Used PI to create a bigbang simulator from scratch with no user generated code (only generated by the LLM) and a local instance of Qwen3.5 9B. Testing the Pi agentic coding harness from @mariozechner (`npm install -g @mariozechner/pi-coding-agent`)
Updated 2026-04-11 01:38:33 -04:00
History of my changes to ollama to enable SYCL support
Updated 2026-04-07 16:21:38 -04:00
configuration scripts/tool to allow for localized LLMs (currently supporting Ollama LLMs) to utilize bash tools and MCP servers via Python
Updated 2026-01-29 10:33:40 -05:00
Transformers Latent-Space reasoning Auto Train Once -- a framework meant to utilize a latent-space reasoning model with 'Auto Train Once' pruning ( https://arxiv.org/abs/2403.14729 ) technique to ensure maximum performance with minimum neuron activation/creation/utilization in the LLM.
Updated 2025-10-03 16:00:26 -04:00
ollama-chat.nvim
Template
Neovim plugin which enables inline ollama chat within the neovim IDE without having to call any external tools. Utilizes curl, lua, vimscript, and various other hooks/interfaces
Updated 2025-10-03 15:46:21 -04:00
ARM Cortex M55 Firmware created for a qemu instantiation of a Cortex M55 SoC in Arch Linux
Updated 2025-10-03 15:35:25 -04:00