Wasm Backend: info.backend.decComplexRef is not a function

[Moumouls]

Please make sure that this is a bug. As per our
GitHub Policy,
we only address code/doc bugs, performance issues, feature requests and
build/installation issues on GitHub. tag:bug_template

System information

• Have I written custom code (as opposed to using a stock example script provided in TensorFlow.js): yes
• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Macos Bigsur
• Mobile device (e.g. iPhone 8, Pixel 2, Samsung Galaxy) if the issue happens on mobile device:
• TensorFlow.js installed from (npm or script link): yarn
• TensorFlow.js version (use command below): 3.0.0
• Browser version: None
• Tensorflow.js Converter Version:

Describe the current behavior

Try to run wasm backend on bidirectional GRU. Wasm won't create the model.

import * as tf from '@tensorflow/tfjs'
import '@tensorflow/tfjs-backend-wasm'
import { getWindowedData, WindowedData } from './utils'

const createModel = (trainData: tf.Tensor3D) => {
	const model = tf.sequential()
	const units = 64
	model.add(
		tf.layers.bidirectional({
			// @ts-ignore
			layer: tf.layers.gru({ units, activation: 'relu', returnSequences: true }),
			inputShape: [trainData.shape[1], trainData.shape[2]],
		}),
	)
	model.add(
		tf.layers.bidirectional({
			// @ts-ignore
			layer: tf.layers.gru({ units, activation: 'relu', returnSequences: true }),
		}),
	)
	model.add(
		tf.layers.bidirectional({
			// @ts-ignore
			layer: tf.layers.gru({ units, activation: 'relu' }),
		}),
	)
	model.add(tf.layers.dense({ units: 1 }))
	model.compile({ optimizer: tf.train.adam(0.001), loss: 'meanSquaredError' })
	return model
}

export const train = async (key: string, arr: any[], window: number = 20) => {
	await tf.setBackend('wasm')
	const modelPath = `file://${__dirname}/../trainedModels/Metrics_${key}`
	const data = getWindowedData(
		arr,
		(items) => items.map((item) => [item[key]]),
		(items, nextItem) => nextItem[key],
		window,
	)
	const trainData = tf.tidy(() => buildTensors(data))
	let model = createModel(trainData.xTensor)
	let bestValLoss = Number.MAX_SAFE_INTEGER
	await model.fit(trainData.xTensor, trainData.yTensor, {
		epochs: 500,
		batchSize: 32,
		validationSplit: 0.8,
		callbacks: {
			onEpochEnd: async function (epoch, logs) {
				if (logs.val_loss < bestValLoss) {
					bestValLoss = logs.val_loss
					await model.save(modelPath)
				}
			},
		},
	})
	return tf.loadLayersModel(`${modelPath}/model.json`)
}

Describe the expected behavior

Should create the model, start fit.

Standalone code to reproduce the issue
Provide a reproducible test case that is the bare minimum necessary to generate
the problem. If possible, please share a link to Colab/CodePen/any notebook.

Other info / logs Include any logs or source code that would be helpful to
diagnose the problem. If including tracebacks, please include the full
traceback. Large logs and files should be attached.

RuntimeError: abort(TypeError: info.backend.decComplexRef is not a function). Build with -s ASSERTIONS=1 for more info.
    at process.abort (/Users/xxxx/Desktop/Repos/xxxx/xxxx/node_modules/@tensorflow/tfjs-backend-wasm/wasm-out/tfjs-backend-wasm.js:9:9928)
    at process.emit (node:events:327:20)
    at process.EventEmitter.emit (node:domain:486:12)
    at process.emit (/Users/xxxx/Desktop/Repos/xxxx/xxxx/node_modules/source-map-support/source-map-support.js:495:21)
    at processPromiseRejections (node:internal/process/promises:214:33)
    at processTicksAndRejections (node:internal/process/task_queues:94:32)

[Moumouls]

Okay i found that you can't have this in your package.json

	"dependencies": {
		"@tensorflow/tfjs": "^3.0.0",
		"@tensorflow/tfjs-backend-wasm": "^3.0.0",
		"@tensorflow/tfjs-node": "^2.8.2",
		"@tensorflow/tfjs-node-gpu": "^3.0.0",

Error is gone after removing other @tensorflow packages

	"dependencies": {
		"@tensorflow/tfjs": "^3.0.0",
		"@tensorflow/tfjs-backend-wasm": "^3.0.0",

[rthadur]

In your package.json, if you have include the union package '@tensorflow/tfjs',
you don't need to include other individual packages (tfjs-node,tfjs-node-gpu etc)

[google-ml-butler]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed in 7 days if no further activity occurs. Thank you.

[google-ml-butler]

Closing as stale. Please @mention us if this needs more attention.

[google-ml-butler]

Are you satisfied with the resolution of your issue?
Yes
No