Skip to content

Instantly share code, notes, and snippets.

View manbehindthemadness's full-sized avatar

Kevin Eales manbehindthemadness

View GitHub Profile
@willccbb
willccbb / grpo_demo.py
Last active October 19, 2025 08:14
GRPO Llama-1B
# train_grpo.py
#
# See https://github.com/willccbb/verifiers for ongoing developments
#
"""
citation:
@misc{brown2025grpodemo,
title={Granular Format Rewards for Eliciting Mathematical Reasoning Capabilities in Small Language Models},
author={Brown, William},
@GrayHatGuy
GrayHatGuy / iridium-sdr.md
Last active November 1, 2024 23:01
Iridium satellite cheat sheet for hackrf SDR capture using gr-iridium iridium-toolkit and iridium live

Description

This details how to capture voice text and rings associated with iridium satellite rebroadcasts of the Aircraft Communications Addressing and Report System (ACARS)

Prerequisites:

  • Start installation in home directory preferred
    cd ~
    
  • Install the following repos preferably from source or using apt in home directory or /usr/scr/:
@todbot
todbot / eyeballs_dual_gc9a01.py
Last active April 24, 2025 06:28
dual display GC9A01 round TFT LCDs with creepy eyeball animation in CircuitPython
# eyeballs_dual_gc9a01.py - dual display GC9A01 round TFT LCDs with creepy eyeball animation
# 17 May 2022 - @todbot / Tod Kurt
#
# requires recompiled CircuitPython with "#define CIRCUITPY_DISPLAY_LIMIT (2)" in "mpconfigboard.h"
#
import time, math, random
import board, busio
import displayio, terminalio
import gc9a01
@r7vme
r7vme / 01-README.md
Last active May 23, 2024 03:17
Flash Jetson Xavier AGX from docker (sdkmanager)

TESTED for Jetpack 4.5.1

Official docker image had few utilities missed, which caused flashing to fail with Jetson AGX Xavier. To fix that I added them to docker image.

  1. Get NVIDIA official docker image.

  2. Rebuild docker image

mkdir sdkmanager; cd sdkmanager
# create dockefile from below gist
@victor-shepardson
victor-shepardson / pytorch-glumpy.py
Last active February 23, 2025 17:12
using pycuda and glumpy to draw pytorch GPU tensors to the screen without copying to host memory
from contextlib import contextmanager
import numpy as np
import torch
from torch import Tensor, ByteTensor
import torch.nn.functional as F
from torch.autograd import Variable
import pycuda.driver
from pycuda.gl import graphics_map_flags
from glumpy import app, gloo, gl