"use strict";
|
/**
|
* @license
|
* Copyright 2018 Google Inc. All Rights Reserved.
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
* you may not use this file except in compliance with the License.
|
* You may obtain a copy of the License at
|
*
|
* http://www.apache.org/licenses/LICENSE-2.0
|
*
|
* Unless required by applicable law or agreed to in writing, software
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
* See the License for the specific language governing permissions and
|
* limitations under the License.
|
* =============================================================================
|
*/
|
var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
|
return new (P || (P = Promise))(function (resolve, reject) {
|
function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
|
function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
|
function step(result) { result.done ? resolve(result.value) : new P(function (resolve) { resolve(result.value); }).then(fulfilled, rejected); }
|
step((generator = generator.apply(thisArg, _arguments || [])).next());
|
});
|
};
|
var __generator = (this && this.__generator) || function (thisArg, body) {
|
var _ = { label: 0, sent: function() { if (t[0] & 1) throw t[1]; return t[1]; }, trys: [], ops: [] }, f, y, t, g;
|
return g = { next: verb(0), "throw": verb(1), "return": verb(2) }, typeof Symbol === "function" && (g[Symbol.iterator] = function() { return this; }), g;
|
function verb(n) { return function (v) { return step([n, v]); }; }
|
function step(op) {
|
if (f) throw new TypeError("Generator is already executing.");
|
while (_) try {
|
if (f = 1, y && (t = op[0] & 2 ? y["return"] : op[0] ? y["throw"] || ((t = y["return"]) && t.call(y), 0) : y.next) && !(t = t.call(y, op[1])).done) return t;
|
if (y = 0, t) op = [op[0] & 2, t.value];
|
switch (op[0]) {
|
case 0: case 1: t = op; break;
|
case 4: _.label++; return { value: op[1], done: false };
|
case 5: _.label++; y = op[1]; op = [0]; continue;
|
case 7: op = _.ops.pop(); _.trys.pop(); continue;
|
default:
|
if (!(t = _.trys, t = t.length > 0 && t[t.length - 1]) && (op[0] === 6 || op[0] === 2)) { _ = 0; continue; }
|
if (op[0] === 3 && (!t || (op[1] > t[0] && op[1] < t[3]))) { _.label = op[1]; break; }
|
if (op[0] === 6 && _.label < t[1]) { _.label = t[1]; t = op; break; }
|
if (t && _.label < t[2]) { _.label = t[2]; _.ops.push(op); break; }
|
if (t[2]) _.ops.pop();
|
_.trys.pop(); continue;
|
}
|
op = body.call(thisArg, _);
|
} catch (e) { op = [6, e]; y = 0; } finally { f = t = 0; }
|
if (op[0] & 5) throw op[1]; return { value: op[0] ? op[1] : void 0, done: true };
|
}
|
};
|
var _this = this;
|
Object.defineProperty(exports, "__esModule", { value: true });
|
var tf = require("../index");
|
var jasmine_util_1 = require("../jasmine_util");
|
var test_util_1 = require("../test_util");
|
var optimizer_1 = require("./optimizer");
|
var sgd_optimizer_1 = require("./sgd_optimizer");
|
jasmine_util_1.describeWithFlags('optimizer', jasmine_util_1.ALL_ENVS, function () {
|
it('basic', function () { return __awaiter(_this, void 0, void 0, function () {
|
var learningRate, optimizer, x, bias, strayVariable, numTensors, f, cost, expectedX1, expectedBias1, _a, _b, _c, _d, expectedX2, expectedBias2, _e, _f, _g;
|
return __generator(this, function (_h) {
|
switch (_h.label) {
|
case 0:
|
learningRate = .1;
|
optimizer = tf.train.sgd(learningRate);
|
x = tf.scalar(4).variable();
|
bias = tf.scalar(1).variable();
|
strayVariable = tf.scalar(-1).variable();
|
numTensors = tf.memory().numTensors;
|
f = function () { return x.square().addStrict(bias); };
|
cost = optimizer.minimize(f, /* returnCost */ true);
|
// Cost should be the only additional arrays.
|
expect(tf.memory().numTensors).toBe(numTensors + 1);
|
expectedX1 = -2 * 4 * learningRate + 4;
|
expectedBias1 = -1 * learningRate + 1;
|
_a = test_util_1.expectArraysClose;
|
return [4 /*yield*/, x.data()];
|
case 1:
|
_a.apply(void 0, [_h.sent(), [expectedX1]]);
|
_b = test_util_1.expectArraysClose;
|
return [4 /*yield*/, bias.data()];
|
case 2:
|
_b.apply(void 0, [_h.sent(), [expectedBias1]]);
|
_c = test_util_1.expectArraysClose;
|
return [4 /*yield*/, cost.data()];
|
case 3:
|
_c.apply(void 0, [_h.sent(), [Math.pow(4, 2) + 1]]);
|
// The stray variable should remain unchanged.
|
_d = test_util_1.expectArraysClose;
|
return [4 /*yield*/, strayVariable.data()];
|
case 4:
|
// The stray variable should remain unchanged.
|
_d.apply(void 0, [_h.sent(), [-1]]);
|
cost.dispose();
|
numTensors = tf.memory().numTensors;
|
cost = optimizer.minimize(f, /* returnCost */ false);
|
// There should be no new additional Tensors.
|
expect(tf.memory().numTensors).toBe(numTensors);
|
expectedX2 = -2 * expectedX1 * learningRate + expectedX1;
|
expectedBias2 = -learningRate + expectedBias1;
|
_e = test_util_1.expectArraysClose;
|
return [4 /*yield*/, x.data()];
|
case 5:
|
_e.apply(void 0, [_h.sent(), [expectedX2]]);
|
_f = test_util_1.expectArraysClose;
|
return [4 /*yield*/, bias.data()];
|
case 6:
|
_f.apply(void 0, [_h.sent(), [expectedBias2]]);
|
expect(cost).toBe(null);
|
// The stray variable should remain unchanged.
|
_g = test_util_1.expectArraysClose;
|
return [4 /*yield*/, strayVariable.data()];
|
case 7:
|
// The stray variable should remain unchanged.
|
_g.apply(void 0, [_h.sent(), [-1]]);
|
optimizer.dispose();
|
x.dispose();
|
bias.dispose();
|
strayVariable.dispose();
|
// The only tensors remaining are the arguments to variable().
|
expect(tf.memory().numTensors).toBe(3);
|
return [2 /*return*/];
|
}
|
});
|
}); });
|
it('varList array of all variables', function () { return __awaiter(_this, void 0, void 0, function () {
|
var learningRate, optimizer, x, bias, strayVariable, varList, f, cost, expectedX1, expectedBias1, _a, _b, _c, _d, expectedX2, expectedBias2, _e, _f, _g;
|
return __generator(this, function (_h) {
|
switch (_h.label) {
|
case 0:
|
learningRate = .1;
|
optimizer = new sgd_optimizer_1.SGDOptimizer(learningRate);
|
x = tf.scalar(4).variable();
|
bias = tf.scalar(1).variable();
|
strayVariable = tf.scalar(-1).variable();
|
varList = [x, bias];
|
f = function () { return x.square().addStrict(bias); };
|
cost = optimizer.minimize(f, /* returnCost */ true, varList);
|
expectedX1 = -2 * 4 * learningRate + 4;
|
expectedBias1 = -1 * learningRate + 1;
|
_a = test_util_1.expectArraysClose;
|
return [4 /*yield*/, x.data()];
|
case 1:
|
_a.apply(void 0, [_h.sent(), [expectedX1]]);
|
_b = test_util_1.expectArraysClose;
|
return [4 /*yield*/, bias.data()];
|
case 2:
|
_b.apply(void 0, [_h.sent(), [expectedBias1]]);
|
_c = test_util_1.expectArraysClose;
|
return [4 /*yield*/, cost.data()];
|
case 3:
|
_c.apply(void 0, [_h.sent(), [Math.pow(4, 2) + 1]]);
|
// The stray variable should remain unchanged.
|
_d = test_util_1.expectArraysClose;
|
return [4 /*yield*/, strayVariable.data()];
|
case 4:
|
// The stray variable should remain unchanged.
|
_d.apply(void 0, [_h.sent(), [-1]]);
|
cost = optimizer.minimize(f, /* returnCost */ false, varList);
|
expectedX2 = -2 * expectedX1 * learningRate + expectedX1;
|
expectedBias2 = -learningRate + expectedBias1;
|
_e = test_util_1.expectArraysClose;
|
return [4 /*yield*/, x.data()];
|
case 5:
|
_e.apply(void 0, [_h.sent(), [expectedX2]]);
|
_f = test_util_1.expectArraysClose;
|
return [4 /*yield*/, bias.data()];
|
case 6:
|
_f.apply(void 0, [_h.sent(), [expectedBias2]]);
|
// The stray variable should remain unchanged.
|
_g = test_util_1.expectArraysClose;
|
return [4 /*yield*/, strayVariable.data()];
|
case 7:
|
// The stray variable should remain unchanged.
|
_g.apply(void 0, [_h.sent(), [-1]]);
|
expect(cost).toBe(null);
|
return [2 /*return*/];
|
}
|
});
|
}); });
|
it('varList empty array of variables throws error', function () {
|
var learningRate = .1;
|
var optimizer = new sgd_optimizer_1.SGDOptimizer(learningRate);
|
var x = tf.scalar(4).variable();
|
var bias = tf.scalar(1).variable();
|
// Stray variable.
|
tf.scalar(-1).variable();
|
var varList = [];
|
var f = function () { return x.square().addStrict(bias); };
|
expect(function () { return optimizer.minimize(f, /* returnCost */ true, varList); })
|
.toThrowError();
|
});
|
it('varList subset of variables update', function () { return __awaiter(_this, void 0, void 0, function () {
|
var learningRate, optimizer, x, bias, strayVariable, varList, f, cost, expectedValue1, _a, _b, _c, _d, expectedValue2, _e, _f, _g;
|
return __generator(this, function (_h) {
|
switch (_h.label) {
|
case 0:
|
learningRate = .1;
|
optimizer = new sgd_optimizer_1.SGDOptimizer(learningRate);
|
x = tf.scalar(4).variable();
|
bias = tf.scalar(1).variable();
|
strayVariable = tf.scalar(-1).variable();
|
varList = [x];
|
f = function () { return x.square().addStrict(bias); };
|
cost = optimizer.minimize(f, /* returnCost */ true, varList);
|
expectedValue1 = -2 * 4 * learningRate + 4;
|
_a = test_util_1.expectArraysClose;
|
return [4 /*yield*/, x.data()];
|
case 1:
|
_a.apply(void 0, [_h.sent(), [expectedValue1]]);
|
// bias should remain unchanged.
|
_b = test_util_1.expectArraysClose;
|
return [4 /*yield*/, bias.data()];
|
case 2:
|
// bias should remain unchanged.
|
_b.apply(void 0, [_h.sent(), [1]]);
|
_c = test_util_1.expectArraysClose;
|
return [4 /*yield*/, cost.data()];
|
case 3:
|
_c.apply(void 0, [_h.sent(), [Math.pow(4, 2) + 1]]);
|
// The stray variable should remain unchanged.
|
_d = test_util_1.expectArraysClose;
|
return [4 /*yield*/, strayVariable.data()];
|
case 4:
|
// The stray variable should remain unchanged.
|
_d.apply(void 0, [_h.sent(), [-1]]);
|
cost = optimizer.minimize(f, /* returnCost */ false, varList);
|
expectedValue2 = -2 * expectedValue1 * learningRate + expectedValue1;
|
_e = test_util_1.expectArraysClose;
|
return [4 /*yield*/, x.data()];
|
case 5:
|
_e.apply(void 0, [_h.sent(), [expectedValue2]]);
|
// Bias still should remain unchanged.
|
_f = test_util_1.expectArraysClose;
|
return [4 /*yield*/, bias.data()];
|
case 6:
|
// Bias still should remain unchanged.
|
_f.apply(void 0, [_h.sent(), [1]]);
|
expect(cost).toBe(null);
|
// The stray variable should remain unchanged.
|
_g = test_util_1.expectArraysClose;
|
return [4 /*yield*/, strayVariable.data()];
|
case 7:
|
// The stray variable should remain unchanged.
|
_g.apply(void 0, [_h.sent(), [-1]]);
|
return [2 /*return*/];
|
}
|
});
|
}); });
|
it('only bias trainable', function () { return __awaiter(_this, void 0, void 0, function () {
|
var learningRate, optimizer, trainable, x, bias, strayVariable, f, cost, _a, expectedBias1, _b, _c, _d, _e, expectedBias2, _f, _g;
|
return __generator(this, function (_h) {
|
switch (_h.label) {
|
case 0:
|
learningRate = .1;
|
optimizer = new sgd_optimizer_1.SGDOptimizer(learningRate);
|
trainable = false;
|
x = tf.scalar(4).variable(trainable);
|
bias = tf.scalar(1).variable();
|
strayVariable = tf.scalar(-1).variable();
|
f = function () { return x.square().addStrict(bias); };
|
cost = optimizer.minimize(f, /* returnCost */ true);
|
// x should not have been updated.
|
_a = test_util_1.expectArraysClose;
|
return [4 /*yield*/, x.data()];
|
case 1:
|
// x should not have been updated.
|
_a.apply(void 0, [_h.sent(), [4]]);
|
expectedBias1 = -1 * learningRate + 1;
|
_b = test_util_1.expectArraysClose;
|
return [4 /*yield*/, bias.data()];
|
case 2:
|
_b.apply(void 0, [_h.sent(), [expectedBias1]]);
|
_c = test_util_1.expectArraysClose;
|
return [4 /*yield*/, cost.data()];
|
case 3:
|
_c.apply(void 0, [_h.sent(), [Math.pow(4, 2) + 1]]);
|
// The stray variable should remain unchanged.
|
_d = test_util_1.expectArraysClose;
|
return [4 /*yield*/, strayVariable.data()];
|
case 4:
|
// The stray variable should remain unchanged.
|
_d.apply(void 0, [_h.sent(), [-1]]);
|
cost = optimizer.minimize(f, /* returnCost */ false);
|
// x should not have been updated.
|
_e = test_util_1.expectArraysClose;
|
return [4 /*yield*/, x.data()];
|
case 5:
|
// x should not have been updated.
|
_e.apply(void 0, [_h.sent(), [4]]);
|
expectedBias2 = -learningRate + expectedBias1;
|
_f = test_util_1.expectArraysClose;
|
return [4 /*yield*/, bias.data()];
|
case 6:
|
_f.apply(void 0, [_h.sent(), [expectedBias2]]);
|
expect(cost).toBe(null);
|
// The stray variable should remain unchanged.
|
_g = test_util_1.expectArraysClose;
|
return [4 /*yield*/, strayVariable.data()];
|
case 7:
|
// The stray variable should remain unchanged.
|
_g.apply(void 0, [_h.sent(), [-1]]);
|
return [2 /*return*/];
|
}
|
});
|
}); });
|
it('only bias trainable, only x in varList throws error', function () {
|
var learningRate = .1;
|
var optimizer = new sgd_optimizer_1.SGDOptimizer(learningRate);
|
var trainable = false;
|
var x = tf.scalar(4).variable(trainable);
|
var bias = tf.scalar(1).variable();
|
// stray variable.
|
tf.scalar(-1).variable();
|
var varList = [x];
|
var f = function () { return x.square().addStrict(bias); };
|
expect(function () { return optimizer.minimize(f, /* returnCost */ true, varList); })
|
.toThrowError();
|
});
|
it('instanceof Optimizer', function () {
|
var learningRate = .1;
|
var optimizer = new sgd_optimizer_1.SGDOptimizer(learningRate);
|
expect(optimizer instanceof optimizer_1.Optimizer).toBe(true);
|
});
|
it('throws error when f returns a non-scalar', function () {
|
var learningRate = .1;
|
var optimizer = new sgd_optimizer_1.SGDOptimizer(learningRate);
|
var x = tf.tensor1d([1, 2]).variable();
|
var f = function () { return x.square(); };
|
// tslint:disable-next-line:no-any
|
expect(function () { return optimizer.minimize(f); }).toThrowError();
|
});
|
});
|
//# sourceMappingURL=optimizer_test.js.map
|
