Skip to content

Fix: aten::matmul converter behavior with 1d tensors #2450

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 1 commit into from
Dec 19, 2023
Merged

Fix: aten::matmul converter behavior with 1d tensors #2450

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
39 changes: 34 additions & 5 deletions core/conversion/converters/impl/matrix_multiply.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,20 +16,49 @@ auto mm_registrations TORCHTRT_UNUSED =
[](ConversionCtx* ctx, const torch::jit::Node* n, args& args) -> bool {
auto self = args[0].ITensorOrFreeze(ctx);
auto other = args[1].ITensorOrFreeze(ctx);

auto selfDims = self->getDimensions().nbDims;
auto otherDims = other->getDimensions().nbDims;

bool squeezeFront = false;
bool squeezeBack = false;

if (selfDims == 1 && selfDims < otherDims) {
squeezeFront = true;
} else if (otherDims == 1 && otherDims < selfDims) {
// Append a 1 to the end of the shape before padding front to match self
other = addPadding(ctx, n, other, 2, true, false);
otherDims = other->getDimensions().nbDims;
squeezeBack = true;
}

// Ensure self and other tensors have same nbDims by expanding the dimensions (from 0 axis) if
// necessary.
if (self->getDimensions().nbDims < other->getDimensions().nbDims) {
self = addPadding(ctx, n, self, other->getDimensions().nbDims, false, false);
} else {
other = addPadding(ctx, n, other, self->getDimensions().nbDims, false, false);
if (selfDims < otherDims) {
self = addPadding(ctx, n, self, otherDims, false, false);
} else if (otherDims < selfDims) {
other = addPadding(ctx, n, other, selfDims, false, false);
}

auto mm_layer = ctx->net->addMatrixMultiply(
*self, nvinfer1::MatrixOperation::kNONE, *other, nvinfer1::MatrixOperation::kNONE);

TORCHTRT_CHECK(mm_layer, "Unable to create matrix multiplication node: " << *n);
mm_layer->setName(util::node_info(n).c_str());
auto out_tensor = ctx->AssociateValueAndTensor(n->outputs()[0], mm_layer->getOutput(0));
auto out = mm_layer->getOutput(0);

if (squeezeFront || squeezeBack) {
auto squeezeDimOffset = squeezeFront ? 2 : 1;
auto reshapeDims =
util::squeezeDims(out->getDimensions(), out->getDimensions().nbDims - squeezeDimOffset);
auto shuffle_layer = ctx->net->addShuffle(*out);
LOG_DEBUG("Squeezing matmul output for 1d correction: " << reshapeDims);
TORCHTRT_CHECK(shuffle_layer, "Unable to create shuffle layer from node: " << *n);
shuffle_layer->setReshapeDimensions(reshapeDims);
shuffle_layer->setName((util::node_info(n) + "_squeeze").c_str());
out = shuffle_layer->getOutput(0);
}
auto out_tensor = ctx->AssociateValueAndTensor(n->outputs()[0], out);

LOG_DEBUG("Output tensor shape: " << out_tensor->getDimensions());
return true;
Expand Down
132 changes: 126 additions & 6 deletions tests/core/conversion/converters/test_matrix_multiply.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,9 +21,8 @@ TEST(Converters, ATenMMConvertsCorrectly) {

params = torch_tensorrt::core::ir::get_static_params(g->inputs(), {});
auto trt_results = torch_tensorrt::tests::util::RunGraphEngine(g, params, {in1, in2});
auto trt = trt_results[0].reshape_as(jit_results[0]);

ASSERT_TRUE(torch_tensorrt::tests::util::almostEqual(jit_results[0], trt, 2e-6));
ASSERT_TRUE(torch_tensorrt::tests::util::almostEqual(jit_results[0], trt_results[0], 2e-6));
}

TEST(Converters, ATenMMWithDiffShapesConvertsCorrectly) {
Expand All @@ -42,9 +41,131 @@ TEST(Converters, ATenMMWithDiffShapesConvertsCorrectly) {

params = torch_tensorrt::core::ir::get_static_params(g->inputs(), {});
auto trt_results = torch_tensorrt::tests::util::RunGraphEngine(g, params, {in1, in2});
auto trt = trt_results[0].reshape_as(jit_results[0]);

ASSERT_TRUE(torch_tensorrt::tests::util::almostEqual(jit_results[0], trt, 2e-6));
ASSERT_TRUE(torch_tensorrt::tests::util::almostEqual(jit_results[0], trt_results[0], 2e-6));
}

TEST(Converters, ATenMM1d2dConvertsCorrectly) {
const auto graph = R"IR(
graph(%0 : Tensor, %1 : Tensor):
%2 : Tensor = aten::matmul(%0, %1)
return (%2))IR";

auto g = std::make_shared<torch::jit::Graph>();
torch::jit::parseIR(graph, g.get());

auto in1 = at::randint(0, 5, {10}, {at::kCUDA});
auto in2 = at::randint(0, 5, {10, 1}, {at::kCUDA});
auto params = torch_tensorrt::core::ir::get_static_params(g->inputs(), {});
auto jit_results = torch_tensorrt::tests::util::RunGraph(g, params, {in1, in2});

params = torch_tensorrt::core::ir::get_static_params(g->inputs(), {});
auto trt_results = torch_tensorrt::tests::util::RunGraphEngine(g, params, {in1, in2});

ASSERT_TRUE(torch_tensorrt::tests::util::almostEqual(jit_results[0], trt_results[0], 2e-6));
}

TEST(Converters, ATenMM1d3dConvertsCorrectly) {
const auto graph = R"IR(
graph(%0 : Tensor, %1 : Tensor):
%2 : Tensor = aten::matmul(%0, %1)
return (%2))IR";

auto g = std::make_shared<torch::jit::Graph>();
torch::jit::parseIR(graph, g.get());

auto in1 = at::randint(0, 5, {10}, {at::kCUDA});
auto in2 = at::randint(0, 5, {2, 10, 8}, {at::kCUDA});
auto params = torch_tensorrt::core::ir::get_static_params(g->inputs(), {});
auto jit_results = torch_tensorrt::tests::util::RunGraph(g, params, {in1, in2});

params = torch_tensorrt::core::ir::get_static_params(g->inputs(), {});
auto trt_results = torch_tensorrt::tests::util::RunGraphEngine(g, params, {in1, in2});

ASSERT_TRUE(torch_tensorrt::tests::util::almostEqual(jit_results[0], trt_results[0], 2e-6));
}

TEST(Converters, ATenMM1d4dConvertsCorrectly) {
const auto graph = R"IR(
graph(%0 : Tensor, %1 : Tensor):
%2 : Tensor = aten::matmul(%0, %1)
return (%2))IR";

auto g = std::make_shared<torch::jit::Graph>();
torch::jit::parseIR(graph, g.get());

auto in1 = at::randint(0, 5, {10}, {at::kCUDA});
auto in2 = at::randint(0, 5, {2, 3, 10, 8}, {at::kCUDA});
auto params = torch_tensorrt::core::ir::get_static_params(g->inputs(), {});
auto jit_results = torch_tensorrt::tests::util::RunGraph(g, params, {in1, in2});

params = torch_tensorrt::core::ir::get_static_params(g->inputs(), {});
auto trt_results = torch_tensorrt::tests::util::RunGraphEngine(g, params, {in1, in2});

ASSERT_TRUE(torch_tensorrt::tests::util::almostEqual(jit_results[0], trt_results[0], 2e-6));
}

TEST(Converters, ATenMM3d1dConvertsCorrectly) {
const auto graph = R"IR(
graph(%0 : Tensor, %1 : Tensor):
%2 : Tensor = aten::matmul(%0, %1)
return (%2))IR";

auto g = std::make_shared<torch::jit::Graph>();
torch::jit::parseIR(graph, g.get());

auto in1 = at::randint(0, 5, {2, 10, 8}, {at::kCUDA});
auto in2 = at::randint(0, 5, {8}, {at::kCUDA});

auto params = torch_tensorrt::core::ir::get_static_params(g->inputs(), {});
auto jit_results = torch_tensorrt::tests::util::RunGraph(g, params, {in1, in2});

params = torch_tensorrt::core::ir::get_static_params(g->inputs(), {});
auto trt_results = torch_tensorrt::tests::util::RunGraphEngine(g, params, {in1, in2});

ASSERT_TRUE(torch_tensorrt::tests::util::almostEqual(jit_results[0], trt_results[0], 2e-6));
}

TEST(Converters, ATenMM2d1dConvertsCorrectly) {
const auto graph = R"IR(
graph(%0 : Tensor, %1 : Tensor):
%2 : Tensor = aten::matmul(%0, %1)
return (%2))IR";

auto g = std::make_shared<torch::jit::Graph>();
torch::jit::parseIR(graph, g.get());

auto in1 = at::randint(0, 5, {1, 10}, {at::kCUDA});
auto in2 = at::randint(0, 5, {10}, {at::kCUDA});

auto params = torch_tensorrt::core::ir::get_static_params(g->inputs(), {});
auto jit_results = torch_tensorrt::tests::util::RunGraph(g, params, {in1, in2});

params = torch_tensorrt::core::ir::get_static_params(g->inputs(), {});
auto trt_results = torch_tensorrt::tests::util::RunGraphEngine(g, params, {in1, in2});

ASSERT_TRUE(torch_tensorrt::tests::util::almostEqual(jit_results[0], trt_results[0], 2e-6));
}

TEST(Converters, ATenMM4d1dConvertsCorrectly) {
const auto graph = R"IR(
graph(%0 : Tensor, %1 : Tensor):
%2 : Tensor = aten::matmul(%0, %1)
return (%2))IR";

auto g = std::make_shared<torch::jit::Graph>();
torch::jit::parseIR(graph, g.get());

auto in1 = at::randint(0, 5, {2, 3, 10, 8}, {at::kCUDA});
auto in2 = at::randint(0, 5, {8}, {at::kCUDA});

auto params = torch_tensorrt::core::ir::get_static_params(g->inputs(), {});
auto jit_results = torch_tensorrt::tests::util::RunGraph(g, params, {in1, in2});

params = torch_tensorrt::core::ir::get_static_params(g->inputs(), {});
auto trt_results = torch_tensorrt::tests::util::RunGraphEngine(g, params, {in1, in2});

ASSERT_TRUE(torch_tensorrt::tests::util::almostEqual(jit_results[0], trt_results[0], 2e-6));
}

TEST(Converters, ATenBMMConvertsCorrectly) {
Expand All @@ -63,9 +184,8 @@ TEST(Converters, ATenBMMConvertsCorrectly) {

params = torch_tensorrt::core::ir::get_static_params(g->inputs(), {});
auto trt_results = torch_tensorrt::tests::util::RunGraphEngine(g, params, {in1, in2});
auto trt = trt_results[0].reshape_as(jit_results[0]);

ASSERT_TRUE(torch_tensorrt::tests::util::almostEqual(jit_results[0], trt, 2e-6));
ASSERT_TRUE(torch_tensorrt::tests::util::almostEqual(jit_results[0], trt_results[0], 2e-6));
}

TEST(Converters, ATenBADDBMMConvertsCorrectly) {
Expand Down