saymrwulf/onnxruntime
1
0
Fork 0
mirror of https://github.com/saymrwulf/onnxruntime.git synced 2026-07-17 18:40:28 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Remove onnxruntime/core/protobuf (#8617)

Browse source 
* remove onnxruntime/core/protobuf

* Update How_To_Update_ONNX_Dev_Notes.md
This commit is contained in:
Changming Sun 2021-08-10 09:36:27 -07:00 committed by GitHub
parent f0073308d0
commit ed17ca3595
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
10 changed files with 7 additions and 3044 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

15
docs/How_To_Update_ONNX_Dev_Notes.md
View file

@ -18,20 +18,19 @@ This file should be generated. See [cgmanifests/README](/cgmanifests/README.md)
3. Update [tools/ci_build/github/linux/docker/scripts/install_onnx.sh](/tools/ci_build/github/linux/docker/scripts/install_onnx.sh).
Search 'for version2tag', update the commit hashes. The list should contain every release version from ONNX 1.2, and the latest one in our cmake/external/onnx folder.
4. If there is any change to `cmake/external/onnx/onnx/*.in.proto`, update onnxruntime/core/protobuf as follows :
```
- Apply these changes to onnxruntime/core/protobuf/*.in.proto
- Copy cmake/external/onnx/onnx/gen_proto.py to onnxruntime/core/protobuf and use this script to generate the new \*.proto and \*.proto3 files
- Regenerate csharp/test/Microsoft.ML.OnnxRuntime.Tests/OnnxMl.cs
```
4. If there is any change to `cmake/external/onnx/onnx/*.in.proto`, you need to re-regenerate OnnxMl.cs. Please build onnxruntime on Windows with csharp enabled, then the file will be auto-updated.
5. Send you PR, and run the CI builds.
6. If you are updating ONNX from a released tag to a new commit, please tell Changming deploying the new test data along with other test models to our CI build machines. This is to ensure that our tests cover every ONNX opset.
5. Send you PR, and **manually** queue a build for every packaging pipeline for your branch.
6. If there is a build failure in stage "Check out of dated documents" in WebAssembly CI pipeline, update ONNX Runtime Web WebGL operator support document:
- Make sure Node.js is installed (see [Prerequisites](../js/README.md#Prerequisites) for instructions).
- Follow step 1 in [js/Build](../js/README.md#Build-2) to install dependencies).
- Follow instructions in [Generate document](../js/README.md#Generating-Document) to update document. Commit changes applied to file `docs/operators.md`.
7. If there is any unitest failure, caught by onnx_test_runner. Please also update
7. Usually there would be some unitest failures, because you introduced new test cases. Then you may need to update
- [onnxruntime/test/onnx/main.cc](/onnxruntime/test/onnx/main.cc)
- [onnxruntime/test/providers/cpu/model_tests.cc](/onnxruntime/test/providers/cpu/model_tests.cc)
- [csharp/test/Microsoft.ML.OnnxRuntime.Tests/InferenceTest.cs](/csharp/test/Microsoft.ML.OnnxRuntime.Tests/InferenceTest.cs)
- [onnxruntime/test/testdata/onnx_backend_test_series_filters.jsonc](/onnxruntime/test/testdata/onnx_backend_test_series_filters.jsonc)

101
onnxruntime/core/protobuf/onnx-data.in.proto
View file

@ -1,101 +0,0 @@
// Copyright (c) ONNX Project Contributors.
// Licensed under the MIT license.
syntax = "proto2";
package {PACKAGE_NAME};
// #if ONNX-ML
import "onnx-ml.proto";
// #else
import "onnx.proto";
// #endif
// This file contains the proto definitions for MapProto and
// SequenceProto. These protos are used to represent the data structures
// of maps and sequence for use in test data or ModelProto.
// Sequences
//
// Defines a dense, ordered, collection of elements that are of homogeneous types.
// Sequences can be made out of tensors, maps, or sequences.
//
// If a sequence is made out of tensors, the tensors must have the same element
// type (i.e. int32). In some cases, the tensors in a sequence can have different
// shapes. Whether the tensors can have different shapes or not depends on the
// type/shape associated with the corresponding "ValueInfo". For example,
// "Sequence<Tensor<float, [M,N]>" means that all tensors have same shape. However,
// "Sequence<Tensor<float, [omitted,omitted]>" means they can have different
// shapes (all of rank 2), where "omitted" means the corresponding dimension has
// no symbolic/constant value. Finally, "Sequence<Tensor<float, omitted>>" means
// that the different tensors can have different ranks, when the "shape" itself
// is omitted from the tensor-type. For a more complete description, refer to
// https://github.com/onnx/onnx/blob/master/docs/IR.md#static-tensor-shapes.
//
message SequenceProto {
optional string name = 1;
enum DataType {
UNDEFINED = 0;
TENSOR = 1;
SPARSE_TENSOR = 2;
SEQUENCE = 3;
MAP = 4;
}
// The data type of the element.
// This field MUST have a valid SequenceProto.DataType value
optional int32 elem_type = 2;
// For TensorProto values.
// When this field is present, the elem_type field MUST be TENSOR.
repeated TensorProto tensor_values = 3;
// For SparseTensorProto values.
// When this field is present, the elem_type field MUST be SPARSE_TENSOR.
repeated SparseTensorProto sparse_tensor_values = 4;
// For SequenceProto values, allowing sequences to be of themselves.
// When this field is present, the elem_type field MUST be SEQUENCE.
repeated SequenceProto sequence_values = 5;
// For MapProto values.
// When this field is present, the elem_type field MUST be MAP.
repeated MapProto map_values = 6;
}
// Maps
//
// Specifies an associative table, defined by keys and values.
// MapProto is formed with a repeated field of keys (of type INT8, INT16, INT32,
// INT64, UINT8, UINT16, UINT32, UINT64, or STRING) and values (of type TENSOR,
// SPARSE_TENSOR, SEQUENCE, or MAP). Key types and value types have to remain
// the same throughout the instantiation of the MapProto.
//
message MapProto {
optional string name = 1;
// All MapProto data types must have the same length of keys and values.
// The data type of the key.
// This field MUST have a valid TensorProto.DataType value of
// INT8, INT16, INT32, INT64, UINT8, UINT16, UINT32, UINT64, or STRING
optional int32 key_type = 2;
// Every element of keys has to be one of the following data types
// INT8, INT16, INT32, INT64, UINT8, UINT16, UINT32, UINT64, or STRING.
// The integer cases are represented by the repeated int64 field keys below.
repeated int64 keys = 3;
// If keys are strings, they are represented by the repeated bytes field
// string_keys below.
repeated bytes string_keys = 4;
// MapProto values are represented in a SequenceProto of the same length as the
// repeated keys field and have to be one of the following data types
// TENSOR, SPARSE_TENSOR, MAP, SEQUENCE.
optional SequenceProto values = 5;
}

106
onnxruntime/core/protobuf/onnx-data.proto
View file

@ -1,106 +0,0 @@
//
// WARNING: This file is automatically generated! Please edit onnx.in.proto.
//
// Copyright (c) ONNX Project Contributors.
// Licensed under the MIT license.
syntax = "proto2";
package onnx;
import "onnx-ml.proto";
// This file contains the proto definitions for MapProto and
// SequenceProto. These protos are used to represent the data structures
// of maps and sequence for use in test data or ModelProto.
// Sequences
//
// Defines a dense, ordered, collection of elements that are of homogeneous types.
// Sequences can be made out of tensors, maps, or sequences.
//
// If a sequence is made out of tensors, the tensors must have the same element
// type (i.e. int32). In some cases, the tensors in a sequence can have different
// shapes. Whether the tensors can have different shapes or not depends on the
// type/shape associated with the corresponding "ValueInfo". For example,
// "Sequence<Tensor<float, [M,N]>" means that all tensors have same shape. However,
// "Sequence<Tensor<float, [omitted,omitted]>" means they can have different
// shapes (all of rank 2), where "omitted" means the corresponding dimension has
// no symbolic/constant value. Finally, "Sequence<Tensor<float, omitted>>" means
// that the different tensors can have different ranks, when the "shape" itself
// is omitted from the tensor-type. For a more complete description, refer to
// https://github.com/onnx/onnx/blob/master/docs/IR.md#static-tensor-shapes.
//
message SequenceProto {
optional string name = 1;
enum DataType {
UNDEFINED = 0;
TENSOR = 1;
SPARSE_TENSOR = 2;
SEQUENCE = 3;
MAP = 4;
}
// The data type of the element.
// This field MUST have a valid SequenceProto.DataType value
optional int32 elem_type = 2;
// For TensorProto values.
// When this field is present, the elem_type field MUST be TENSOR.
repeated TensorProto tensor_values = 3;
// For SparseTensorProto values.
// When this field is present, the elem_type field MUST be SPARSE_TENSOR.
repeated SparseTensorProto sparse_tensor_values = 4;
// For SequenceProto values, allowing sequences to be of themselves.
// When this field is present, the elem_type field MUST be SEQUENCE.
repeated SequenceProto sequence_values = 5;
// For MapProto values.
// When this field is present, the elem_type field MUST be MAP.
repeated MapProto map_values = 6;
}
// Maps
//
// Specifies an associative table, defined by keys and values.
// MapProto is formed with a repeated field of keys (of type INT8, INT16, INT32,
// INT64, UINT8, UINT16, UINT32, UINT64, or STRING) and values (of type TENSOR,
// SPARSE_TENSOR, SEQUENCE, or MAP). Key types and value types have to remain
// the same throughout the instantiation of the MapProto.
//
message MapProto {
optional string name = 1;
// All MapProto data types must have the same length of keys and values.
// The data type of the key.
// This field MUST have a valid TensorProto.DataType value of
// INT8, INT16, INT32, INT64, UINT8, UINT16, UINT32, UINT64, or STRING
optional int32 key_type = 2;
// Every element of keys has to be one of the following data types
// INT8, INT16, INT32, INT64, UINT8, UINT16, UINT32, UINT64, or STRING.
// The integer cases are represented by the repeated int64 field keys below.
repeated int64 keys = 3;
// If keys are strings, they are represented by the repeated bytes field
// string_keys below.
repeated bytes string_keys = 4;
// MapProto values are represented in a SequenceProto of the same length as the
// repeated keys field and have to be one of the following data types
// TENSOR, SPARSE_TENSOR, MAP, SEQUENCE.
optional SequenceProto values = 5;
}
// For using protobuf-lite
option optimize_for = LITE_RUNTIME;

106
onnxruntime/core/protobuf/onnx-data.proto3
View file

@ -1,106 +0,0 @@
//
// WARNING: This file is automatically generated! Please edit onnx.in.proto.
//
// Copyright (c) ONNX Project Contributors.
// Licensed under the MIT license.
syntax = "proto3";
package onnx;
import "onnx-ml.proto3";
// This file contains the proto definitions for MapProto and
// SequenceProto. These protos are used to represent the data structures
// of maps and sequence for use in test data or ModelProto.
// Sequences
//
// Defines a dense, ordered, collection of elements that are of homogeneous types.
// Sequences can be made out of tensors, maps, or sequences.
//
// If a sequence is made out of tensors, the tensors must have the same element
// type (i.e. int32). In some cases, the tensors in a sequence can have different
// shapes. Whether the tensors can have different shapes or not depends on the
// type/shape associated with the corresponding "ValueInfo". For example,
// "Sequence<Tensor<float, [M,N]>" means that all tensors have same shape. However,
// "Sequence<Tensor<float, [omitted,omitted]>" means they can have different
// shapes (all of rank 2), where "omitted" means the corresponding dimension has
// no symbolic/constant value. Finally, "Sequence<Tensor<float, omitted>>" means
// that the different tensors can have different ranks, when the "shape" itself
// is omitted from the tensor-type. For a more complete description, refer to
// https://github.com/onnx/onnx/blob/master/docs/IR.md#static-tensor-shapes.
//
message SequenceProto {
string name = 1;
enum DataType {
UNDEFINED = 0;
TENSOR = 1;
SPARSE_TENSOR = 2;
SEQUENCE = 3;
MAP = 4;
}
// The data type of the element.
// This field MUST have a valid SequenceProto.DataType value
int32 elem_type = 2;
// For TensorProto values.
// When this field is present, the elem_type field MUST be TENSOR.
repeated TensorProto tensor_values = 3;
// For SparseTensorProto values.
// When this field is present, the elem_type field MUST be SPARSE_TENSOR.
repeated SparseTensorProto sparse_tensor_values = 4;
// For SequenceProto values, allowing sequences to be of themselves.
// When this field is present, the elem_type field MUST be SEQUENCE.
repeated SequenceProto sequence_values = 5;
// For MapProto values.
// When this field is present, the elem_type field MUST be MAP.
repeated MapProto map_values = 6;
}
// Maps
//
// Specifies an associative table, defined by keys and values.
// MapProto is formed with a repeated field of keys (of type INT8, INT16, INT32,
// INT64, UINT8, UINT16, UINT32, UINT64, or STRING) and values (of type TENSOR,
// SPARSE_TENSOR, SEQUENCE, or MAP). Key types and value types have to remain
// the same throughout the instantiation of the MapProto.
//
message MapProto {
string name = 1;
// All MapProto data types must have the same length of keys and values.
// The data type of the key.
// This field MUST have a valid TensorProto.DataType value of
// INT8, INT16, INT32, INT64, UINT8, UINT16, UINT32, UINT64, or STRING
int32 key_type = 2;
// Every element of keys has to be one of the following data types
// INT8, INT16, INT32, INT64, UINT8, UINT16, UINT32, UINT64, or STRING.
// The integer cases are represented by the repeated int64 field keys below.
repeated int64 keys = 3;
// If keys are strings, they are represented by the repeated bytes field
// string_keys below.
repeated bytes string_keys = 4;
// MapProto values are represented in a SequenceProto of the same length as the
// repeated keys field and have to be one of the following data types
// TENSOR, SPARSE_TENSOR, MAP, SEQUENCE.
SequenceProto values = 5;
}
// For using protobuf-lite
option optimize_for = LITE_RUNTIME;

774
onnxruntime/core/protobuf/onnx-ml.proto
View file

@ -1,774 +0,0 @@
//
// WARNING: This file is automatically generated! Please edit onnx.in.proto.
//
// Copyright (c) ONNX Project Contributors.
// Licensed under the MIT license.
syntax = "proto2";
package onnx;
// Overview
//
// ONNX is an open specification that is comprised of the following components:
//
// 1) A definition of an extensible computation graph model.
// 2) Definitions of standard data types.
// 3) Definitions of built-in operators.
//
// This document describes the syntax of models and their computation graphs,
// as well as the standard data types. Together, they are referred to as the ONNX
// Intermediate Representation, or 'IR' for short.
//
// The normative semantic specification of the ONNX IR is found in docs/IR.md.
// Definitions of the built-in neural network operators may be found in docs/Operators.md.
// Definitions of the built-in classical machine learning operators may be found in
// docs/Operators-ml.md.
// Notes
//
// Release
//
// We are still in the very early stage of defining ONNX. The current
// version of ONNX is a starting point. While we are actively working
// towards a complete spec, we would like to get the community involved
// by sharing our working version of ONNX.
//
// Protobuf compatibility
//
// To simplify framework compatibility, ONNX is defined using the subset of protobuf
// that is compatible with both protobuf v2 and v3. This means that we do not use any
// protobuf features that are only available in one of the two versions.
//
// Here are the most notable contortions we have to carry out to work around
// these limitations:
//
// - No 'map' (added protobuf 3.0). We instead represent mappings as lists
// of key-value pairs, where order does not matter and duplicates
// are not allowed.
// Versioning
//
// ONNX versioning is specified in docs/IR.md and elaborated on in docs/Versioning.md
//
// To be compatible with both proto2 and proto3, we will use a version number
// that is not defined by the default value but an explicit enum number.
enum Version {
// proto3 requires the first enum value to be zero.
// We add this just to appease the compiler.
_START_VERSION = 0;
// The version field is always serialized and we will use it to store the
// version that the graph is generated from. This helps us set up version
// control.
// For the IR, we are using simple numbers starting with 0x00000001,
// which was the version we published on Oct 10, 2017.
IR_VERSION_2017_10_10 = 0x0000000000000001;
// IR_VERSION 2 published on Oct 30, 2017
// - Added type discriminator to AttributeProto to support proto3 users
IR_VERSION_2017_10_30 = 0x0000000000000002;
// IR VERSION 3 published on Nov 3, 2017
// - For operator versioning:
// - Added new message OperatorSetIdProto
// - Added opset_import in ModelProto
// - For vendor extensions, added domain in NodeProto
IR_VERSION_2017_11_3 = 0x0000000000000003;
// IR VERSION 4 published on Jan 22, 2019
// - Relax constraint that initializers should be a subset of graph inputs
// - Add type BFLOAT16
IR_VERSION_2019_1_22 = 0x0000000000000004;
// IR VERSION 5 published on March 18, 2019
// - Add message TensorAnnotation.
// - Add quantization annotation in GraphProto to map tensor with its scale and zero point quantization parameters.
IR_VERSION_2019_3_18 = 0x0000000000000005;
// IR VERSION 6 published on Sep 19, 2019
// - Add support for sparse tensor constants stored in model.
// - Add message SparseTensorProto
// - Add sparse initializers
IR_VERSION_2019_9_19 = 0x0000000000000006;
// IR VERSION 7 published on <TBD>
// - Add support to allow function body graph to rely on multiple external opreator sets.
// - Add a list to promote inference graph's initializers to global and
// mutable variables. Global variables are visible in all graphs of the
// stored models.
// - Add message TrainingInfoProto to store initialization
// method and training algorithm. The execution of TrainingInfoProto
// can modify the values of mutable variables.
// - Make inference graph callable from TrainingInfoProto via GraphCall operator.
IR_VERSION = 0x0000000000000007;
}
// Attributes
//
// A named attribute containing either singular float, integer, string, graph,
// and tensor values, or repeated float, integer, string, graph, and tensor values.
// An AttributeProto MUST contain the name field, and *only one* of the
// following content fields, effectively enforcing a C/C++ union equivalent.
message AttributeProto {
// Note: this enum is structurally identical to the OpSchema::AttrType
// enum defined in schema.h. If you rev one, you likely need to rev the other.
enum AttributeType {
UNDEFINED = 0;
FLOAT = 1;
INT = 2;
STRING = 3;
TENSOR = 4;
GRAPH = 5;
SPARSE_TENSOR = 11;
FLOATS = 6;
INTS = 7;
STRINGS = 8;
TENSORS = 9;
GRAPHS = 10;
SPARSE_TENSORS = 12;
}
// The name field MUST be present for this version of the IR.
optional string name = 1; // namespace Attribute
// if ref_attr_name is not empty, ref_attr_name is the attribute name in parent function.
// In this case, this AttributeProto does not contain data, and it's a reference of attribute
// in parent scope.
// NOTE: This should ONLY be used in function (sub-graph). It's invalid to be used in main graph.
optional string ref_attr_name = 21;
// A human-readable documentation for this attribute. Markdown is allowed.
optional string doc_string = 13;
// The type field MUST be present for this version of the IR.
// For 0.0.1 versions of the IR, this field was not defined, and
// implementations needed to use has_field heuristics to determine
// which value field was in use. For IR_VERSION 0.0.2 or later, this
// field MUST be set and match the f|i|s|t|... field in use. This
// change was made to accommodate proto3 implementations.
optional AttributeType type = 20; // discriminator that indicates which field below is in use
// Exactly ONE of the following fields must be present for this version of the IR
optional float f = 2; // float
optional int64 i = 3; // int
optional bytes s = 4; // UTF-8 string
optional TensorProto t = 5; // tensor value
optional GraphProto g = 6; // graph
optional SparseTensorProto sparse_tensor = 22; // sparse tensor value
// Do not use field below, it's deprecated.
// optional ValueProto v = 12; // value - subsumes everything but graph
repeated float floats = 7; // list of floats
repeated int64 ints = 8; // list of ints
repeated bytes strings = 9; // list of UTF-8 strings
repeated TensorProto tensors = 10; // list of tensors
repeated GraphProto graphs = 11; // list of graph
repeated SparseTensorProto sparse_tensors = 23; // list of sparse tensors
}
// Defines information on value, including the name, the type, and
// the shape of the value.
message ValueInfoProto {
// This field MUST be present in this version of the IR.
optional string name = 1; // namespace Value
// This field MUST be present in this version of the IR for
// inputs and outputs of the top-level graph.
optional TypeProto type = 2;
// A human-readable documentation for this value. Markdown is allowed.
optional string doc_string = 3;
}
// Nodes
//
// Computation graphs are made up of a DAG of nodes, which represent what is
// commonly called a "layer" or "pipeline stage" in machine learning frameworks.
//
// For example, it can be a node of type "Conv" that takes in an image, a filter
// tensor and a bias tensor, and produces the convolved output.
message NodeProto {
repeated string input = 1; // namespace Value
repeated string output = 2; // namespace Value
// An optional identifier for this node in a graph.
// This field MAY be absent in ths version of the IR.
optional string name = 3; // namespace Node
// The symbolic identifier of the Operator to execute.
optional string op_type = 4; // namespace Operator
// The domain of the OperatorSet that specifies the operator named by op_type.
optional string domain = 7; // namespace Domain
// Additional named attributes.
repeated AttributeProto attribute = 5;
// A human-readable documentation for this node. Markdown is allowed.
optional string doc_string = 6;
}
// Training information
// TrainingInfoProto stores information for training a model.
// In particular, this defines two functionalities: an initialization-step
// and a training-algorithm-step. Initialization resets the model
// back to its original state as if no training has been consumed.
// Training algorithm improves the model based on input data.
//
// The semantics of the initialization-step is that the initializers
// in ModelProto.graph and in TrainingInfoProto.algorithm are first
// initialized as specified by the initializers in the graph, and then
// updated by the "initialization_binding" in every instance in
// ModelProto.training_info.
//
// The field "algorithm" defines a computation graph which represents a
// training algorithm's step. After the execution of a
// TrainingInfoProto.algorithm, the initializers specified by "update_binding"
// may be immediately updated. If the targeted training algorithm contains
// consecutive update stages (such as block coordinate descent methods),
// the user needs to create a TrainingInfoProto for each stage.
message TrainingInfoProto {
// This field describes a graph to compute the initial tensors
// upon starting the training process. Initialization graph has no input
// and can have multiple outputs. Usually, trainable tensors in neural
// networks are randomly initialized. To achieve that, for each tensor,
// the user can put a random number operator such as RandomNormal or
// RandomUniform in TrainingInfoProto.initialization.node and assign its
// random output to the specific tensor using "initialization_binding".
// This graph can also set the initializers in "algorithm" in the same
// TrainingInfoProto; a use case is resetting the number of training
// iteration to zero.
//
// By default, this field is an empty graph and its evaluation does not
// produce any output.
optional GraphProto initialization = 1;
// This field represents a training algorithm step. Given required inputs,
// it computes outputs to update initializers in its own or inference graph's
// initializer lists. In general, this graph contains loss node, gradient node,
// optimizer node, increment of iteration count, and some calls to the inference
// graph.
//
// The field algorithm.node is the only place the user can use GraphCall
// operator. The only callable graph is the one stored in ModelProto.graph.
//
// By default, this field is an empty graph and its evaluation does not
// produce any output.
optional GraphProto algorithm = 2;
// This field specifies the bindings from the outputs of "initialization" to
// some initializers in "ModelProto.graph.initializer" and
// the "algorithm.initializer" in the same TrainingInfoProto.
// See "update_binding" below for details.
//
// By default, this field is empty and no initializer would be changed
// by the execution of "initialization".
repeated StringStringEntryProto initialization_binding = 3;
// Gradient-based training is usually an iterative procedure. In one gradient
// descent iteration, we apply
//
// x = x - r * g
//
// where "x" is the optimized tensor, "r" stands for learning rate, and "g" is
// gradient of "x" with respect to a chosen loss. To avoid adding assignments
// into the training graph, we split the update equation into
//
// y = x - r * g
// x = y
//
// The user needs to save "y = x - r * g" into TrainingInfoProto.algorithm. To
// tell that "y" should be assigned to "x", the field "update_binding" may
// contain a key-value pair of strings, "x" (key of StringStringEntryProto)
// and "y" (value of StringStringEntryProto).
// For a neural network with multiple trainable (mutable) tensors, there can
// be multiple key-value pairs in "update_binding".
//
// The initializers appears as keys in "update_binding" are considered
// mutable and globally-visible variables. This implies some behaviors
// as described below.
//
// 1. We have only unique keys in all "update_binding"s so that two global
// variables may not have the same name. This ensures that one
// global variable is assigned up to once.
// 2. The keys must appear in names of "ModelProto.graph.initializer" or
// "TrainingInfoProto.algorithm.initializer".
// 3. The values must be output names of "algorithm".
// 4. If an optional input of a graph is omitted when using GraphCall, the
// global variable with the same name may be used.
// 5. When using GraphCall, the users always can pass values to optional
// inputs of the called graph even if the associated initializers appears
// as keys in "update_binding"s.
// 6. The graphs in TrainingInfoProto's can use global variables as
// their operator inputs.
// 7. Mutable variables are initialized to the value specified by the
// corresponding initializer, and then potentially updated by
// "initializer_binding"s and "update_binding"s in "TrainingInfoProto"s.
//
// This field usually contains names of trainable tensors
// (in ModelProto.graph), optimizer states such as momentums in advanced
// stochastic gradient methods (in TrainingInfoProto.graph),
// and number of training iterations (in TrainingInfoProto.graph).
//
// By default, this field is empty and no initializer would be changed
// by the execution of "algorithm".
repeated StringStringEntryProto update_binding = 4;
}
// Models
//
// ModelProto is a top-level file/container format for bundling a ML model and
// associating its computation graph with metadata.
//
// The semantics of the model are described by the associated GraphProto's.
message ModelProto {
// The version of the IR this model targets. See Version enum above.
// This field MUST be present.
optional int64 ir_version = 1;
// The OperatorSets this model relies on.
// All ModelProtos MUST have at least one entry that
// specifies which version of the ONNX OperatorSet is
// being imported.
//
// All nodes in the ModelProto's graph will bind against the operator
// with the same-domain/same-op_type operator with the HIGHEST version
// in the referenced operator sets.
repeated OperatorSetIdProto opset_import = 8;
// The name of the framework or tool used to generate this model.
// This field SHOULD be present to indicate which implementation/tool/framework
// emitted the model.
optional string producer_name = 2;
// The version of the framework or tool used to generate this model.
// This field SHOULD be present to indicate which implementation/tool/framework
// emitted the model.
optional string producer_version = 3;
// Domain name of the model.
// We use reverse domain names as name space indicators. For example:
// `com.facebook.fair` or `com.microsoft.cognitiveservices`
//
// Together with `model_version` and GraphProto.name, this forms the unique identity of
// the graph.
optional string domain = 4;
// The version of the graph encoded. See Version enum below.
optional int64 model_version = 5;
// A human-readable documentation for this model. Markdown is allowed.
optional string doc_string = 6;
// The parameterized graph that is evaluated to execute the model.
optional GraphProto graph = 7;
// kezhan: This field is not in ONNX, and will be pushed into ONNX with good use cases in microsoft.
repeated FunctionProto functions = 100;
// Named metadata values; keys should be distinct.
repeated StringStringEntryProto metadata_props = 14;
// Training-specific information. Sequentially executing all stored
// `TrainingInfoProto.algorithm`s and assigning their outputs following
// the corresponding `TrainingInfoProto.update_binding`s is one training
// iteration. Similarly, to initialize the model
// (as if training hasn't happened), the user should sequentially execute
// all stored `TrainingInfoProto.initialization`s and assigns their outputs
// using `TrainingInfoProto.initialization_binding`s.
//
// If this field is empty, the training behavior of the model is undefined.
repeated TrainingInfoProto training_info = 20;
};
// StringStringEntryProto follows the pattern for cross-proto-version maps.
// See https://developers.google.com/protocol-buffers/docs/proto3#maps
message StringStringEntryProto {
optional string key = 1;
optional string value= 2;
};
message TensorAnnotation {
optional string tensor_name = 1;
// <key, value> pairs to annotate tensor specified by <tensor_name> above.
// The keys used in the mapping below must be pre-defined in ONNX spec.
// For example, for 8-bit linear quantization case, 'SCALE_TENSOR', 'ZERO_POINT_TENSOR' will be pre-defined as
// quantization parameter keys.
repeated StringStringEntryProto quant_parameter_tensor_names = 2;
}
// Graphs
//
// A graph defines the computational logic of a model and is comprised of a parameterized
// list of nodes that form a directed acyclic graph based on their inputs and outputs.
// This is the equivalent of the "network" or "graph" in many deep learning
// frameworks.
message GraphProto {
// The nodes in the graph, sorted topologically.
repeated NodeProto node = 1;
// The name of the graph.
optional string name = 2; // namespace Graph
// A list of named tensor values, used to specify constant inputs of the graph.
// Each TensorProto entry must have a distinct name (within the list) that
// MAY also appear in the input list.
repeated TensorProto initializer = 5;
// Initializers (see above) stored in sparse format.
repeated SparseTensorProto sparse_initializer = 15;
// A human-readable documentation for this graph. Markdown is allowed.
optional string doc_string = 10;
// The inputs and outputs of the graph.
repeated ValueInfoProto input = 11;
repeated ValueInfoProto output = 12;
// Information for the values in the graph. The ValueInfoProto.name's
// must be distinct. It is optional for a value to appear in value_info list.
repeated ValueInfoProto value_info = 13;
// This field carries information to indicate the mapping among a tensor and its
// quantization parameter tensors. For example:
// For tensor 'a', it may have {'SCALE_TENSOR', 'a_scale'} and {'ZERO_POINT_TENSOR', 'a_zero_point'} annotated,
// which means, tensor 'a_scale' and tensor 'a_zero_point' are scale and zero point of tensor 'a' in the model.
repeated TensorAnnotation quantization_annotation = 14;
// DO NOT USE the following fields, they were deprecated from earlier versions.
// repeated string input = 3;
// repeated string output = 4;
// optional int64 ir_version = 6;
// optional int64 producer_version = 7;
// optional string producer_tag = 8;
// optional string domain = 9;
}
// Tensors
//
// A serialized tensor value.
message TensorProto {
enum DataType {
UNDEFINED = 0;
// Basic types.
FLOAT = 1; // float
UINT8 = 2; // uint8_t
INT8 = 3; // int8_t
UINT16 = 4; // uint16_t
INT16 = 5; // int16_t
INT32 = 6; // int32_t
INT64 = 7; // int64_t
STRING = 8; // string
BOOL = 9; // bool
// IEEE754 half-precision floating-point format (16 bits wide).
// This format has 1 sign bit, 5 exponent bits, and 10 mantissa bits.
FLOAT16 = 10;
DOUBLE = 11;
UINT32 = 12;
UINT64 = 13;
COMPLEX64 = 14; // complex with float32 real and imaginary components
COMPLEX128 = 15; // complex with float64 real and imaginary components
// Non-IEEE floating-point format based on IEEE754 single-precision
// floating-point number truncated to 16 bits.
// This format has 1 sign bit, 8 exponent bits, and 7 mantissa bits.
BFLOAT16 = 16;
// Future extensions go here.
}
// The shape of the tensor.
repeated int64 dims = 1;
// The data type of the tensor.
// This field MUST have a valid TensorProto.DataType value
optional int32 data_type = 2;
// For very large tensors, we may want to store them in chunks, in which
// case the following fields will specify the segment that is stored in
// the current TensorProto.
message Segment {
optional int64 begin = 1;
optional int64 end = 2;
}
optional Segment segment = 3;
// Tensor content must be organized in row-major order.
//
// Depending on the data_type field, exactly one of the fields below with
// name ending in _data is used to store the elements of the tensor.
// For float and complex64 values
// Complex64 tensors are encoded as a single array of floats,
// with the real components appearing in odd numbered positions,
// and the corresponding imaginary component appearing in the
// subsequent even numbered position. (e.g., [1.0 + 2.0i, 3.0 + 4.0i]
// is encoded as [1.0, 2.0 ,3.0 ,4.0]
// When this field is present, the data_type field MUST be FLOAT or COMPLEX64.
repeated float float_data = 4 [packed = true];
// For int32, uint8, int8, uint16, int16, bool, and float16 values
// float16 values must be bit-wise converted to an uint16_t prior
// to writing to the buffer.
// When this field is present, the data_type field MUST be
// INT32, INT16, INT8, UINT16, UINT8, BOOL, or FLOAT16
repeated int32 int32_data = 5 [packed = true];
// For strings.
// Each element of string_data is a UTF-8 encoded Unicode
// string. No trailing null, no leading BOM. The protobuf "string"
// scalar type is not used to match ML community conventions.
// When this field is present, the data_type field MUST be STRING
repeated bytes string_data = 6;
// For int64.
// When this field is present, the data_type field MUST be INT64
repeated int64 int64_data = 7 [packed = true];
// Optionally, a name for the tensor.
optional string name = 8; // namespace Value
// A human-readable documentation for this tensor. Markdown is allowed.
optional string doc_string = 12;
// Serializations can either use one of the fields above, or use this
// raw bytes field. The only exception is the string case, where one is
// required to store the content in the repeated bytes string_data field.
//
// When this raw_data field is used to store tensor value, elements MUST
// be stored in as fixed-width, little-endian order.
// Floating-point data types MUST be stored in IEEE 754 format.
// Complex64 elements must be written as two consecutive FLOAT values, real component first.
// Complex128 elements must be written as two consecutive DOUBLE values, real component first.
// Boolean type MUST be written one byte per tensor element (00000001 for true, 00000000 for false).
//
// Note: the advantage of specific field rather than the raw_data field is
// that in some cases (e.g. int data), protobuf does a better packing via
// variable length storage, and may lead to smaller binary footprint.
// When this field is present, the data_type field MUST NOT be STRING or UNDEFINED
optional bytes raw_data = 9;
// Data can be stored inside the protobuf file using type-specific fields or raw_data.
// Alternatively, raw bytes data can be stored in an external file, using the external_data field.
// external_data stores key-value pairs describing data location. Recognized keys are:
// - "location" (required) - POSIX filesystem path relative to the directory where the ONNX
// protobuf model was stored
// - "offset" (optional) - position of byte at which stored data begins. Integer stored as string.
// Offset values SHOULD be multiples 4096 (page size) to enable mmap support.
// - "length" (optional) - number of bytes containing data. Integer stored as string.
// - "checksum" (optional) - SHA1 digest of file specified in under 'location' key.
repeated StringStringEntryProto external_data = 13;
// Location of the data for this tensor. MUST be one of:
// - DEFAULT - data stored inside the protobuf message. Data is stored in raw_data (if set) otherwise in type-specified field.
// - EXTERNAL - data stored in an external location as described by external_data field.
enum DataLocation {
DEFAULT = 0;
EXTERNAL = 1;
}
// If value not set, data is stored in raw_data (if set) otherwise in type-specified field.
optional DataLocation data_location = 14;
// For double
// Complex128 tensors are encoded as a single array of doubles,
// with the real components appearing in odd numbered positions,
// and the corresponding imaginary component appearing in the
// subsequent even numbered position. (e.g., [1.0 + 2.0i, 3.0 + 4.0i]
// is encoded as [1.0, 2.0 ,3.0 ,4.0]
// When this field is present, the data_type field MUST be DOUBLE or COMPLEX128
repeated double double_data = 10 [packed = true];
// For uint64 and uint32 values
// When this field is present, the data_type field MUST be
// UINT32 or UINT64
repeated uint64 uint64_data = 11 [packed = true];
}
// A serialized sparse-tensor value
message SparseTensorProto {
// The sequence of non-default values are encoded as a tensor of shape [NNZ].
// The default-value is zero for numeric tensors, and empty-string for string tensors.
optional TensorProto values = 1;
// The indices of the non-default values, which may be stored in one of two formats.
// (a) Indices can be a tensor of shape [NNZ, rank] with the [i,j]-th value
// corresponding to the j-th index of the i-th value (in the values tensor).
// (b) Indices can be a tensor of shape [NNZ], in which case the i-th value
// must be the linearized-index of the i-th value (in the values tensor).
// The linearized-index can be converted into an index tuple (k_1,...,k_rank)
// using the shape provided below.
// The indices must appear in ascending order without duplication.
// In the first format, the ordering is lexicographic-ordering:
// e.g., index-value [1,4] must appear before [2,1]
optional TensorProto indices = 2;
// The shape of the underlying dense-tensor: [dim_1, dim_2, ... dim_rank]
repeated int64 dims = 3;
}
// Defines a tensor shape. A dimension can be either an integer value
// or a symbolic variable. A symbolic variable represents an unknown
// dimension.
message TensorShapeProto {
message Dimension {
oneof value {
int64 dim_value = 1;
string dim_param = 2; // namespace Shape
};
// Standard denotation can optionally be used to denote tensor
// dimensions with standard semantic descriptions to ensure
// that operations are applied to the correct axis of a tensor.
// Refer to https://github.com/onnx/onnx/blob/master/docs/DimensionDenotation.md#denotation-definition
// for pre-defined dimension denotations.
optional string denotation = 3;
};
repeated Dimension dim = 1;
}
// Types
//
// The standard ONNX data types.
message TypeProto {
message Tensor {
// This field MUST NOT have the value of UNDEFINED
// This field MUST have a valid TensorProto.DataType value
// This field MUST be present for this version of the IR.
optional int32 elem_type = 1;
optional TensorShapeProto shape = 2;
}
// repeated T
message Sequence {
// The type and optional shape of each element of the sequence.
// This field MUST be present for this version of the IR.
optional TypeProto elem_type = 1;
};
// map<K,V>
message Map {
// This field MUST have a valid TensorProto.DataType value
// This field MUST be present for this version of the IR.
// This field MUST refer to an integral type ([U]INT{8|16|32|64}) or STRING
optional int32 key_type = 1;
// This field MUST be present for this version of the IR.
optional TypeProto value_type = 2;
};
message SparseTensor {
// This field MUST NOT have the value of UNDEFINED
// This field MUST have a valid TensorProto.DataType value
// This field MUST be present for this version of the IR.
optional int32 elem_type = 1;
optional TensorShapeProto shape = 2;
}
message Opaque {
// When missing, the domain is the same as the model's.
optional string domain = 1;
// The name is optional but significant when provided.
optional string name = 2;
// parameters that help defining the type
// DEPRECATED do not use.
// repeated TypeProto parameters = 3;
}
oneof value {
// The type of a tensor.
Tensor tensor_type = 1;
// NOTE: DNN-only implementations of ONNX MAY elect to not support non-tensor values
// as input and output to graphs and nodes. These types are needed to naturally
// support classical ML operators. DNN operators SHOULD restrict their input
// and output types to tensors.
// The type of a sequence.
Sequence sequence_type = 4;
// The type of a map.
Map map_type = 5;
SparseTensor sparse_tensor_type = 8;
Opaque opaque_type = 7;
}
// An optional denotation can be used to denote the whole
// type with a standard semantic description as to what is
// stored inside. Refer to https://github.com/onnx/onnx/blob/master/docs/TypeDenotation.md#type-denotation-definition
// for pre-defined type denotations.
optional string denotation = 6;
}
// Operator Sets
//
// OperatorSets are uniquely identified by a (domain, opset_version) pair.
message OperatorSetIdProto {
// The domain of the operator set being identified.
// The empty string ("") or absence of this field implies the operator
// set that is defined as part of the ONNX specification.
// This field MUST be present in this version of the IR when referring to any other operator set.
optional string domain = 1;
// The version of the operator set being identified.
// This field MUST be present in this version of the IR.
optional int64 version = 2;
}
// Operator/function status.
enum OperatorStatus {
EXPERIMENTAL = 0;
STABLE = 1;
}
message FunctionProto {
// The name of the function, similar usage of op_type in OperatorProto.
optional string name = 1;
// The first version of a function set which contains this function.
// When there's any breaking change for this function, the function set
// contains the function needs to bump its version, and since_version of
// the updated function will be changed to the updated function set version.
optional int64 since_version = 2;
// This field indicates whether the syntax, semantics, or presence
// of this function is in an experimental or stable stage. Once an
// function is published as STABLE, its syntax and semantics MUST NOT
// change in subsequent versions of the operator set.
// When a function is published as EXPERIMENTAL, the syntax and semantics
// of the function MAY change across operator set versions.
// Functions "become" stable by deprecating the experimental version and
// introducing a new stable function with the same name.
optional OperatorStatus status = 3;
// The inputs and outputs of the function.
repeated string input = 4;
repeated string output = 5;
// The attributes of the function.
repeated string attribute= 6;
// The nodes in the function.
repeated NodeProto node = 7;
// A human-readable documentation for this function. Markdown is allowed.
optional string doc_string = 8;
// The OperatorSets this function body (graph) relies on.
// A FunctionProto body (graph) may implicitly rely on the OperatorSet that
// this function belongs to. It can also explicitly rely on more OperatorSets
// with this field specified.
//
// All nodes in the function body (graph) will bind against the operator
// with the same-domain/same-op_type operator with the HIGHEST version
// in the referenced operator sets. This means at most one version can be relied
// for one domain.
repeated OperatorSetIdProto opset_import = 9;
}

779
onnxruntime/core/protobuf/onnx-ml.proto3
View file

@ -1,779 +0,0 @@
//
// WARNING: This file is automatically generated! Please edit onnx.in.proto.
//
// Copyright (c) ONNX Project Contributors.
// Licensed under the MIT license.
syntax = "proto3";
package onnx;
// Overview
//
// ONNX is an open specification that is comprised of the following components:
//
// 1) A definition of an extensible computation graph model.
// 2) Definitions of standard data types.
// 3) Definitions of built-in operators.
//
// This document describes the syntax of models and their computation graphs,
// as well as the standard data types. Together, they are referred to as the ONNX
// Intermediate Representation, or 'IR' for short.
//
// The normative semantic specification of the ONNX IR is found in docs/IR.md.
// Definitions of the built-in neural network operators may be found in docs/Operators.md.
// Definitions of the built-in classical machine learning operators may be found in
// docs/Operators-ml.md.
// Notes
//
// Release
//
// We are still in the very early stage of defining ONNX. The current
// version of ONNX is a starting point. While we are actively working
// towards a complete spec, we would like to get the community involved
// by sharing our working version of ONNX.
//
// Protobuf compatibility
//
// To simplify framework compatibility, ONNX is defined using the subset of protobuf
// that is compatible with both protobuf v2 and v3. This means that we do not use any
// protobuf features that are only available in one of the two versions.
//
// Here are the most notable contortions we have to carry out to work around
// these limitations:
//
// - No 'map' (added protobuf 3.0). We instead represent mappings as lists
// of key-value pairs, where order does not matter and duplicates
// are not allowed.
// Versioning
//
// ONNX versioning is specified in docs/IR.md and elaborated on in docs/Versioning.md
//
// To be compatible with both proto2 and proto3, we will use a version number
// that is not defined by the default value but an explicit enum number.
enum Version {
// proto3 requires the first enum value to be zero.
// We add this just to appease the compiler.
_START_VERSION = 0;
// The version field is always serialized and we will use it to store the
// version that the graph is generated from. This helps us set up version
// control.
// For the IR, we are using simple numbers starting with 0x00000001,
// which was the version we published on Oct 10, 2017.
IR_VERSION_2017_10_10 = 0x0000000000000001;
// IR_VERSION 2 published on Oct 30, 2017
// - Added type discriminator to AttributeProto to support proto3 users
IR_VERSION_2017_10_30 = 0x0000000000000002;
// IR VERSION 3 published on Nov 3, 2017
// - For operator versioning:
// - Added new message OperatorSetIdProto
// - Added opset_import in ModelProto
// - For vendor extensions, added domain in NodeProto
IR_VERSION_2017_11_3 = 0x0000000000000003;
// IR VERSION 4 published on Jan 22, 2019
// - Relax constraint that initializers should be a subset of graph inputs
// - Add type BFLOAT16
IR_VERSION_2019_1_22 = 0x0000000000000004;
// IR VERSION 5 published on March 18, 2019
// - Add message TensorAnnotation.
// - Add quantization annotation in GraphProto to map tensor with its scale and zero point quantization parameters.
IR_VERSION_2019_3_18 = 0x0000000000000005;
// IR VERSION 6 published on Sep 19, 2019
// - Add support for sparse tensor constants stored in model.
// - Add message SparseTensorProto
// - Add sparse initializers
IR_VERSION_2019_9_19 = 0x0000000000000006;
// IR VERSION 7 published on <TBD>
// - Add support to allow function body graph to rely on multiple external opreator sets.
// - Add a list to promote inference graph's initializers to global and
// mutable variables. Global variables are visible in all graphs of the
// stored models.
// - Add message TrainingInfoProto to store initialization
// method and training algorithm. The execution of TrainingInfoProto
// can modify the values of mutable variables.
// - Make inference graph callable from TrainingInfoProto via GraphCall operator.
IR_VERSION = 0x0000000000000007;
}
// Attributes
//
// A named attribute containing either singular float, integer, string, graph,
// and tensor values, or repeated float, integer, string, graph, and tensor values.
// An AttributeProto MUST contain the name field, and *only one* of the
// following content fields, effectively enforcing a C/C++ union equivalent.
message AttributeProto {
// Note: this enum is structurally identical to the OpSchema::AttrType
// enum defined in schema.h. If you rev one, you likely need to rev the other.
enum AttributeType {
UNDEFINED = 0;
FLOAT = 1;
INT = 2;
STRING = 3;
TENSOR = 4;
GRAPH = 5;
SPARSE_TENSOR = 11;
FLOATS = 6;
INTS = 7;
STRINGS = 8;
TENSORS = 9;
GRAPHS = 10;
SPARSE_TENSORS = 12;
}
// The name field MUST be present for this version of the IR.
string name = 1; // namespace Attribute
// if ref_attr_name is not empty, ref_attr_name is the attribute name in parent function.
// In this case, this AttributeProto does not contain data, and it's a reference of attribute
// in parent scope.
// NOTE: This should ONLY be used in function (sub-graph). It's invalid to be used in main graph.
string ref_attr_name = 21;
// A human-readable documentation for this attribute. Markdown is allowed.
string doc_string = 13;
// The type field MUST be present for this version of the IR.
// For 0.0.1 versions of the IR, this field was not defined, and
// implementations needed to use has_field heuristics to determine
// which value field was in use. For IR_VERSION 0.0.2 or later, this
// field MUST be set and match the f|i|s|t|... field in use. This
// change was made to accommodate proto3 implementations.
AttributeType type = 20; // discriminator that indicates which field below is in use
// Exactly ONE of the following fields must be present for this version of the IR
float f = 2; // float
int64 i = 3; // int
bytes s = 4; // UTF-8 string
TensorProto t = 5; // tensor value
GraphProto g = 6; // graph
SparseTensorProto sparse_tensor = 22; // sparse tensor value
// Do not use field below, it's deprecated.
// optional ValueProto v = 12; // value - subsumes everything but graph
repeated float floats = 7; // list of floats
repeated int64 ints = 8; // list of ints
repeated bytes strings = 9; // list of UTF-8 strings
repeated TensorProto tensors = 10; // list of tensors
repeated GraphProto graphs = 11; // list of graph
repeated SparseTensorProto sparse_tensors = 23; // list of sparse tensors
}
// Defines information on value, including the name, the type, and
// the shape of the value.
message ValueInfoProto {
// This field MUST be present in this version of the IR.
string name = 1; // namespace Value
// This field MUST be present in this version of the IR for
// inputs and outputs of the top-level graph.
TypeProto type = 2;
// A human-readable documentation for this value. Markdown is allowed.
string doc_string = 3;
}
// Nodes
//
// Computation graphs are made up of a DAG of nodes, which represent what is
// commonly called a "layer" or "pipeline stage" in machine learning frameworks.
//
// For example, it can be a node of type "Conv" that takes in an image, a filter
// tensor and a bias tensor, and produces the convolved output.
message NodeProto {
repeated string input = 1; // namespace Value
repeated string output = 2; // namespace Value
// An optional identifier for this node in a graph.
// This field MAY be absent in ths version of the IR.
string name = 3; // namespace Node
// The symbolic identifier of the Operator to execute.
string op_type = 4; // namespace Operator
// The domain of the OperatorSet that specifies the operator named by op_type.
string domain = 7; // namespace Domain
// Additional named attributes.
repeated AttributeProto attribute = 5;
// A human-readable documentation for this node. Markdown is allowed.
string doc_string = 6;
}
// Training information
// TrainingInfoProto stores information for training a model.
// In particular, this defines two functionalities: an initialization-step
// and a training-algorithm-step. Initialization resets the model
// back to its original state as if no training has been consumed.
// Training algorithm improves the model based on input data.
//
// The semantics of the initialization-step is that the initializers
// in ModelProto.graph and in TrainingInfoProto.algorithm are first
// initialized as specified by the initializers in the graph, and then
// updated by the "initialization_binding" in every instance in
// ModelProto.training_info.
//
// The field "algorithm" defines a computation graph which represents a
// training algorithm's step. After the execution of a
// TrainingInfoProto.algorithm, the initializers specified by "update_binding"
// may be immediately updated. If the targeted training algorithm contains
// consecutive update stages (such as block coordinate descent methods),
// the user needs to create a TrainingInfoProto for each stage.
message TrainingInfoProto {
// This field describes a graph to compute the initial tensors
// upon starting the training process. Initialization graph has no input
// and can have multiple outputs. Usually, trainable tensors in neural
// networks are randomly initialized. To achieve that, for each tensor,
// the user can put a random number operator such as RandomNormal or
// RandomUniform in TrainingInfoProto.initialization.node and assign its
// random output to the specific tensor using "initialization_binding".
// This graph can also set the initializers in "algorithm" in the same
// TrainingInfoProto; a use case is resetting the number of training
// iteration to zero.
//
// By default, this field is an empty graph and its evaluation does not
// produce any output.
GraphProto initialization = 1;
// This field represents a training algorithm step. Given required inputs,
// it computes outputs to update initializers in its own or inference graph's
// initializer lists. In general, this graph contains loss node, gradient node,
// optimizer node, increment of iteration count, and some calls to the inference
// graph.
//
// The field algorithm.node is the only place the user can use GraphCall
// operator. The only callable graph is the one stored in ModelProto.graph.
//
// By default, this field is an empty graph and its evaluation does not
// produce any output.
GraphProto algorithm = 2;
// This field specifies the bindings from the outputs of "initialization" to
// some initializers in "ModelProto.graph.initializer" and
// the "algorithm.initializer" in the same TrainingInfoProto.
// See "update_binding" below for details.
//
// By default, this field is empty and no initializer would be changed
// by the execution of "initialization".
repeated StringStringEntryProto initialization_binding = 3;
// Gradient-based training is usually an iterative procedure. In one gradient
// descent iteration, we apply
//
// x = x - r * g
//
// where "x" is the optimized tensor, "r" stands for learning rate, and "g" is
// gradient of "x" with respect to a chosen loss. To avoid adding assignments
// into the training graph, we split the update equation into
//
// y = x - r * g
// x = y
//
// The user needs to save "y = x - r * g" into TrainingInfoProto.algorithm. To
// tell that "y" should be assigned to "x", the field "update_binding" may
// contain a key-value pair of strings, "x" (key of StringStringEntryProto)
// and "y" (value of StringStringEntryProto).
// For a neural network with multiple trainable (mutable) tensors, there can
// be multiple key-value pairs in "update_binding".
//
// The initializers appears as keys in "update_binding" are considered
// mutable and globally-visible variables. This implies some behaviors
// as described below.
//
// 1. We have only unique keys in all "update_binding"s so that two global
// variables may not have the same name. This ensures that one
// global variable is assigned up to once.
// 2. The keys must appear in names of "ModelProto.graph.initializer" or
// "TrainingInfoProto.algorithm.initializer".
// 3. The values must be output names of "algorithm".
// 4. If an optional input of a graph is omitted when using GraphCall, the
// global variable with the same name may be used.
// 5. When using GraphCall, the users always can pass values to optional
// inputs of the called graph even if the associated initializers appears
// as keys in "update_binding"s.
// 6. The graphs in TrainingInfoProto's can use global variables as
// their operator inputs.
// 7. Mutable variables are initialized to the value specified by the
// corresponding initializer, and then potentially updated by
// "initializer_binding"s and "update_binding"s in "TrainingInfoProto"s.
//
// This field usually contains names of trainable tensors
// (in ModelProto.graph), optimizer states such as momentums in advanced
// stochastic gradient methods (in TrainingInfoProto.graph),
// and number of training iterations (in TrainingInfoProto.graph).
//
// By default, this field is empty and no initializer would be changed
// by the execution of "algorithm".
repeated StringStringEntryProto update_binding = 4;
}
// Models
//
// ModelProto is a top-level file/container format for bundling a ML model and
// associating its computation graph with metadata.
//
// The semantics of the model are described by the associated GraphProto's.
message ModelProto {
// The version of the IR this model targets. See Version enum above.
// This field MUST be present.
int64 ir_version = 1;
// The OperatorSets this model relies on.
// All ModelProtos MUST have at least one entry that
// specifies which version of the ONNX OperatorSet is
// being imported.
//
// All nodes in the ModelProto's graph will bind against the operator
// with the same-domain/same-op_type operator with the HIGHEST version
// in the referenced operator sets.
repeated OperatorSetIdProto opset_import = 8;
// The name of the framework or tool used to generate this model.
// This field SHOULD be present to indicate which implementation/tool/framework
// emitted the model.
string producer_name = 2;
// The version of the framework or tool used to generate this model.
// This field SHOULD be present to indicate which implementation/tool/framework
// emitted the model.
string producer_version = 3;
// Domain name of the model.
// We use reverse domain names as name space indicators. For example:
// `com.facebook.fair` or `com.microsoft.cognitiveservices`
//
// Together with `model_version` and GraphProto.name, this forms the unique identity of
// the graph.
string domain = 4;
// The version of the graph encoded. See Version enum below.
int64 model_version = 5;
// A human-readable documentation for this model. Markdown is allowed.
string doc_string = 6;
// The parameterized graph that is evaluated to execute the model.
GraphProto graph = 7;
// kezhan: This field is not in ONNX, and will be pushed into ONNX with good use cases in microsoft.
repeated FunctionProto functions = 100;
// Named metadata values; keys should be distinct.
repeated StringStringEntryProto metadata_props = 14;
// Training-specific information. Sequentially executing all stored
// `TrainingInfoProto.algorithm`s and assigning their outputs following
// the corresponding `TrainingInfoProto.update_binding`s is one training
// iteration. Similarly, to initialize the model
// (as if training hasn't happened), the user should sequentially execute
// all stored `TrainingInfoProto.initialization`s and assigns their outputs
// using `TrainingInfoProto.initialization_binding`s.
//
// If this field is empty, the training behavior of the model is undefined.
repeated TrainingInfoProto training_info = 20;
};
// StringStringEntryProto follows the pattern for cross-proto-version maps.
// See https://developers.google.com/protocol-buffers/docs/proto3#maps
message StringStringEntryProto {
string key = 1;
string value= 2;
};
message TensorAnnotation {
string tensor_name = 1;
// <key, value> pairs to annotate tensor specified by <tensor_name> above.
// The keys used in the mapping below must be pre-defined in ONNX spec.
// For example, for 8-bit linear quantization case, 'SCALE_TENSOR', 'ZERO_POINT_TENSOR' will be pre-defined as
// quantization parameter keys.
repeated StringStringEntryProto quant_parameter_tensor_names = 2;
}
// Graphs
//
// A graph defines the computational logic of a model and is comprised of a parameterized
// list of nodes that form a directed acyclic graph based on their inputs and outputs.
// This is the equivalent of the "network" or "graph" in many deep learning
// frameworks.
message GraphProto {
// The nodes in the graph, sorted topologically.
repeated NodeProto node = 1;
// The name of the graph.
string name = 2; // namespace Graph
// A list of named tensor values, used to specify constant inputs of the graph.
// Each TensorProto entry must have a distinct name (within the list) that
// MAY also appear in the input list.
repeated TensorProto initializer = 5;
// Initializers (see above) stored in sparse format.
repeated SparseTensorProto sparse_initializer = 15;
// A human-readable documentation for this graph. Markdown is allowed.
string doc_string = 10;
// The inputs and outputs of the graph.
repeated ValueInfoProto input = 11;
repeated ValueInfoProto output = 12;
// Information for the values in the graph. The ValueInfoProto.name's
// must be distinct. It is optional for a value to appear in value_info list.
repeated ValueInfoProto value_info = 13;
// This field carries information to indicate the mapping among a tensor and its
// quantization parameter tensors. For example:
// For tensor 'a', it may have {'SCALE_TENSOR', 'a_scale'} and {'ZERO_POINT_TENSOR', 'a_zero_point'} annotated,
// which means, tensor 'a_scale' and tensor 'a_zero_point' are scale and zero point of tensor 'a' in the model.
repeated TensorAnnotation quantization_annotation = 14;
// DO NOT USE the following fields, they were deprecated from earlier versions.
// repeated string input = 3;
// repeated string output = 4;
// optional int64 ir_version = 6;
// optional int64 producer_version = 7;
// optional string producer_tag = 8;
// optional string domain = 9;
}
// Tensors
//
// A serialized tensor value.
message TensorProto {
enum DataType {
UNDEFINED = 0;
// Basic types.
FLOAT = 1; // float
UINT8 = 2; // uint8_t
INT8 = 3; // int8_t
UINT16 = 4; // uint16_t
INT16 = 5; // int16_t
INT32 = 6; // int32_t
INT64 = 7; // int64_t
STRING = 8; // string
BOOL = 9; // bool
// IEEE754 half-precision floating-point format (16 bits wide).
// This format has 1 sign bit, 5 exponent bits, and 10 mantissa bits.
FLOAT16 = 10;
DOUBLE = 11;
UINT32 = 12;
UINT64 = 13;
COMPLEX64 = 14; // complex with float32 real and imaginary components
COMPLEX128 = 15; // complex with float64 real and imaginary components
// Non-IEEE floating-point format based on IEEE754 single-precision
// floating-point number truncated to 16 bits.
// This format has 1 sign bit, 8 exponent bits, and 7 mantissa bits.
BFLOAT16 = 16;
// Future extensions go here.
}
// The shape of the tensor.
repeated int64 dims = 1;
// The data type of the tensor.
// This field MUST have a valid TensorProto.DataType value
int32 data_type = 2;
// For very large tensors, we may want to store them in chunks, in which
// case the following fields will specify the segment that is stored in
// the current TensorProto.
message Segment {
int64 begin = 1;
int64 end = 2;
}
Segment segment = 3;
// Tensor content must be organized in row-major order.
//
// Depending on the data_type field, exactly one of the fields below with
// name ending in _data is used to store the elements of the tensor.
// For float and complex64 values
// Complex64 tensors are encoded as a single array of floats,
// with the real components appearing in odd numbered positions,
// and the corresponding imaginary component appearing in the
// subsequent even numbered position. (e.g., [1.0 + 2.0i, 3.0 + 4.0i]
// is encoded as [1.0, 2.0 ,3.0 ,4.0]
// When this field is present, the data_type field MUST be FLOAT or COMPLEX64.
repeated float float_data = 4 [packed = true];
// For int32, uint8, int8, uint16, int16, bool, and float16 values
// float16 values must be bit-wise converted to an uint16_t prior
// to writing to the buffer.
// When this field is present, the data_type field MUST be
// INT32, INT16, INT8, UINT16, UINT8, BOOL, or FLOAT16
repeated int32 int32_data = 5 [packed = true];
// For strings.
// Each element of string_data is a UTF-8 encoded Unicode
// string. No trailing null, no leading BOM. The protobuf "string"
// scalar type is not used to match ML community conventions.
// When this field is present, the data_type field MUST be STRING
repeated bytes string_data = 6;
// For int64.
// When this field is present, the data_type field MUST be INT64
repeated int64 int64_data = 7 [packed = true];
// Optionally, a name for the tensor.
string name = 8; // namespace Value
// A human-readable documentation for this tensor. Markdown is allowed.
string doc_string = 12;
// Serializations can either use one of the fields above, or use this
// raw bytes field. The only exception is the string case, where one is
// required to store the content in the repeated bytes string_data field.
//
// When this raw_data field is used to store tensor value, elements MUST
// be stored in as fixed-width, little-endian order.
// Floating-point data types MUST be stored in IEEE 754 format.
// Complex64 elements must be written as two consecutive FLOAT values, real component first.
// Complex128 elements must be written as two consecutive DOUBLE values, real component first.
// Boolean type MUST be written one byte per tensor element (00000001 for true, 00000000 for false).
//
// Note: the advantage of specific field rather than the raw_data field is
// that in some cases (e.g. int data), protobuf does a better packing via
// variable length storage, and may lead to smaller binary footprint.
// When this field is present, the data_type field MUST NOT be STRING or UNDEFINED
bytes raw_data = 9;
// Data can be stored inside the protobuf file using type-specific fields or raw_data.
// Alternatively, raw bytes data can be stored in an external file, using the external_data field.
// external_data stores key-value pairs describing data location. Recognized keys are:
// - "location" (required) - POSIX filesystem path relative to the directory where the ONNX
// protobuf model was stored
// - "offset" (optional) - position of byte at which stored data begins. Integer stored as string.
// Offset values SHOULD be multiples 4096 (page size) to enable mmap support.
// - "length" (optional) - number of bytes containing data. Integer stored as string.
// - "checksum" (optional) - SHA1 digest of file specified in under 'location' key.
repeated StringStringEntryProto external_data = 13;
// Location of the data for this tensor. MUST be one of:
// - DEFAULT - data stored inside the protobuf message. Data is stored in raw_data (if set) otherwise in type-specified field.
// - EXTERNAL - data stored in an external location as described by external_data field.
enum DataLocation {
DEFAULT = 0;
EXTERNAL = 1;
}
// If value not set, data is stored in raw_data (if set) otherwise in type-specified field.
DataLocation data_location = 14;
// For double
// Complex128 tensors are encoded as a single array of doubles,
// with the real components appearing in odd numbered positions,
// and the corresponding imaginary component appearing in the
// subsequent even numbered position. (e.g., [1.0 + 2.0i, 3.0 + 4.0i]
// is encoded as [1.0, 2.0 ,3.0 ,4.0]
// When this field is present, the data_type field MUST be DOUBLE or COMPLEX128
repeated double double_data = 10 [packed = true];
// For uint64 and uint32 values
// When this field is present, the data_type field MUST be
// UINT32 or UINT64
repeated uint64 uint64_data = 11 [packed = true];
}
// A serialized sparse-tensor value
message SparseTensorProto {
// The sequence of non-default values are encoded as a tensor of shape [NNZ].
// The default-value is zero for numeric tensors, and empty-string for string tensors.
TensorProto values = 1;
// The indices of the non-default values, which may be stored in one of two formats.
// (a) Indices can be a tensor of shape [NNZ, rank] with the [i,j]-th value
// corresponding to the j-th index of the i-th value (in the values tensor).
// (b) Indices can be a tensor of shape [NNZ], in which case the i-th value
// must be the linearized-index of the i-th value (in the values tensor).
// The linearized-index can be converted into an index tuple (k_1,...,k_rank)
// using the shape provided below.
// The indices must appear in ascending order without duplication.
// In the first format, the ordering is lexicographic-ordering:
// e.g., index-value [1,4] must appear before [2,1]
TensorProto indices = 2;
// The shape of the underlying dense-tensor: [dim_1, dim_2, ... dim_rank]
repeated int64 dims = 3;
}
// Defines a tensor shape. A dimension can be either an integer value
// or a symbolic variable. A symbolic variable represents an unknown
// dimension.
message TensorShapeProto {
message Dimension {
oneof value {
int64 dim_value = 1;
string dim_param = 2; // namespace Shape
};
// Standard denotation can optionally be used to denote tensor
// dimensions with standard semantic descriptions to ensure
// that operations are applied to the correct axis of a tensor.
// Refer to https://github.com/onnx/onnx/blob/master/docs/DimensionDenotation.md#denotation-definition
// for pre-defined dimension denotations.
string denotation = 3;
};
repeated Dimension dim = 1;
}
// Types
//
// The standard ONNX data types.
message TypeProto {
message Tensor {
// This field MUST NOT have the value of UNDEFINED
// This field MUST have a valid TensorProto.DataType value
// This field MUST be present for this version of the IR.
int32 elem_type = 1;
TensorShapeProto shape = 2;
}
// repeated T
message Sequence {
// The type and optional shape of each element of the sequence.
// This field MUST be present for this version of the IR.
TypeProto elem_type = 1;
};
// map<K,V>
message Map {
// This field MUST have a valid TensorProto.DataType value
// This field MUST be present for this version of the IR.
// This field MUST refer to an integral type ([U]INT{8|16|32|64}) or STRING
int32 key_type = 1;
// This field MUST be present for this version of the IR.
TypeProto value_type = 2;
};
message SparseTensor {
// This field MUST NOT have the value of UNDEFINED
// This field MUST have a valid TensorProto.DataType value
// This field MUST be present for this version of the IR.
int32 elem_type = 1;
TensorShapeProto shape = 2;
}
message Opaque {
// When missing, the domain is the same as the model's.
string domain = 1;
// The name is optional but significant when provided.
string name = 2;
// parameters that help defining the type
// DEPRECATED do not use.
// repeated TypeProto parameters = 3;
}
oneof value {
// The type of a tensor.
Tensor tensor_type = 1;
// NOTE: DNN-only implementations of ONNX MAY elect to not support non-tensor values
// as input and output to graphs and nodes. These types are needed to naturally
// support classical ML operators. DNN operators SHOULD restrict their input
// and output types to tensors.
// The type of a sequence.
Sequence sequence_type = 4;
// The type of a map.
Map map_type = 5;
SparseTensor sparse_tensor_type = 8;
Opaque opaque_type = 7;
}
// An optional denotation can be used to denote the whole
// type with a standard semantic description as to what is
// stored inside. Refer to https://github.com/onnx/onnx/blob/master/docs/TypeDenotation.md#type-denotation-definition
// for pre-defined type denotations.
string denotation = 6;
}
// Operator Sets
//
// OperatorSets are uniquely identified by a (domain, opset_version) pair.
message OperatorSetIdProto {
// The domain of the operator set being identified.
// The empty string ("") or absence of this field implies the operator
// set that is defined as part of the ONNX specification.
// This field MUST be present in this version of the IR when referring to any other operator set.
string domain = 1;
// The version of the operator set being identified.
// This field MUST be present in this version of the IR.
int64 version = 2;
}
// Operator/function status.
enum OperatorStatus {
EXPERIMENTAL = 0;
STABLE = 1;
}
message FunctionProto {
// The name of the function, similar usage of op_type in OperatorProto.
string name = 1;
// The first version of a function set which contains this function.
// When there's any breaking change for this function, the function set
// contains the function needs to bump its version, and since_version of
// the updated function will be changed to the updated function set version.
int64 since_version = 2;
// This field indicates whether the syntax, semantics, or presence
// of this function is in an experimental or stable stage. Once an
// function is published as STABLE, its syntax and semantics MUST NOT
// change in subsequent versions of the operator set.
// When a function is published as EXPERIMENTAL, the syntax and semantics
// of the function MAY change across operator set versions.
// Functions "become" stable by deprecating the experimental version and
// introducing a new stable function with the same name.
OperatorStatus status = 3;
// The inputs and outputs of the function.
repeated string input = 4;
repeated string output = 5;
// The attributes of the function.
repeated string attribute= 6;
// The nodes in the function.
repeated NodeProto node = 7;
// A human-readable documentation for this function. Markdown is allowed.
string doc_string = 8;
// The OperatorSets this function body (graph) relies on.
// A FunctionProto body (graph) may implicitly rely on the OperatorSet that
// this function belongs to. It can also explicitly rely on more OperatorSets
// with this field specified.
//
// All nodes in the function body (graph) will bind against the operator
// with the same-domain/same-op_type operator with the HIGHEST version
// in the referenced operator sets. This means at most one version can be relied
// for one domain.
repeated OperatorSetIdProto opset_import = 9;
}

131
onnxruntime/core/protobuf/onnx-operators-ml.proto
View file

@ -1,131 +0,0 @@
//
// WARNING: This file is automatically generated! Please edit onnx.in.proto.
//
// Copyright (c) ONNX Project Contributors.
// Licensed under the MIT license.
syntax = "proto2";
package onnx;
import "onnx-ml.proto";
//
// This file contains the proto definitions for OperatorSetProto and
// OperatorProto. OperatorSetProtos are used to describe a versioned
// set of operators that can be used by a ModelProto.
//
// Like ModelProto, OperatorSetProto is defined as a top-level file/wire
// format, however their usage is different.
//
// ModelProto files are used to describe executable graphs that can be
// executed directly by a framework, runtime, or engine.
//
// OperatorSetProto files are used to describe a set of operators that are
// available in a given environment. The file TBD.TBD is the OperatorSetProto
// that describes the ONNX standard operators.
//
// An OperatorProto represents the immutable specification of the signature
// and semantics of an operator.
//
// Operators are declared as part of an OperatorSet, which also defines the
// domain name for the set.
//
// Operators are uniquely identified by a three part identifier
// (domain, op_type, since_version)
// where
// *domain* is the domain of an operator set that
// contains this operator specification.
//
// *op_type* is the name of the operator as referenced by a
// NodeProto.op_type
//
// *since_version* is the version of the operator set that
// this operator was initially declared in.
//
message OperatorProto {
// The name of the operator within a domain.
// This field MUST be present in this version of the IR.
optional string op_type = 1;
// The version of the operator set that first introduced this
// operator. This value MUST be the same value as the
// opset_version of the operator set that first published this operator.
// Subsequent versions of the operator set MUST NOT alter the signature
// or semantics of the operator once published as STABLE.
// This field MUST be present in this version of the IR.
optional int64 since_version = 2;
// This field indicates whether the syntax, semantics, or presence
// of this operator is in an experimental or stable stage. Once an
// operator is published as STABLE, it's syntax and semantics MUST NOT
// change in subsequent versions of the operator set.
// When an operator is published as EXPERIMENTAL, the syntax and semantics
// of the operator MAY change across operator set versions.
// Operators "become" stable by deprecating the experimental version and
// introducing a new stable operator with the same op_type.
optional OperatorStatus status = 3;
// Eventually we will declare the signature of the operator here
// A human-readable documentation for this operator. Markdown is allowed.
optional string doc_string = 10;
}
// An OperatorSetProto represents an immutable set of immutable operator
// specifications.
//
// The domain of the set (OperatorSetProto.domain) is a reverse-DNS name
// that disambiguates operator sets defined by independent entities.
//
// The version of the set (opset_version) is a monotonically increasing
// integer that indicates changes to the membership of the operator set.
//
//
// Operator sets are uniquely identified by a two part identifier (domain, opset_version)
//
// Like ModelProto, OperatorSetProto is intended as a top-level file/wire format,
// and thus has the standard format headers in addition to the operator set information.
//
message OperatorSetProto {
// All OperatorSetProtos start with a distingushed byte sequence to disambiguate
// protobuf files containing OperatorSets from other content.
// This field MUST be "ONNXOPSET"
// This field MUST be present in this version of the IR
optional string magic = 1;
// All OperatorSetProtos indicate the version of the IR syntax and semantics
// they adhere to. It is always IR_VERSION.
// This field MUST be present in this version of the IR
optional int64 ir_version = 2;
// The prerelease component of the SemVer of the IR.
// This field MAY be absent in this version of the IR
optional string ir_version_prerelease = 3;
// The build metadata component of the SemVer of the IR.
// This field MAY be absent in this version of the IR
optional string ir_build_metadata = 7;
// Domain name of the operator set, in reverse DNS form (e.g., com.acme.dnnops).
optional string domain = 4;
// The version of the set of operators. This is a simple int value
// that is monotonically increasing as new versions of operator set
// are published. All operators in this set MUST have version
// numbers no greater than opset_version.
optional int64 opset_version = 5;
// A human-readable documentation for this set of operators. Markdown is allowed.
optional string doc_string = 6;
// The operators specified by this operator set.
// The (name, version) MUST be unique across all OperatorProtos in operator
repeated OperatorProto operator = 8;
// The functions specified by this operator set.
// The (name, version) MUST be unique across all OperatorProtos/FunctionProtos in operator/functions
repeated FunctionProto functions = 9;
}

133
onnxruntime/core/protobuf/onnx-operators-ml.proto3
View file

@ -1,133 +0,0 @@
//
// WARNING: This file is automatically generated! Please edit onnx.in.proto.
//
// Copyright (c) ONNX Project Contributors.
// Licensed under the MIT license.
syntax = "proto3";
package onnx;
import "onnx-ml.proto3";
//
// This file contains the proto definitions for OperatorSetProto and
// OperatorProto. OperatorSetProtos are used to describe a versioned
// set of operators that can be used by a ModelProto.
//
// Like ModelProto, OperatorSetProto is defined as a top-level file/wire
// format, however their usage is different.
//
// ModelProto files are used to describe executable graphs that can be
// executed directly by a framework, runtime, or engine.
//
// OperatorSetProto files are used to describe a set of operators that are
// available in a given environment. The file TBD.TBD is the OperatorSetProto
// that describes the ONNX standard operators.
//
// An OperatorProto represents the immutable specification of the signature
// and semantics of an operator.
//
// Operators are declared as part of an OperatorSet, which also defines the
// domain name for the set.
//
// Operators are uniquely identified by a three part identifier
// (domain, op_type, since_version)
// where
// *domain* is the domain of an operator set that
// contains this operator specification.
//
// *op_type* is the name of the operator as referenced by a
// NodeProto.op_type
//
// *since_version* is the version of the operator set that
// this operator was initially declared in.
//
message OperatorProto {
// The name of the operator within a domain.
// This field MUST be present in this version of the IR.
string op_type = 1;
// The version of the operator set that first introduced this
// operator. This value MUST be the same value as the
// opset_version of the operator set that first published this operator.
// Subsequent versions of the operator set MUST NOT alter the signature
// or semantics of the operator once published as STABLE.
// This field MUST be present in this version of the IR.
int64 since_version = 2;
// This field indicates whether the syntax, semantics, or presence
// of this operator is in an experimental or stable stage. Once an
// operator is published as STABLE, it's syntax and semantics MUST NOT
// change in subsequent versions of the operator set.
// When an operator is published as EXPERIMENTAL, the syntax and semantics
// of the operator MAY change across operator set versions.
// Operators "become" stable by deprecating the experimental version and
// introducing a new stable operator with the same op_type.
OperatorStatus status = 3;
// Eventually we will declare the signature of the operator here
// A human-readable documentation for this operator. Markdown is allowed.
string doc_string = 10;
}
// An OperatorSetProto represents an immutable set of immutable operator
// specifications.
//
// The domain of the set (OperatorSetProto.domain) is a reverse-DNS name
// that disambiguates operator sets defined by independent entities.
//
// The version of the set (opset_version) is a monotonically increasing
// integer that indicates changes to the membership of the operator set.
//
//
// Operator sets are uniquely identified by a two part identifier (domain, opset_version)
//
// Like ModelProto, OperatorSetProto is intended as a top-level file/wire format,
// and thus has the standard format headers in addition to the operator set information.
//
message OperatorSetProto {
// All OperatorSetProtos start with a distingushed byte sequence to disambiguate
// protobuf files containing OperatorSets from other content.
// This field MUST be "ONNXOPSET"
// This field MUST be present in this version of the IR
string magic = 1;
// All OperatorSetProtos indicate the version of the IR syntax and semantics
// they adhere to. It is always IR_VERSION.
// This field MUST be present in this version of the IR
int64 ir_version = 2;
// The prerelease component of the SemVer of the IR.
// This field MAY be absent in this version of the IR
string ir_version_prerelease = 3;
// The build metadata component of the SemVer of the IR.
// This field MAY be absent in this version of the IR
string ir_build_metadata = 7;
// Domain name of the operator set, in reverse DNS form (e.g., com.acme.dnnops).
string domain = 4;
// The version of the set of operators. This is a simple int value
// that is monotonically increasing as new versions of operator set
// are published. All operators in this set MUST have version
// numbers no greater than opset_version.
int64 opset_version = 5;
// A human-readable documentation for this set of operators. Markdown is allowed.
string doc_string = 6;
// The operators specified by this operator set.
// The (name, version) MUST be unique across all OperatorProtos in operator
repeated OperatorProto operator = 8;
// The functions specified by this operator set.
// The (name, version) MUST be unique across all OperatorProtos/FunctionProtos in operator/functions
repeated FunctionProto functions = 9;
}

130
onnxruntime/core/protobuf/onnx-operators.in.proto
View file

@ -1,130 +0,0 @@
// Copyright (c) ONNX Project Contributors.
// Licensed under the MIT license.
syntax = "proto2";
package {PACKAGE_NAME};
// #if ONNX-ML
import "onnx-ml.proto";
// #else
import "onnx.proto";
// #endif
//
// This file contains the proto definitions for OperatorSetProto and
// OperatorProto. OperatorSetProtos are used to describe a versioned
// set of operators that can be used by a ModelProto.
//
// Like ModelProto, OperatorSetProto is defined as a top-level file/wire
// format, however their usage is different.
//
// ModelProto files are used to describe executable graphs that can be
// executed directly by a framework, runtime, or engine.
//
// OperatorSetProto files are used to describe a set of operators that are
// available in a given environment. The file TBD.TBD is the OperatorSetProto
// that describes the ONNX standard operators.
//
// An OperatorProto represents the immutable specification of the signature
// and semantics of an operator.
//
// Operators are declared as part of an OperatorSet, which also defines the
// domain name for the set.
//
// Operators are uniquely identified by a three part identifier
// (domain, op_type, since_version)
// where
// *domain* is the domain of an operator set that
// contains this operator specification.
//
// *op_type* is the name of the operator as referenced by a
// NodeProto.op_type
//
// *since_version* is the version of the operator set that
// this operator was initially declared in.
//
message OperatorProto {
// The name of the operator within a domain.
// This field MUST be present in this version of the IR.
optional string op_type = 1;
// The version of the operator set that first introduced this
// operator. This value MUST be the same value as the
// opset_version of the operator set that first published this operator.
// Subsequent versions of the operator set MUST NOT alter the signature
// or semantics of the operator once published as STABLE.
// This field MUST be present in this version of the IR.
optional int64 since_version = 2;
// This field indicates whether the syntax, semantics, or presence
// of this operator is in an experimental or stable stage. Once an
// operator is published as STABLE, it's syntax and semantics MUST NOT
// change in subsequent versions of the operator set.
// When an operator is published as EXPERIMENTAL, the syntax and semantics
// of the operator MAY change across operator set versions.
// Operators "become" stable by deprecating the experimental version and
// introducing a new stable operator with the same op_type.
optional OperatorStatus status = 3;
// Eventually we will declare the signature of the operator here
// A human-readable documentation for this operator. Markdown is allowed.
optional string doc_string = 10;
}
// An OperatorSetProto represents an immutable set of immutable operator
// specifications.
//
// The domain of the set (OperatorSetProto.domain) is a reverse-DNS name
// that disambiguates operator sets defined by independent entities.
//
// The version of the set (opset_version) is a monotonically increasing
// integer that indicates changes to the membership of the operator set.
//
//
// Operator sets are uniquely identified by a two part identifier (domain, opset_version)
//
// Like ModelProto, OperatorSetProto is intended as a top-level file/wire format,
// and thus has the standard format headers in addition to the operator set information.
//
message OperatorSetProto {
// All OperatorSetProtos start with a distingushed byte sequence to disambiguate
// protobuf files containing OperatorSets from other content.
// This field MUST be "ONNXOPSET"
// This field MUST be present in this version of the IR
optional string magic = 1;
// All OperatorSetProtos indicate the version of the IR syntax and semantics
// they adhere to. It is always IR_VERSION.
// This field MUST be present in this version of the IR
optional int64 ir_version = 2;
// The prerelease component of the SemVer of the IR.
// This field MAY be absent in this version of the IR
optional string ir_version_prerelease = 3;
// The build metadata component of the SemVer of the IR.
// This field MAY be absent in this version of the IR
optional string ir_build_metadata = 7;
// Domain name of the operator set, in reverse DNS form (e.g., com.acme.dnnops).
optional string domain = 4;
// The version of the set of operators. This is a simple int value
// that is monotonically increasing as new versions of operator set
// are published. All operators in this set MUST have version
// numbers no greater than opset_version.
optional int64 opset_version = 5;
// A human-readable documentation for this set of operators. Markdown is allowed.
optional string doc_string = 6;
// The operators specified by this operator set.
// The (name, version) MUST be unique across all OperatorProtos in operator
repeated OperatorProto operator = 8;
// The functions specified by this operator set.
// The (name, version) MUST be unique across all OperatorProtos/FunctionProtos in operator/functions
repeated FunctionProto functions = 9;
}

776
onnxruntime/core/protobuf/onnx.in.proto
View file

@ -1,776 +0,0 @@
// Copyright (c) ONNX Project Contributors.
// Licensed under the MIT license.
syntax = "proto2";
package {PACKAGE_NAME};
// Overview
//
// ONNX is an open specification that is comprised of the following components:
//
// 1) A definition of an extensible computation graph model.
// 2) Definitions of standard data types.
// 3) Definitions of built-in operators.
//
// This document describes the syntax of models and their computation graphs,
// as well as the standard data types. Together, they are referred to as the ONNX
// Intermediate Representation, or 'IR' for short.
//
// The normative semantic specification of the ONNX IR is found in docs/IR.md.
// Definitions of the built-in neural network operators may be found in docs/Operators.md.
// #if ONNX-ML
// Definitions of the built-in classical machine learning operators may be found in
// docs/Operators-ml.md.
// #endif
// Notes
//
// Release
//
// We are still in the very early stage of defining ONNX. The current
// version of ONNX is a starting point. While we are actively working
// towards a complete spec, we would like to get the community involved
// by sharing our working version of ONNX.
//
// Protobuf compatibility
//
// To simplify framework compatibility, ONNX is defined using the subset of protobuf
// that is compatible with both protobuf v2 and v3. This means that we do not use any
// protobuf features that are only available in one of the two versions.
//
// Here are the most notable contortions we have to carry out to work around
// these limitations:
//
// - No 'map' (added protobuf 3.0). We instead represent mappings as lists
// of key-value pairs, where order does not matter and duplicates
// are not allowed.
// Versioning
//
// ONNX versioning is specified in docs/IR.md and elaborated on in docs/Versioning.md
//
// To be compatible with both proto2 and proto3, we will use a version number
// that is not defined by the default value but an explicit enum number.
enum Version {
// proto3 requires the first enum value to be zero.
// We add this just to appease the compiler.
_START_VERSION = 0;
// The version field is always serialized and we will use it to store the
// version that the graph is generated from. This helps us set up version
// control.
// For the IR, we are using simple numbers starting with 0x00000001,
// which was the version we published on Oct 10, 2017.
IR_VERSION_2017_10_10 = 0x0000000000000001;
// IR_VERSION 2 published on Oct 30, 2017
// - Added type discriminator to AttributeProto to support proto3 users
IR_VERSION_2017_10_30 = 0x0000000000000002;
// IR VERSION 3 published on Nov 3, 2017
// - For operator versioning:
// - Added new message OperatorSetIdProto
// - Added opset_import in ModelProto
// - For vendor extensions, added domain in NodeProto
IR_VERSION_2017_11_3 = 0x0000000000000003;
// IR VERSION 4 published on Jan 22, 2019
// - Relax constraint that initializers should be a subset of graph inputs
// - Add type BFLOAT16
IR_VERSION_2019_1_22 = 0x0000000000000004;
// IR VERSION 5 published on March 18, 2019
// - Add message TensorAnnotation.
// - Add quantization annotation in GraphProto to map tensor with its scale and zero point quantization parameters.
IR_VERSION_2019_3_18 = 0x0000000000000005;
// IR VERSION 6 published on Sep 19, 2019
// - Add support for sparse tensor constants stored in model.
// - Add message SparseTensorProto
// - Add sparse initializers
IR_VERSION_2019_9_19 = 0x0000000000000006;
// IR VERSION 7 published on <TBD>
// - Add support to allow function body graph to rely on multiple external opreator sets.
// - Add a list to promote inference graph's initializers to global and
// mutable variables. Global variables are visible in all graphs of the
// stored models.
// - Add message TrainingInfoProto to store initialization
// method and training algorithm. The execution of TrainingInfoProto
// can modify the values of mutable variables.
// - Make inference graph callable from TrainingInfoProto via GraphCall operator.
IR_VERSION = 0x0000000000000007;
}
// Attributes
//
// A named attribute containing either singular float, integer, string, graph,
// and tensor values, or repeated float, integer, string, graph, and tensor values.
// An AttributeProto MUST contain the name field, and *only one* of the
// following content fields, effectively enforcing a C/C++ union equivalent.
message AttributeProto {
// Note: this enum is structurally identical to the OpSchema::AttrType
// enum defined in schema.h. If you rev one, you likely need to rev the other.
enum AttributeType {
UNDEFINED = 0;
FLOAT = 1;
INT = 2;
STRING = 3;
TENSOR = 4;
GRAPH = 5;
SPARSE_TENSOR = 11;
FLOATS = 6;
INTS = 7;
STRINGS = 8;
TENSORS = 9;
GRAPHS = 10;
SPARSE_TENSORS = 12;
}
// The name field MUST be present for this version of the IR.
optional string name = 1; // namespace Attribute
// if ref_attr_name is not empty, ref_attr_name is the attribute name in parent function.
// In this case, this AttributeProto does not contain data, and it's a reference of attribute
// in parent scope.
// NOTE: This should ONLY be used in function (sub-graph). It's invalid to be used in main graph.
optional string ref_attr_name = 21;
// A human-readable documentation for this attribute. Markdown is allowed.
optional string doc_string = 13;
// The type field MUST be present for this version of the IR.
// For 0.0.1 versions of the IR, this field was not defined, and
// implementations needed to use has_field heuristics to determine
// which value field was in use. For IR_VERSION 0.0.2 or later, this
// field MUST be set and match the f|i|s|t|... field in use. This
// change was made to accommodate proto3 implementations.
optional AttributeType type = 20; // discriminator that indicates which field below is in use
// Exactly ONE of the following fields must be present for this version of the IR
optional float f = 2; // float
optional int64 i = 3; // int
optional bytes s = 4; // UTF-8 string
optional TensorProto t = 5; // tensor value
optional GraphProto g = 6; // graph
optional SparseTensorProto sparse_tensor = 22; // sparse tensor value
// Do not use field below, it's deprecated.
// optional ValueProto v = 12; // value - subsumes everything but graph
repeated float floats = 7; // list of floats
repeated int64 ints = 8; // list of ints
repeated bytes strings = 9; // list of UTF-8 strings
repeated TensorProto tensors = 10; // list of tensors
repeated GraphProto graphs = 11; // list of graph
repeated SparseTensorProto sparse_tensors = 23; // list of sparse tensors
}
// Defines information on value, including the name, the type, and
// the shape of the value.
message ValueInfoProto {
// This field MUST be present in this version of the IR.
optional string name = 1; // namespace Value
// This field MUST be present in this version of the IR for
// inputs and outputs of the top-level graph.
optional TypeProto type = 2;
// A human-readable documentation for this value. Markdown is allowed.
optional string doc_string = 3;
}
// Nodes
//
// Computation graphs are made up of a DAG of nodes, which represent what is
// commonly called a "layer" or "pipeline stage" in machine learning frameworks.
//
// For example, it can be a node of type "Conv" that takes in an image, a filter
// tensor and a bias tensor, and produces the convolved output.
message NodeProto {
repeated string input = 1; // namespace Value
repeated string output = 2; // namespace Value
// An optional identifier for this node in a graph.
// This field MAY be absent in ths version of the IR.
optional string name = 3; // namespace Node
// The symbolic identifier of the Operator to execute.
optional string op_type = 4; // namespace Operator
// The domain of the OperatorSet that specifies the operator named by op_type.
optional string domain = 7; // namespace Domain
// Additional named attributes.
repeated AttributeProto attribute = 5;
// A human-readable documentation for this node. Markdown is allowed.
optional string doc_string = 6;
}
// Training information
// TrainingInfoProto stores information for training a model.
// In particular, this defines two functionalities: an initialization-step
// and a training-algorithm-step. Initialization resets the model
// back to its original state as if no training has been consumed.
// Training algorithm improves the model based on input data.
//
// The semantics of the initialization-step is that the initializers
// in ModelProto.graph and in TrainingInfoProto.algorithm are first
// initialized as specified by the initializers in the graph, and then
// updated by the "initialization_binding" in every instance in
// ModelProto.training_info.
//
// The field "algorithm" defines a computation graph which represents a
// training algorithm's step. After the execution of a
// TrainingInfoProto.algorithm, the initializers specified by "update_binding"
// may be immediately updated. If the targeted training algorithm contains
// consecutive update stages (such as block coordinate descent methods),
// the user needs to create a TrainingInfoProto for each stage.
message TrainingInfoProto {
// This field describes a graph to compute the initial tensors
// upon starting the training process. Initialization graph has no input
// and can have multiple outputs. Usually, trainable tensors in neural
// networks are randomly initialized. To achieve that, for each tensor,
// the user can put a random number operator such as RandomNormal or
// RandomUniform in TrainingInfoProto.initialization.node and assign its
// random output to the specific tensor using "initialization_binding".
// This graph can also set the initializers in "algorithm" in the same
// TrainingInfoProto; a use case is resetting the number of training
// iteration to zero.
//
// By default, this field is an empty graph and its evaluation does not
// produce any output.
optional GraphProto initialization = 1;
// This field represents a training algorithm step. Given required inputs,
// it computes outputs to update initializers in its own or inference graph's
// initializer lists. In general, this graph contains loss node, gradient node,
// optimizer node, increment of iteration count, and some calls to the inference
// graph.
//
// The field algorithm.node is the only place the user can use GraphCall
// operator. The only callable graph is the one stored in ModelProto.graph.
//
// By default, this field is an empty graph and its evaluation does not
// produce any output.
optional GraphProto algorithm = 2;
// This field specifies the bindings from the outputs of "initialization" to
// some initializers in "ModelProto.graph.initializer" and
// the "algorithm.initializer" in the same TrainingInfoProto.
// See "update_binding" below for details.
//
// By default, this field is empty and no initializer would be changed
// by the execution of "initialization".
repeated StringStringEntryProto initialization_binding = 3;
// Gradient-based training is usually an iterative procedure. In one gradient
// descent iteration, we apply
//
// x = x - r * g
//
// where "x" is the optimized tensor, "r" stands for learning rate, and "g" is
// gradient of "x" with respect to a chosen loss. To avoid adding assignments
// into the training graph, we split the update equation into
//
// y = x - r * g
// x = y
//
// The user needs to save "y = x - r * g" into TrainingInfoProto.algorithm. To
// tell that "y" should be assigned to "x", the field "update_binding" may
// contain a key-value pair of strings, "x" (key of StringStringEntryProto)
// and "y" (value of StringStringEntryProto).
// For a neural network with multiple trainable (mutable) tensors, there can
// be multiple key-value pairs in "update_binding".
//
// The initializers appears as keys in "update_binding" are considered
// mutable and globally-visible variables. This implies some behaviors
// as described below.
//
// 1. We have only unique keys in all "update_binding"s so that two global
// variables may not have the same name. This ensures that one
// global variable is assigned up to once.
// 2. The keys must appear in names of "ModelProto.graph.initializer" or
// "TrainingInfoProto.algorithm.initializer".
// 3. The values must be output names of "algorithm".
// 4. If an optional input of a graph is omitted when using GraphCall, the
// global variable with the same name may be used.
// 5. When using GraphCall, the users always can pass values to optional
// inputs of the called graph even if the associated initializers appears
// as keys in "update_binding"s.
// 6. The graphs in TrainingInfoProto's can use global variables as
// their operator inputs.
// 7. Mutable variables are initialized to the value specified by the
// corresponding initializer, and then potentially updated by
// "initializer_binding"s and "update_binding"s in "TrainingInfoProto"s.
//
// This field usually contains names of trainable tensors
// (in ModelProto.graph), optimizer states such as momentums in advanced
// stochastic gradient methods (in TrainingInfoProto.graph),
// and number of training iterations (in TrainingInfoProto.graph).
//
// By default, this field is empty and no initializer would be changed
// by the execution of "algorithm".
repeated StringStringEntryProto update_binding = 4;
}
// Models
//
// ModelProto is a top-level file/container format for bundling a ML model and
// associating its computation graph with metadata.
//
// The semantics of the model are described by the associated GraphProto's.
message ModelProto {
// The version of the IR this model targets. See Version enum above.
// This field MUST be present.
optional int64 ir_version = 1;
// The OperatorSets this model relies on.
// All ModelProtos MUST have at least one entry that
// specifies which version of the ONNX OperatorSet is
// being imported.
//
// All nodes in the ModelProto's graph will bind against the operator
// with the same-domain/same-op_type operator with the HIGHEST version
// in the referenced operator sets.
repeated OperatorSetIdProto opset_import = 8;
// The name of the framework or tool used to generate this model.
// This field SHOULD be present to indicate which implementation/tool/framework
// emitted the model.
optional string producer_name = 2;
// The version of the framework or tool used to generate this model.
// This field SHOULD be present to indicate which implementation/tool/framework
// emitted the model.
optional string producer_version = 3;
// Domain name of the model.
// We use reverse domain names as name space indicators. For example:
// `com.facebook.fair` or `com.microsoft.cognitiveservices`
//
// Together with `model_version` and GraphProto.name, this forms the unique identity of
// the graph.
optional string domain = 4;
// The version of the graph encoded. See Version enum below.
optional int64 model_version = 5;
// A human-readable documentation for this model. Markdown is allowed.
optional string doc_string = 6;
// The parameterized graph that is evaluated to execute the model.
optional GraphProto graph = 7;
// kezhan: This field is not in ONNX, and will be pushed into ONNX with good use cases in microsoft.
repeated FunctionProto functions = 100;
// Named metadata values; keys should be distinct.
repeated StringStringEntryProto metadata_props = 14;
// Training-specific information. Sequentially executing all stored
// `TrainingInfoProto.algorithm`s and assigning their outputs following
// the corresponding `TrainingInfoProto.update_binding`s is one training
// iteration. Similarly, to initialize the model
// (as if training hasn't happened), the user should sequentially execute
// all stored `TrainingInfoProto.initialization`s and assigns their outputs
// using `TrainingInfoProto.initialization_binding`s.
//
// If this field is empty, the training behavior of the model is undefined.
repeated TrainingInfoProto training_info = 20;
};
// StringStringEntryProto follows the pattern for cross-proto-version maps.
// See https://developers.google.com/protocol-buffers/docs/proto3#maps
message StringStringEntryProto {
optional string key = 1;
optional string value= 2;
};
message TensorAnnotation {
optional string tensor_name = 1;
// <key, value> pairs to annotate tensor specified by <tensor_name> above.
// The keys used in the mapping below must be pre-defined in ONNX spec.
// For example, for 8-bit linear quantization case, 'SCALE_TENSOR', 'ZERO_POINT_TENSOR' will be pre-defined as
// quantization parameter keys.
repeated StringStringEntryProto quant_parameter_tensor_names = 2;
}
// Graphs
//
// A graph defines the computational logic of a model and is comprised of a parameterized
// list of nodes that form a directed acyclic graph based on their inputs and outputs.
// This is the equivalent of the "network" or "graph" in many deep learning
// frameworks.
message GraphProto {
// The nodes in the graph, sorted topologically.
repeated NodeProto node = 1;
// The name of the graph.
optional string name = 2; // namespace Graph
// A list of named tensor values, used to specify constant inputs of the graph.
// Each TensorProto entry must have a distinct name (within the list) that
// MAY also appear in the input list.
repeated TensorProto initializer = 5;
// Initializers (see above) stored in sparse format.
repeated SparseTensorProto sparse_initializer = 15;
// A human-readable documentation for this graph. Markdown is allowed.
optional string doc_string = 10;
// The inputs and outputs of the graph.
repeated ValueInfoProto input = 11;
repeated ValueInfoProto output = 12;
// Information for the values in the graph. The ValueInfoProto.name's
// must be distinct. It is optional for a value to appear in value_info list.
repeated ValueInfoProto value_info = 13;
// This field carries information to indicate the mapping among a tensor and its
// quantization parameter tensors. For example:
// For tensor 'a', it may have {'SCALE_TENSOR', 'a_scale'} and {'ZERO_POINT_TENSOR', 'a_zero_point'} annotated,
// which means, tensor 'a_scale' and tensor 'a_zero_point' are scale and zero point of tensor 'a' in the model.
repeated TensorAnnotation quantization_annotation = 14;
// DO NOT USE the following fields, they were deprecated from earlier versions.
// repeated string input = 3;
// repeated string output = 4;
// optional int64 ir_version = 6;
// optional int64 producer_version = 7;
// optional string producer_tag = 8;
// optional string domain = 9;
}
// Tensors
//
// A serialized tensor value.
message TensorProto {
enum DataType {
UNDEFINED = 0;
// Basic types.
FLOAT = 1; // float
UINT8 = 2; // uint8_t
INT8 = 3; // int8_t
UINT16 = 4; // uint16_t
INT16 = 5; // int16_t
INT32 = 6; // int32_t
INT64 = 7; // int64_t
STRING = 8; // string
BOOL = 9; // bool
// IEEE754 half-precision floating-point format (16 bits wide).
// This format has 1 sign bit, 5 exponent bits, and 10 mantissa bits.
FLOAT16 = 10;
DOUBLE = 11;
UINT32 = 12;
UINT64 = 13;
COMPLEX64 = 14; // complex with float32 real and imaginary components
COMPLEX128 = 15; // complex with float64 real and imaginary components
// Non-IEEE floating-point format based on IEEE754 single-precision
// floating-point number truncated to 16 bits.
// This format has 1 sign bit, 8 exponent bits, and 7 mantissa bits.
BFLOAT16 = 16;
// Future extensions go here.
}
// The shape of the tensor.
repeated int64 dims = 1;
// The data type of the tensor.
// This field MUST have a valid TensorProto.DataType value
optional int32 data_type = 2;
// For very large tensors, we may want to store them in chunks, in which
// case the following fields will specify the segment that is stored in
// the current TensorProto.
message Segment {
optional int64 begin = 1;
optional int64 end = 2;
}
optional Segment segment = 3;
// Tensor content must be organized in row-major order.
//
// Depending on the data_type field, exactly one of the fields below with
// name ending in _data is used to store the elements of the tensor.
// For float and complex64 values
// Complex64 tensors are encoded as a single array of floats,
// with the real components appearing in odd numbered positions,
// and the corresponding imaginary component appearing in the
// subsequent even numbered position. (e.g., [1.0 + 2.0i, 3.0 + 4.0i]
// is encoded as [1.0, 2.0 ,3.0 ,4.0]
// When this field is present, the data_type field MUST be FLOAT or COMPLEX64.
repeated float float_data = 4 [packed = true];
// For int32, uint8, int8, uint16, int16, bool, and float16 values
// float16 values must be bit-wise converted to an uint16_t prior
// to writing to the buffer.
// When this field is present, the data_type field MUST be
// INT32, INT16, INT8, UINT16, UINT8, BOOL, or FLOAT16
repeated int32 int32_data = 5 [packed = true];
// For strings.
// Each element of string_data is a UTF-8 encoded Unicode
// string. No trailing null, no leading BOM. The protobuf "string"
// scalar type is not used to match ML community conventions.
// When this field is present, the data_type field MUST be STRING
repeated bytes string_data = 6;
// For int64.
// When this field is present, the data_type field MUST be INT64
repeated int64 int64_data = 7 [packed = true];
// Optionally, a name for the tensor.
optional string name = 8; // namespace Value
// A human-readable documentation for this tensor. Markdown is allowed.
optional string doc_string = 12;
// Serializations can either use one of the fields above, or use this
// raw bytes field. The only exception is the string case, where one is
// required to store the content in the repeated bytes string_data field.
//
// When this raw_data field is used to store tensor value, elements MUST
// be stored in as fixed-width, little-endian order.
// Floating-point data types MUST be stored in IEEE 754 format.
// Complex64 elements must be written as two consecutive FLOAT values, real component first.
// Complex128 elements must be written as two consecutive DOUBLE values, real component first.
// Boolean type MUST be written one byte per tensor element (00000001 for true, 00000000 for false).
//
// Note: the advantage of specific field rather than the raw_data field is
// that in some cases (e.g. int data), protobuf does a better packing via
// variable length storage, and may lead to smaller binary footprint.
// When this field is present, the data_type field MUST NOT be STRING or UNDEFINED
optional bytes raw_data = 9;
// Data can be stored inside the protobuf file using type-specific fields or raw_data.
// Alternatively, raw bytes data can be stored in an external file, using the external_data field.
// external_data stores key-value pairs describing data location. Recognized keys are:
// - "location" (required) - POSIX filesystem path relative to the directory where the ONNX
// protobuf model was stored
// - "offset" (optional) - position of byte at which stored data begins. Integer stored as string.
// Offset values SHOULD be multiples 4096 (page size) to enable mmap support.
// - "length" (optional) - number of bytes containing data. Integer stored as string.
// - "checksum" (optional) - SHA1 digest of file specified in under 'location' key.
repeated StringStringEntryProto external_data = 13;
// Location of the data for this tensor. MUST be one of:
// - DEFAULT - data stored inside the protobuf message. Data is stored in raw_data (if set) otherwise in type-specified field.
// - EXTERNAL - data stored in an external location as described by external_data field.
enum DataLocation {
DEFAULT = 0;
EXTERNAL = 1;
}
// If value not set, data is stored in raw_data (if set) otherwise in type-specified field.
optional DataLocation data_location = 14;
// For double
// Complex128 tensors are encoded as a single array of doubles,
// with the real components appearing in odd numbered positions,
// and the corresponding imaginary component appearing in the
// subsequent even numbered position. (e.g., [1.0 + 2.0i, 3.0 + 4.0i]
// is encoded as [1.0, 2.0 ,3.0 ,4.0]
// When this field is present, the data_type field MUST be DOUBLE or COMPLEX128
repeated double double_data = 10 [packed = true];
// For uint64 and uint32 values
// When this field is present, the data_type field MUST be
// UINT32 or UINT64
repeated uint64 uint64_data = 11 [packed = true];
}
// A serialized sparse-tensor value
message SparseTensorProto {
// The sequence of non-default values are encoded as a tensor of shape [NNZ].
// The default-value is zero for numeric tensors, and empty-string for string tensors.
optional TensorProto values = 1;
// The indices of the non-default values, which may be stored in one of two formats.
// (a) Indices can be a tensor of shape [NNZ, rank] with the [i,j]-th value
// corresponding to the j-th index of the i-th value (in the values tensor).
// (b) Indices can be a tensor of shape [NNZ], in which case the i-th value
// must be the linearized-index of the i-th value (in the values tensor).
// The linearized-index can be converted into an index tuple (k_1,...,k_rank)
// using the shape provided below.
// The indices must appear in ascending order without duplication.
// In the first format, the ordering is lexicographic-ordering:
// e.g., index-value [1,4] must appear before [2,1]
optional TensorProto indices = 2;
// The shape of the underlying dense-tensor: [dim_1, dim_2, ... dim_rank]
repeated int64 dims = 3;
}
// Defines a tensor shape. A dimension can be either an integer value
// or a symbolic variable. A symbolic variable represents an unknown
// dimension.
message TensorShapeProto {
message Dimension {
oneof value {
int64 dim_value = 1;
string dim_param = 2; // namespace Shape
};
// Standard denotation can optionally be used to denote tensor
// dimensions with standard semantic descriptions to ensure
// that operations are applied to the correct axis of a tensor.
// Refer to https://github.com/onnx/onnx/blob/master/docs/DimensionDenotation.md#denotation-definition
// for pre-defined dimension denotations.
optional string denotation = 3;
};
repeated Dimension dim = 1;
}
// Types
//
// The standard ONNX data types.
message TypeProto {
message Tensor {
// This field MUST NOT have the value of UNDEFINED
// This field MUST have a valid TensorProto.DataType value
// This field MUST be present for this version of the IR.
optional int32 elem_type = 1;
optional TensorShapeProto shape = 2;
}
// repeated T
message Sequence {
// The type and optional shape of each element of the sequence.
// This field MUST be present for this version of the IR.
optional TypeProto elem_type = 1;
};
// map<K,V>
message Map {
// This field MUST have a valid TensorProto.DataType value
// This field MUST be present for this version of the IR.
// This field MUST refer to an integral type ([U]INT{8|16|32|64}) or STRING
optional int32 key_type = 1;
// This field MUST be present for this version of the IR.
optional TypeProto value_type = 2;
};
// #if ONNX-ML
message SparseTensor {
// This field MUST NOT have the value of UNDEFINED
// This field MUST have a valid TensorProto.DataType value
// This field MUST be present for this version of the IR.
optional int32 elem_type = 1;
optional TensorShapeProto shape = 2;
}
message Opaque {
// When missing, the domain is the same as the model's.
optional string domain = 1;
// The name is optional but significant when provided.
optional string name = 2;
// parameters that help defining the type
// DEPRECATED do not use.
// repeated TypeProto parameters = 3;
}
// #endif
oneof value {
// The type of a tensor.
Tensor tensor_type = 1;
// NOTE: DNN-only implementations of ONNX MAY elect to not support non-tensor values
// as input and output to graphs and nodes. These types are needed to naturally
// support classical ML operators. DNN operators SHOULD restrict their input
// and output types to tensors.
// The type of a sequence.
Sequence sequence_type = 4;
// The type of a map.
Map map_type = 5;
// #if ONNX-ML
SparseTensor sparse_tensor_type = 8;
Opaque opaque_type = 7;
// #endif
}
// An optional denotation can be used to denote the whole
// type with a standard semantic description as to what is
// stored inside. Refer to https://github.com/onnx/onnx/blob/master/docs/TypeDenotation.md#type-denotation-definition
// for pre-defined type denotations.
optional string denotation = 6;
}
// Operator Sets
//
// OperatorSets are uniquely identified by a (domain, opset_version) pair.
message OperatorSetIdProto {
// The domain of the operator set being identified.
// The empty string ("") or absence of this field implies the operator
// set that is defined as part of the ONNX specification.
// This field MUST be present in this version of the IR when referring to any other operator set.
optional string domain = 1;
// The version of the operator set being identified.
// This field MUST be present in this version of the IR.
optional int64 version = 2;
}
// Operator/function status.
enum OperatorStatus {
EXPERIMENTAL = 0;
STABLE = 1;
}
message FunctionProto {
// The name of the function, similar usage of op_type in OperatorProto.
optional string name = 1;
// The first version of a function set which contains this function.
// When there's any breaking change for this function, the function set
// contains the function needs to bump its version, and since_version of
// the updated function will be changed to the updated function set version.
optional int64 since_version = 2;
// This field indicates whether the syntax, semantics, or presence
// of this function is in an experimental or stable stage. Once an
// function is published as STABLE, its syntax and semantics MUST NOT
// change in subsequent versions of the operator set.
// When a function is published as EXPERIMENTAL, the syntax and semantics
// of the function MAY change across operator set versions.
// Functions "become" stable by deprecating the experimental version and
// introducing a new stable function with the same name.
optional OperatorStatus status = 3;
// The inputs and outputs of the function.
repeated string input = 4;
repeated string output = 5;
// The attributes of the function.
repeated string attribute= 6;
// The nodes in the function.
repeated NodeProto node = 7;
// A human-readable documentation for this function. Markdown is allowed.
optional string doc_string = 8;
// The OperatorSets this function body (graph) relies on.
// A FunctionProto body (graph) may implicitly rely on the OperatorSet that
// this function belongs to. It can also explicitly rely on more OperatorSets
// with this field specified.
//
// All nodes in the function body (graph) will bind against the operator
// with the same-domain/same-op_type operator with the HIGHEST version
// in the referenced operator sets. This means at most one version can be relied
// for one domain.
repeated OperatorSetIdProto opset_import = 9;
}