Skip to content

Instantly share code, notes, and snippets.

View cmdr2's full-sized avatar

cmdr2

209 followers · 1 following
View GitHub Profile
@cmdr2
cmdr2 / iss-brightness.py
Last active February 4, 2020 09:35
Calculate brightness magnitude of ISS (and whether Earth's shadow is falling on it) using pyephem
# by cmdr2, anyone is free to use and modify this, without the need to give credit
import math
import ephem
import time
from datetime import datetime, timedelta
# consts
degrees_per_rad = 180.0 / math.pi
au = 149597892
@cmdr2
cmdr2 / test.js
Last active April 28, 2020 11:24
<script src="https://cdnjs.cloudflare.com/ajax/libs/processing.js/1.6.0/processing.min.js"></script>
<body>
<script type="application/processing">
void setup(){
size(200, 200);
}
void draw(){
background(64);
ellipse(40, 40, 20, 20);
@cmdr2
cmdr2 / gaussian_random_standard_deviation.js
Created June 24, 2021 17:54
From https://stackoverflow.com/a/35599181
// returns a gaussian random function with the given mean and stdev.
function gaussian(mean, stdev) {
var y2;
var use_last = false;
return function() {
var y1;
if (use_last) {
y1 = y2;
use_last = false;
} else {
@cmdr2
cmdr2 / transform-3d-world-local.py
Created July 8, 2021 07:28
3D world-to-local and local-to-world transform using Python (numpy and transforms3d)
# pip install transforms3d numpy
import transforms3d as t3d
import numpy as np
scene_transform = {"pos": [0,0,0], "rot": [0,0,0,1], "scale": [1,1,1]}
# https://forum.unity.com/threads/whats-the-math-behind-transform-transformpoint.107401/#post-709765
def local_to_world(p):
v = np.multiply(p, scene_transform["scale"])
@cmdr2
cmdr2 / gist:0e4ad366d1fe7b2524511d6bddd15e2a
Last active December 5, 2021 14:12
Disable discrete GPU on MacBook Pro
Follow the steps on http://www.angelofarina.it/DisableDiscrete.htm
if this doesn't work:
`sudo nvram fa4ce28d-b62f-4c99-9cc3-6815686e30f9:gpu-power-prefs=%01%00%00%00`
then try `sudo nvram 000fa4ce-b62f-4c99-9cc3-6815686e30f9:gpu-power-prefs=%01%00%00%00`
Also check if any new NVidia or AMD drivers got created in /System/Library/Extensions, like GeForce* or NVDA* or NVIDIA* or CUDA*, which will need to be moved away
@cmdr2
cmdr2 / user_paths_blender_xr.diff
Created May 5, 2022 11:56
Diff for exposing user_path and user_path_other in the Blender XrEventData
diff --git a/source/blender/makesrna/intern/rna_xr.c b/source/blender/makesrna/intern/rna_xr.c
index 6f43afe23fe..52efa1b6653 100644
--- a/source/blender/makesrna/intern/rna_xr.c
+++ b/source/blender/makesrna/intern/rna_xr.c
@@ -1195,6 +1195,50 @@ static int rna_XrEventData_action_length(PointerRNA *ptr)
# endif
}
+static void rna_XrEventData_user_path_get(PointerRNA *ptr, char *r_value)
+{
@cmdr2
cmdr2 / madrang-img-buttons.plugin.js
Last active October 20, 2022 08:33
// contributed by @madrang
PLUGINS['IMAGE_INFO_BUTTONS'].push({ text: 'Double Size (Madrang)', on_click: getStartNewTaskHandler('img2img_X2') })
PLUGINS['IMAGE_INFO_BUTTONS'].push({ text: 'Redo (Madrang)', on_click: getStartNewTaskHandler('img2img') })
function getStartNewTaskHandler(mode) {
return function(reqBody, img) {
const newTaskRequest = getCurrentUserRequest()
switch (mode) {
case 'img2img':
@cmdr2
cmdr2 / sd1-vs-sd2.diff
Created November 24, 2022 04:39
Diff between https://github.com/CompVis/stable-diffusion (sd1) and https://github.com/Stability-AI/stablediffusion (sd2). Ran `diff -w -x '*.gif' -x '*.png' -x '*.jpg' -x LICENSE -x '*.md' -x '*.git' -r sd1 sd2 > d.txt`
Only in sd2: LICENSE-MODEL
Only in sd2/assets: stable-inpainting
Only in sd2/assets/stable-samples: depth2img
Only in sd2/assets/stable-samples/txt2img: 768
Only in sd2/assets/stable-samples: upscaling
Only in sd1/configs: autoencoder
Only in sd1/configs: latent-diffusion
Only in sd1/configs: retrieval-augmented-diffusion
Only in sd1/configs/stable-diffusion: v1-inference.yaml
Only in sd2/configs/stable-diffusion: v2-inference-v.yaml
@cmdr2
cmdr2 / v2.1-inference.yaml
Last active December 24, 2022 21:48
Custom model config for Stable Diffusion 2.0 models, adds an "extra: {attn_precision: 'fp32'}" field to the end of the model config. This "extra" field is recognized by SDKit (https://github.com/easydiffusion/sdkit)
model:
base_learning_rate: 1.0e-4
target: ldm.models.diffusion.ddpm.LatentDiffusion
params:
parameterization: "v"
linear_start: 0.00085
linear_end: 0.0120
num_timesteps_cond: 1
log_every_t: 200
timesteps: 1000
@cmdr2
cmdr2 / run_everything_diffusers.py
Created March 1, 2023 18:13
"""
Runs the desired Stable Diffusion models against the desired samplers. Saves the output images
to disk, along with the peak RAM and VRAM usage, as well as the sampling performance.
"""
import argparse
import os
import time
import traceback
from pathlib import Path