public static Tensor with_dependencies(Operation[] dependencies, Tensor output_tensor, string name = "")
{
var values = new List <object>();
values.AddRange(dependencies);
values.Add(output_tensor);
return(Python.with <ops.name_scope, Tensor>(new ops.name_scope(name, "control_dependency", values), scope =>
{
name = scope;
return Python.with <_ControlDependenciesController, Tensor>(ops.control_dependencies(dependencies), ctl =>
{
output_tensor = ops.convert_to_tensor_or_composite(output_tensor);
return _Identity(output_tensor, name: name);
});
}));
}
private static Operation _GroupControlDeps(string dev, Operation[] deps, string name = "")
{
Operation result = null;
Python.with(ops.control_dependencies(deps), delegate
{
if (string.IsNullOrEmpty(dev))
{
result = gen_control_flow_ops.no_op(name);
}
else
{
result = gen_control_flow_ops.no_op(name);
}
});
return(result);
}
public static Tensor[] internal_convert_n_to_tensor_or_indexed_slices(Tensor[] values, TF_DataType dtype = TF_DataType.DtInvalid, string name = null, bool as_ref = false)
{
var ret = new List <Tensor>();
foreach (var(i, value) in Python.enumerate(values))
{
if (value == null)
{
ret.Add(value);
}
else
{
var n = string.IsNullOrEmpty(name) ? "" : $"{name}_{i}";
ret.Add(internal_convert_to_tensor_or_indexed_slices(value, dtype: dtype, name: n, as_ref: as_ref));
}
}
return(ret.ToArray());
}
public static Tensor random_normal(int[] shape,
float mean = 0.0f,
float stddev = 1.0f,
TF_DataType dtype = TF_DataType.TF_FLOAT,
int?seed = null,
string name = "")
{
return(Python.with <ops.name_scope, Tensor>(new ops.name_scope(name, "random_normal", new object[] { shape, mean, stddev }), scope =>
{
var shape_tensor = _ShapeTensor(shape);
var mean_tensor = ops.convert_to_tensor(mean, dtype: dtype, name: "mean");
var stddev_tensor = ops.convert_to_tensor(stddev, dtype: dtype, name = "stddev");
var(seed1, seed2) = random_seed.get_seed(seed);
var rnd = gen_random_ops.random_standard_normal(shape_tensor, dtype: dtype, seed: seed1, seed2: seed2);
var mul = rnd * stddev_tensor;
var value = math_ops.add(mul, mean_tensor, name: name);
return value;
}));
}
public static Tensor rank_internal(Tensor input, string name = "", bool optimize = true)
{
return(Python.with <ops.name_scope, Tensor>(new ops.name_scope(name, "Rank", new List <Tensor> {
input
}), scope =>
{
name = scope;
var input_tensor = ops.convert_to_tensor(input);
var input_shape = tensor_util.to_shape(input_tensor.shape);
if (optimize && input_shape.NDim == null)
{
return constant_op.constant(input_shape.NDim);
}
else
{
return gen_array_ops.rank(input, name);
}
}));
}
private static Tensor shape_internal(Tensor input, string name = "", bool optimize = true, TF_DataType out_type = TF_DataType.TF_INT32)
{
return(Python.with <ops.name_scope, Tensor>(new ops.name_scope(name, "Shape", new Tensor[] { input }), scope =>
{
name = scope;
if (!tf.context.executing_eagerly())
{
var input_tensor = ops.convert_to_tensor(input);
var input_shape = tensor_util.to_shape(input_tensor.shape);
if (optimize && input_shape.is_fully_defined())
{
var nd = np.array(input_tensor.shape, out_type.as_numpy_datatype());
return constant_op.constant(nd, name: name);
}
}
return gen_array_ops.shape(input);
}));
}
/// <summary>
/// A context manager that lifts ops out of control-flow scopes and function-building graphs.
/// </summary>
/// <returns></returns>
public static void init_scope()
{
// Retrieve the active name scope: entering an `init_scope` preserves
// the name scope of the current context.
var default_graph = get_default_graph();
var scope = default_graph.get_name_scope();
if (!String.IsNullOrEmpty(scope) && !scope.EndsWith("/"))
{
// Names that end with trailing slashes are treated by `name_scope` as
// absolute.
scope += "/";
}
// inner_device_stack = default_graph._device_function_stack
// var outer_context = default_graph.as_default;
Python.with(ops.control_dependencies(null), delegate
{
var outer_graph = get_default_graph();
// outer_device_stack = None
});
}
public static Tensor zeros_like(Tensor tensor, TF_DataType dtype = TF_DataType.DtInvalid, string name = "", bool optimize = true)
{
return(Python.with <ops.name_scope, Tensor>(new ops.name_scope(name, "zeros_like", new Tensor[] { tensor }), scope =>
{
name = scope;
tensor = ops.convert_to_tensor(tensor, name: "tensor");
// is_fully_defined return unexpected value.
if (optimize && tensor_util.to_shape(tensor.shape).is_fully_defined() && dtype != TF_DataType.TF_VARIANT)
{
}
if (dtype != TF_DataType.DtInvalid && dtype != tensor.dtype && dtype != TF_DataType.TF_VARIANT)
{
throw new NotImplementedException("zeros_like");
// return zeros(shape_internal(tensor, optimize: optimize), dtype: dtype, name: name);
}
else
{
return gen_array_ops.zeros_like(tensor, name: name);
}
}));
}
public static Tensor range(object start, object limit = null, object delta = null, TF_DataType dtype = TF_DataType.DtInvalid, string name = "range")
{
if (limit == null)
{
limit = start;
start = 0;
}
if (delta == null)
{
delta = 1;
}
return(Python.with <ops.name_scope, Tensor>(new ops.name_scope(name, "Range", new object[] { start, limit, delta }), scope =>
{
name = scope;
var start1 = ops.convert_to_tensor(start, name: "start");
var limit1 = ops.convert_to_tensor(limit, name: "limit");
var delta1 = ops.convert_to_tensor(delta, name: "delta");
return gen_math_ops.range(start1, limit1, delta1, name);
}));
}
public virtual SaverDef _build_internal(RefVariable[] names_to_saveables,
bool reshape = false,
bool sharded = false,
int max_to_keep = 5,
float keep_checkpoint_every_n_hours = 10000,
string name = "",
bool restore_sequentially = false,
string filename = "model",
bool build_save = true,
bool build_restore = true)
{
if (!build_save || !build_restore)
{
throw new ValueError("save and restore operations need to be built together " +
" when eager execution is not enabled.");
}
var saveables = saveable_object_util.validate_and_slice_inputs(names_to_saveables);
if (max_to_keep < 0)
{
max_to_keep = 0;
}
Tensor save_tensor = null;
Operation restore_op = null;
return(Python.with <ops.name_scope, SaverDef>(new ops.name_scope(name, "save", saveables.Select(x => x.op).ToArray()), scope =>
{
name = scope;
// Add a placeholder string tensor for the filename.
var filename_tensor = array_ops.placeholder_with_default(string.IsNullOrEmpty(filename) ? "model" : filename, shape: new int[0], name: "filename");
// Keep the name "Const" for backwards compatibility.
filename_tensor = gen_array_ops.placeholder_with_default(filename_tensor, shape: new int[0], name: "Const");
// Add the save ops.
if (sharded)
{
}
else
{
if (build_save)
{
save_tensor = _AddSaveOps(filename_tensor, saveables);
}
if (build_restore)
{
restore_op = _AddRestoreOps(filename_tensor, saveables, restore_sequentially, reshape);
}
}
var graph = ops.get_default_graph();
var check_collection_list = graph.get_all_collection_keys();
foreach (var collection_type in check_collection_list)
{
var cols = graph.get_collection(collection_type);
switch (cols)
{
case List <RefVariable> values:
foreach (var element in values)
{
;
}
break;
case List <ITensorOrOperation> values:
foreach (var element in values)
{
;
}
break;
default:
throw new NotImplementedException("_build_internal.check_collection_list");
}
}
return new SaverDef()
{
FilenameTensorName = filename_tensor.name,
SaveTensorName = save_tensor.name,
RestoreOpName = restore_op.name,
MaxToKeep = max_to_keep,
Sharded = sharded,
KeepCheckpointEveryNHours = keep_checkpoint_every_n_hours,
Version = _write_version
};
}));
}
public static Tensor[] _GradientsHelper(Tensor[] ys,
Tensor[] xs,
Tensor[] grad_ys = null,
string name = "gradients",
bool colocate_gradients_with_ops = false,
bool gate_gradients = false,
int aggregation_method = 0,
Tensor[] stop_gradients = null,
Graph src_graph = null)
{
if (src_graph == null)
{
src_graph = ops.get_default_graph();
}
// If src_graph is a _FuncGraph (i.e. a function body), gather it and all
// ancestor graphs. This is necessary for correctly handling captured values.
var curr_graph = src_graph;
if (stop_gradients == null)
{
stop_gradients = new Tensor[0];
}
if (grad_ys == null)
{
grad_ys = new Tensor[ys.Length];
}
var all = new List <Tensor>();
all.AddRange(ys);
all.AddRange(xs);
all.AddRange(stop_gradients);
all.AddRange(grad_ys);
// Iterate over the collected ops.
/**
* grads: op => list of gradients received on each output endpoint of the
* op. The gradients for each endpoint are initially collected as a list.
* When it is time to call the op's gradient function, for each endpoint we
* aggregate the list of received gradients into a Add() Operation if there
* is more than one.
**/
var grads = new Dictionary <string, Tensor[][]>();
with(ops.name_scope(name, "gradients", values: all), scope =>
{
string grad_scope = scope;
// Get a uid for this call to gradients that can be used to help
// cluster ops for compilation.
var gradient_uid = ops.get_default_graph().unique_name("uid");
ys = ops.convert_n_to_tensor_or_indexed_slices(ys, name: "y");
xs = ops.internal_convert_n_to_tensor_or_indexed_slices(xs, name: "x", as_ref: true);
grad_ys = _DefaultGradYs(grad_ys, ys, colocate_gradients_with_ops, gradient_uid);
/**
* The approach we take here is as follows: Create a list of all ops in the
* subgraph between the ys and xs. Visit these ops in reverse order of ids
* to ensure that when we visit an op the gradients w.r.t its outputs have
* been collected. Then aggregate these gradients if needed, call the op's
* gradient function, and add the generated gradients to the gradients for
* its input.
**/
// Initialize the pending count for ops in the connected subgraph from ys
// to the xs.
var to_ops = ys.Select(x => x.op).ToList();
var from_ops = xs.Select(x => x.op).ToList();
var stop_gradient_ops = stop_gradients.Select(x => x.op).ToList();
(var reachable_to_ops, var pending_count, var loop_state) = _PendingCount(to_ops, from_ops, colocate_gradients_with_ops, new List <object>(), xs);
foreach (var(y, grad_y) in Python.zip(ys, grad_ys))
{
_SetGrad(grads, y, grad_y);
}
// Initialize queue with to_ops.
var queue = new Queue <Operation>();
// Add the ops in 'to_ops' into the queue.
var to_ops_set = new List <Operation>();
foreach (var op in to_ops)
{
// 'ready' handles the case where one output gradient relies on
// another output's gradient.
if (!pending_count.ContainsKey(op.name))
{
pending_count[op.name] = 0;
}
bool ready = pending_count[op.name] == 0;
if (ready && !to_ops_set.Contains(op) && reachable_to_ops.Contains(op))
{
to_ops_set.Add(op);
queue.Enqueue(op);
}
}
var stop_ops = _StopOps(from_ops, stop_gradient_ops, pending_count, xs);
while (queue.Count > 0)
{
// generate gradient subgraph for op.
var op = queue.Dequeue();
_maybe_colocate_with(op, gradient_uid, colocate_gradients_with_ops);
//if (loop_state != null)
//loop_state.EnterGradWhileContext(op, before: true);
var out_grads = _AggregatedGrads(grads, op, gradient_uid, loop_state, aggregation_method);
Tensor[] in_grads = null;
var is_partitioned_call = _IsPartitionedCall(op);
var is_func_call = false;
var has_out_grads = true;
if (has_out_grads && !stop_ops.Contains(op))
{
if (is_func_call)
{
}
else
{
// A grad_fn must be defined, either as a function or as None
// for ops that do not have gradients.
var grad_fn = ops.get_gradient_function(op);
foreach (var(i, out_grad) in enumerate(out_grads))
{
if (out_grad == null)
{
if (loop_state != null)
{
;
}
else
{
out_grads[i] = control_flow_ops.ZerosLikeOutsideLoop(op, i);
}
}
}
with(ops.name_scope(op.name + "_grad"), scope1 =>
{
string name1 = scope1;
if (grad_fn != null)
{
in_grads = _MaybeCompile(grad_scope, op, out_grads, null, grad_fn);
_VerifyGeneratedGradients(in_grads, op);
}
if (gate_gradients && in_grads.Count(x => x != null) > 1)
{
ops._colocate_with_for_gradient(null, gradient_uid, ignore_existing: true);
in_grads = control_flow_ops.tuple(in_grads);
}
});
}
}
else
{
in_grads = new Tensor[_NonEagerInputs(op, xs).Count()];
}
var inputs = _NonEagerInputs(op, xs).ToList();
foreach (var(t_in, in_grad) in Python.zip(inputs, in_grads))
{
if (in_grad != null)
{
if (in_grad is Tensor && t_in.dtype != TF_DataType.TF_RESOURCE)
{
in_grad.shape = t_in.shape;
}
_SetGrad(grads, t_in, in_grad);
}
}
// Update pending count for the inputs of op and enqueue ready ops.
_UpdatePendingAndEnqueueReady(grads, op, queue, pending_count, loop_state, xs);
}
});
public Operation apply_gradients(Tuple <Tensor, RefVariable>[] grads_and_vars, Tensor global_step = null, string name = null)
{
// No DistributionStrategy case.
var converted_grads_and_vars = new List <Tuple <Tensor, RefVariable, _OptimizableVariable> >();
foreach (var(g, v) in grads_and_vars)
{
if (g != null)
{
// Convert the grad to Tensor or IndexedSlices if necessary.
var gR = ops.convert_to_tensor_or_indexed_slices(g);
var p = _get_processor(v);
converted_grads_and_vars.Add(new Tuple <Tensor, RefVariable, _OptimizableVariable>(gR, v, p));
}
}
var var_list = converted_grads_and_vars.Where(x => x.Item1 != null).Select(x => x.Item2).ToArray();
if (var_list.Length == 0)
{
throw new ValueError($"No gradients provided for any variable");
}
ops.init_scope();
_create_slots(var_list);
var update_ops = new List <Operation>();
return(Python.with <ops.name_scope, Operation>(new ops.name_scope(name, Name), scope =>
{
name = scope;
_prepare();
foreach (var(grad, var, processor) in converted_grads_and_vars)
{
if (grad == null)
{
continue;
}
var scope_name = var.op.name;
Python.with <ops.name_scope>(new ops.name_scope("update_" + scope_name), scope2 =>
{
update_ops.Add(processor.update_op(this, grad));
});
}
Operation apply_updates = null;
if (global_step == null)
{
apply_updates = _finish(update_ops.ToArray(), name);
}
else
{
}
if (!tf.context.executing_eagerly())
{
var train_op = ops.get_collection_ref(ops.GraphKeys.TRAIN_OP) as List <object>;
if (!train_op.Contains(apply_updates))
{
train_op.Add(apply_updates);
}
}
return apply_updates;
}));
private void _init_from_args(object initial_value,
bool trainable = true,
List <string> collections = null,
bool validate_shape = true,
string caching_device = "",
string name = null,
TF_DataType dtype = TF_DataType.DtInvalid)
{
if (initial_value is null)
{
throw new ValueError("initial_value must be specified.");
}
var init_from_fn = false;
if (collections == null)
{
collections = new List <string> {
ops.GraphKeys.GLOBAL_VARIABLES
};
}
// Store the graph key so optimizers know how to only retrieve variables from
// this graph.
_graph_key = ops.get_default_graph()._graph_key;
_trainable = trainable;
if (!collections.Contains(ops.GraphKeys.TRAINABLE_VARIABLES))
{
collections.Add(ops.GraphKeys.TRAINABLE_VARIABLES);
}
ops.init_scope();
var values = init_from_fn ? new List <object>() : new List <object> {
initial_value
};
Python.with <ops.name_scope>(new ops.name_scope(name, "Variable", values), scope =>
{
if (init_from_fn)
{
}
// Or get the initial value from a Tensor or Python object.
else
{
_initial_value = ops.convert_to_tensor(initial_value, name: "initial_value");
var shape = _initial_value.shape;
dtype = _initial_value.dtype;
_variable = gen_state_ops.variable_v2(shape, dtype.as_base_dtype(), scope);
}
// Manually overrides the variable's shape with the initial value's.
if (validate_shape)
{
var initial_value_shape = _initial_value.shape;
}
// If 'initial_value' makes use of other variables, make sure we don't
// have an issue if these other variables aren't initialized first by
// using their initialized_value() method.
var _initial_value2 = _try_guard_against_uninitialized_dependencies(_initial_value);
_initializer_op = gen_state_ops.assign(_variable, _initial_value2, validate_shape).op;
if (!String.IsNullOrEmpty(caching_device))
{
}
else
{
ops.colocate_with(_initializer_op);
_snapshot = gen_array_ops.identity(_variable, name = "read");
}
ops.add_to_collections(collections, this);
});
}
public Operation _apply_op_helper(string op_type_name, string name = "", dynamic args = null)
{
Dictionary <string, object> keywords = ConvertToDict(args);
var g = ops.get_default_graph();
var op_def = g.GetOpDef(op_type_name);
// Default name if not specified.
if (String.IsNullOrEmpty(name))
{
name = op_type_name;
}
// Check for deprecation
if (op_def.Deprecation != null && op_def.Deprecation.Version > 0)
{
}
var default_type_attr_map = new Dictionary <string, object>();
foreach (var attr_def in op_def.Attr)
{
if (attr_def.Type != "type")
{
continue;
}
var key = attr_def.Name;
if (attr_def.DefaultValue != null)
{
default_type_attr_map[key] = attr_def.DefaultValue.Type;
}
}
var attrs = new Dictionary <string, object>();
var inputs = new List <Tensor>();
var input_types = new List <TF_DataType>();
dynamic values = null;
return(Python.with <ops.name_scope, Operation>(new ops.name_scope(name), scope =>
{
var inferred_from = new Dictionary <string, object>();
var base_types = new List <TF_DataType>();
var types = new List <TF_DataType>();
// Perform input type inference
foreach (var input_arg in op_def.InputArg)
{
var input_name = input_arg.Name;
if (keywords.ContainsKey(input_name))
{
values = keywords[input_name];
}
else if (keywords.ContainsKey(input_name + "_"))
{
input_name += "_";
values = keywords[input_name];
}
else
{
throw new TypeError("No argument for input " + input_name);
}
// Goals:
// * Convert values to Tensors if it contains constants.
// * Verify that values is a list if that matches the input_arg's
// type.
// * If the input_arg's type is determined by attrs, either set
// those attrs and validate those attr values are legal (if
// they have not yet been set) or validate the input matches
// the type indicated by the attrs (if they have already been
// inferred via an earlier input).
// * If the input_arg has an explicit type, make sure the input
// conforms.
DataType dtype = DataType.DtInvalid;
DataType default_dtype = DataType.DtInvalid;
if (_IsListParameter(input_arg))
{
if (!_IsListValue(values))
{
throw new TypeError($"Expected list for '{input_name}' argument to '{op_type_name}' Op, not {values}.");
}
if (input_arg.Type != DataType.DtInvalid)
{
dtype = input_arg.Type;
}
else if (!String.IsNullOrEmpty(input_arg.NumberAttr))
{
if (attrs.ContainsKey(input_arg.TypeAttr))
{
dtype = (DataType)attrs[input_arg.TypeAttr];
}
else
if (values is Tensor[] values1)
{
dtype = values1[0].dtype.as_datatype_enum();
}
if (dtype == DataType.DtInvalid && default_type_attr_map.ContainsKey(input_arg.TypeAttr))
{
default_dtype = (DataType)default_type_attr_map[input_arg.TypeAttr];
}
}
if (input_arg.IsRef && dtype != DataType.DtInvalid)
{
dtype = dtype.as_base_dtype();
}
values = ops.internal_convert_n_to_tensor(values,
name: input_arg.Name,
dtype: dtype.as_tf_dtype(),
preferred_dtype: default_dtype.as_tf_dtype(),
as_ref: input_arg.IsRef);
}
else
{
if (input_arg.Type != DataType.DtInvalid)
{
dtype = input_arg.Type;
}
else if (attrs.ContainsKey(input_arg.TypeAttr))
{
dtype = (DataType)attrs[input_arg.TypeAttr];
}
else if (default_type_attr_map.ContainsKey(input_arg.TypeAttr))
{
default_dtype = (DataType)default_type_attr_map[input_arg.TypeAttr];
}
values = ops.internal_convert_to_tensor(values,
name: input_name,
dtype: dtype.as_tf_dtype(),
as_ref: input_arg.IsRef,
preferred_dtype: default_dtype.as_tf_dtype());
//if (!String.IsNullOrEmpty(input_arg.TypeAttr))
//attrs[input_arg.TypeAttr] = values.dtype;
values = new Tensor[] { values };
}
if (values is Tensor[] values2)
{
types = values2.Select(x => x.dtype).ToList();
inputs.AddRange(values2);
base_types = values2.Select(x => x.dtype.as_base_dtype()).ToList();
}
else
{
throw new NotImplementedException("_IsListParameter");
}
SetAttrs(op_type_name,
input_arg,
op_def,
attrs,
inferred_from,
types,
base_types,
input_types,
values);
}
// Process remaining attrs
foreach (var attr in op_def.Attr)
{
if (keywords.ContainsKey(attr.Name))
{
attrs[attr.Name] = keywords[attr.Name];
}
}
// Convert attr values to AttrValue protos.
var attr_protos = new Dictionary <string, AttrValue>();
foreach (var attr_def in op_def.Attr)
{
var key = attr_def.Name;
var value = attrs[key];
if (!attrs.ContainsKey(key))
{
Console.WriteLine($"_apply_op_helper: key '{key}' is not found in '{op_def.Name}' operation's attr_def.");
}
attr_protos[key] = SetAttrValue(op_def, attr_def, value);
}
attrs.Clear();
// Determine output types (possibly using attrs)
var output_types = new List <TF_DataType>();
foreach (var arg in op_def.OutputArg)
{
types = new List <TF_DataType>();
if (!string.IsNullOrEmpty(arg.NumberAttr))
{
}
else if (!string.IsNullOrEmpty(arg.TypeAttr))
{
types = new List <TF_DataType>()
{
(TF_DataType)attr_protos[arg.TypeAttr].Type
};
}
if (arg.IsRef)
{
types = types.Select(x => x.as_ref()).ToList();
}
output_types.AddRange(types);
}
// Add Op to graph
var op = g.create_op(op_type_name, inputs.ToArray(), output_types.ToArray(),
name: scope,
input_types: input_types.ToArray(),
attrs: attr_protos,
op_def: op_def);
return op;
}));
}
public Operation _apply_op_helper(string op_type_name, string name = "", dynamic args = null)
{
var keywords = ConvertToDict(args);
var g = ops.get_default_graph();
var op_def = g.GetOpDef(op_type_name);
// Default name if not specified.
if (String.IsNullOrEmpty(name))
{
name = op_type_name;
}
// Check for deprecation
if (op_def.Deprecation != null && op_def.Deprecation.Version > 0)
{
}
var default_type_attr_map = new Dictionary <string, object>();
foreach (var attr_def in op_def.Attr)
{
if (attr_def.Type != "type")
{
continue;
}
var key = attr_def.Name;
if (attr_def.DefaultValue != null)
{
default_type_attr_map[key] = attr_def.DefaultValue.Type;
}
}
var attrs = new Dictionary <string, object>();
var inputs = new List <Tensor>();
var input_types = new List <TF_DataType>();
Operation op = null;
Python.with <ops.name_scope>(new ops.name_scope(name), scope =>
{
// Perform input type inference
foreach (var input_arg in op_def.InputArg)
{
var input_name = input_arg.Name;
if (keywords[input_name] is double int_value)
{
keywords[input_name] = constant_op.constant(int_value, input_name);
}
if (keywords[input_name] is Tensor value)
{
if (keywords.ContainsKey(input_name))
{
inputs.Add(value);
}
if (!String.IsNullOrEmpty(input_arg.TypeAttr))
{
attrs[input_arg.TypeAttr] = value.dtype;
}
if (input_arg.IsRef)
{
}
else
{
var base_type = value.dtype.as_base_dtype();
input_types.Add(base_type);
}
}
}
// Process remaining attrs
foreach (var attr in op_def.Attr)
{
if (keywords.ContainsKey(attr.Name))
{
attrs[attr.Name] = keywords[attr.Name];
}
}
// Convert attr values to AttrValue protos.
var attr_protos = new Dictionary <string, AttrValue>();
foreach (var attr_def in op_def.Attr)
{
var key = attr_def.Name;
var value = attrs[key];
var attr_value = new AttrValue();
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 "bool":
attr_value.B = (bool)value;
break;
case "shape":
attr_value.Shape = value == null ?
attr_def.DefaultValue.Shape :
tensor_util.as_shape((long[])value);
break;
default:
throw new InvalidDataException($"attr_def.Type {attr_def.Type}");
}
attr_protos[key] = attr_value;
}
// Determine output types (possibly using attrs)
var output_types = new List <TF_DataType>();
foreach (var arg in op_def.OutputArg)
{
if (!String.IsNullOrEmpty(arg.NumberAttr))
{
}
else if (!String.IsNullOrEmpty(arg.TypeAttr))
{
output_types.Add((TF_DataType)attr_protos[arg.TypeAttr].Type);
}
}
// Add Op to graph
op = g.create_op(op_type_name, inputs, output_types.ToArray(),
name: scope,
input_types: input_types.ToArray(),
attrs: attr_protos,
op_def: op_def);
});
return(op);
}