tfjs VS deep-high-resolution-net.pytorch

TensorFlow.js

Many people seem to be unaware of tensorflow.js, an official JS implementation of TF

https://github.com/tensorflow/tfjs

I'd love to see PyTorch in JS, but I think unless you get it running on the GPU it won't be able to do much.

TensorflowJS GitHub Repository

Also, Tensorflow.js WebGPU backend has been in the works for quite some time: https://github.com/tensorflow/tfjs/tree/master/tfjs-backend-...

It's a pity that tfjs never truly developed any decent ops. E.g. you need lgamma to implement the cap for zero-inflated poisson regression and tfjs simply doesn't have that: https://github.com/tensorflow/tfjs/issues/2011

People have been doing it for long with WebGL, see eg https://github.com/tensorflow/tfjs and https://cloudblogs.microsoft.com/opensource/2021/09/02/onnx-...

thanks, no not unity, going to show it as a demo with threejs + tensorflow on the web. I found a github request to add face orientation https://github.com/tensorflow/tfjs/issues/3835 looks like they assigned someone to add it but doesn't look like its available yet, but there's some posts about the math I can use to get rotations based on some of the landmarks

There already exist a wide variety of neural network inference engines that run in web browsers (e.g. TensorFlow.js and, my personal favorite for use with PyTorch models, ONNX Runtime Web), but pre- and post-processing has always required imperative manipulations on flat buffers rather than a clean ndarray interface.

Full docs and code: https://github.com/tensorflow/tfjs/tree/master/e2e/benchmarks/local-benchmark

I am sure you have some great ideas in mind for your model design, but a quick tip (from having worked on a similar project) - it may be very challenging to simply take raw videos as input (especially with unconstrained camera viewpoints) without doing some feature extraction/keypoint detection first. For example, open-source plug-and-play 2D keypoint detectors are extremely good nowadays (e.g HRNet), and explicitly detecting 2D keypoints before classifying deadlift videos as "good" or "bad" will likely make the task more feasible!

The repo is well maintained on GitHub and I have replicated the author's work. But I didn't get much understanding of how things are working there. I didn't see any results.