diffrax VS Pion WebRTC

Optimistix has high-level APIs for minimisation, least-squares, root-finding, and fixed-point iteration and was written to take care of these kinds of subroutines in Diffrax.

Diffrax (https://github.com/patrick-kidger/diffrax).

Here is the GitHub: https://github.com/patrick-kidger/optimistix

The elevator pitch is Optimistix is really fast, especially to compile. It

Sure. So I've got some PyTorch benchmarks here. The main take-away so far has been that for a neural ODE, the backward pass takes about 50% longer in PyTorch, and the forward (inference) pass takes an incredible 100x longer.

FWIW this worked for me. :D My full-time job is now writing JAX libraries at Google. Equinox for neural networks, Diffrax for differential equation solvers, etc.

It's relatively early days yet, but JAX is in the process of developing its nascent scientific computing / scientific machine learning ecosystem. Mostly because of its strong autodifferentiation capabilities, excellent JIT compiler etc. (E.g. to show off one of my own projects, Diffrax is the library of diffeq solvers for JAX.)

Diffrax - solving ODEs with Jax and computing it's derivatives automatically functools - love partial and lru_cache fastprogress - simpler progress bar than tqdm

At least prior to this announcement: JAX was much faster than PyTorch for differentiable physics. (Better JIT compiler; reduced Python-level overhead.)

E.g for numerical ODE simulation, I've found that Diffrax (https://github.com/patrick-kidger/diffrax) is ~100 times faster than torchdiffeq on the forward pass. The backward pass is much closer, and for this Diffrax is about 1.5 times faster.

It remains to be seen how PyTorch 2.0 will compare, or course!

Right now my job is actually building out the scientific computing ecosystem in JAX, so feel free to ping me with any other questions.

https://github.com/patrick-kidger/diffrax

Which are neural network and differential equation libraries for JAX.

[Obligatory I-am-googler-my-opinions-do-not-represent- your-employer...]

Autodifferentiable programming!

Neural networks are the famous example of this, of course -- but this can be extended to all of scientific computing. ODE/SDE solvers, root-finding algorithms, LQP, molecular dynamics, ...

These days I'm doing all my work in JAX. (E.g. see Equinox or Diffrax: https://github.com/patrick-kidger/equinox, https://github.com/patrick-kidger/diffrax). A lot of modern work is now based around hybridising such techniques with neural networks.

I'd really encourage anyone interested to learn how JAX works under-the-hood as well. (Look up "autodidax") Lots of clever/novel ideas in its design.

I am really excited about https://github.com/sepfy/libpeer. It has examples ready for ESP32 etc....

When working on KVS I wasn't familiar with the embedded space at all. I saw 'heavyweight' embedded where you were running on Linux. Then you had RTOS/No OS at all. I wasn't prepared for these devices at all. If we can make WebRTC work in the embedded space I think it will really accelerate what developers are able to build!

Remotely driven cars, security cameras, robots in hospitals that bring iPads to infectious patients etc... Creative people are building amazing things. The WebRTC/video space needs to work harder and support them :)

-----

I love how diverse the WebRTC space is now. Outside of this implementation you have plenty of other options!

* https://github.com/shinyoshiaki/werift-webrtc (Typescript)

* https://github.com/pion/webrtc (Golang)

* https://github.com/webrtc-rs/webrtc (Rust)

* https://github.com/algesten/str0m (Rust)

* hhttps://github.com/sepfy/libpeer (C/Embedded)

* https://webrtc.googlesource.com/src/ (C++)

* https://github.com/sipsorcery-org/sipsorcery (C#)

* https://github.com/paullouisageneau/libdatachannel (C++)

* https://github.com/elixir-webrtc (Elixir)

* https://github.com/aiortc/aiortc (Python)

* GStreamer’s webrtcbin (C)

See https://github.com/sipsorcery/webrtc-echoes for examples of some running against each other.

I have worked four jobs related to https://github.com/pion/webrtc and one for https://webrtcforthecurious.com

Two companies used Pion. The other two were just using the protocol (WebRTC)

In my experience Go has been relatively approachable for people that are good at PHP. It has a great standard library and a pretty solid ecosystem, though frameworks aren’t as popular in Go. There are some well regarded libraries for things like WebRTC via https://github.com/pion/webrtc WebSicket via https://github.com/nhooyr/websocket

Don't try to make RTC yourself, it looks easy, but in fact, it's a really hard problem to solve. Use https://pion.ly/ it's a pretty solid package they also have a discord/slack channel with a lot of helpful people there.

Maybe you can reuse some of this code: https://github.com/pion/webrtc/blob/master/examples/data-channels/main.go