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

Contrib ops for TRT plugin: Disentangled Attention Plugin #11287

Merged
merged 8 commits into from
May 8, 2022
Merged

Contrib ops for TRT plugin: Disentangled Attention Plugin #11287

Show file tree
Hide file tree
Changes from 4 commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
e98993d
Add disentangled attention TRT plugin as contrib op
symphonylyh Apr 7, 2022
b5aa92a
update plugin name & remove null character
symphonylyh Apr 14, 2022
f28014a
update onnx-tensorrt submodule with my beta version
symphonylyh Apr 20, 2022
2cd6a3c
Merge branch 'master' into deberta_trt_plugin
symphonylyh Apr 20, 2022
b1d0428
use suggested plugin name & simpler shape propagation
symphonylyh Apr 22, 2022
1bbc014
update onnx-tensorrt gitsubmodule to temporary fork
symphonylyh Apr 22, 2022
a18c96a
update onnx-tensorrt to temporary commit
symphonylyh Apr 22, 2022
86cffed
redirect submodule back to latest 8.2-GA release of onnx-tensorrt repo
symphonylyh May 5, 2022
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 cmake/external/onnx-tensorrt
Open in desktop
Submodule onnx-tensorrt updated 7 files
+4 −0 ModelImporter.cpp
+4 −4 README.md
+7 −89 builtin_op_importers.cpp
+6 −6 docs/faq.md
+1 −1 docs/operators.md
+15 −11 onnx2trt_utils.cpp
+0 −25 trt_utils.hpp
40 changes: 39 additions & 1 deletion onnxruntime/core/graph/contrib_ops/contrib_defs.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2339,6 +2339,44 @@ void RegisterContribSchemas() {
updateOutputShape(ctx, 0, output_shape);
});

static const char* DisentangledAttention_TRT_ver1_doc =
R"DOC(Disentangled Attention TensorRT Plugin.)DOC";

ONNX_CONTRIB_OPERATOR_SCHEMA(DisentangledAttentionPlugin)
symphonylyh marked this conversation as resolved.
Show resolved Hide resolved
.SetDomain(kOnnxDomain)
.SinceVersion(1)
.SetDoc(DisentangledAttention_TRT_ver1_doc)
.Input(0, "c2c_attention", "content-to-content attention tensor, QcKc^T.", "T")
.Input(1, "c2p_attention", "content-to-position attention tensor, QcKr^T.", "T")
.Input(2, "p2c_attention", "position-to-content attention tensor, KcQr^T.", "T")
.Output(0, "disentangled_attention", "The disentangled attention output tensor.", "T")
.TypeConstraint("T", {"tensor(float)", "tensor(float16)"}, "Constrain input and output types to float tensors.")
.Attr("span", "Maximum relative distance, k.", AttributeProto::INT)
.Attr("factor", "Scaling factor applied to attention values, 1/sqrt(3d). d is hidden size per head = H/N. H is hidden size, N is number of heads.", AttributeProto::FLOAT)
.TypeAndShapeInferenceFunction([](ONNX_NAMESPACE::InferenceContext& ctx) {
// Type inference
using namespace ONNX_NAMESPACE;
propagateElemTypeFromInputToOutput(ctx, 0, 0);

// Shape Inference
if (!hasInputShape(ctx, 0)) {
return;
}

auto& input0_shape = getInputShape(ctx, 0);
auto& input0_dims = input0_shape.dim();
if (input0_dims.size() != 3) {
fail_shape_inference("Input 0 shall be 3 dimensions");
}

// output dims is same as input[0] dims, i.e., regular c2c attention dims
symphonylyh marked this conversation as resolved.
Show resolved Hide resolved
ONNX_NAMESPACE::TensorShapeProto disentangled_attention_shape;
for (auto& dim : input0_dims) {
*disentangled_attention_shape.add_dim() = dim;
}
updateOutputShape(ctx, 0, disentangled_attention_shape);
});

#ifndef _OPSCHEMA_LIB_
// Register the NCHWc schemas if supported by the platform.
if (MlasNchwcGetBlockSize() > 1) {
Expand All @@ -2352,4 +2390,4 @@ void RegisterContribSchemas() {
}

} // namespace contrib
} // namespace onnxruntime
} // namespace onnxruntime