/// <summary>
/// Doesn't work because the op can't be found on binary
/// </summary>
/// <returns></returns>
private static OpDef op_NearestNeighbors()
{
var def = new OpDef
{
Name = "NearestNeighbors"
};
def.InputArg.Add(new ArgDef {
Name = "points", Type = DataType.DtFloat
});
def.InputArg.Add(new ArgDef {
Name = "centers", Type = DataType.DtFloat
});
def.InputArg.Add(new ArgDef {
Name = "k", Type = DataType.DtInt64
});
def.OutputArg.Add(new ArgDef {
Name = "nearest_center_indices", Type = DataType.DtInt64
});
def.OutputArg.Add(new ArgDef {
Name = "nearest_center_distances", Type = DataType.DtFloat
});
return(def);
}
/// <summary>
/// Creates a TF_Operation.
/// </summary>
/// <param name="graph">a `Graph`.</param>
/// <param name="node_def">`node_def_pb2.NodeDef` for the operation to create.</param>
/// <param name="inputs">
/// A list of `Tensor`s (corresponding to scalar inputs) and lists of
/// `Tensor`s (corresponding to sequence inputs, e.g. "int64 * N",
/// "list(int64)"). The length of the list should be equal to the number of
/// inputs specified by this operation's op def.
/// </param>
/// <param name="control_inputs">A list of `Operation`s to set as control dependencies.</param>
/// <returns>A wrapped TF_Operation*.</returns>
public static (IntPtr, OperationDescription) _create_c_op(Graph graph, NodeDef node_def, Tensor[] inputs, Operation[] control_inputs,
OpDef op_def = null)
{
if (op_def == null)
{
op_def = graph.GetOpDef(node_def.Op);
}
var input_tensors = _reconstruct_sequence_inputs(op_def, inputs, node_def.Attr);
lock (Locks.ProcessWide)
{
var op_desc = graph.NewOperation(node_def.Op, node_def.Name);
if (!string.IsNullOrEmpty(node_def.Device))
{
c_api.TF_SetDevice(op_desc, node_def.Device);
}
// Add inputs
foreach (var op_input in input_tensors)
{
if (op_input.IsList)
{
c_api.TF_AddInputList(op_desc, op_input.Select(x => x._as_tf_output()).ToArray(), op_input.Count());
}
else if (op_input.Count() == 1)
{
c_api.TF_AddInput(op_desc, op_input[0]._as_tf_output());
}
}
var status = tf.Status;
// Add control inputs
foreach (var control_input in control_inputs)
{
c_api.TF_AddControlInput(op_desc, control_input);
}
// Add attrs
foreach (var attr in node_def.Attr)
{
var bytes = attr.Value.ToByteArray(); //TODO: we can use attr.Value.WriteTo with a memory stream.
var protoHandle = Marshal.AllocHGlobal(bytes.Length);
Marshal.Copy(bytes, 0, protoHandle, bytes.Length);
uint len = (uint)bytes.Length;
c_api.TF_SetAttrValueProto(op_desc, attr.Key, protoHandle, proto_len: len, status: status.Handle);
status.Check(true);
Marshal.FreeHGlobal(protoHandle);
}
var c_op = c_api.TF_FinishOperation(op_desc, status.Handle);
status.Check(true);
return(c_op, op_desc);
}
}
/// <summary>
/// Creates a TF_Operation.
/// </summary>
/// <param name="graph">a `Graph`.</param>
/// <param name="node_def">`node_def_pb2.NodeDef` for the operation to create.</param>
/// <param name="inputs">
/// A list of `Tensor`s (corresponding to scalar inputs) and lists of
/// `Tensor`s (corresponding to sequence inputs, e.g. "int64 * N",
/// "list(int64)"). The length of the list should be equal to the number of
/// inputs specified by this operation's op def.
/// </param>
/// <param name="control_inputs">A list of `Operation`s to set as control dependencies.</param>
/// <returns>A wrapped TF_Operation*.</returns>
public static (IntPtr, OperationDescription) _create_c_op(Graph graph, NodeDef node_def, Tensor[] inputs, Operation[] control_inputs,
OpDef op_def = null)
{
if (op_def == null)
{
op_def = graph.GetOpDef(node_def.Op);
}
var input_tensors = _reconstruct_sequence_inputs(op_def, inputs, node_def.Attr);
var op_desc = graph.NewOperation(node_def.Op, node_def.Name);
if (!string.IsNullOrEmpty(node_def.Device))
{
c_api.TF_SetDevice(op_desc, node_def.Device);
}
// Add inputs
foreach (var op_input in input_tensors)
{
if (op_input.IsList)
{
c_api.TF_AddInputList(op_desc, op_input.Select(x => x._as_tf_output()).ToArray(), op_input.Count());
}
else if (op_input.Count() == 1)
{
c_api.TF_AddInput(op_desc, op_input[0]._as_tf_output());
}
}
var status = tf.Status;
// Add control inputs
foreach (var control_input in control_inputs)
{
c_api.TF_AddControlInput(op_desc, control_input);
}
// Add attrs
foreach (var attr in node_def.Attr)
{
var bytes = attr.Value.ToByteArray();
c_api.TF_SetAttrValueProto(op_desc, attr.Key, bytes, proto_len: bytes.Length, status: status.Handle);
status.Check(true);
}
var c_op = op_desc.FinishOperation(status);
status.Check(true);
return(c_op, op_desc);
}
private static void _SetDefaultAttrValues(NodeDef node_def, OpDef op_def)
{
foreach (var attr_def in op_def.Attr)
{
var key = attr_def.Name;
if (attr_def.DefaultValue != null)
{
var value = node_def.Attr[key];
if (value == null)
{
node_def.Attr[key] = attr_def.DefaultValue;
}
}
}
}
private object[] _reconstruct_sequence_inputs(OpDef op_def, Tensor[] inputs, MapField <string, AttrValue> attrs)
{
var grouped_inputs = new List <object>();
int i = 0;
int input_len = 0;
bool is_sequence = false;
foreach (var input_arg in op_def.InputArg)
{
if (!string.IsNullOrEmpty(input_arg.NumberAttr))
{
input_len = (int)attrs[input_arg.NumberAttr].I;
is_sequence = true;
}
else if (!string.IsNullOrEmpty(input_arg.TypeListAttr))
{
input_len = attrs[input_arg.TypeListAttr].List.Type.Count;
is_sequence = true;
}
else
{
input_len = 1;
is_sequence = false;
}
if (is_sequence)
{
grouped_inputs.Add(inputs.Skip(i).Take(input_len).ToArray());
}
else
{
grouped_inputs.Add(inputs[i]);
}
i += input_len;
}
return(grouped_inputs.ToArray());
}
/// <summary>
/// Creates an `Operation`.
/// </summary>
/// <param name="node_def">`node_def_pb2.NodeDef`. `NodeDef` for the `Operation`.</param>
/// <param name="g">`Graph`. The parent graph.</param>
/// <param name="inputs">list of `Tensor` objects. The inputs to this `Operation`.</param>
/// <param name="output_types">list of `DType` objects.</param>
/// <param name="control_inputs">
/// list of operations or tensors from which to have a
/// control dependency.
/// </param>
/// <param name="input_types">
/// List of `DType` objects representing the
/// types of the tensors accepted by the `Operation`. By default
/// uses `[x.dtype.base_dtype for x in inputs]`. Operations that expect
/// reference-typed inputs must specify these explicitly.
/// </param>
/// <param name="original_op"></param>
/// <param name="op_def"></param>
public Operation(NodeDef node_def, Graph g, Tensor[] inputs = null, TF_DataType[] output_types = null, ITensorOrOperation[] control_inputs = null, TF_DataType[] input_types = null, string original_op = "", OpDef op_def = null)
{
_graph = g;
// Build the list of control inputs.
var control_input_ops = new List <Operation>();
if (control_inputs != null)
{
foreach (var c in control_inputs)
{
switch (c)
{
case Operation c1:
control_input_ops.Add(c1);
break;
case Tensor tensor:
control_input_ops.Add(tensor.op);
break;
// TODO: IndexedSlices don't yet exist, but once they do, this needs to be uncommented
//case IndexedSlices islices:
// control_input_ops.Add(islices.op);
// break;
default:
throw new NotImplementedException($"Control input must be an Operation, a Tensor, or IndexedSlices: {c}");
}
}
}
// Dict mapping op name to file and line information for op colocation
// context managers.
_control_flow_context = graph._get_control_flow_context();
// This will be set by self.inputs.
if (op_def == null)
{
op_def = g.GetOpDef(node_def.Op);
}
var grouped_inputs = _reconstruct_sequence_inputs(op_def, inputs, node_def.Attr);
(_handle, _operDesc) = ops._create_c_op(g, node_def, grouped_inputs, control_input_ops.ToArray());
// Initialize self._outputs.
output_types = new TF_DataType[NumOutputs];
for (int i = 0; i < NumOutputs; i++)
{
output_types[i] = OutputType(i);
}
_outputs = new Tensor[NumOutputs];
for (int i = 0; i < NumOutputs; i++)
{
_outputs[i] = new Tensor(this, i, OutputType(i));
}
graph._add_op(this);
if (_handle != IntPtr.Zero)
{
_control_flow_post_processing();
}
}
public void TestInception()
{
using (Tensor imageTensor = ImageIO.ReadTensorFromImageFile <float>("grace_hopper.jpg", 224, 224, 128.0f, 1.0f))
using (Inception inceptionGraph = new Inception())
{
bool processCompleted = false;
inceptionGraph.OnDownloadCompleted += (sender, e) =>
{
HashSet <string> opNames = new HashSet <string>();
HashSet <string> couldBeInputs = new HashSet <string>();
HashSet <string> couldBeOutputs = new HashSet <string>();
foreach (Operation op in inceptionGraph.Graph)
{
String name = op.Name;
opNames.Add(name);
if (op.NumInputs == 0 && op.OpType.Equals("Placeholder"))
{
couldBeInputs.Add(op.Name);
AttrMetadata dtypeMeta = op.GetAttrMetadata("dtype");
AttrMetadata shapeMeta = op.GetAttrMetadata("shape");
DataType type = op.GetAttrType("dtype");
Int64[] shape = op.GetAttrShape("shape");
Buffer valueBuffer = op.GetAttrValueProto("shape");
Buffer shapeBuffer = op.GetAttrTensorShapeProto("shape");
Tensorflow.TensorShapeProto shapeProto =
Tensorflow.TensorShapeProto.Parser.ParseFrom(shapeBuffer.Data);
}
if (op.OpType.Equals("Const"))
{
AttrMetadata dtypeMeta = op.GetAttrMetadata("dtype");
AttrMetadata valueMeta = op.GetAttrMetadata("value");
using (Tensor valueTensor = op.GetAttrTensor("value"))
{
var dim = valueTensor.Dim;
}
}
if (op.OpType.Equals("Conv2D"))
{
AttrMetadata stridesMeta = op.GetAttrMetadata("strides");
AttrMetadata paddingMeta = op.GetAttrMetadata("padding");
AttrMetadata boolMeta = op.GetAttrMetadata("use_cudnn_on_gpu");
Int64[] strides = op.GetAttrIntList("strides");
bool useCudnn = op.GetAttrBool("use_cudnn_on_gpu");
String padding = op.GetAttrString("padding");
}
foreach (Output output in op.Outputs)
{
int[] shape = inceptionGraph.Graph.GetTensorShape(output);
if (output.NumConsumers == 0)
{
couldBeOutputs.Add(name);
}
}
Buffer buffer = inceptionGraph.Graph.GetOpDef(op.OpType);
Tensorflow.OpDef opDef = Tensorflow.OpDef.Parser.ParseFrom(buffer.Data);
}
using (Buffer versionDef = inceptionGraph.Graph.Versions())
{
int l = versionDef.Length;
}
Inception.RecognitionResult[] results = inceptionGraph.Recognize(imageTensor);
Trace.WriteLine(String.Format("Object is {0} with {1}% probability", results[0].Label, results[0].Probability * 100));
processCompleted = true;
};
inceptionGraph.Init();
while (!processCompleted)
{
Thread.Sleep(1000);
}
}
}
/// <summary>
/// Creates an `Operation`.
/// </summary>
/// <param name="node_def">`node_def_pb2.NodeDef`. `NodeDef` for the `Operation`.</param>
/// <param name="g">`Graph`. The parent graph.</param>
/// <param name="inputs">list of `Tensor` objects. The inputs to this `Operation`.</param>
/// <param name="output_types">list of `DType` objects.</param>
/// <param name="control_inputs">
/// list of operations or tensors from which to have a
/// control dependency.
/// </param>
/// <param name="input_types">
/// List of `DType` objects representing the
/// types of the tensors accepted by the `Operation`. By default
/// uses `[x.dtype.base_dtype for x in inputs]`. Operations that expect
/// reference-typed inputs must specify these explicitly.
/// </param>
/// <param name="original_op"></param>
/// <param name="op_def"></param>
public Operation(NodeDef node_def, Graph g, Tensor[] inputs = null, TF_DataType[] output_types = null, ITensorOrOperation[] control_inputs = null, TF_DataType[] input_types = null, string original_op = "", OpDef op_def = null)
{
_graph = g;
// Build the list of control inputs.
var control_input_ops = new List <Operation>();
if (control_inputs != null)
{
foreach (var c in control_inputs)
{
switch (c)
{
case Operation c1:
control_input_ops.Add(c1);
break;
default:
throw new NotImplementedException($"Control input must be an Operation, a Tensor, or IndexedSlices: {c}");
}
}
}
// This will be set by self.inputs.
if (op_def == null)
{
op_def = g.GetOpDef(node_def.Op);
}
var grouped_inputs = _reconstruct_sequence_inputs(op_def, inputs, node_def.Attr);
_handle = ops._create_c_op(g, node_def, grouped_inputs, control_input_ops.ToArray());
// Initialize self._outputs.
output_types = new TF_DataType[NumOutputs];
for (int i = 0; i < NumOutputs; i++)
{
output_types[i] = OutputType(i);
}
_outputs = new Tensor[NumOutputs];
for (int i = 0; i < NumOutputs; i++)
{
_outputs[i] = new Tensor(this, i, OutputType(i));
}
graph._add_op(this);
if (_handle != IntPtr.Zero)
{
_control_flow_post_processing();
}
}
private static AttrDef _FindAttrInOpDef(string name, OpDef op_def)
{
return(op_def.Attr.FirstOrDefault(x => x.Name == name));
}
public async Task TestResnet()
{
using (Tensor imageTensor = ImageIO.ReadTensorFromImageFile <float>("surfers.jpg", 224, 224, 0, 1.0f / 255.0f))
using (Resnet resnet = new Resnet())
{
await resnet.Init();
MetaGraphDef metaGraphDef = MetaGraphDef.Parser.ParseFrom(resnet.MetaGraphDefBuffer.Data);
var signatureDef = metaGraphDef.SignatureDef["serving_default"];
var inputNode = signatureDef.Inputs;
var outputNode = signatureDef.Outputs;
HashSet <string> opNames = new HashSet <string>();
HashSet <string> couldBeInputs = new HashSet <string>();
HashSet <string> couldBeOutputs = new HashSet <string>();
foreach (Operation op in resnet.Graph)
{
String name = op.Name;
opNames.Add(name);
if (op.NumInputs == 0 && op.OpType.Equals("Placeholder"))
{
couldBeInputs.Add(op.Name);
AttrMetadata dtypeMeta = op.GetAttrMetadata("dtype");
AttrMetadata shapeMeta = op.GetAttrMetadata("shape");
DataType type = op.GetAttrType("dtype");
Int64[] shape = op.GetAttrShape("shape");
Buffer valueBuffer = op.GetAttrValueProto("shape");
Buffer shapeBuffer = op.GetAttrTensorShapeProto("shape");
Tensorflow.TensorShapeProto shapeProto =
Tensorflow.TensorShapeProto.Parser.ParseFrom(shapeBuffer.Data);
}
if (op.OpType.Equals("Const"))
{
AttrMetadata dtypeMeta = op.GetAttrMetadata("dtype");
AttrMetadata valueMeta = op.GetAttrMetadata("value");
using (Tensor valueTensor = op.GetAttrTensor("value"))
{
var dim = valueTensor.Dim;
}
}
if (op.OpType.Equals("Conv2D"))
{
AttrMetadata stridesMeta = op.GetAttrMetadata("strides");
AttrMetadata paddingMeta = op.GetAttrMetadata("padding");
AttrMetadata boolMeta = op.GetAttrMetadata("use_cudnn_on_gpu");
Int64[] strides = op.GetAttrIntList("strides");
bool useCudnn = op.GetAttrBool("use_cudnn_on_gpu");
String padding = op.GetAttrString("padding");
}
foreach (Output output in op.Outputs)
{
int[] shape = resnet.Graph.GetTensorShape(output);
if (output.NumConsumers == 0)
{
couldBeOutputs.Add(name);
}
}
Buffer buffer = resnet.Graph.GetOpDef(op.OpType);
Tensorflow.OpDef opDef = Tensorflow.OpDef.Parser.ParseFrom(buffer.Data);
}
using (Buffer versionDef = resnet.Graph.Versions())
{
int l = versionDef.Length;
}
Resnet.RecognitionResult[][] results = resnet.Recognize(imageTensor);
}
}
public unsafe Operation create_op(string op_type, List <Tensor> inputs, TF_DataType[] dtypes,
TF_DataType[] input_types = null, string name = "",
Dictionary <string, AttrValue> attrs = null, OpDef op_def = null)
{
if (String.IsNullOrEmpty(name))
{
name = op_type;
}
name = name.EndsWith("/") ? ops._name_from_scope_name(name) : unique_name(name);
var node_def = ops._NodeDef(op_type, name, device: "", attrs: attrs);
var op = new Operation(node_def,
this,
inputs: inputs,
output_types: dtypes,
control_inputs: new object[] { },
input_types: input_types,
original_op: null,
op_def: op_def);
_create_op_helper(op, true);
return(op);
}
public void add_op(OpDef op_def)
{
_ops[op_def.Name] = op_def;
}
public Operation(NodeDef node_def, Graph g, List <Tensor> inputs = null, TF_DataType[] output_types = null, object control_inputs = null, TF_DataType[] input_types = null, string original_op = "", OpDef op_def = null)
{
_graph = g;
_id_value = _graph._next_id();
_c_op = ops._create_c_op(g, node_def, inputs);
var num_outputs = c_api.TF_OperationNumOutputs(_c_op);
_outputs = new Tensor[num_outputs];
for (int i = 0; i < num_outputs; i++)
{
_outputs[i] = new Tensor(this, i, TF_DataType.TF_FLOAT);
}
_graph._add_op(this);
}
public Operation(NodeDef node_def, Graph g, List <Tensor> inputs = null, TF_DataType[] output_types = null, object control_inputs = null, TF_DataType[] input_types = null, string original_op = "", OpDef op_def = null)
{
Graph = g;
_id_value = Graph._next_id();
if (op_def == null)
{
op_def = g.GetOpDef(node_def.Op);
}
_handle = ops._create_c_op(g, node_def, inputs);
_outputs = new Tensor[NumOutputs];
output_types = new TF_DataType[NumOutputs];
for (int i = 0; i < NumOutputs; i++)
{
output_types[i] = OutputType(i);
_outputs[i] = new Tensor(this, i, output_types[i]);
}
Graph._add_op(this);
}
private void SetAttrs(string op_type_name,
ArgDef input_arg,
OpDef op_def,
Dictionary <string, object> attrs,
Dictionary <string, object> inferred_from,
List <TF_DataType> types,
List <TF_DataType> base_types,
List <TF_DataType> input_types,
dynamic values)
{
var input_name = input_arg.Name;
if (!string.IsNullOrEmpty(input_arg.NumberAttr))
{
if (attrs.ContainsKey(input_arg.NumberAttr))
{
}
else
{
attrs[input_arg.NumberAttr] = (values as Tensor[]).Length;
inferred_from[input_arg.NumberAttr] = input_name;
var num_attr = op_def.Attr.First(x => x.Name == input_arg.NumberAttr);
if (num_attr.HasMinimum && (values as Tensor[]).Length < num_attr.Minimum)
{
throw new ValueError($"List argument '{input_name}' to '{op_type_name}' Op with length {(values as Tensor[]).Length} shorter " +
$"than minimum length {num_attr.Minimum}");
}
}
// All tensors must have the same base type.
if (input_arg.Type != DataType.DtInvalid)
{
}
else
{
attrs[input_arg.TypeAttr] = base_types[0];
inferred_from[input_arg.TypeAttr] = input_name;
var type_attr = op_def.Attr.First(x => x.Name == input_arg.TypeAttr);
}
}
else if (!string.IsNullOrEmpty(input_arg.TypeAttr))
{
var attr_value = base_types[0];
if (attrs.ContainsKey(input_arg.TypeAttr))
{
}
else
{
attrs[input_arg.TypeAttr] = attr_value;
inferred_from[input_arg.TypeAttr] = input_name;
}
}
else if (!string.IsNullOrEmpty(input_arg.TypeListAttr))
{
var attr_value = base_types;
if (attrs.ContainsKey(input_arg.TypeListAttr))
{
}
else
{
attrs[input_arg.TypeListAttr] = attr_value;
inferred_from[input_arg.TypeListAttr] = input_name;
}
}
if (input_arg.IsRef)
{
input_types.AddRange(types);
}
else
{
input_types.AddRange(base_types);
}
}
private AttrValue SetAttrValue(OpDef op_def, AttrDef attr_def, object value)
{
var attr_value = new AttrValue();
if (attr_def.Type.StartsWith("list("))
{
if (attr_def.HasMinimum)
{
;
}
attr_value.List = new AttrValue.Types.ListValue();
}
switch (attr_def.Type)
{
case "string":
attr_value.S = Google.Protobuf.ByteString.CopyFromUtf8((string)value);
break;
case "type":
attr_value.Type = _MakeType((TF_DataType)value, attr_def);
break;
case "list(type)":
attr_value.List.Type.AddRange((value as IList <TF_DataType>).Select(x => _MakeType(x, attr_def)));
break;
case "list(int)":
attr_value.List.I.AddRange((value as int[]).Select(x => Convert.ToInt64(x)));
break;
case "bool":
attr_value.B = (bool)value;
break;
case "float":
attr_value.F = (float)value;
break;
case "int":
attr_value.I = (int)value;
if (attr_def.HasMinimum && attr_value.I < attr_def.Minimum)
{
throw new ValueError($"Attr '{attr_def.Name}' of '{op_def.Name}' Op passed {attr_value.I} less than minimum {attr_def.Minimum}.");
}
break;
case "shape":
if (value == null && attr_def.DefaultValue != null)
{
attr_value.Shape = attr_def.DefaultValue.Shape;
}
if (value is TensorShape val1)
{
attr_value.Shape = val1.as_proto();
}
else if (value is long[] val2)
{
attr_value.Shape = tensor_util.as_shape(val2);
}
else if (value is int[] val3)
{
attr_value.Shape = tensor_util.as_shape(val3);
}
break;
default:
throw new TypeError($"SetAttrValue: can't not convert attr_def.Type '{attr_def.Type}' to protos.");
}
return(attr_value);
}
private AttrValue SetAttrValue(OpDef op_def, AttrDef attr_def, object value)
{
var attr_value = new AttrValue();
if (attr_def.Type.StartsWith("list("))
{
if (attr_def.HasMinimum)
#pragma warning disable CS0642 // Possible mistaken empty statement
{
;
}
#pragma warning restore CS0642 // Possible mistaken empty statement
attr_value.List = new AttrValue.Types.ListValue();
}
switch (attr_def.Type)
{
case "string":
attr_value.S = _MakeStr((string)value, attr_def);
break;
case "type":
attr_value.Type = _MakeType((TF_DataType)value, attr_def);
break;
case "list(type)":
attr_value.List.Type.AddRange((value as IList <TF_DataType>).Select(x => _MakeType(x, attr_def)));
break;
case "list(int)":
if (value != null)
{
attr_value.List.I.AddRange((value as int[]).Select(x => Convert.ToInt64(x)));
}
break;
case "bool":
attr_value.B = (bool)value;
break;
case "float":
attr_value.F = (float)value;
break;
case "int":
if (value is long value_long)
{
attr_value.I = value_long;
}
else
{
attr_value.I = Convert.ToInt64(value);
}
if (attr_def.HasMinimum && attr_value.I < attr_def.Minimum)
{
throw new ValueError($"Attr '{attr_def.Name}' of '{op_def.Name}' Op passed {attr_value.I} less than minimum {attr_def.Minimum}.");
}
break;
case "shape":
if (value == null && attr_def.DefaultValue != null)
{
attr_value.Shape = attr_def.DefaultValue.Shape;
}
if (value is Shape val1)
{
attr_value.Shape = val1.as_proto();
}
else if (value is long[] val2)
{
attr_value.Shape = tensor_util.as_shape(val2);
}
else if (value is int[] val3)
{
attr_value.Shape = tensor_util.as_shape(val3);
}
break;
case "list(shape)":
attr_value.List.Shape.AddRange((value as Shape[]).Select(x => _MakeShape(x, attr_def)));
break;
default:
throw new TypeError($"SetAttrValue: can't not convert attr_def.Type '{attr_def.Type}' to protos.");
}
return(attr_value);
}