Face-api.js: Slow in nodejs project

[galinkyuchukov]

Hello!

I am trying to setup simple face detection in nodejs env.
Out of the box everything run fine with the exception of speed. It takes 12+ seconds for a single face detection.

require('@tensorflow/tfjs-node')

is included!

On execution i see:

cpu backend was already registered. Reusing existing backend

then

Hi there 👋. Looks like you are running TensorFlow.js in Node.js. To speed things up
dramatically, install our node backend, which binds to TensorFlow C++, by running npm i
@tensorflow/tfjs-node, or npm i @tensorflow/tfjs-node-gpu if you have CUDA. Then call
require('@tensorflow/tfjs-node'); (-gpu suffix for CUDA) at the start of your program. Visit
https://github.com/tensorflow/tfjs-node for more details.

When i install older versions of the two packages:

npm install @tensorflow/[email protected] [email protected]

there is significant speed improvement, BUT i need age and gender detection , which is not present below face-api.js < 0.20.0

I also tryed:
@tensorflow/tfjs-node version 1.0.2

but then i get:

(node:29723) UnhandledPromiseRejectionWarning: TypeError: tfc.ENV.findBackend is not a
function
at nodeBackend (/home/pi/testproject/node_modules/@tensorflow/tfjs-
node/dist/ops/op_utils.js:25:28)

How can i use age and gender recognition and get faster speed under nodejs?

Environment
Raspberry Pi
Node: v8.10.0
face-api.js: 0.20.0 / 0.14.3
@tensorflow/tfjs-node: 1.1.2 / 1.0.2 / 0.3.0

[justadudewhohacks]

@tensorflow/tfjs-node 1.0.2 + @tensorflow/tfjs-core 1.0.3 should work with face-api.js 0.20.0. This is the setup that the nodejs examples use. The error message you posted might indicate, that you are not using tfjs-core 1.0.3.

[singhania1408]

Hi @justadudewhohacks, I am also facing the same issue. I have run the face-api.js on browser for facial expression and landmark detection, which was comparatively very fast, but in nodejs, it was damn too slow, (i.e approx 1 frame in 5-6sec). Please help me out how I can do to fix this issue.

[justadudewhohacks]

As I answered above, make sure the package versions line up correctly.

[billdai998]

I also had the problem of agreeing and spent about 17 seconds in nodejs. Please see if my reference package is correct.

[galinkyuchukov]

Thank you for the help @justadudewhohacks!
More than 10 times faster when

@tensorflow/tfjs-node 1.0.2 + @tensorflow/tfjs-core 1.0.3

[ramesh586]

I'm getting following error, please help to resolve

cpu backend was already registered. Reusing existing backend factory.
Platform node has already been set. Overwriting the platform with [object Object].
node-pre-gyp info This Node instance does not support builds for N-API version 4
node-pre-gyp info This Node instance does not support builds for N-API version 4
module.js:549
throw err;
^

Error: Cannot find module 'D:\bench_work_space\GCH\face-expressions\face-expressions\facetest\node_modules\@tensorflow\tfjs-node\lib\napi-v3\tfjs_binding.node'
at Function.Module._resolveFilename (module.js:547:15)
at Function.Module._load (module.js:474:25)
at Module.require (module.js:596:17)
at require (internal/module.js:11:18)
at Object. (D:\bench_work_space\GCH\face-expressions\face-expressions\facetest\node_modules\@tensorflow\tfjs-node\dist\index.js:44:16)
at Module._compile (module.js:652:30)
at Object.Module._extensions..js (module.js:663:10)
at Module.load (module.js:565:32)
at tryModuleLoad (module.js:505:12)
at Function.Module._load (module.js:497:3)

[0x6c23]

I'm getting following error, please help to resolve

cpu backend was already registered. Reusing existing backend factory.
Platform node has already been set. Overwriting the platform with [object Object].
node-pre-gyp info This Node instance does not support builds for N-API version 4
node-pre-gyp info This Node instance does not support builds for N-API version 4
module.js:549
throw err;
^

Error: Cannot find module 'D:\bench_work_space\GCH\face-expressions\face-expressions\facetest\node_modules@tensorflow\tfjs-node\lib\napi-v3\tfjs_binding.node'
at Function.Module._resolveFilename (module.js:547:15)
at Function.Module._load (module.js:474:25)
at Module.require (module.js:596:17)
at require (internal/module.js:11:18)
at Object. (D:\bench_work_space\GCH\face-expressions\face-expressions\facetest\node_modules@tensorflow\tfjs-node\dist\index.js:44:16)
at Module._compile (module.js:652:30)
at Object.Module._extensions..js (module.js:663:10)
at Module.load (module.js:565:32)
at tryModuleLoad (module.js:505:12)
at Function.Module._load (module.js:497:3)

Same issue here. Impossible to include @tensorflow. It installs without errors. Tried many versions. Still not able to include.

[bamwani]

@0x6c23 Did you find any solution for this error?