Twitter Beğeni Hilesi

Portable RL models in C for WASM x.com/i/broadcasts/1…

I've been sitting on this really awesome secret project for a few months now and I couldn't bring myself to release it... perf wasn't quite there, was a bit too difficult to install for enough people to see it ... well I'm finally porting it to C + wasm today

... well uh ... 2200 lines of C in on the port, all of which needs to be painstakingly tested... and probably 1000 more before it's done. Several core data structures need to be rethought entirely... going to be an adventure and a half... Why did I use 3 separate algos for PCG?

Simplex noise working -- didn't write this one, just put vec_noise's C implementation into a single standalone file and wrapped it separately + raylib render

Accidentally made abstract art

Much better. Proper multicolor flood fill

Cellular automata for filling in gaps... 3/3 ported to C and tested!

Ta-da! Terrain gen! Simplex + flood fill + cellular automata

2000 lines ported to C today, ~500 tested correct. Pretty sure I can hear code at this point... going to get some rest, will really grind this release next week!

PufferAI is officially open for business! Puff up your company's reinforcement learning efforts with extended support, priority features, and extra services. Visit puffer(.)ai or DM for details. All our tools are still free and open source -- that's not changing!

Portable RL models in C for WASM x.com/i/broadcasts/1…

Portable RL models in C for WASM x.com/i/broadcasts/1…

Portable RL models in C for WASM x.com/i/broadcasts/1…

Portable RL models in C for WASM x.com/i/broadcasts/1…

PufferLib has hit 1000 stars! Thank you all so much for the support. To say thanks, today, I'm releasing puffernet, a single-file C implementation of mlp/conv2d/lstm layers, now available in pufferlib/dev. This will be powering our web demos where RL agents run live in browser!

Portable RL models in C for WASM x.com/i/broadcasts/1…

Congrats on the release!

Portable RL models in C for WASM x.com/i/broadcasts/1…

PufferAI is currently looking for a couple more academic labs to support with reinforcement learning tools and infra. It's all free and open source. Our focus is on high-perf envs and automated experimentation. DM me and let's chat. For industry labs, we sell support packages!

Portable RL models in C for WASM x.com/i/broadcasts/1…