public ONNXTensor Reshape(long[] onnxShape)
{
var symbolicShape = ONNXLayout.ConvertSymbolicShapeToBarracuda(onnxShape, "?");
var reshapedData = m_Data.Reshape(symbolicShape);
return(new ONNXTensor(reshapedData, onnxShape));
}
internal void Reduce(ModelBuilder net, ONNXNodeWrapper node, Layer.Type reduceType)
{
node.UnsupportedAttribute("keepdims", 1);
// @TODO: extract into separate function and reuse for other Reduce ops
var features = node.Input0Features;
var rank = node.Input0Rank;
object input = node.Input0;
foreach (var onnxAxis in node.Axes)
{
var axis = ONNXLayout.ConvertAxisToBarracuda(onnxAxis, onnxRank: rank, onnxLayout: "NCHW");
input = net.Reduce(reduceType, $"{node.Name}__axis{axis}", input, axis);
bool lastAxis = (axis == -1 || axis == node.Input0Rank - 1); // last axis in Barracuda is feature axis
if (lastAxis)
{
features = 1; // if reducing over the last feature axis, then operation will collapse all features to 1
}
rank--; // rank will be reduced after this operation
Output(name, features: features, rank: rank);
}
net.Identity(node.Name, input);
}
internal static Tensor Permute(Tensor inTensor, int[] permutations) // TODO: unify Permute() arguments
{
// See: https://stackoverflow.com/a/32034565
Profiler.BeginSample("ONNXTensor.Permute");
var outTensor = new Tensor(ONNXLayout.Permute(inTensor.shape.ToArray(), permutations));
Debug.Assert(outTensor.length == inTensor.length);
// {0, 2, 3, 1} => {0, 3, 1, 2}
// {2, 3, 1, 0} => {3, 2, 0, 1}
// => {find_index(0), find_index(1), find_index(2), find_index(3)}
var reversePermute = new int[permutations.Length];
for (var i = 0; i < permutations.Length; ++i)
{
reversePermute[i] = Array.IndexOf(permutations, i);
}
// outTensor strides
var outStrideC = outTensor.channels;
var outStrideWC = outStrideC * outTensor.width;
var outStrideHWC = outStrideWC * outTensor.height;
var outStride = new int[reversePermute.Length];
for (var i = 0; i < reversePermute.Length; ++i)
{
outStride[i] = new[] { 0, outStrideHWC, outStrideWC, outStrideC, 1 }
}
public ONNXTensor Permute(int[] permutations)
{
// transpose both data & shape
var transposedData = Permute(m_Data, permutations);
var transposedShape = ONNXLayout.Permute(m_Shape, permutations);
return(new ONNXTensor(transposedData, transposedShape));
}
public ONNXTensor Reshape(long[] onnxShape)
{
var symbolicShape = ONNXLayout.ConvertSymbolicShapeToBarracuda(onnxShape, "?");
var reshapedData = m_Data.Reshape(symbolicShape);
for (var i = 0; i < onnxShape.Length; ++i)
{
if (onnxShape[i] < 0)
{
onnxShape[i] = reshapedData.shape[i];
}
Debug.Assert(onnxShape[i] == reshapedData.shape[i]);
}
return(new ONNXTensor(reshapedData, onnxShape));
}
public Tensor ToBarracuda(string onnxLayout)
{
Profiler.BeginSample("ONNXTensor.ToBarracuda");
if (onnxLayout == "?")
{
throw new OnnxLayerImportException("Unknown ONNX layout in not supported when converting constant tensor to Barracuda");
}
Debug.Assert(m_Shape.All(v => v > 0));
var permutations = ONNXLayout.AxisPermutationsForMappingONNXLayoutToBarracuda(rank, onnxLayout);
Debug.Assert(rank <= permutations.Length);
var outTensor = Permute(m_Data, permutations);
Profiler.EndSample();
return(outTensor);
}
public ONNXTensor(TensorProto onnxTensor)
{
// read shape
var onnxShape = onnxTensor.Dims.ToArray();
var shape = ONNXLayout.ConvertShapeToBarracuda(onnxShape, onnxLayout: "?");
// read data
float[] data;
if ((onnxTensor.RawData != null) && (!onnxTensor.RawData.IsEmpty))
{
var byteArray = new byte[onnxTensor.RawData.Length];
onnxTensor.RawData.CopyTo(byteArray, 0);
// Float32
if (onnxTensor.DataType == (int)TensorProto.Types.DataType.Float)
{
data = new float[shape.length];
Debug.Assert((sizeof(float) * shape.length) == onnxTensor.RawData.Length);
Buffer.BlockCopy(byteArray, 0, data, 0, byteArray.Length);
}
// Float16
else if (onnxTensor.DataType == (int)TensorProto.Types.DataType.Float16)
{
var typedData = new UInt16[shape.length];
Debug.Assert((sizeof(UInt16) * shape.length) == onnxTensor.RawData.Length);
Buffer.BlockCopy(byteArray, 0, typedData, 0, byteArray.Length);
data = typedData.Select(x => HalfHelper.HalfToSingle(x)).ToArray();
}
// Int32
else if (onnxTensor.DataType == (int)TensorProto.Types.DataType.Int32)
{
var typedData = new int[shape.length];
Debug.Assert((sizeof(int) * shape.length) == onnxTensor.RawData.Length);
Buffer.BlockCopy(byteArray, 0, typedData, 0, byteArray.Length);
data = typedData.Select(x => (float)x).ToArray();
}
// Int64
else if (onnxTensor.DataType == (int)TensorProto.Types.DataType.Int64)
{
var typedData = new long[shape.length];
Debug.Assert((sizeof(long) * shape.length) == onnxTensor.RawData.Length);
Buffer.BlockCopy(byteArray, 0, typedData, 0, byteArray.Length);
data = typedData.Select(x => (float)x).ToArray();
}
else
{
throw new OnnxLayerImportException($"Tensor data type {(TensorProto.Types.DataType)onnxTensor.DataType} is not supported.");
}
}
// Float32
else if ((onnxTensor.FloatData != null) && (onnxTensor.FloatData.Count != 0))
{
Debug.Assert(shape.length == onnxTensor.FloatData.Count);
data = new float[shape.length];
onnxTensor.FloatData.CopyTo(data, 0);
}
// Int32
else if ((onnxTensor.Int32Data != null) && (onnxTensor.Int32Data.Count != 0))
{
Debug.Assert(shape.length == onnxTensor.Int32Data.Count);
data = onnxTensor.Int32Data.Select(x => (float)x).ToArray();
}
// Int64
else if ((onnxTensor.Int64Data != null) && (onnxTensor.Int64Data.Count != 0))
{
Debug.Assert(shape.length == onnxTensor.Int64Data.Count);
data = onnxTensor.Int64Data.Select(x => (float)x).ToArray();
}
else
{
throw new OnnxLayerImportException("Could not read tensor data for constant tensor.");
}
m_Data = new Tensor(shape, new SharedArrayTensorData(data));
m_Shape = onnxShape;
}
private Model ConvertOnnxModel(ModelProto onnxModel)
{
var model = new Model();
var modelBuilder = new ModelBuilder(model);
// Convert graph inputs & outputs
var initializersByName = onnxModel.Graph.Initializer.ToDictionary(i => i.Name, i => true);
foreach (ValueInfoProto i in onnxModel.Graph.Input)
{
// skip input tensors that have initializer data, they are constant tensors not global inputs
if (initializersByName.ContainsKey(i.Name))
{
continue;
}
if (m_OverrideGlobalInputs.ContainsKey(i.Name))
{
Const(i.Name, m_OverrideGlobalInputs[i.Name]);
continue;
}
modelBuilder.Input(i.Name, ONNXLayout.ConvertSymbolicShapeToBarracuda(i.Type.TensorType.Shape, onnxLayout: "NCHW"));
Output(i.Name, onnxShape: i.Type.TensorType.Shape.Dim.Select(d => d.DimValue).ToArray(), onnxLayout: "NCHW");
}
foreach (ValueInfoProto o in onnxModel.Graph.Output)
{
modelBuilder.Output(o.Name);
}
// TODO: process model (recurrent nodes) memories
// Read constants from initializer list
foreach (TensorProto initializer in onnxModel.Graph.Initializer)
{
Const(initializer.Name, new ONNXTensor(initializer));
}
// Convert graph nodes
foreach (NodeProto onnxNode in onnxModel.Graph.Node)
{
var node = new ONNXNodeWrapper(onnxNode, m_ModelTensors, model.Warnings);
var nodeId = node.Name;
var opType = node.OperatorType;
Output(node);
bool injectDummy = false;
if (m_NodeImporters.ContainsKey(opType))
{
try
{
if (node.AreAllInputsConst && !m_ShouldNotBeBaked.Contains(opType))
{
Profiler.BeginSample($"Bake {opType} {node.Name}");
var bakedTensor = BakeNodeIntoConstant(m_NodeImporters[opType], node);
Const(node.Name, bakedTensor);
var printTensor = bakedTensor.ToBarracuda("NCHW");
D.Log($"Baked node {nodeId} into constant of shape {printTensor.shape} and values: {printTensor.DataToString()}");
Profiler.EndSample();
}
else
{
Profiler.BeginSample($"Import {opType} {node.Name}");
m_NodeImporters[opType](modelBuilder, node);
Profiler.EndSample();
}
}
catch (Exception e)
{
// We support the layer but something went wrong while importing it
// We log the problem and insert an identity layer
string message = $"Unexpected error while parsing layer {nodeId} of type {opType}.\n{e.Message}\n\nJson: {onnxNode}\n{e.StackTrace}\n";
Warn(model, nodeId, message);
injectDummy = true;
}
}
else
{
//We don't support this type of layer
//We log the problem and insert an identity layer
string message = $"Unknown type encountered while parsing layer {nodeId} of type {opType}. We replace by an identity layer.";
Warn(model, nodeId, message);
injectDummy = true;
}
if (injectDummy)
{
var originalLayerHadInputs = (node.InputCount > 0);
if (originalLayerHadInputs)
{
modelBuilder.Identity(nodeId, node.Input0);
}
else // if errorneous layer had no inputs, inject dummy constant which does not require any inputs
{
modelBuilder.Const(nodeId, new Tensor());
}
}
m_ModelTensors.CompleteUninitializedFields(node);
}
// Convert constant tensors
int insertionIndex = 0;
foreach (var entry in constantTensors)
{
modelBuilder.Const(entry.Key, entry.Value.ToBarracuda(onnxLayout: "CONST"),
insertionIndex++);
}
// Model should not contain any broken links in the end
var unconnectedInputs = ModelAnalyzer.FindBrokenLinks(model);
Debug.Assert(unconnectedInputs.Length == 0);
if (unconnectedInputs.Length > 0)
{
var message = $"Broken links: {string.Join(", ", unconnectedInputs)}";
Warn(model, "", message);
}
// Parse meta data
var irVersion = onnxModel.IrVersion; // legacy
if (onnxModel.OpsetImport?.Count > 0)
{
irVersion = onnxModel.OpsetImport[0].Version;
}
model.ProducerName = $"{onnxModel.ProducerName} v{onnxModel.ProducerVersion}";
model.IrSource = "ONNX";
model.IrVersion = $"{irVersion}";
// strip :0 at the end of string name for TF import
if (patchRemoveTrailingTFExportCharacters)
{
model.inputs = model.inputs.Select(i => { i.name = i.name.EndsWith(":0") ? i.name.Remove(i.name.Length - 2) : i.name;
return(i); }).ToList();
model.outputs = model.outputs.Select(o => { o = o.EndsWith(":0") ? o.Remove(o.Length - 2) : o;
return(o); }).ToList();
model.memories = model.memories.Select(m => { m.input = m.input.EndsWith(":0") ? m.input.Remove(m.input.Length - 2) : m.input;
m.output = m.output.EndsWith(":0") ? m.output.Remove(m.output.Length - 2) : m.output;
return(m); }).ToList();
model.layers = model.layers.Select(l => { l.name = l.name.EndsWith(":0") ? l.name.Remove(l.name.Length - 2) : l.name;
for (int i = 0; i < l.datasets.Length; i++)
{
l.datasets[i].name = l.datasets[i].name.EndsWith(":0") ? l.datasets[i].name.Remove(l.datasets[i].name.Length - 2) : l.datasets[i].name;
}
for (int i = 0; i < l.inputs.Length; i++)
{
l.inputs[i] = l.inputs[i].EndsWith(":0") ? l.inputs[i].Remove(l.inputs[i].Length - 2) : l.inputs[i];
}
return(l); }).ToList();
}
return(model);
}
// ONNX parser declaration
// Add operator handlers here
public ONNXModelImporter()
{
// TODO: setup m_NodeImporters via initializer list
// TODO: simplify code to avoid passing node.Name over and over again
Add("Constant", (net, node) => {
node.UnsupportedAttribute("sparse_value");
Const(node, node.ValueAsTensor);
});
Add("Reshape", (net, node) => {
long[] onnxShape;
if (node.InputCount > 1) // Reshape-5
{
onnxShape = node.Input1Constant(onnxLayout: "C", name: "shape").AsLongs();
}
else // Reshape-1
{
onnxShape = node.Shape;
}
if (node.IsInput0Const)
{ // reshape constant source tensor and store it as the new constant
var reshapedTensor = constantTensors[node.Input0].Reshape(onnxShape);
Const(node, reshapedTensor);
}
else
{
var symbolicShape = ONNXLayout.ConvertSymbolicShapeToBarracuda(onnxShape, "NCHW");
if (patchReshapeToSupportBatchSize)
{
symbolicShape[0] = 0; // force keep batch size
}
net.Reshape(node.Name, node.Input0, symbolicShape);
Output(node, rank: symbolicShape.Length);
}
});
Add("Shape", (net, node) => {
// @TODO: dynamic implementation that would return real shape during execution of the model
if (!node.IsInput0Const)
{
throw new OnnxLayerImportException(
$"Currently only constant inputs for node of type {node.OperatorType} are supported. Instead input of {node.Name} is pointing to non-constant node {node.Input0}.");
}
var shapeTensor = new ONNXTensor(
data: new Tensor(4, 1, new[] { -1f, (float)node.Input0Rank, -1f, -1f }),
onnxShape: new [] { 4L });
Const(node, shapeTensor);
Output(node, rank: 1);
});
Add("Unsqueeze", (net, node) => {
if (node.IsInput0Const)
{
var unsqueezed = constantTensors[node.Input0].Unsqueeze(node.Axes);
Const(node, unsqueezed);
}
else
{
// NOTE: axis=0 or 1 will require Transpose between channels and other spatial dimensions when converted to Barracuda layout.
// As we have different layouts between ONNX and Barracuda, Unsqueeze might require actual Transpose not just Reshape!
// ONNX pseudocode here:
// a = Tensor [2, 10] # NC -> barracuda N11C
// b = Unsqueeze(a, axis=0)
// # b is now Tensor [1, 2, 10] # NCHW -> barrada NHWC
// Because ONNX is NCHW, but generally hell knows what goes where and Barracuda is strict NHWC. We end up with:
// `a` would be [2, 1, 1, 10], but `b` would have to be [1, 10, 1, 2]. Note the actual data swap in channels!
bool mightNeedTranspose = node.Axes.Any(axis => axis <= 1);
if (mightNeedTranspose)
{
Warn(net, node, $"Unsqeeze on axis next to batch dimension for non-constant tensors might lead to unexpected results.");
}
net.Identity(node.Name, node.Input0);
}
});
Add("Squeeze", (net, node) => {
if (node.IsInput0Const)
{
var squeezed = constantTensors[node.Input0].Squeeze(node.Axes);
Const(node, squeezed);
}
else
{
// See Unsqueeze above for explanation
bool mightNeedTranspose = node.Axes.Any(axis => axis <= 1);
if (mightNeedTranspose)
{
Warn(net, node, $"Sqeeze on any axis next to batch dimension for non-constant tensors might lead to unexpected results.");
}
net.Identity(node.Name, node.Input0);
}
});
Add("Flatten", (net, node) => {
node.UnsupportedAttribute("axis", 1);
net.Flatten(node.Name, node.Input0);
Output(node, rank: 2);
});
Add("Concat", (net, node) => {
int axis = node.AxisOptional(0);
// TODO: write dedicated ONNXTensor.Concat() so that output shape is exact to ONNX
// if (node.AreAllInputsConst) Const(node, ONNXTensor.Concat(node.Inputs.Select(i => constantTensors[i]), axis));
axis = ONNXLayout.ConvertAxisToBarracuda(axis, onnxRank: node.Input0Rank, onnxLayout: "NCHW");
net.Concat(node.Name, node.Inputs, axis);
bool lastAxis = (axis == -1 || axis == node.Input0Rank - 1); // last axis in Barracuda is feature axis
if (lastAxis)
{
var featuresConcatenated = node.Inputs.Sum(i => variableTensors[i].features);
Output(node, features: featuresConcatenated);
}
});
Add("Slice", (net, node) => {
int[] starts, ends, axes, steps;
if (node.InputCount > 1) // Slice-10
{
var constStarts = node.Input1Constant(onnxLayout: "C", name: "starts");
var constEnds = node.Input2Constant(onnxLayout: "C", name: "ends");
var defaultAxes = new Tensor(constStarts.shape, Enumerable.Range(0, constStarts.length).Select(v => (float)v).ToArray());
var constAxes = node.Input3ConstantOptional(defaultAxes, onnxLayout: "C", name: "axes");
var constSteps = node.Input4ConstantOptional(constStarts.shape, 1.0f, onnxLayout: "C", name: "steps");
starts = constStarts.AsInts();
ends = constEnds.AsInts();
axes = constAxes.AsInts();
steps = constSteps.AsInts();
}
else // Slice-1
{
starts = node.Starts;
ends = node.Ends;
axes = node.AxesOptional(Enumerable.Range(0, starts.Length).ToArray());
steps = Enumerable.Repeat(1, starts.Length).ToArray();
}
Debug.Assert(starts.Length == ends.Length);
var onnxRank = node.Input0Rank;
var onnxLast = (long)int.MaxValue;
var onnxStarts = Enumerable.Repeat(0L, onnxRank).ToArray();
var onnxEnds = Enumerable.Repeat(onnxLast, onnxRank).ToArray(); // by default copy the whole axis till the end
var onnxSteps = Enumerable.Repeat(1L, onnxRank).ToArray();
// NOTE: begin=0, end=0, stride=1 <= full range from existing axis
// begin=0, end=inf,stride=1 <= full range from existing axis
// begin=0, end=X, stride=1 <= full range from existing axis, if X==last element on this axis
// begin=0, end=0, stride=0 <= new axis OR shrink axis to single 1st element
// begin=N, end=N, stride=0 <= shrink axis to single Nth element
// These notes are copied from TensorExtensions.ApplyStridedSlice(...)
for (int i = 0; i < axes.Length; ++i)
{
var axis = axes[i];
if (axis < 0)
{
axis += onnxRank;
}
axis = Math.Min(Math.Max(axis, 0), onnxRank);
onnxStarts[axis] = starts[i];
onnxEnds[axis] = ends[i];
onnxSteps[axis] = steps[i];
}
if (node.IsInput0Const)
{
var slicedTensor = constantTensors[node.Input0].Slice(
starts: onnxStarts.Select(x => (int)x).ToArray(),
ends: onnxEnds.Select(x => (int)x).ToArray(),
steps: onnxSteps.Select(x => (int)x).ToArray());
Const(node, slicedTensor);
}
else
{
net.StridedSlice(node.Name, node.Input0,
starts: ONNXLayout.ConvertSymbolicShapeToBarracuda(onnxStarts, onnxLayout: "NCHW"),
ends: ONNXLayout.ConvertSymbolicShapeToBarracuda(onnxEnds, onnxLayout: "NCHW"),
strides: ONNXLayout.ConvertSymbolicShapeToBarracuda(onnxSteps, onnxLayout: "NCHW"));
}
});
Add("Gather", (net, node) =>
{
int axis = node.AxisOptional(0);
if (node.IsInput0Const)
{
var indices = node.Input1Constant(onnxLayout: "C", name: "indices").AsInts();
var gatheredTensor = constantTensors[node.Input0].Gather(axis, indices);
Const(node, gatheredTensor);
}
else
{
axis = ONNXLayout.ConvertAxisToBarracuda(axis, onnxRank: node.Input0Rank, onnxLayout: "NCHW");
net.Gather(node.Name, node.Input0, node.Input1, axis);
}
});
Add("OneHot", (net, node) => {
node.UnsupportedAttribute("axis", -1);
var defaultOffOn = new Tensor(1, 1, 1, 2, new float[] { 0, 1 });
var depth = (int)node.Input1Constant(onnxLayout: "C", name: "depth")[0];
var offon = node.Input2ConstantOptional(defaultOffOn, onnxLayout: "C", name: "values");
net.OneHot(node.Name, node.Input0, depth, (int)offon[1], (int)offon[0]);
});
// Activation ops
Add("Relu", (net, node) => { net.Relu(node.Name, node.Input0); });
Add("Softmax", (net, node) => { net.Softmax(node.Name, node.Input0); node.UnsupportedAttribute("axis", 1); });
Add("Tanh", (net, node) => { net.Tanh(node.Name, node.Input0); });
Add("Sqrt", (net, node) => { net.Sqrt(node.Name, node.Input0); });
Add("Sigmoid", (net, node) => { net.Sigmoid(node.Name, node.Input0); });
Add("Elu", (net, node) => { net.Elu(node.Name, node.Input0, node.AlphaOptional(1f)); });
Add("LeakyRelu", (net, node) => { net.LeakyRelu(node.Name, node.Input0, node.AlphaOptional(0.01f)); });
Add("Selu", (net, node) => { net.Selu(node.Name, node.Input0, node.AlphaOptional(1.67326f), node.GammaOptional(1.0507f)); });
Add("Swish", (net, node) => { net.Swish(node.Name, node.Input0); });
Add("PRelu", (net, node) => { net.PRelu(node.Name, node.Input0, node.Input1); });
Add("LogSoftmax", (net, node) => { net.LogSoftmax(node.Name, node.Input0); node.UnsupportedAttribute("axis", 1); });
// TODO: Add("Hardmax", (net, node) => { net.Hardmax(node.Name, node.Input0); node.UnsupportedAttribute("axis", 1); });
// TODO: Add("Softplus", (net, node) => { net.Softplus(node.Name, node.Input0); });
// TODO: Add("Softsign", (net, node) => { net.Softsign(node.Name, node.Input0); });
// TODO: Add("HardSigmoid", (net, node) => { net.HardSigmoid(node.Name, node.Input0, node.AlphaOptional(0.2f), node.BetaOptional(0.5f)); });
Add("Exp", (net, node) => { net.Exp(node.Name, node.Input0); });
Add("Log", (net, node) => { net.Log(node.Name, node.Input0); });
Add("Reciprocal", (net, node) => { net.Reciprocal(node.Name, node.Input0); });
Add("Abs", (net, node) => { net.Abs(node.Name, node.Input0); });
Add("Neg", (net, node) => { net.Neg(node.Name, node.Input0); });
Add("Ceil", (net, node) => { net.Ceil(node.Name, node.Input0); });
Add("Floor", (net, node) => { net.Floor(node.Name, node.Input0); });
Add("Round", (net, node) => { net.Round(node.Name, node.Input0); });
Add("Clip", (net, node) => { net.Clip(node.Name, node.Input0, node.MinOptional(float.MinValue), node.MaxOptional(float.MaxValue)); });
// Broadcast ops
Add("Add", (net, node) => { net.Add(node.Name, node.Inputs); });
Add("Sum", (net, node) => { net.Add(node.Name, node.Inputs); }); // Sum is implemented via Add
Add("Sub", (net, node) => { net.Sub(node.Name, node.Inputs); });
Add("Mul", (net, node) => { net.Mul(node.Name, node.Inputs); });
Add("Div", (net, node) => { net.Div(node.Name, node.Inputs); });
Add("Pow", (net, node) => { net.Pow(node.Name, node.Inputs); });
Add("Min", (net, node) => { net.Min(node.Name, node.Inputs); });
Add("Max", (net, node) => { net.Max(node.Name, node.Inputs); });
Add("Mean", (net, node) => { net.Mean(node.Name, node.Inputs); });
// Logical ops
Add("Greater", (net, node) => { net.Greater(node.Name, node.Input0, node.Input1); });
Add("Less", (net, node) => { net.Less(node.Name, node.Input0, node.Input1); });
Add("Equal", (net, node) => { net.Equal(node.Name, node.Input0, node.Input1); });
Add("Or", (net, node) => { net.LogicalOr(node.Name, node.Input0, node.Input1); });
Add("And", (net, node) => { net.LogicalAnd(node.Name, node.Input0, node.Input1); });
Add("Not", (net, node) => { net.LogicalNot(node.Name, node.Input0); });
Add("Xor", (net, node) => { net.LogicalXor(node.Name, node.Input0, node.Input1); });
// Padding ops
Add("Pad", (net, node) =>
{
var mode = node.GetOptionalString("mode", "constant");
switch (mode)
{
case "constant": net.Border2D(node.Name, node.Input0, node.Pads, node.GetOptionalFloat("value", 0.0f)); break;
case "reflect": net.Pad2DReflect(node.Name, node.Input0, node.Pads); break;
case "edge": net.Pad2DEdge(node.Name, node.Input0, node.Pads); break;
}
});
// Pooling ops
Add("AveragePool", (net, node) => {
node.UnsupportedAttribute("ceil_mode", 0);
node.UnsupportedAttribute("count_include_pad", 0);
net.AvgPool2D(node.Name, node.Input0, node.KernelShape, node.Strides, node.Pads);
});
Add("MaxPool", (net, node) => {
node.UnsupportedAttribute("ceil_mode", 0);
node.UnsupportedAttribute("dilations", new[] { 1, 1 });
node.UnsupportedAttribute("storage_order", 0);
net.MaxPool2D(node.Name, node.Input0, node.KernelShape, node.Strides, node.Pads);
});
Add("GlobalAveragePool", (net, node) => { net.GlobalAvgPool2D(node.Name, node.Input0); });
Add("GlobalMaxPool", (net, node) => { net.GlobalMaxPool2D(node.Name, node.Input0); });
Add("Upsample", (net, node) => {
node.UnsupportedAttribute("mode", "nearest");
float[] scales;
if (node.InputCount > 1)
{
scales = node.Input1Constant(onnxLayout: "C", name: "scales").AsFloats();
}
else
{
scales = node.Scales;
}
if (scales.Length < 2 || scales.Length > 5)
{
throw new OnnxLayerImportException(
$"Input scales of unsupported length {scales.Length} in {node.Name} ot fype {node.OperatorType}.");
}
if ((scales[0] != 1) || (scales[1] != 1))
{
Warn(net, node, $"Unsupported scaling, only H and W scaling are supported. Value will be ignored and defaulted to 1.");
}
// skip NC from onnx NCHW layout
scales = scales.Skip(2).ToArray();
if (scales.Length == 1) // append default H, if 1D
{
scales = new[] { 1f, scales[0] }
}
;
Resize2D(net, node, scales);
});
Add("Resize", (net, node) => {
node.UnsupportedAttribute("mode", "nearest");
float[] scales;
if (node.InputCount > 2) // Resize-11
{
node.UnsupportedAttribute("coordinate_transformation_mode", "half_pixel");
node.UnsupportedAttribute("cubic_coeff_a", -0.75f);
node.UnsupportedAttribute("exclude_outside", 0);
node.UnsupportedAttribute("extrapolation_value", 0f);
node.UnsupportedAttribute("nearest_mode", "round_prefer_floor");
// Inputs (3 - 4)
// X : T1
// roi : T2, It only takes effect when coordinate_transformation_mode is "tf_crop_and_resize"
// scales : tensor(float)
// sizes (optional) : tensor(int64)
// TODO: cropping via roi input
// TODO: support sizes
scales = node.Input2Constant(onnxLayout: "C", name: "scales").AsFloats();
}
else // Resize-10
{
scales = node.Input1Constant(onnxLayout: "C", name: "scales").AsFloats();
}
if (scales.Length < 2 || scales.Length > 5)
{
throw new OnnxLayerImportException(
$"Input scales of unsupported length {scales.Length} in {node.Name} ot fype {node.OperatorType}.");
}
if ((scales[0] != 1) || (scales[1] != 1))
{
Warn(net, node, $"Unsupported scaling, only H and W scaling are supported. Value will be ignored and defaulted to 1.");
}
// skip NC from onnx NCHW layout
scales = scales.Skip(2).ToArray();
Resize2D(net, node, scales);
});
Add("Transpose", (net, node) =>
{
if (!node.IsInput0Const)
{
throw new OnnxLayerImportException(
$"Currently only constant inputs for node of type {node.OperatorType} are supported. Instead input of {node.Name} is pointing to non-constant node {node.Input0}.");
}
var permutations = node.GetOptionalIntArray("perm", new [] { 3, 2, 1, 0 });
var transposedTensor = constantTensors[node.Input0].Permute(permutations);
Const(node, transposedTensor);
});
// Tensor ops
Add("Gemm", (net, node) => {
node.UnsupportedAttribute("alpha", 1.0f);
node.UnsupportedAttribute("beta", 1.0f);
node.UnsupportedAttribute("transA", 0);
var onnxLayout = node.TransBOptional() ? "KC" : "CK";
var weights = node.Input1Constant(onnxLayout, name: "B");
var biases = node.Input2ConstantOptional(Bias(weights.shape), 0.0f, onnxLayout: "C", name: "C");
// Change data layout from "channels first" to "channels last"
weights = SwapSpatialDimensionsAndFeaturesInMatMulWeights(weights, node.Input0Features);
net.Dense(node.Name, node.Input0, weights, biases);
Output(node, features: weights.channels, rank: 2); // Gemm forces flatten of the input to rank 2
});
Add("MatMul", (net, node) => {
var weights = node.Input1Constant(onnxLayout: "CK", name: "B");
var biases = node.DefaultTensor(Bias(weights.shape), 0.0f);
// Change data layout from "channels first" to "channels last"
weights = SwapSpatialDimensionsAndFeaturesInMatMulWeights(weights, node.Input0Features);
net.Dense(node.Name, node.Input0, weights, biases);
Output(node, features: weights.channels, rank: 2); // MatMul forces flatten of the input to rank 2
});
Add("Conv", (net, node) => {
node.UnsupportedAttribute("dilations", new[] { 1, 1 });
node.IgnoredAttribute("kernel_shape", "Kernel shape is derived from K tensor weights instead");
var kernels = node.Input1Constant(onnxLayout: "KCHW", name: "W");
var biases = node.Input2ConstantOptional(Bias(kernels.shape), 0.0f, onnxLayout: "C", name: "B");
if (node.GroupOptional() > 1)
{
net.DepthwiseConv2D(node.Name, node.Input0, node.Strides, node.Pads, kernels, biases);
}
else
{
net.Conv2D(node.Name, node.Input0, node.Strides, node.Pads, kernels, biases);
}
Output(node, features: kernels.channels);
});
Add("ConvTranspose", (net, node) => {
node.UnsupportedAttribute("dilations", new[] { 1, 1 });
node.UnsupportedAttribute("group", 1);
node.UnsupportedAttribute("output_shape", new int[0]);
node.IgnoredAttribute("kernel_shape", "Kernel shape is derived from K tensor weights instead");
var kernels = node.Input1Constant(onnxLayout: "CKHW", name: "W");
var biases = node.Input2ConstantOptional(Bias(kernels.shape), 0.0f, onnxLayout: "C", name: "B");
net.Conv2DTrans(node.Name, node.Input0, node.Strides, node.Pads, node.OutputPadding, kernels, biases);
Output(node, features: kernels.channels);
});
Add("BatchNormalization", (net, node) => {
var variance = node.Input4Constant(onnxLayout: "C", name: "var");
var scale = node.Input1ConstantOptional(variance.shape, 1.0f, onnxLayout: "C", name: "scale");
var bias = node.Input2ConstantOptional(variance.shape, 0.0f, onnxLayout: "C", name: "B");
var mean = node.Input3ConstantOptional(variance.shape, 0.0f, onnxLayout: "C", name: "mean");
var fusedData = FuseBatchNormWeights(scale, bias, mean, variance, node.EpsilonOptional());
net.ScaleBias(node.Name, node.Input0, fusedData.Item1, fusedData.Item2);
});
Add("InstanceNormalization", (net, node) => {
var scale = node.Input1Constant(onnxLayout: "C", name: "scale");
var bias = node.Input2ConstantOptional(scale.shape, 0.0f, onnxLayout: "C", name: "B");
net.Normalization(node.Name, node.Input0, scale, bias, node.EpsilonOptional());
});
// random ops
Add("RandomNormal", (net, node) => {
float mean = node.GetOptionalFloat("mean", 1.0f);
float scale = node.GetOptionalFloat("scale", 0.0f);
float seed = node.GetOptionalFloat("seed", 0.0f);
int[] shape = node.GetRequiredIntArray("shape");
net.RandomNormal(node.Name, mean, scale, seed, shape);
});
Add("RandomNormalLike", (net, node) => {
float mean = node.GetOptionalFloat("mean", 1.0f);
float scale = node.GetOptionalFloat("scale", 0.0f);
float seed = node.GetOptionalFloat("seed", 0.0f);
net.RandomNormal(node.Name, mean, scale, seed, node.Input0);
});
Add("RandomUniform", (net, node) => {
float high = node.GetOptionalFloat("high", 1.0f);
float low = node.GetOptionalFloat("low", 0.0f);
float seed = node.GetOptionalFloat("seed", 0.0f);
int[] shape = node.GetRequiredIntArray("shape");
net.RandomUniform(node.Name, low, high, seed, shape);
});
Add("RandomUniformLike", (net, node) => {
float high = node.GetOptionalFloat("high", 1.0f);
float low = node.GetOptionalFloat("low", 0.0f);
float seed = node.GetOptionalFloat("seed", 0.0f);
net.RandomUniform(node.Name, low, high, seed, node.Input0);
});
Add("ReduceMax", (net, node) => {
Reduce(net, node, Layer.Type.ReduceMax);
});
Add("ReduceMean", (net, node) => {
Reduce(net, node, Layer.Type.ReduceMean);
});
Add("ReduceMin", (net, node) => {
Reduce(net, node, Layer.Type.ReduceMin);
});
Add("ReduceProd", (net, node) => {
Reduce(net, node, Layer.Type.ReduceProd);
});
Add("ReduceSum", (net, node) => {
Reduce(net, node, Layer.Type.ReduceSum);
});
// Ignore, noop during inference
Add("Identity", (net, node) => { net.Identity(node.Name, node.Input0); });
Add("Cast", (net, node) => { net.Identity(node.Name, node.Input0); });
Add("Dropout", (net, node) => { net.Identity(node.Name, node.Input0); });
}
// Complex attribute helpers
private int[] GetPads()
{
var noPadding = new[] { 0, 0, 0, 0 };
if (SupportsAutoPad)
{
// known_paddings = {
// 'VALID' : [0,0,0,0],
// 'SAME_UPPER' : [-1],
// 'SAME_LOWER' : [-2],
// }
var autoPad = GetOptionalString("auto_pad", "NOTSET");
if (autoPad == "VALID")
{
return(noPadding);
}
else if (autoPad == "SAME_UPPER")
{
return new[] { -1 }
}
;
else if (autoPad == "SAME_LOWER")
{
return new[] { -2 }
}
;
else
{
} // TODO: Assert NOTSET
}
// NOTE: ONNX has pad layout of [z, y, x ...] while Barracuda is opposite [x, y, z ...]
var pads = GetOptionalIntArray("pads", noPadding);
if (SupportsSpatialOnlyPads)
{
// See: https://github.com/onnx/onnx/blob/master/docs/Operators.md#AveragePool
// Padding for the beginning and ending along each spatial axis, it can take any value greater than or equal to 0.
// The value represent the number of pixels added to the beginning and end part of the corresponding axis.
// `pads` format should be as follow [x1_begin, x2_begin...x1_end, x2_end,...],
// where xi_begin the number of pixels added at the beginning of axis `i` and xi_end,
// the number of pixels added at the end of axis `i`.
switch (pads.Length)
{
case 2: return(new [] { pads[0], 0, pads[1], 0 }); // 1D WW => W_W_
case 4: return(new [] { pads[1], pads[0], pads[3], pads[2] }); // 2D HWHW => WHWH
case 6: Warn("3D pads are not supported yet!");
return(new [] { pads[2], pads[1], pads[0], pads[3], pads[4], pads[5] }); // TODO: 3D DHWDHW => WHDWHD
default:
throw new OnnxLayerImportException(
$"Attribute pads of unsupported length {pads.Length} in {Name} ot fype {OperatorType}.");
}
}
// See: https://github.com/onnx/onnx/blob/master/docs/Operators.md#Pad
// `pads` should be a 1D tensor of shape [2 * input_rank].
// `pads` format should be: [x1_begin, x2_begin,...,x1_end, x2_end,...],
// where xi_begin is the number of pad values added at the beginning of axis `i` and xi_end,
// the number of pad values added at the end of axis `i
Debug.Assert(pads.Length % 2 == 0);
long[] onnxStarts = new long[pads.Length / 2];//TODO make the semantic diff between Permute(int[] and Permute(long[] more clear.
long[] onnxEnds = new long[pads.Length / 2];
for (int i = 0; i < onnxStarts.Length; ++i)
{
onnxStarts[i] = (long)pads[i];
onnxEnds[i] = (long)pads[i + onnxStarts.Length];
}
var starts = ONNXLayout.Permute(onnxStarts, "NCHW");
var ends = ONNXLayout.Permute(onnxEnds, "NCHW");
if ((starts[0] != 0) || (starts[3] != 0) || (ends[0] != 0) || (ends[3] != 0))
{
Warn($"Unsupported padding, only H and W padding are supported. Value will be ignored and defaulted to 0.");
}
return(new int[] { starts[2], starts[1], ends[2], ends[1] });
}