mirror of
https://github.com/wassname/keras-js.git
synced 2026-09-10 12:15:12 +08:00
add IMDB bidirectional LSTM demo
This commit is contained in:
Leon Chen
@@ -0,0 +1,133 @@
|
||||
@import './_variables.css';
|
||||
|
||||
.demo.imdb-bidirectional-lstm {
|
||||
.column {
|
||||
display: flex;
|
||||
align-items: center;
|
||||
justify-content: center;
|
||||
}
|
||||
|
||||
.column.input-column {
|
||||
justify-content: center;
|
||||
|
||||
.input-container {
|
||||
text-align: right;
|
||||
margin: 5px 5px 5px 20px;
|
||||
position: relative;
|
||||
|
||||
.input-label {
|
||||
font-family: $font-3;
|
||||
font-size: 18px;
|
||||
color: $color-2;
|
||||
text-align: left;
|
||||
}
|
||||
|
||||
.mdl-textfield {
|
||||
width: 550px;
|
||||
|
||||
textarea {
|
||||
color: $color-3;
|
||||
font-family: $font-1;
|
||||
font-size: 18px;
|
||||
padding: 10px;
|
||||
}
|
||||
}
|
||||
|
||||
.input-clear {
|
||||
display: flex;
|
||||
align-items: center;
|
||||
justify-content: flex-end;
|
||||
color: $color-2;
|
||||
transition: color 0.2s ease-in;
|
||||
|
||||
&:hover {
|
||||
color: $color-1-lighter;
|
||||
cursor: pointer;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
.column.output-column {
|
||||
justify-content: center;
|
||||
|
||||
.output {
|
||||
height: 160px;
|
||||
display: flex;
|
||||
flex-direction: row;
|
||||
align-items: flex-end;
|
||||
justify-content: center;
|
||||
user-select: none;
|
||||
cursor: default;
|
||||
|
||||
.output-class {
|
||||
display: flex;
|
||||
flex-direction: column;
|
||||
align-items: center;
|
||||
justify-content: center;
|
||||
padding: 0 6px;
|
||||
border-bottom: 2px solid $color-1-lighter;
|
||||
|
||||
.output-label {
|
||||
font-family: $font-2;
|
||||
font-size: 1.5rem;
|
||||
color: $color-2;
|
||||
}
|
||||
|
||||
.output-bar {
|
||||
width: 8px;
|
||||
background: #EEEEEE;
|
||||
transition: height 0.2s ease-out;
|
||||
}
|
||||
}
|
||||
|
||||
.output-class.predicted {
|
||||
border-bottom-color: $color-1;
|
||||
|
||||
.output-label {
|
||||
color: $color-1;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
.layer-results-container {
|
||||
position: relative;
|
||||
|
||||
.layer-result {
|
||||
position: relative;
|
||||
z-index: 1;
|
||||
margin: 30px 20px;
|
||||
background: white;
|
||||
border-radius: 10px;
|
||||
padding: 20px;
|
||||
overflow-x: auto;
|
||||
|
||||
.layer-result-heading {
|
||||
font-size: 1rem;
|
||||
color: #999999;
|
||||
margin-bottom: 10px;
|
||||
display: flex;
|
||||
flex-direction: column;
|
||||
font-size: 12px;
|
||||
|
||||
span.layer-class {
|
||||
color: $color-1;
|
||||
font-size: 14px;
|
||||
font-weight: bold;
|
||||
}
|
||||
}
|
||||
|
||||
.layer-result-canvas-container {
|
||||
display: inline-flex;
|
||||
flex-wrap: wrap;
|
||||
background: white;
|
||||
|
||||
canvas {
|
||||
border: 1px solid lightgray;
|
||||
margin: 1px;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,215 @@
|
||||
/* global Vue */
|
||||
import './imdb-bidirectional-lstm.css'
|
||||
|
||||
import debounce from 'lodash/debounce'
|
||||
import * as utils from './utils'
|
||||
|
||||
const MODEL_FILEPATHS_DEV = {
|
||||
model: '/demos/data/imdb_bidirectional_lstm/imdb_bidirectional_lstm.json',
|
||||
weights: '/demos/data/imdb_bidirectional_lstm/imdb_bidirectional_lstm_weights.buf',
|
||||
metadata: '/demos/data/imdb_bidirectional_lstm/imdb_bidirectional_lstm_metadata.json'
|
||||
}
|
||||
const MODEL_FILEPATHS_PROD = {
|
||||
model: 'demos/data/imdb_bidirectional_lstm/imdb_bidirectional_lstm.json',
|
||||
weights: 'https://transcranial.github.io/keras-js-demos-data/imdb_bidirectional_lstm/imdb_bidirectional_lstm_weights.buf',
|
||||
metadata: 'demos/data/imdb_bidirectional_lstm/imdb_bidirectional_lstm_metadata.json'
|
||||
}
|
||||
const MODEL_CONFIG = {
|
||||
filepaths: (process.env.NODE_ENV === 'production') ? MODEL_FILEPATHS_PROD : MODEL_FILEPATHS_DEV
|
||||
}
|
||||
|
||||
const LAYER_DISPLAY_CONFIG = {
|
||||
}
|
||||
|
||||
/**
|
||||
*
|
||||
* VUE COMPONENT
|
||||
*
|
||||
*/
|
||||
export const ImdbBidirectionalLstm = Vue.extend({
|
||||
props: ['hasWebgl'],
|
||||
|
||||
template: require('raw!./imdb-bidirectional-lstm.template.html'),
|
||||
|
||||
data: function () {
|
||||
return {
|
||||
model: new KerasJS.Model(Object.assign({ gpu: this.hasWebgl }, MODEL_CONFIG)),
|
||||
modelLoading: true,
|
||||
input: new Float32Array(200),
|
||||
output: new Float32Array(1),
|
||||
layerResultImages: [],
|
||||
layerDisplayConfig: LAYER_DISPLAY_CONFIG,
|
||||
drawing: false,
|
||||
strokes: [],
|
||||
useGpu: this.hasWebgl
|
||||
}
|
||||
},
|
||||
|
||||
computed: {
|
||||
loadingProgress: function () {
|
||||
return this.model.getLoadingProgress()
|
||||
}
|
||||
},
|
||||
|
||||
ready: function () {
|
||||
this.model.ready().then(() => {
|
||||
this.modelLoading = false
|
||||
this.$nextTick(function () {
|
||||
this.getIntermediateResults()
|
||||
})
|
||||
})
|
||||
},
|
||||
|
||||
methods: {
|
||||
|
||||
toggleGpu: function () {
|
||||
this.model.toggleGpu(!this.useGpu)
|
||||
},
|
||||
|
||||
clear: function (e) {
|
||||
this.clearIntermediateResults()
|
||||
const ctx = document.getElementById('input-canvas').getContext('2d')
|
||||
ctx.clearRect(0, 0, ctx.canvas.width, ctx.canvas.height)
|
||||
const ctxCenterCrop = document.getElementById('input-canvas-centercrop').getContext('2d')
|
||||
ctxCenterCrop.clearRect(0, 0, ctxCenterCrop.canvas.width, ctxCenterCrop.canvas.height)
|
||||
const ctxScaled = document.getElementById('input-canvas-scaled').getContext('2d')
|
||||
ctxScaled.clearRect(0, 0, ctxScaled.canvas.width, ctxScaled.canvas.height)
|
||||
this.output = new Float32Array(10)
|
||||
this.drawing = false
|
||||
this.strokes = []
|
||||
},
|
||||
|
||||
activateDraw: function (e) {
|
||||
this.drawing = true
|
||||
this.strokes.push([])
|
||||
let points = this.strokes[this.strokes.length - 1]
|
||||
points.push(utils.getCoordinates(e))
|
||||
},
|
||||
|
||||
draw: function (e) {
|
||||
if (!this.drawing) return
|
||||
|
||||
const ctx = document.getElementById('input-canvas').getContext('2d')
|
||||
|
||||
ctx.lineWidth = 20
|
||||
ctx.lineJoin = ctx.lineCap = 'round'
|
||||
ctx.strokeStyle = '#393E46'
|
||||
|
||||
ctx.clearRect(0, 0, ctx.canvas.width, ctx.canvas.height)
|
||||
|
||||
let points = this.strokes[this.strokes.length - 1]
|
||||
points.push(utils.getCoordinates(e))
|
||||
|
||||
// draw individual strokes
|
||||
for (let s = 0, slen = this.strokes.length; s < slen; s++) {
|
||||
points = this.strokes[s]
|
||||
|
||||
let p1 = points[0]
|
||||
let p2 = points[1]
|
||||
ctx.beginPath()
|
||||
ctx.moveTo(...p1)
|
||||
|
||||
// draw points in stroke
|
||||
// quadratic bezier curve
|
||||
for (let i = 1, len = points.length; i < len; i++) {
|
||||
ctx.quadraticCurveTo(...p1, ...utils.getMidpoint(p1, p2))
|
||||
p1 = points[i]
|
||||
p2 = points[i + 1]
|
||||
}
|
||||
ctx.lineTo(...p1)
|
||||
ctx.stroke()
|
||||
}
|
||||
},
|
||||
|
||||
deactivateDrawAndPredict: debounce(function () {
|
||||
if (!this.drawing) return
|
||||
this.drawing = false
|
||||
|
||||
const ctx = document.getElementById('input-canvas').getContext('2d')
|
||||
|
||||
// center crop
|
||||
const imageDataCenterCrop = utils.centerCrop(ctx.getImageData(0, 0, ctx.canvas.width, ctx.canvas.height))
|
||||
const ctxCenterCrop = document.getElementById('input-canvas-centercrop').getContext('2d')
|
||||
ctxCenterCrop.canvas.width = imageDataCenterCrop.width
|
||||
ctxCenterCrop.canvas.height = imageDataCenterCrop.height
|
||||
ctxCenterCrop.putImageData(imageDataCenterCrop, 0, 0)
|
||||
|
||||
// scaled to 28 x 28
|
||||
const ctxScaled = document.getElementById('input-canvas-scaled').getContext('2d')
|
||||
ctxScaled.save()
|
||||
ctxScaled.scale(28 / ctxCenterCrop.canvas.width, 28 / ctxCenterCrop.canvas.height)
|
||||
ctxScaled.clearRect(0, 0, ctxCenterCrop.canvas.width, ctxCenterCrop.canvas.height)
|
||||
ctxScaled.drawImage(document.getElementById('input-canvas-centercrop'), 0, 0)
|
||||
const imageDataScaled = ctxScaled.getImageData(0, 0, ctxScaled.canvas.width, ctxScaled.canvas.height)
|
||||
ctxScaled.restore()
|
||||
|
||||
// process image data for model input
|
||||
const { data } = imageDataScaled
|
||||
this.input = new Float32Array(784)
|
||||
for (let i = 0, len = data.length; i < len; i += 4) {
|
||||
this.input[i / 4] = data[i + 3] / 255
|
||||
}
|
||||
|
||||
this.model.predict({ input: this.input }).then(outputData => {
|
||||
this.output = outputData.output
|
||||
this.getIntermediateResults()
|
||||
})
|
||||
}, 200, { leading: true, trailing: true }),
|
||||
|
||||
getIntermediateResults: function () {
|
||||
let results = []
|
||||
for (let [name, layer] of this.model.modelLayersMap.entries()) {
|
||||
if (name === 'input') continue
|
||||
|
||||
const layerClass = layer.layerClass || ''
|
||||
|
||||
let images = []
|
||||
if (layer.result && layer.result.tensor.shape.length === 3) {
|
||||
images = utils.unroll3Dtensor(layer.result.tensor)
|
||||
} else if (layer.result && layer.result.tensor.shape.length === 2) {
|
||||
images = [utils.image2Dtensor(layer.result.tensor)]
|
||||
} else if (layer.result && layer.result.tensor.shape.length === 1) {
|
||||
images = [utils.image1Dtensor(layer.result.tensor)]
|
||||
}
|
||||
results.push({
|
||||
name,
|
||||
layerClass,
|
||||
images
|
||||
})
|
||||
}
|
||||
this.layerResultImages = results
|
||||
setTimeout(() => {
|
||||
this.showIntermediateResults()
|
||||
}, 0)
|
||||
},
|
||||
|
||||
showIntermediateResults: function () {
|
||||
this.layerResultImages.forEach((result, layerNum) => {
|
||||
const scalingFactor = this.layerDisplayConfig[result.name].scalingFactor
|
||||
result.images.forEach((image, imageNum) => {
|
||||
const ctx = document.getElementById(`intermediate-result-${layerNum}-${imageNum}`).getContext('2d')
|
||||
ctx.putImageData(image, 0, 0)
|
||||
const ctxScaled = document.getElementById(`intermediate-result-${layerNum}-${imageNum}-scaled`).getContext('2d')
|
||||
ctxScaled.save()
|
||||
ctxScaled.scale(scalingFactor, scalingFactor)
|
||||
ctxScaled.clearRect(0, 0, ctxScaled.canvas.width, ctxScaled.canvas.height)
|
||||
ctxScaled.drawImage(document.getElementById(`intermediate-result-${layerNum}-${imageNum}`), 0, 0)
|
||||
ctxScaled.restore()
|
||||
})
|
||||
})
|
||||
},
|
||||
|
||||
clearIntermediateResults: function () {
|
||||
this.layerResultImages.forEach((result, layerNum) => {
|
||||
const scalingFactor = this.layerDisplayConfig[result.name].scalingFactor
|
||||
result.images.forEach((image, imageNum) => {
|
||||
const ctxScaled = document.getElementById(`intermediate-result-${layerNum}-${imageNum}-scaled`).getContext('2d')
|
||||
ctxScaled.save()
|
||||
ctxScaled.scale(scalingFactor, scalingFactor)
|
||||
ctxScaled.clearRect(0, 0, ctxScaled.canvas.width, ctxScaled.canvas.height)
|
||||
ctxScaled.restore()
|
||||
})
|
||||
})
|
||||
}
|
||||
}
|
||||
})
|
||||
@@ -0,0 +1,63 @@
|
||||
<div class="demo imdb-bidirectional-lstm">
|
||||
<div class="title">
|
||||
<span>Bidirectional LSTM for IMDB sentiment classification</span>
|
||||
<mdl-spinner v-if="modelLoading && loadingProgress < 100"></mdl-spinner>
|
||||
</div>
|
||||
<div class="loading-progress" v-if="modelLoading && loadingProgress < 100">
|
||||
Loading...{{ loadingProgress }}%
|
||||
</div>
|
||||
<div class="info-panel">
|
||||
<div class="info-panel-text">
|
||||
|
||||
</div>
|
||||
<div class="info-panel-close">
|
||||
<div class="info-panel-close-btn"><i class="material-icons">close</i>CLOSE</div>
|
||||
</div>
|
||||
</div>
|
||||
<div class="columns input-output" v-if="!modelLoading">
|
||||
<div class="column input-column">
|
||||
<div class="input-container">
|
||||
<mdl-textfield
|
||||
floating-label="input text"
|
||||
:value.sync="inputText"
|
||||
spellcheck="false"
|
||||
textarea
|
||||
rows="7"
|
||||
@change="inputChanged"
|
||||
></mdl-textfield>
|
||||
<div class="input-clear" v-on:click="clear">
|
||||
<i class="material-icons">clear</i>CLEAR
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
<div class="column output-column">
|
||||
<div class="output">
|
||||
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
<div class="layer-results-container" v-if="!modelLoading">
|
||||
<div
|
||||
v-for="layerResult in layerResultImages"
|
||||
class="layer-result"
|
||||
>
|
||||
<div class="layer-result-heading">
|
||||
<span class="layer-class">{{ layerResult.layerClass }}</span>
|
||||
<span> {{ layerDisplayConfig[layerResult.name].heading }}</span>
|
||||
</div>
|
||||
<div class="layer-result-canvas-container">
|
||||
<canvas v-for="image in layerResult.images"
|
||||
id="intermediate-result-{{ $parent.$index }}-{{ $index }}"
|
||||
width="{{ image.width }}"
|
||||
height="{{ image.height }}"
|
||||
style="display:none;"
|
||||
></canvas>
|
||||
<canvas v-for="image in layerResult.images"
|
||||
id="intermediate-result-{{ $parent.$index }}-{{ $index }}-scaled"
|
||||
width="{{ layerDisplayConfig[layerResult.name].scalingFactor * image.width }}"
|
||||
height="{{ layerDisplayConfig[layerResult.name].scalingFactor * image.height }}"
|
||||
></canvas>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
Reference in New Issue
Block a user