceIR],
name: str,
input: TRTTensor,
pad: Sequence[int],
pad: Sequence[Union[int, TRTTensor]],
value: Union[int, float] = 0,
) -> TRTTensor:

Expand Down
Loading