@tensorflow/tfjs-core
Version:
Hardware-accelerated JavaScript library for machine intelligence
189 lines • 9.34 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 } from '../test_util';
describeWithFlags('transpose', ALL_ENVS, () => {
it('of scalar is no-op', async () => {
const a = tf.scalar(3);
expectArraysClose(await tf.transpose(a).data(), [3]);
});
it('of 1D is no-op', async () => {
const a = tf.tensor1d([1, 2, 3]);
expectArraysClose(await tf.transpose(a).data(), [1, 2, 3]);
});
it('of scalar with perm of incorrect rank throws error', () => {
const a = tf.scalar(3);
const perm = [0]; // Should be empty array.
expect(() => tf.transpose(a, perm)).toThrowError();
});
it('of 1d with perm out of bounds throws error', () => {
const a = tf.tensor1d([1, 2, 3]);
const perm = [1];
expect(() => tf.transpose(a, perm)).toThrowError();
});
it('of 1d with perm incorrect rank throws error', () => {
const a = tf.tensor1d([1, 2, 3]);
const perm = [0, 0]; // Should be of length 1.
expect(() => tf.transpose(a, perm)).toThrowError();
});
it('2D (no change)', async () => {
const t = tf.tensor2d([1, 11, 2, 22, 3, 33, 4, 44], [2, 4]);
const t2 = tf.transpose(t, [0, 1]);
expect(t2.shape).toEqual(t.shape);
expectArraysClose(await t2.array(), await t.array());
});
it('2D (transpose)', async () => {
const t = tf.tensor2d([1, 11, 2, 22, 3, 33, 4, 44], [2, 4]);
const t2 = tf.transpose(t, [1, 0]);
expect(t2.shape).toEqual([4, 2]);
expectArraysClose(await t2.data(), [1, 3, 11, 33, 2, 4, 22, 44]);
});
it('2D, shape has ones', async () => {
const t = tf.tensor2d([1, 2, 3, 4], [1, 4]);
const t2 = tf.transpose(t, [1, 0]);
expect(t2.shape).toEqual([4, 1]);
expectArraysClose(await t2.data(), [1, 2, 3, 4]);
});
it('3D [r, c, d] => [d, r, c]', async () => {
const t = tf.tensor3d([1, 11, 2, 22, 3, 33, 4, 44], [2, 2, 2]);
const t2 = tf.transpose(t, [2, 0, 1]);
expect(t2.shape).toEqual([2, 2, 2]);
expectArraysClose(await t2.data(), [1, 2, 3, 4, 11, 22, 33, 44]);
});
it('3D [r, c, d] => [d, c, r]', async () => {
const t = tf.tensor3d([1, 11, 2, 22, 3, 33, 4, 44], [2, 2, 2]);
const t2 = tf.transpose(t, [2, 1, 0]);
expect(t2.shape).toEqual([2, 2, 2]);
expectArraysClose(await t2.data(), [1, 3, 2, 4, 11, 33, 22, 44]);
});
it('3D [r, c, d] => [d, r, c], shape has ones', async () => {
const perm = [2, 0, 1];
const t = tf.tensor3d([1, 2, 3, 4], [2, 1, 2]);
const tt = tf.transpose(t, perm);
expect(tt.shape).toEqual([2, 2, 1]);
expectArraysClose(await tt.data(), [1, 3, 2, 4]);
const t2 = tf.tensor3d([1, 2, 3, 4], [2, 2, 1]);
const tt2 = tf.transpose(t2, perm);
expect(tt2.shape).toEqual([1, 2, 2]);
expectArraysClose(await tt2.data(), [1, 2, 3, 4]);
const t3 = tf.tensor3d([1, 2, 3, 4], [1, 2, 2]);
const tt3 = tf.transpose(t3, perm);
expect(tt3.shape).toEqual([2, 1, 2]);
expectArraysClose(await tt3.data(), [1, 3, 2, 4]);
});
it('3D [r, c, d] => [r, d, c]', async () => {
const perm = [0, 2, 1];
const t = tf.tensor3d([1, 2, 3, 4, 5, 6, 7, 8], [2, 2, 2]);
const tt = tf.transpose(t, perm);
expect(tt.shape).toEqual([2, 2, 2]);
expectArraysClose(await tt.data(), [1, 3, 2, 4, 5, 7, 6, 8]);
});
it('5D [r, c, d, e, f] => [r, c, d, f, e]', async () => {
const t = tf.tensor5d(new Array(32).fill(0).map((x, i) => i + 1), [2, 2, 2, 2, 2]);
const t2 = tf.transpose(t, [0, 1, 2, 4, 3]);
expect(t2.shape).toEqual([2, 2, 2, 2, 2]);
expectArraysClose(await t2.data(), [
1, 3, 2, 4, 5, 7, 6, 8, 9, 11, 10, 12, 13, 15, 14, 16,
17, 19, 18, 20, 21, 23, 22, 24, 25, 27, 26, 28, 29, 31, 30, 32
]);
});
it('4D [r, c, d, e] => [c, r, d, e]', async () => {
const t = tf.tensor4d(new Array(16).fill(0).map((x, i) => i + 1), [2, 2, 2, 2]);
const t2 = tf.transpose(t, [1, 0, 2, 3]);
expect(t2.shape).toEqual([2, 2, 2, 2]);
expectArraysClose(await t2.data(), [1, 2, 3, 4, 9, 10, 11, 12, 5, 6, 7, 8, 13, 14, 15, 16]);
});
it('4D [r, c, d, e] => [c, r, e, d]', async () => {
const t = tf.tensor4d(new Array(16).fill(0).map((x, i) => i + 1), [2, 2, 2, 2]);
const t2 = tf.transpose(t, [1, 0, 3, 2]);
expect(t2.shape).toEqual([2, 2, 2, 2]);
expectArraysClose(await t2.data(), [1, 3, 2, 4, 9, 11, 10, 12, 5, 7, 6, 8, 13, 15, 14, 16]);
});
it('4D [r, c, d, e] => [e, r, c, d]', async () => {
const t = tf.tensor4d(new Array(16).fill(0).map((x, i) => i + 1), [2, 2, 2, 2]);
const t2 = tf.transpose(t, [3, 0, 1, 2]);
expect(t2.shape).toEqual([2, 2, 2, 2]);
expectArraysClose(await t2.data(), [1, 3, 5, 7, 9, 11, 13, 15, 2, 4, 6, 8, 10, 12, 14, 16]);
});
it('4D [r, c, d, e] => [d, c, e, r]', async () => {
const t = tf.tensor4d(new Array(16).fill(0).map((x, i) => i + 1), [2, 2, 2, 2]);
const t2 = tf.transpose(t, [2, 1, 3, 0]);
expect(t2.shape).toEqual([2, 2, 2, 2]);
expectArraysClose(await t2.data(), [1, 9, 2, 10, 5, 13, 6, 14, 3, 11, 4, 12, 7, 15, 8, 16]);
});
it('5D [r, c, d, e, f] => [c, r, d, e, f]', async () => {
const t = tf.tensor5d(new Array(32).fill(0).map((x, i) => i + 1), [2, 2, 2, 2, 2]);
const t2 = tf.transpose(t, [1, 0, 2, 3, 4]);
expect(t2.shape).toEqual([2, 2, 2, 2, 2]);
expectArraysClose(await t2.data(), [
1, 2, 3, 4, 5, 6, 7, 8, 17, 18, 19, 20, 21, 22, 23, 24,
9, 10, 11, 12, 13, 14, 15, 16, 25, 26, 27, 28, 29, 30, 31, 32
]);
});
it('6D [r, c, d, e, f] => [r, c, d, f, e]', async () => {
const t = tf.tensor6d(new Array(64).fill(0).map((x, i) => i + 1), [2, 2, 2, 2, 2, 2]);
const t2 = tf.transpose(t, [0, 1, 2, 3, 5, 4]);
expect(t2.shape).toEqual([2, 2, 2, 2, 2, 2]);
expectArraysClose(await t2.data(), [
1, 3, 2, 4, 5, 7, 6, 8, 9, 11, 10, 12, 13, 15, 14, 16,
17, 19, 18, 20, 21, 23, 22, 24, 25, 27, 26, 28, 29, 31, 30, 32,
33, 35, 34, 36, 37, 39, 38, 40, 41, 43, 42, 44, 45, 47, 46, 48,
49, 51, 50, 52, 53, 55, 54, 56, 57, 59, 58, 60, 61, 63, 62, 64
]);
});
it('6D [r, c, d, e, f, g] => [c, r, d, e, f, g]', async () => {
const t = tf.tensor6d(new Array(64).fill(0).map((x, i) => i + 1), [2, 2, 2, 2, 2, 2]);
const t2 = tf.transpose(t, [1, 0, 2, 3, 4, 5]);
expect(t2.shape).toEqual([2, 2, 2, 2, 2, 2]);
expectArraysClose(await t2.data(), [
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,
33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48,
17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32,
49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64
]);
});
it('gradient 3D [r, c, d] => [d, c, r]', async () => {
const t = tf.tensor3d([1, 11, 2, 22, 3, 33, 4, 44], [2, 2, 2]);
const perm = [2, 1, 0];
const dy = tf.tensor3d([111, 211, 121, 221, 112, 212, 122, 222], [2, 2, 2]);
const dt = tf.grad(t => t.transpose(perm))(t, dy);
expect(dt.shape).toEqual(t.shape);
expect(dt.dtype).toEqual('float32');
expectArraysClose(await dt.data(), [111, 112, 121, 122, 211, 212, 221, 222]);
});
it('gradient with clones', async () => {
const t = tf.tensor3d([1, 11, 2, 22, 3, 33, 4, 44], [2, 2, 2]);
const perm = [2, 1, 0];
const dy = tf.tensor3d([111, 211, 121, 221, 112, 212, 122, 222], [2, 2, 2]);
const dt = tf.grad(t => t.clone().transpose(perm).clone())(t, dy);
expect(dt.shape).toEqual(t.shape);
expect(dt.dtype).toEqual('float32');
expectArraysClose(await dt.data(), [111, 112, 121, 122, 211, 212, 221, 222]);
});
it('throws when passed a non-tensor', () => {
expect(() => tf.transpose({}))
.toThrowError(/Argument 'x' passed to 'transpose' must be a Tensor/);
});
it('accepts a tensor-like object', async () => {
const t = [[1, 11, 2, 22], [3, 33, 4, 44]];
const res = tf.transpose(t, [1, 0]);
expect(res.shape).toEqual([4, 2]);
expectArraysClose(await res.data(), [1, 3, 11, 33, 2, 4, 22, 44]);
});
});
//# sourceMappingURL=transpose_test.js.map