Oleksandr Kuvshynov

slowllama. Finetune llama2-70b and codellama on MacBook Air without quantization

b63. Micro-benchmarking library for C and C++ with PMU counters tracking

llama-sandbox. A collection of experiments related to LLM inference with llama.cpp/mlx

cubestat. Horizon chart for CPU/GPU/Neural Engine utilization monitoring. Supports Apple M1-M4, Nvidia GPUs, AMD GPUs

flamereport. FlameGraph visualization for terminal

fewlines. Create histograms, time series charts and dashboards in command-line output and log files.

qna. AI pretends to be paper/textbook author, you can ask it questions about the paper as a whole, specific parts of it right in the PDF viewing app (e.g. Apple Preview) using annotations and see the replies there.

vimqq. local LLM chat for vim with dynamic warmup, forks and multi-model support

dotfiles. Lua

experiments. C++

golem. Python

bioproxy. Go

rlscout. In progress attempt to use Deep RL to solve some game