@tensorflow/tfjs-core
Version:
Hardware-accelerated JavaScript library for machine intelligence
264 lines • 11.2 kB
JavaScript
/**
* @license
* Copyright 2017 Google LLC. 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.
* =============================================================================
*/
import * as tf from '../index';
import { ALL_ENVS, describeWithFlags } from '../jasmine_util';
import { expectArraysClose, expectArraysEqual } from '../test_util';
describeWithFlags('clone', ALL_ENVS, () => {
it('returns a tensor with the same shape and value', async () => {
const a = tf.tensor2d([1, 2, 3, 4, 5, 6, 7, 8, 9], [3, 3]);
const aPrime = tf.clone(a);
expect(aPrime.shape).toEqual(a.shape);
expectArraysClose(await aPrime.data(), await a.data());
expect(aPrime.shape).toEqual(a.shape);
});
it('accepts a tensor-like object', async () => {
const res = tf.clone([[1, 2, 3], [4, 5, 6]]);
expect(res.dtype).toBe('float32');
expect(res.shape).toEqual([2, 3]);
expectArraysClose(await res.data(), [1, 2, 3, 4, 5, 6]);
});
});
describeWithFlags('clone', ALL_ENVS, () => {
it('1D default dtype', async () => {
const a = tf.tensor1d([1, 2, 3]);
const b = tf.clone(a);
expect(b.dtype).toBe('float32');
expect(b.shape).toEqual([3]);
expectArraysClose(await b.data(), [1, 2, 3]);
});
it('1D float32 dtype', async () => {
const a = tf.tensor1d([1, 2, 3], 'float32');
const b = tf.clone(a);
expect(b.dtype).toBe('float32');
expect(b.shape).toEqual([3]);
expectArraysClose(await b.data(), [1, 2, 3]);
});
it('1D int32 dtype', async () => {
const a = tf.tensor1d([1, 2, 3], 'int32');
const b = tf.clone(a);
expect(b.dtype).toBe('int32');
expect(b.shape).toEqual([3]);
expectArraysEqual(await b.data(), [1, 2, 3]);
});
it('1D bool dtype', async () => {
const a = tf.tensor1d([1, 1, 0], 'bool');
const b = tf.clone(a);
expect(b.dtype).toBe('bool');
expect(b.shape).toEqual([3]);
expectArraysEqual(await b.data(), [1, 1, 0]);
});
it('1D complex64 dtype', async () => {
const a = tf.complex([1], [1]);
const b = tf.clone(a);
expect(b.dtype).toBe('complex64');
expect(b.shape).toEqual([1]);
expectArraysEqual(await b.data(), [1, 1]);
});
it('1D string dtype', async () => {
const a = tf.tensor1d(['a', 'b', 'c'], 'string');
const b = tf.clone(a);
expect(b.dtype).toBe('string');
expect(b.shape).toEqual([3]);
expectArraysEqual(await b.data(), ['a', 'b', 'c']);
});
it('2D default dtype', async () => {
const a = tf.tensor2d([1, 2, 3, 4], [2, 2]);
const b = tf.clone(a);
expect(b.dtype).toBe('float32');
expect(b.shape).toEqual([2, 2]);
expectArraysClose(await b.data(), [1, 2, 3, 4]);
});
it('2D float32 dtype', async () => {
const a = tf.tensor2d([1, 2, 3, 4], [2, 2], 'float32');
const b = tf.clone(a);
expect(b.dtype).toBe('float32');
expect(b.shape).toEqual([2, 2]);
expectArraysClose(await b.data(), [1, 2, 3, 4]);
});
it('2D int32 dtype', async () => {
const a = tf.tensor2d([1, 2, 3, 4], [2, 2], 'int32');
const b = tf.clone(a);
expect(b.dtype).toBe('int32');
expect(b.shape).toEqual([2, 2]);
expectArraysEqual(await b.data(), [1, 2, 3, 4]);
});
it('2D bool dtype', async () => {
const a = tf.tensor2d([1, 1, 1, 0], [2, 2], 'bool');
const b = tf.clone(a);
expect(b.dtype).toBe('bool');
expect(b.shape).toEqual([2, 2]);
expectArraysEqual(await b.data(), [1, 1, 1, 0]);
});
it('2D complex64 dtype', async () => {
const a = tf.complex([[1, 3], [5, 7]], [[2, 4], [6, 8]]);
const b = tf.clone(a);
expect(b.dtype).toBe('complex64');
expect(b.shape).toEqual([2, 2]);
expectArraysEqual(await b.data(), [1, 2, 3, 4, 5, 6, 7, 8]);
});
it('2D string dtype', async () => {
const a = tf.tensor2d(['a', 'b', 'c', 'd'], [2, 2], 'string');
const b = tf.clone(a);
expect(b.dtype).toBe('string');
expect(b.shape).toEqual([2, 2]);
expectArraysEqual(await b.data(), ['a', 'b', 'c', 'd']);
});
it('3D default dtype', async () => {
const a = tf.tensor3d([1, 2, 3, 4], [2, 2, 1]);
const b = tf.clone(a);
expect(b.dtype).toBe('float32');
expect(b.shape).toEqual([2, 2, 1]);
expectArraysClose(await b.data(), [1, 2, 3, 4]);
});
it('3D float32 dtype', async () => {
const a = tf.tensor3d([1, 2, 3, 4], [2, 2, 1], 'float32');
const b = tf.clone(a);
expect(b.dtype).toBe('float32');
expect(b.shape).toEqual([2, 2, 1]);
expectArraysClose(await b.data(), [1, 2, 3, 4]);
});
it('3D int32 dtype', async () => {
const a = tf.tensor3d([1, 2, 3, 4], [2, 2, 1], 'int32');
const b = tf.clone(a);
expect(b.dtype).toBe('int32');
expect(b.shape).toEqual([2, 2, 1]);
expectArraysEqual(await b.data(), [1, 2, 3, 4]);
});
it('3D bool dtype', async () => {
const a = tf.tensor3d([1, 1, 1, 0], [2, 2, 1], 'bool');
const b = tf.clone(a);
expect(b.dtype).toBe('bool');
expect(b.shape).toEqual([2, 2, 1]);
expectArraysEqual(await b.data(), [1, 1, 1, 0]);
});
it('3D complex64 dtype', async () => {
const a = tf.complex([[[1], [3]], [[5], [7]]], [[[2], [4]], [[6], [8]]]);
const b = tf.clone(a);
expect(b.dtype).toBe('complex64');
expect(b.shape).toEqual([2, 2, 1]);
expectArraysEqual(await b.data(), [1, 2, 3, 4, 5, 6, 7, 8]);
});
it('3D string dtype', async () => {
const a = tf.tensor3d(['a', 'b', 'c', 'd'], [2, 2, 1], 'string');
const b = tf.clone(a);
expect(b.dtype).toBe('string');
expect(b.shape).toEqual([2, 2, 1]);
expectArraysEqual(await b.data(), ['a', 'b', 'c', 'd']);
});
it('4D default dtype', async () => {
const a = tf.tensor4d([1, 2, 3, 4], [2, 2, 1, 1]);
const b = tf.clone(a);
expect(b.dtype).toBe('float32');
expect(b.shape).toEqual([2, 2, 1, 1]);
expectArraysClose(await b.data(), [1, 2, 3, 4]);
});
it('4D float32 dtype', async () => {
const a = tf.tensor4d([1, 2, 3, 4], [2, 2, 1, 1], 'float32');
const b = tf.clone(a);
expect(b.dtype).toBe('float32');
expect(b.shape).toEqual([2, 2, 1, 1]);
expectArraysClose(await b.data(), [1, 2, 3, 4]);
});
it('4D int32 dtype', async () => {
const a = tf.tensor4d([1, 2, 3, 4], [2, 2, 1, 1], 'int32');
const b = tf.clone(a);
expect(b.dtype).toBe('int32');
expect(b.shape).toEqual([2, 2, 1, 1]);
expectArraysEqual(await b.data(), [1, 2, 3, 4]);
});
it('4D bool dtype', async () => {
const a = tf.tensor4d([1, 1, 1, 0], [2, 2, 1, 1], 'bool');
const b = tf.clone(a);
expect(b.dtype).toBe('bool');
expect(b.shape).toEqual([2, 2, 1, 1]);
expectArraysEqual(await b.data(), [1, 1, 1, 0]);
});
it('4D complex64 dtype', async () => {
const a = tf.complex([[[[1]], [[3]]], [[[5]], [[7]]]], [[[[2]], [[4]]], [[[6]], [[8]]]]);
const b = tf.clone(a);
expect(b.dtype).toBe('complex64');
expect(b.shape).toEqual([2, 2, 1, 1]);
expectArraysEqual(await b.data(), [1, 2, 3, 4, 5, 6, 7, 8]);
});
it('4D string dtype', async () => {
const a = tf.tensor4d(['a', 'b', 'c', 'd'], [2, 2, 1, 1], 'string');
const b = tf.clone(a);
expect(b.dtype).toBe('string');
expect(b.shape).toEqual([2, 2, 1, 1]);
expectArraysEqual(await b.data(), ['a', 'b', 'c', 'd']);
});
it('gradient: 1D', async () => {
const a = tf.tensor1d([1, 2, 3]);
const dy = tf.tensor1d([4, 5, 6]);
const da = tf.grad(x => tf.clone(x))(a, dy);
expect(da.dtype).toBe('float32');
expect(da.shape).toEqual([3]);
expectArraysClose(await da.data(), [4, 5, 6]);
});
it('gradient with clones', async () => {
const a = tf.tensor1d([1, 2, 3]);
const dy = tf.tensor1d([4, 5, 6]);
const da = tf.grad(x => tf.clone(x.clone()).clone())(a, dy);
expect(da.dtype).toBe('float32');
expect(da.shape).toEqual([3]);
expectArraysClose(await da.data(), [4, 5, 6]);
});
it('gradient: 1D string throws error with string dy', () => {
const a = tf.tensor1d(['a', 'b', 'c'], 'string');
const dy = tf.tensor1d(['d', 'e', 'f']);
expect(() => tf.grad(x => tf.clone(x))(a, dy)).toThrowError();
});
it('gradient: 1D string throws error with bool dy', () => {
const a = tf.tensor1d(['a', 'b', 'c'], 'string');
const dy = tf.tensor1d([false, true, false], 'bool');
expect(() => tf.grad(x => tf.clone(x))(a, dy)).toThrowError();
});
it('gradient: 1D string throws error with int32 dy', () => {
const a = tf.tensor1d(['a', 'b', 'c'], 'string');
const dy = tf.tensor1d([4, 5, 6], 'int32');
expect(() => tf.grad(x => tf.clone(x))(a, dy)).toThrowError();
});
it('gradient: 1D string works with float32 dy', async () => {
const a = tf.tensor1d(['a', 'b', 'c'], 'string');
const dy = tf.tensor1d([4, 5, 6]);
const da = tf.grad(x => tf.clone(x))(a, dy);
expect(da.dtype).toBe('float32');
expect(da.shape).toEqual([3]);
expectArraysClose(await da.data(), [4, 5, 6]);
});
it('gradient: 2D int32', async () => {
const a = tf.tensor2d([1, 2, 3, 4], [2, 2], 'int32');
const dy = tf.tensor2d([5, 6, 7, 8], [2, 2], 'float32');
const da = tf.grad(x => tf.clone(x))(a, dy);
expect(da.dtype).toBe('float32');
expect(da.shape).toEqual([2, 2]);
expectArraysEqual(await da.data(), [5, 6, 7, 8]);
});
it('gradient: 4D bool', async () => {
const a = tf.tensor4d([1, 1, 1, 0], [2, 2, 1, 1], 'bool');
const dy = tf.tensor4d([5, 6, 7, 8], [2, 2, 1, 1], 'float32');
const da = tf.grad(x => tf.clone(x))(a, dy);
expect(da.dtype).toBe('float32');
expect(da.shape).toEqual([2, 2, 1, 1]);
expectArraysEqual(await da.data(), [5, 6, 7, 8]);
});
it('throws when passed a non-tensor', () => {
expect(() => tf.clone({}))
.toThrowError(/Argument 'x' passed to 'clone' must be a Tensor/);
});
});
//# sourceMappingURL=clone_test.js.map