@tensorflow-models/coco-ssd
Version:
Object detection model (coco-ssd) in TensorFlow.js
1,407 lines (1,264 loc) • 61.3 kB
JavaScript
/*eslint-disable block-scoped-var, no-redeclare, no-control-regex, no-prototype-builtins*/
"use strict";
var $protobuf = require("protobufjs/minimal");
var $Reader = $protobuf.Reader, $util = $protobuf.util;
var $root = $protobuf.roots["default"] || ($protobuf.roots["default"] = {});
$root.tensorflow = (function() {
var tensorflow = {};
tensorflow.Any = (function() {
function Any(p) {
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
Any.prototype.typeUrl = "";
Any.prototype.value = $util.newBuffer([]);
Any.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.Any();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
m.typeUrl = r.string();
break;
case 2:
m.value = r.bytes();
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return Any;
})();
tensorflow.DataType = (function() {
var valuesById = {}, values = Object.create(valuesById);
values[valuesById[0] = "DT_INVALID"] = 0;
values[valuesById[1] = "DT_FLOAT"] = 1;
values[valuesById[2] = "DT_DOUBLE"] = 2;
values[valuesById[3] = "DT_INT32"] = 3;
values[valuesById[4] = "DT_UINT8"] = 4;
values[valuesById[5] = "DT_INT16"] = 5;
values[valuesById[6] = "DT_INT8"] = 6;
values[valuesById[7] = "DT_STRING"] = 7;
values[valuesById[8] = "DT_COMPLEX64"] = 8;
values[valuesById[9] = "DT_INT64"] = 9;
values[valuesById[10] = "DT_BOOL"] = 10;
values[valuesById[11] = "DT_QINT8"] = 11;
values[valuesById[12] = "DT_QUINT8"] = 12;
values[valuesById[13] = "DT_QINT32"] = 13;
values[valuesById[14] = "DT_BFLOAT16"] = 14;
values[valuesById[101] = "DT_FLOAT_REF"] = 101;
values[valuesById[102] = "DT_DOUBLE_REF"] = 102;
values[valuesById[103] = "DT_INT32_REF"] = 103;
values[valuesById[104] = "DT_UINT8_REF"] = 104;
values[valuesById[105] = "DT_INT16_REF"] = 105;
values[valuesById[106] = "DT_INT8_REF"] = 106;
values[valuesById[107] = "DT_STRING_REF"] = 107;
values[valuesById[108] = "DT_COMPLEX64_REF"] = 108;
values[valuesById[109] = "DT_INT64_REF"] = 109;
values[valuesById[110] = "DT_BOOL_REF"] = 110;
values[valuesById[111] = "DT_QINT8_REF"] = 111;
values[valuesById[112] = "DT_QUINT8_REF"] = 112;
values[valuesById[113] = "DT_QINT32_REF"] = 113;
values[valuesById[114] = "DT_BFLOAT16_REF"] = 114;
return values;
})();
tensorflow.TensorShape = (function() {
function TensorShape(p) {
this.dim = [];
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
TensorShape.prototype.dim = $util.emptyArray;
TensorShape.prototype.unknownRank = false;
TensorShape.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.TensorShape();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 2:
if (!(m.dim && m.dim.length))
m.dim = [];
m.dim.push($root.tensorflow.TensorShape.Dim.decode(r, r.uint32()));
break;
case 3:
m.unknownRank = r.bool();
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
TensorShape.Dim = (function() {
function Dim(p) {
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
Dim.prototype.size = $util.Long ? $util.Long.fromBits(0,0,false) : 0;
Dim.prototype.name = "";
Dim.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.TensorShape.Dim();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
m.size = r.int64();
break;
case 2:
m.name = r.string();
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return Dim;
})();
return TensorShape;
})();
tensorflow.Tensor = (function() {
function Tensor(p) {
this.floatVal = [];
this.doubleVal = [];
this.intVal = [];
this.stringVal = [];
this.scomplexVal = [];
this.int64Val = [];
this.boolVal = [];
this.uint32Val = [];
this.uint64Val = [];
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
Tensor.prototype.dtype = 0;
Tensor.prototype.tensorShape = null;
Tensor.prototype.versionNumber = 0;
Tensor.prototype.tensorContent = $util.newBuffer([]);
Tensor.prototype.floatVal = $util.emptyArray;
Tensor.prototype.doubleVal = $util.emptyArray;
Tensor.prototype.intVal = $util.emptyArray;
Tensor.prototype.stringVal = $util.emptyArray;
Tensor.prototype.scomplexVal = $util.emptyArray;
Tensor.prototype.int64Val = $util.emptyArray;
Tensor.prototype.boolVal = $util.emptyArray;
Tensor.prototype.uint32Val = $util.emptyArray;
Tensor.prototype.uint64Val = $util.emptyArray;
Tensor.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.Tensor();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
m.dtype = r.int32();
break;
case 2:
m.tensorShape = $root.tensorflow.TensorShape.decode(r, r.uint32());
break;
case 3:
m.versionNumber = r.int32();
break;
case 4:
m.tensorContent = r.bytes();
break;
case 5:
if (!(m.floatVal && m.floatVal.length))
m.floatVal = [];
if ((t & 7) === 2) {
var c2 = r.uint32() + r.pos;
while (r.pos < c2)
m.floatVal.push(r.float());
} else
m.floatVal.push(r.float());
break;
case 6:
if (!(m.doubleVal && m.doubleVal.length))
m.doubleVal = [];
if ((t & 7) === 2) {
var c2 = r.uint32() + r.pos;
while (r.pos < c2)
m.doubleVal.push(r.double());
} else
m.doubleVal.push(r.double());
break;
case 7:
if (!(m.intVal && m.intVal.length))
m.intVal = [];
if ((t & 7) === 2) {
var c2 = r.uint32() + r.pos;
while (r.pos < c2)
m.intVal.push(r.int32());
} else
m.intVal.push(r.int32());
break;
case 8:
if (!(m.stringVal && m.stringVal.length))
m.stringVal = [];
m.stringVal.push(r.bytes());
break;
case 9:
if (!(m.scomplexVal && m.scomplexVal.length))
m.scomplexVal = [];
if ((t & 7) === 2) {
var c2 = r.uint32() + r.pos;
while (r.pos < c2)
m.scomplexVal.push(r.float());
} else
m.scomplexVal.push(r.float());
break;
case 10:
if (!(m.int64Val && m.int64Val.length))
m.int64Val = [];
if ((t & 7) === 2) {
var c2 = r.uint32() + r.pos;
while (r.pos < c2)
m.int64Val.push(r.int64());
} else
m.int64Val.push(r.int64());
break;
case 11:
if (!(m.boolVal && m.boolVal.length))
m.boolVal = [];
if ((t & 7) === 2) {
var c2 = r.uint32() + r.pos;
while (r.pos < c2)
m.boolVal.push(r.bool());
} else
m.boolVal.push(r.bool());
break;
case 16:
if (!(m.uint32Val && m.uint32Val.length))
m.uint32Val = [];
if ((t & 7) === 2) {
var c2 = r.uint32() + r.pos;
while (r.pos < c2)
m.uint32Val.push(r.uint32());
} else
m.uint32Val.push(r.uint32());
break;
case 17:
if (!(m.uint64Val && m.uint64Val.length))
m.uint64Val = [];
if ((t & 7) === 2) {
var c2 = r.uint32() + r.pos;
while (r.pos < c2)
m.uint64Val.push(r.uint64());
} else
m.uint64Val.push(r.uint64());
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return Tensor;
})();
tensorflow.AttrValue = (function() {
function AttrValue(p) {
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
AttrValue.prototype.list = null;
AttrValue.prototype.s = $util.newBuffer([]);
AttrValue.prototype.i = $util.Long ? $util.Long.fromBits(0,0,false) : 0;
AttrValue.prototype.f = 0;
AttrValue.prototype.b = false;
AttrValue.prototype.type = 0;
AttrValue.prototype.shape = null;
AttrValue.prototype.tensor = null;
AttrValue.prototype.placeholder = "";
AttrValue.prototype.func = null;
var $oneOfFields;
Object.defineProperty(AttrValue.prototype, "value", {
get: $util.oneOfGetter($oneOfFields = ["list", "s", "i", "f", "b", "type", "shape", "tensor", "placeholder", "func"]),
set: $util.oneOfSetter($oneOfFields)
});
AttrValue.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.AttrValue();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
m.list = $root.tensorflow.AttrValue.ListValue.decode(r, r.uint32());
break;
case 2:
m.s = r.bytes();
break;
case 3:
m.i = r.int64();
break;
case 4:
m.f = r.float();
break;
case 5:
m.b = r.bool();
break;
case 6:
m.type = r.int32();
break;
case 7:
m.shape = $root.tensorflow.TensorShape.decode(r, r.uint32());
break;
case 8:
m.tensor = $root.tensorflow.Tensor.decode(r, r.uint32());
break;
case 9:
m.placeholder = r.string();
break;
case 10:
m.func = $root.tensorflow.NameAttrList.decode(r, r.uint32());
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
AttrValue.ListValue = (function() {
function ListValue(p) {
this.s = [];
this.i = [];
this.f = [];
this.b = [];
this.type = [];
this.shape = [];
this.tensor = [];
this.func = [];
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
ListValue.prototype.s = $util.emptyArray;
ListValue.prototype.i = $util.emptyArray;
ListValue.prototype.f = $util.emptyArray;
ListValue.prototype.b = $util.emptyArray;
ListValue.prototype.type = $util.emptyArray;
ListValue.prototype.shape = $util.emptyArray;
ListValue.prototype.tensor = $util.emptyArray;
ListValue.prototype.func = $util.emptyArray;
ListValue.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.AttrValue.ListValue();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 2:
if (!(m.s && m.s.length))
m.s = [];
m.s.push(r.bytes());
break;
case 3:
if (!(m.i && m.i.length))
m.i = [];
if ((t & 7) === 2) {
var c2 = r.uint32() + r.pos;
while (r.pos < c2)
m.i.push(r.int64());
} else
m.i.push(r.int64());
break;
case 4:
if (!(m.f && m.f.length))
m.f = [];
if ((t & 7) === 2) {
var c2 = r.uint32() + r.pos;
while (r.pos < c2)
m.f.push(r.float());
} else
m.f.push(r.float());
break;
case 5:
if (!(m.b && m.b.length))
m.b = [];
if ((t & 7) === 2) {
var c2 = r.uint32() + r.pos;
while (r.pos < c2)
m.b.push(r.bool());
} else
m.b.push(r.bool());
break;
case 6:
if (!(m.type && m.type.length))
m.type = [];
if ((t & 7) === 2) {
var c2 = r.uint32() + r.pos;
while (r.pos < c2)
m.type.push(r.int32());
} else
m.type.push(r.int32());
break;
case 7:
if (!(m.shape && m.shape.length))
m.shape = [];
m.shape.push($root.tensorflow.TensorShape.decode(r, r.uint32()));
break;
case 8:
if (!(m.tensor && m.tensor.length))
m.tensor = [];
m.tensor.push($root.tensorflow.Tensor.decode(r, r.uint32()));
break;
case 9:
if (!(m.func && m.func.length))
m.func = [];
m.func.push($root.tensorflow.NameAttrList.decode(r, r.uint32()));
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return ListValue;
})();
return AttrValue;
})();
tensorflow.NameAttrList = (function() {
function NameAttrList(p) {
this.attr = {};
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
NameAttrList.prototype.name = "";
NameAttrList.prototype.attr = $util.emptyObject;
NameAttrList.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.NameAttrList(), k;
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
m.name = r.string();
break;
case 2:
r.skip().pos++;
if (m.attr === $util.emptyObject)
m.attr = {};
k = r.string();
r.pos++;
m.attr[k] = $root.tensorflow.AttrValue.decode(r, r.uint32());
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return NameAttrList;
})();
tensorflow.NodeDef = (function() {
function NodeDef(p) {
this.input = [];
this.attr = {};
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
NodeDef.prototype.name = "";
NodeDef.prototype.op = "";
NodeDef.prototype.input = $util.emptyArray;
NodeDef.prototype.device = "";
NodeDef.prototype.attr = $util.emptyObject;
NodeDef.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.NodeDef(), k;
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
m.name = r.string();
break;
case 2:
m.op = r.string();
break;
case 3:
if (!(m.input && m.input.length))
m.input = [];
m.input.push(r.string());
break;
case 4:
m.device = r.string();
break;
case 5:
r.skip().pos++;
if (m.attr === $util.emptyObject)
m.attr = {};
k = r.string();
r.pos++;
m.attr[k] = $root.tensorflow.AttrValue.decode(r, r.uint32());
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return NodeDef;
})();
tensorflow.VersionDef = (function() {
function VersionDef(p) {
this.badConsumers = [];
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
VersionDef.prototype.producer = 0;
VersionDef.prototype.minConsumer = 0;
VersionDef.prototype.badConsumers = $util.emptyArray;
VersionDef.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.VersionDef();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
m.producer = r.int32();
break;
case 2:
m.minConsumer = r.int32();
break;
case 3:
if (!(m.badConsumers && m.badConsumers.length))
m.badConsumers = [];
if ((t & 7) === 2) {
var c2 = r.uint32() + r.pos;
while (r.pos < c2)
m.badConsumers.push(r.int32());
} else
m.badConsumers.push(r.int32());
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return VersionDef;
})();
tensorflow.GraphDef = (function() {
function GraphDef(p) {
this.node = [];
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
GraphDef.prototype.node = $util.emptyArray;
GraphDef.prototype.versions = null;
GraphDef.prototype.library = null;
GraphDef.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.GraphDef();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
if (!(m.node && m.node.length))
m.node = [];
m.node.push($root.tensorflow.NodeDef.decode(r, r.uint32()));
break;
case 4:
m.versions = $root.tensorflow.VersionDef.decode(r, r.uint32());
break;
case 2:
m.library = $root.tensorflow.FunctionDefLibrary.decode(r, r.uint32());
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return GraphDef;
})();
tensorflow.CollectionDef = (function() {
function CollectionDef(p) {
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
CollectionDef.prototype.nodeList = null;
CollectionDef.prototype.bytesList = null;
CollectionDef.prototype.int64List = null;
CollectionDef.prototype.floatList = null;
CollectionDef.prototype.anyList = null;
var $oneOfFields;
Object.defineProperty(CollectionDef.prototype, "kind", {
get: $util.oneOfGetter($oneOfFields = ["nodeList", "bytesList", "int64List", "floatList", "anyList"]),
set: $util.oneOfSetter($oneOfFields)
});
CollectionDef.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.CollectionDef();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
m.nodeList = $root.tensorflow.CollectionDef.NodeList.decode(r, r.uint32());
break;
case 2:
m.bytesList = $root.tensorflow.CollectionDef.BytesList.decode(r, r.uint32());
break;
case 3:
m.int64List = $root.tensorflow.CollectionDef.Int64List.decode(r, r.uint32());
break;
case 4:
m.floatList = $root.tensorflow.CollectionDef.FloatList.decode(r, r.uint32());
break;
case 5:
m.anyList = $root.tensorflow.CollectionDef.AnyList.decode(r, r.uint32());
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
CollectionDef.NodeList = (function() {
function NodeList(p) {
this.value = [];
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
NodeList.prototype.value = $util.emptyArray;
NodeList.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.CollectionDef.NodeList();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
if (!(m.value && m.value.length))
m.value = [];
m.value.push(r.string());
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return NodeList;
})();
CollectionDef.BytesList = (function() {
function BytesList(p) {
this.value = [];
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
BytesList.prototype.value = $util.emptyArray;
BytesList.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.CollectionDef.BytesList();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
if (!(m.value && m.value.length))
m.value = [];
m.value.push(r.bytes());
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return BytesList;
})();
CollectionDef.Int64List = (function() {
function Int64List(p) {
this.value = [];
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
Int64List.prototype.value = $util.emptyArray;
Int64List.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.CollectionDef.Int64List();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
if (!(m.value && m.value.length))
m.value = [];
if ((t & 7) === 2) {
var c2 = r.uint32() + r.pos;
while (r.pos < c2)
m.value.push(r.int64());
} else
m.value.push(r.int64());
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return Int64List;
})();
CollectionDef.FloatList = (function() {
function FloatList(p) {
this.value = [];
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
FloatList.prototype.value = $util.emptyArray;
FloatList.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.CollectionDef.FloatList();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
if (!(m.value && m.value.length))
m.value = [];
if ((t & 7) === 2) {
var c2 = r.uint32() + r.pos;
while (r.pos < c2)
m.value.push(r.float());
} else
m.value.push(r.float());
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return FloatList;
})();
CollectionDef.AnyList = (function() {
function AnyList(p) {
this.value = [];
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
AnyList.prototype.value = $util.emptyArray;
AnyList.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.CollectionDef.AnyList();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
if (!(m.value && m.value.length))
m.value = [];
m.value.push($root.tensorflow.Any.decode(r, r.uint32()));
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return AnyList;
})();
return CollectionDef;
})();
tensorflow.SaverDef = (function() {
function SaverDef(p) {
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
SaverDef.prototype.filenameTensorName = "";
SaverDef.prototype.saveTensorName = "";
SaverDef.prototype.restoreOpName = "";
SaverDef.prototype.maxToKeep = 0;
SaverDef.prototype.sharded = false;
SaverDef.prototype.keepCheckpointEveryNHours = 0;
SaverDef.prototype.version = 0;
SaverDef.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.SaverDef();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
m.filenameTensorName = r.string();
break;
case 2:
m.saveTensorName = r.string();
break;
case 3:
m.restoreOpName = r.string();
break;
case 4:
m.maxToKeep = r.int32();
break;
case 5:
m.sharded = r.bool();
break;
case 6:
m.keepCheckpointEveryNHours = r.float();
break;
case 7:
m.version = r.int32();
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
SaverDef.CheckpointFormatVersion = (function() {
var valuesById = {}, values = Object.create(valuesById);
values[valuesById[0] = "LEGACY"] = 0;
values[valuesById[1] = "V1"] = 1;
values[valuesById[2] = "V2"] = 2;
return values;
})();
return SaverDef;
})();
tensorflow.TensorInfo = (function() {
function TensorInfo(p) {
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
TensorInfo.prototype.name = "";
TensorInfo.prototype.cooSparse = null;
TensorInfo.prototype.dtype = 0;
TensorInfo.prototype.tensorShape = null;
var $oneOfFields;
Object.defineProperty(TensorInfo.prototype, "encoding", {
get: $util.oneOfGetter($oneOfFields = ["name", "cooSparse"]),
set: $util.oneOfSetter($oneOfFields)
});
TensorInfo.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.TensorInfo();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
m.name = r.string();
break;
case 4:
m.cooSparse = $root.tensorflow.TensorInfo.CooSparse.decode(r, r.uint32());
break;
case 2:
m.dtype = r.int32();
break;
case 3:
m.tensorShape = $root.tensorflow.TensorShape.decode(r, r.uint32());
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
TensorInfo.CooSparse = (function() {
function CooSparse(p) {
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
CooSparse.prototype.valuesTensorName = "";
CooSparse.prototype.indicesTensorName = "";
CooSparse.prototype.denseShapeTensorName = "";
CooSparse.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.TensorInfo.CooSparse();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
m.valuesTensorName = r.string();
break;
case 2:
m.indicesTensorName = r.string();
break;
case 3:
m.denseShapeTensorName = r.string();
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return CooSparse;
})();
return TensorInfo;
})();
tensorflow.SignatureDef = (function() {
function SignatureDef(p) {
this.inputs = {};
this.outputs = {};
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
SignatureDef.prototype.inputs = $util.emptyObject;
SignatureDef.prototype.outputs = $util.emptyObject;
SignatureDef.prototype.methodName = "";
SignatureDef.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.SignatureDef(), k;
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
r.skip().pos++;
if (m.inputs === $util.emptyObject)
m.inputs = {};
k = r.string();
r.pos++;
m.inputs[k] = $root.tensorflow.TensorInfo.decode(r, r.uint32());
break;
case 2:
r.skip().pos++;
if (m.outputs === $util.emptyObject)
m.outputs = {};
k = r.string();
r.pos++;
m.outputs[k] = $root.tensorflow.TensorInfo.decode(r, r.uint32());
break;
case 3:
m.methodName = r.string();
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return SignatureDef;
})();
tensorflow.AssetFileDef = (function() {
function AssetFileDef(p) {
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
AssetFileDef.prototype.tensorInfo = null;
AssetFileDef.prototype.filename = "";
AssetFileDef.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.AssetFileDef();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
m.tensorInfo = $root.tensorflow.TensorInfo.decode(r, r.uint32());
break;
case 2:
m.filename = r.string();
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return AssetFileDef;
})();
tensorflow.OpDef = (function() {
function OpDef(p) {
this.inputArg = [];
this.outputArg = [];
this.attr = [];
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
OpDef.prototype.name = "";
OpDef.prototype.inputArg = $util.emptyArray;
OpDef.prototype.outputArg = $util.emptyArray;
OpDef.prototype.attr = $util.emptyArray;
OpDef.prototype.deprecation = null;
OpDef.prototype.summary = "";
OpDef.prototype.description = "";
OpDef.prototype.isCommutative = false;
OpDef.prototype.isAggregate = false;
OpDef.prototype.isStateful = false;
OpDef.prototype.allowsUninitializedInput = false;
OpDef.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.OpDef();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
m.name = r.string();
break;
case 2:
if (!(m.inputArg && m.inputArg.length))
m.inputArg = [];
m.inputArg.push($root.tensorflow.OpDef.ArgDef.decode(r, r.uint32()));
break;
case 3:
if (!(m.outputArg && m.outputArg.length))
m.outputArg = [];
m.outputArg.push($root.tensorflow.OpDef.ArgDef.decode(r, r.uint32()));
break;
case 4:
if (!(m.attr && m.attr.length))
m.attr = [];
m.attr.push($root.tensorflow.OpDef.AttrDef.decode(r, r.uint32()));
break;
case 8:
m.deprecation = $root.tensorflow.OpDef.OpDeprecation.decode(r, r.uint32());
break;
case 5:
m.summary = r.string();
break;
case 6:
m.description = r.string();
break;
case 18:
m.isCommutative = r.bool();
break;
case 16:
m.isAggregate = r.bool();
break;
case 17:
m.isStateful = r.bool();
break;
case 19:
m.allowsUninitializedInput = r.bool();
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
OpDef.ArgDef = (function() {
function ArgDef(p) {
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
ArgDef.prototype.name = "";
ArgDef.prototype.description = "";
ArgDef.prototype.type = 0;
ArgDef.prototype.typeAttr = "";
ArgDef.prototype.numberAttr = "";
ArgDef.prototype.typeListAttr = "";
ArgDef.prototype.isRef = false;
ArgDef.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.OpDef.ArgDef();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
m.name = r.string();
break;
case 2:
m.description = r.string();
break;
case 3:
m.type = r.int32();
break;
case 4:
m.typeAttr = r.string();
break;
case 5:
m.numberAttr = r.string();
break;
case 6:
m.typeListAttr = r.string();
break;
case 16:
m.isRef = r.bool();
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return ArgDef;
})();
OpDef.AttrDef = (function() {
function AttrDef(p) {
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
AttrDef.prototype.name = "";
AttrDef.prototype.type = "";
AttrDef.prototype.defaultValue = null;
AttrDef.prototype.description = "";
AttrDef.prototype.hasMinimum = false;
AttrDef.prototype.minimum = $util.Long ? $util.Long.fromBits(0,0,false) : 0;
AttrDef.prototype.allowedValues = null;
AttrDef.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.OpDef.AttrDef();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
m.name = r.string();
break;
case 2:
m.type = r.string();
break;
case 3:
m.defaultValue = $root.tensorflow.AttrValue.decode(r, r.uint32());
break;
case 4:
m.description = r.string();
break;
case 5:
m.hasMinimum = r.bool();
break;
case 6:
m.minimum = r.int64();
break;
case 7:
m.allowedValues = $root.tensorflow.AttrValue.decode(r, r.uint32());
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return AttrDef;
})();
OpDef.OpDeprecation = (function() {
function OpDeprecation(p) {
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
OpDeprecation.prototype.version = 0;
OpDeprecation.prototype.explanation = "";
OpDeprecation.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.OpDef.OpDeprecation();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1:
m.version = r.int32();
break;
case 2:
m.explanation = r.string();
break;
default:
r.skipType(t & 7);
break;
}
}
return m;
};
return OpDeprecation;
})();
return OpDef;
})();
tensorflow.OpList = (function() {
function OpList(p) {
this.op = [];
if (p)
for (var ks = Object.keys(p), i = 0; i < ks.length; ++i)
if (p[ks[i]] != null)
this[ks[i]] = p[ks[i]];
}
OpList.prototype.op = $util.emptyArray;
OpList.decode = function decode(r, l) {
if (!(r instanceof $Reader))
r = $Reader.create(r);
var c = l === undefined ? r.len : r.pos + l, m = new $root.tensorflow.OpList();
while (r.pos < c) {
var t = r.uint32();
switch (t >>> 3) {
case 1: