mseri

Let's hack on llm!

If you have uv installed (and you should!), you can install llm globally in a uv-managed tool environment with:

uv tool install llm

If you want to use models other than OpenAI models, you'll need some extensions:

mseri

import sys
from io import StringIO

import streamlit as st              # pip install streamlit
from code_editor import code_editor # pip install streamlit_code_editor
import ollama as ol                 # pip install ollama

st.set_page_config(layout='wide')
st.title('`Offline code completion`')

mseri

"""
A minimal, fast example generating text with Llama 3.1 in MLX.

To run, install the requirements:

    pip install -U mlx transformers fire

Then generate text with:

    python l3min.py "How tall is K2?"

mseri

package main

import (
	"encoding/json"
	"fmt"
	"io/ioutil"
	"os"
	"os/user"
	"path/filepath"
	"strings"

mseri

#!/bin/bash
# Ollama Model Export Script
# Usage: bash ollama-export.sh vicuna:7b
# License: MIT (https://ncurl.xyz/s/o_o6DVqIR)
# https://gist.github.com/supersonictw/f6cf5e599377132fe5e180b3d495c553

# Interrupt if any error occurred
set -e

# Declare

mseri

Replace natbib by biblatex to get citation links

\usepackage[backend=bibtex, style=authoryear-comp, natbib=true, sortcites=false]{biblatex}
\addbibresource{main.bib}

% optional if you want (Smith 1776) instead of (Smith, 1776)
\renewcommand*{\nameyeardelim}{\addspace}

\begin{document}

mseri

def energy(z, n, m, L):
    return np.cos(n* np.pi *np.real(z) / L) *np.cos(m *np.pi*np.imag(z) / L) - np.cos(m*np.pi*np.real(z)/ L) *np.cos(n* np.pi *np.imag(z)  / L)


class ChladniPlate:
    def __init__(self, n, m, L=1, n_particles=10000):
        self.L = L
        self.n_particles = n_particles
        self.n = n
        self.m = m

mseri

How to setup a practically free CDN

I've been using [Backblaze][bbz] for a while now as my online backup service. I have used a few others in the past. None were particularly satisfactory until Backblaze came along.

It was - still is - keenly priced at a flat $5 (£4) per month for unlimited backup (I've currently got just under half a terabyte backed-up). It has a fast, reliable client. The company itself is [transparent about their operations][trans] and [generous with their knowledge sharing][blog]. To me, this says they understand their customers well. I've never had reliability problems and everything about the outfit exudes a sense of simple, quick, solid quality. The service has even saved the day on a couple of occasions where I've lost files.

Safe to say, I'm a happy customer. If you're not already using Backblaze, [I highly recommend you do][recommend].

Taking on the big boys with B2

mseri

import math

import numpy as np
import random
import matplotlib.pyplot as plt

medium_size = 1024
x = np.linspace(-1., 1., medium_size)
y = np.linspace(-1., 1., medium_size)
x, y = np.meshgrid(x, y, indexing='ij')

mseri

module Trie = struct
  type t = { mutable count : int; children : t option Array.t }

  let[@inline] empty () = { count = 0; children = Array.make 128 None }

  let add buf pos len t =
    let rec loop idx t =
      if idx > pos + len - 1 then t.count <- t.count + 1
      else
        let c = Char.lowercase_ascii (Bytes.get buf idx) in