tfjs | A WebGL accelerated JavaScript library for training and deploying ML models | Graphics library

I am trying to convert my nodejs application into an .exe file. I am stuck with the canvas module(npm install @tensorflow/tfjs canvas). For some reason,nw.js is not able to load this module.

This is the line for which I am getting the error :

So I've been trying to get user input to calculate outer product of two vectors using tensorflow.js but I couldn't figure out how to get an array from html to pass it into javascript. for example inside the calculator function const a has to be something like this tf.tensor1d([1, 2, 3])

Problem:

• a ValueError is saying that conv2d_Conv2D1_input is expecting to have 4 dimension(s), but got array with shape [475,475,3]

However:

• The inputShape is set to [475,475,3]
• when logged, tensors have the shape [475,475,3]

Error: ValueError: Error when checking : expected conv2d_Conv2D1_input to have 4 dimension(s), but got array with shape [475,475,3]

Tensor:

I am trying to use Tensorflow's DANFO.js library inside an Angular web app (with Ionic component's but that is irelevant).

• I simply ran the command: npm install danfojs --save and npm install @tensorflow/tfjs-node --save
• I tried all kinds of imports inside an Angular Service like: import 'danfojs';, import { DataFrame } from 'danfojs'; or even according to only related question to this topic I could find:import { DataFrame } from 'danfojs/dist/core/frame';
• All of the above examples don't work
• I even tried importing the DANFO.js in app.module without any change

And I always get this error message (rest of the error is in edit):

I am currently using tfjs 3.8 to load a segmentation model (loaded as a tf.GraphModel) on the client side. To create the input Tensor, I call browser.fromPixels(imageData), which creates the Tensor on CPU from the ImageData object that is also on CPU. Since I'm using tfjs' webgl backend, the data is sent to the GPU when calling the model.predict(tensor) function. All of this works well, excepted that my ImageData object is created from an image on a canvas with a WebGLRenderingContext, meaning it comes from the GPU. This GPU->CPU->GPU data transfer is slowing down my process, which I am trying to optimize.

I briefly searched tfjs and could not find a way to create a Tensor on GPU to prevent the GPU->CPU data transfer. Is there a way I could keep my data on the GPU?

Ive done a lot of Pose Estimation in Python using OpenCV and Mediapipe a year ago. I recently read an the Tensorflow Blog(https://blog.tensorflow.org/2021/08/3d-pose-detection-with-mediapipe-blazepose-ghum-tfjs.html) about the possibility of 3d Pose estimation. This got me triggered, so i learned a bit of JavaScript to start my journey. After 3 days, i realised that my code doest work, Javascript gives out no errors, though nothing works. I followed the steps closely with a mate who has used JavaScript for a month now, Though he couldnt help me, cause he never worked with AI. I never worked with it, so i got no idea how await works in JS. MAy be a problem lol

I'm trying to use images off the internet to try and train my network. I'm using an Image() object to create the images and pass them to tensorflow. According to my knowledge, Image() returns a HTMLImageElement, however, I'm still getting the following error:

I spent long time trying to figure out how to install and run Tensorboard on a Windows 10 machine, for projects that are using Tensorflow.JS

Has anybody ever made this work? If so, please share the steps.

I have Tensorflow.JS working fine with my GPU under Windows 10, but I cannot find any documentation for using Tensorboard under the combination of Windows + Tensorflow.JS

• Windows 10 64-bit
• Node.JS v14.17.3
• Python v3.8.9
• Tensorflow v2.5.0
• Tensorflow.JS v3.8.0

I'm using @tensorflow/tfjs-node-gpu, with RTX3080Ti, which works fine.

I have created a model which applied Mobilenet V2 for the convolutional base layers in Google colab. Then I converted it by using this command: