A chat interface I created for running LLMs locally using llama.cpp. Entirely self-hosted and can be run as a simple docker container!
View on | |Core maintainer on the repo powering HuggingChat, the open-source alternative to ChatGPT with 500k+ users.
View on | |A set of AI generated decks of tarot cards. Built using a fine-tuned image generation model and Llama 3 for prompt generation, hosted on a SvelteKit website.
View projectA Svelte 5 component for adding a comments section to your website using a Bluesky discussion as a source.
View on | |A 231M parameter transformer model based on GPT-2 and trained on chess games to predict move sequences.
View projectA game project running in the browser, where you can control a spacecraft using an emulated 6502 CPU.
View projectA VSCode extension that automatically activates the right venv using uv and inline script metadata.
View on | |A modular rocket trajectory simulator, using Numba and SciPy solvers to accelerate the resolution of ODEs.
View project