implement UpSampling1D/2D/3D layers, with tests

2026-09-10 12:15:12 +08:00 · 2016-08-26 22:29:22 -04:00
parent 032e290607
commit 9b07e511bf
15 changed files with 1185 additions and 6 deletions
@@ -0,0 +1,82 @@
+/* eslint-env browser, mocha */
+
+describe('convolutional layer: UpSampling2D', function () {
+  const assert = chai.assert
+  const styles = testGlobals.styles
+  const logTime = testGlobals.logTime
+  const stringifyCondensed = testGlobals.stringifyCondensed
+  const approxEquals = KerasJS.testUtils.approxEquals
+  const layers = KerasJS.layers
+
+  before(function () {
+    console.log('\n%cconvolutional layer: UpSampling2D', styles.h1)
+  })
+
+  it(`[convolutional.UpSampling2D.0] size 2x2 upsampling on 3x3x3 input, dimOrdering=tf`, function () {
+    const key = `convolutional.UpSampling2D.0`
+    console.log(`\n%c[${key}] size 2x2 upsampling on 3x3x3 input, dimOrdering=tf`, styles.h3)
+    let testLayer = new layers.UpSampling2D([2, 2], { dimOrdering: 'tf' })
+    let t = new KerasJS.Tensor(TEST_DATA[key].input.data, TEST_DATA[key].input.shape)
+    console.log('%cin', styles.h4, stringifyCondensed(t.tensor))
+    const startTime = performance.now()
+    t = testLayer.call(t)
+    const endTime = performance.now()
+    console.log('%cout', styles.h4, stringifyCondensed(t.tensor))
+    logTime(startTime, endTime)
+    const dataExpected = new Float32Array(TEST_DATA[key].expected.data)
+    const shapeExpected = TEST_DATA[key].expected.shape
+    assert.deepEqual(t.tensor.shape, shapeExpected)
+    assert.isTrue(approxEquals(t.tensor, dataExpected))
+  })
+
+  it(`[convolutional.UpSampling2D.1] size 2x2 upsampling on 3x3x3 input, dimOrdering=th`, function () {
+    const key = `convolutional.UpSampling2D.1`
+    console.log(`\n%c[${key}] size 2x2 upsampling on 3x3x3 input, dimOrdering=th`, styles.h3)
+    let testLayer = new layers.UpSampling2D([2, 2], { dimOrdering: 'th' })
+    let t = new KerasJS.Tensor(TEST_DATA[key].input.data, TEST_DATA[key].input.shape)
+    console.log('%cin', styles.h4, stringifyCondensed(t.tensor))
+    const startTime = performance.now()
+    t = testLayer.call(t)
+    const endTime = performance.now()
+    console.log('%cout', styles.h4, stringifyCondensed(t.tensor))
+    logTime(startTime, endTime)
+    const dataExpected = new Float32Array(TEST_DATA[key].expected.data)
+    const shapeExpected = TEST_DATA[key].expected.shape
+    assert.deepEqual(t.tensor.shape, shapeExpected)
+    assert.isTrue(approxEquals(t.tensor, dataExpected))
+  })
+
+  it(`[convolutional.UpSampling2D.2] size 3x2 upsampling on 4x2x2 input, dimOrdering=tf`, function () {
+    const key = `convolutional.UpSampling2D.2`
+    console.log(`\n%c[${key}] size 3x2 upsampling on 4x2x2 input, dimOrdering=tf`, styles.h3)
+    let testLayer = new layers.UpSampling2D([3, 2], { dimOrdering: 'tf' })
+    let t = new KerasJS.Tensor(TEST_DATA[key].input.data, TEST_DATA[key].input.shape)
+    console.log('%cin', styles.h4, stringifyCondensed(t.tensor))
+    const startTime = performance.now()
+    t = testLayer.call(t)
+    const endTime = performance.now()
+    console.log('%cout', styles.h4, stringifyCondensed(t.tensor))
+    logTime(startTime, endTime)
+    const dataExpected = new Float32Array(TEST_DATA[key].expected.data)
+    const shapeExpected = TEST_DATA[key].expected.shape
+    assert.deepEqual(t.tensor.shape, shapeExpected)
+    assert.isTrue(approxEquals(t.tensor, dataExpected))
+  })
+
+  it(`[convolutional.UpSampling2D.3] size 1x3 upsampling on 4x3x2 input, dimOrdering=th`, function () {
+    const key = `convolutional.UpSampling2D.3`
+    console.log(`\n%c[${key}] size 1x3 upsampling on 4x3x2 input, dimOrdering=th`, styles.h3)
+    let testLayer = new layers.UpSampling2D([1, 3], { dimOrdering: 'th' })
+    let t = new KerasJS.Tensor(TEST_DATA[key].input.data, TEST_DATA[key].input.shape)
+    console.log('%cin', styles.h4, stringifyCondensed(t.tensor))
+    const startTime = performance.now()
+    t = testLayer.call(t)
+    const endTime = performance.now()
+    console.log('%cout', styles.h4, stringifyCondensed(t.tensor))
+    logTime(startTime, endTime)
+    const dataExpected = new Float32Array(TEST_DATA[key].expected.data)
+    const shapeExpected = TEST_DATA[key].expected.shape
+    assert.deepEqual(t.tensor.shape, shapeExpected)
+    assert.isTrue(approxEquals(t.tensor, dataExpected))
+  })
+})