Alternatives
Products that do what Watch a neural net learn to play Snake does
In browser PPO training demo, made possible by tinygrad: TinyJit -> WebGPU kernels. Requires WebGPU.
- 1BP
Hi everyone! After spending hundreds of hours, we're excited to finally share our progress in developing a reinforcement learning system to beat Pokémon Red. Our system successfully completes the game using a policy under 10M parameters, PPO, and a few novel techniques. With the release of Claude Plays Pokémon, now feels like the perfect time to showcase our work. We'd love to get feedback!
2025 · drubinstein.github.io
- 2IB
2024 · graphgame.sabrina.dev
- 3DL
2019 · beta.aifiddle.io
- 4IB
We wanted to do something very challenging to prove to ourselves that we can do anything we put our mind to. The reasoning for why we chose to build a toy TPU specifically is fairly simple: - Building a chip for ML workloads seemed cool - There was no well-documented open source repo for an ML accelerator that performed both inference and training None of us have real professional experience in hardware design, which, in a way, made the TPU even more appealing since we weren't able to estimate exactly how difficult it would be. As we worked on the initial stages of this project, we…
2025 · tinytpu.com
- 5
- 6TA
I recently dug my Nintendo GameCube out of storage to revisit the first Animal Crossing game. Things were mostly as I remembered, but the game's heavy reliance on a clunky on-screen keyboard quickly wore my patience thin. Unwilling to accept this subpar experience, I did what any rational person would do and ordered a rare, Japan-exclusive, keyboard/controller hybrid on eBay, then used a Raspberry Pi Pico to 1. listen for keypresses and 2. send simulated controller events to the GameCube, automating typing in Animal Crossing at a Tool-Assisted Speedrun level. Of course, this oddball…
2025 · github.com
- 7
- 8SO
2012 · snaketron.com
- 9AS
Hi HN, author here. SHARP is Apple's recent single-image 3D Gaussian splatting model (https://arxiv.org/abs/2512.10685). Their reference code is PyTorch + a pretty heavy pipeline; I wanted to see if it could run in a browser with no server hop, so I exported the predictor to ONNX and ran it via onnxruntime-web with the WebGPU EP. What works: drop in an image, get a .ply you can download or preview live, all on your machine — your image never leaves the tab. The model is large (~2.4 GB sidecar) so first load is slow on a cold cache, but inference itself is a few seconds on…
May 2026 · github.com
- 10SI
I've added controls for mobile and fixed the game running too fast in the browser initially because FPS where not limited. Original repo: https://github.com/rapiz1/DungeonRush My fork: https://github.com/midzer/DungeonRush/tree/emscripten
2024 · midzer.de
- 11
- 12MS
2013 · hipstersnake.com
- 13NA
I built over the last two years a human-like neural network chess engine that tries to predict your rating from a single game. It automatically adapts to your play and tries to play like a human at your level would play, giving you a balanced game. At the core I’m using an AlphaZero / Leela Chess Zero style neural network that I have trained on 1 billion human games from the lichess.org open database. Around this network I have built a chess engine in Rust with algorithms that use the outputs from the NN to produce human-like moves at a given rating from beginner to world champion, as…
2022 · noctie.ai
- 14RN
2015 · otoro.net
- 15
- 16ET
2023 · github.com
- 17PI
OP here. Most Deep Learning approaches for TSP rely on pre-training with large-scale datasets. I wanted to see if a solver could learn "on the fly" for a specific instance without any priors from other problems. I built a solver using PPO that learns from scratch per instance. It achieved a 1.66% gap on TSPLIB d1291 in about 5.6 hours on a single A100. The Core Idea: My hypothesis was that while optimal solutions are mostly composed of 'minimum edges' (nearest neighbors), the actual difficulty comes from a small number of 'exception edges' outside of that local scope. Instead of…
Dec 2025
- 18
- 19AI
A chess engine implementation inspired by AlphaZero, using MLX for neural network computations and Monte Carlo Tree Search (MCTS) for move selection.
2025 · github.com
- 20IB
"Creatures" is a stretch given that the environment is Minecraft, but the idea is simple: iteratively add blocks conditionally (tensor convolution) on the current environment (blocks) to maximize some reward. In this case I use PPO RL to train creatures to touch a glowstone block but you can adapt it to use any algorithm and reward (easily, as it uses the Ray framework). What I like about this work: iteratively finding solutions has a long and colorful history of doing things well: gradient boosting, ResNets, Stable Diffusion, etc. We're after some end optimal state and usually try to get…
2024 · github.com
- 21AG
This is a vector index I built that supports insertion and k-nearest neighbors (k-NN) querying, optimized for GPUs. It operates entirely in CUDA and can process queries on half a billion vectors in under 200 milliseconds. The codebase is structured as a standalone library with an HTTP API for remote access. It’s intended for high-performance search tasks—think similarity search, AI model retrieval, or reinforcement learning replay buffers. The codebase is located at https://github.com/rodlaf/BinaryGPUIndex.
2025 · rlafuente.com
- 22LF
We're excited to announce that we've open-sourced LeanRL, a lightweight PyTorch reinforcement learning library that provides recipes for fast RL training using torch.compile and CUDA graphs. By leveraging these tools, we've achieved significant speed-ups compared to the original CleanRL implementations - up to 6x faster! Reinforcement learning is notoriously CPU-bound due to the high frequency of small CPU operations. PyTorch's powerful compiler can help alleviate these issues, but comes with its own costs. LeanRL addresses this challenge by providing simple recipes to accelerate your…
2024 · github.com
- 23NG
2024 · github.com
- 24M3
Grant Sanderson (3Blue1Brown) created Manim, the Python library he uses to make the math animations in his videos. We reimplemented Manim with the same Python API, but the implementation underneath is Rust, connected to Python through PyO3. The Rust code uses wgpu, so rendering happens on the GPU. To run it in the browser, we compiled the Rust parts to WebAssembly so the PyO3 extension loads in Pyodide. In the browser, wgpu targets the WebGPU API, so animations render in real time on your GPU through the browser. The editor is Monaco (the editor that powers VS Code) with a live preview:…
Jul 2026 · studio.academa.ai
Ranked by how close each launch is in meaning, then by votes. Refine with a description →