Skip to content

Instantly share code, notes, and snippets.

View selfup's full-sized avatar
💻
doing the most

Regis Boudinot selfup

💻
doing the most
181 followers · 139 following
View GitHub Profile
@selfup
selfup / run_model.sh
Created May 12, 2026 18:54
#!/usr/bin/env bash
#
# llama-serve.sh — wrap llama-server with sane defaults for M3 Ultra
# that approximate the LM Studio endpoint used by qwaude / SSH tunnel.
#
# Usage:
# ./llama-serve.sh <path-to.gguf>
#
# Env overrides:
# PORT (default 1234)
@selfup
selfup / ubuntu_3060ti_ministral_llama_cpp_bench.sh
Last active May 23, 2026 14:11
Ministral 3 (3B, 8B) Instruct - Ubuntu 3800x 3060ti llama.cpp benchmark
set -eou pipefail
llama-bench \
-m ~/.models/Ministral-3-3B-Instruct-2512-Q4_K_M.gguf \
-m ~/.models/Ministral-3-8B-Instruct-2512-Q4_K_M.gguf \
-p 512 -n 128 -fa 1 -ngl 99 \
-d 0,4096,8192,16384 \
-o md > local-ministral-q4km-${MACHINE}.md 2>/dev/null
@selfup
selfup / ubuntu_llama_cpp_user_install.sh
Last active May 11, 2026 01:49
llama.cpp on Ubuntu x86 (3800x, 32GB RAM, 3060ti)
# wherever you have cloned the repo
cd ~/Documents/Repos/llama.cpp
# if you built relative to repo instead of user
# build in $HOME/.local
rm -rf build && cmake -B build \
-DGGML_CUDA=ON \
-DCMAKE_CUDA_ARCHITECTURES=86 \
-DBUILD_SHARED_LIBS=OFF \
-DCMAKE_INSTALL_PREFIX=$HOME/.local \
@selfup
selfup / llama_bench_gemma.sh
Last active May 11, 2026 22:48
Gemma 4 (E2B and E4B) - M5 Max 36GB (binned) and Ubuntu 3800x 3060ti llama.cpp benchmark
# Apple Silicon llama-bench: Gemma 4 (E2B and E4B) Q4_K_M depth sweep
# Testing for Gemma 4 (E2B and E4B) on Apple Silicon
llama-bench \
-m ~/.lmstudio/models/lmstudio-community/gemma-4-E2B-it-GGUF/gemma-4-E2B-it-Q4_K_M.gguf \
-m ~/.lmstudio/models/lmstudio-community/gemma-4-E4B-it-GGUF/gemma-4-E4B-it-Q4_K_M.gguf \
-p 512 -n 128 -fa 1 -ngl 99 \
-d 0,4096,8192,16384 \
-o md > local-gemma-q4km-m5_max_36gb.md 2>/dev/null
@selfup
selfup / llama_bench_ministral.sh
Last active May 10, 2026 13:49
Ministral 3 (3B, 8B, 14B) Instruct - M3 Ultra 96GB and M5 Max 36GB llama.cpp benchmark
# Apple Silicon and Linux x86 llama-bench: Ministral 3 family Q4_K_M depth sweep
# Testing for the Ministral 3 family (Ministral-3B, Ministral-8B, Ministral-14B) on Apple Silicon
# and (3B, 8B) on Ubuntu x86 with a 3060ti
# Instruct only no reasoning
llama-bench \
-m ~/.lmstudio/models/lmstudio-community/Ministral-3-3B-Instruct-2512-GGUF/Ministral-3-3B-Instruct-2512-Q4_K_M.gguf \
-m ~/.lmstudio/models/lmstudio-community/Ministral-3-8B-Instruct-2512-GGUF/Ministral-3-8B-Instruct-2512-Q4_K_M.gguf \
-m ~/.lmstudio/models/lmstudio-community/Ministral-3-14B-Instruct-2512-GGUF/Ministral-3-14B-Instruct-2512-Q4_K_M.gguf \
-p 512 -n 128 -fa 1 -ngl 99 \
@selfup
selfup / godot_dir_bootstrap.sh
Created February 23, 2025 22:46
if [[ $PROJECT_DIR != "" ]]; then
mkdir -p \
$PROJECT_DIR/models \
$PROJECT_DIR/textures \
$PROJECT_DIR/shaders \
$PROJECT_DIR/sounds \
$PROJECT_DIR/materials \
$PROJECT_DIR/scenes \
$PROJECT_DIR/addons \
$PROJECT_DIR/scripts \
@selfup
selfup / fizz_buzz.rs
Last active January 16, 2023 16:12
Parallel FizzBuzz Vec Builder in Rust
use rayon::prelude::*;
const VEC_SIZE: usize = 1_000_000_001;
const CHUNK_SIZE: usize = 128;
enum FizzBuzz {
FIFTEEN,
THREE,
FIVE,
VALUE(u32),
@selfup
selfup / decrease_memory_using_enums_in_rust.md
Last active November 22, 2020 13:41
layout post
title Decrease Memory Using Enums in Rust
published true

Decrease Memory Using Enums in Rust

I am slowly building a sort of atom/universe generator/visualizer in Rust with friends. The project is called Oxidizy.

@selfup
selfup / ubuntu_random_crashes.sh
Created June 21, 2020 17:32
sudo ls /var/crash
sudo rm /var/crash/*
sudo ls /var/crash
@selfup
selfup / keybase.md
Created June 17, 2020 01:27

Keybase proof

I hereby claim:

  • I am selfup on github.
  • I am selfup (https://keybase.io/selfup) on keybase.
  • I have a public key whose fingerprint is F6BD CA79 8ED7 CF49 B9FF CAA8 2977 288B EE59 5B60

To claim this, I am signing this object: