A light-weight codebase that enables memory-efficient and performant finetuning of Mistral's models. It is based on LoRA, a training paradigm where most weights are frozen and only 1-2% additional weights in the form of low-rank matrix perturbations are trained.
Utilities for Llama.cpp, OpenAI, Anthropic, Mistral-rs. A collection of tools for interacting with various large language models. The code is written in Rust and includes functions for loading models, tokenization, prompting, text generation, and more.
This article explains the use of GitHub Copilot, an AI-powered coding companion that suggests edits and refactors to improve code quality and developer productivity. It includes examples of how to use Copilot effectively to modify and refactor code.
This gist provides instructions on how to create a systemd service that sets the Nvidia power limit after booting up. The power limit can be adjusted to increase longevity or performance, depending on the user's needs. The instructions include checking the current power settings, setting up the systemd service, and an example output after reboot.
The workflow triggers on push events on the 'master', 'main', and 'fix' branches, and runs on Ubuntu. It installs Make, caches the Cosmocc toolchain, sets up Cosmocc and Ape Loader, builds the project, makes a specific Llamafile, executes the Llama CLI CPU, and more.
LlamaFS is a self-organizing file manager that automatically renames and organizes files based on their contents. It supports various file types and even images and audio. It can run in two modes - batch mode and watch mode. In batch mode, LlamaFS suggests a file structure and organizes files. In watch mode, it watches your directory and proactively learns your file organization habits. The project is built on a Python backend and Electron for the frontend.
Discover how GitHub engineers leverage GitHub Copilot to automate repetitive tasks, stay focused, and even explore new technologies.
The GitHub Blog article "4 ways GitHub engineers use GitHub Copilot" discusses how different GitHub engineers leverage GitHub Copilot to enhance efficiency and productivity in their work. The main use cases highlighted in the article are:
1. Semi-automating repetitive tasks: A GitHub engineer uses GitHub Copilot to automate the process of incrementing ID numbers in protobuf definitions, significantly reducing manual work.
2. Avoiding distractions: A colleague develops a regular expression to capture a Markdown code block and extract the language identifier. By prompting GitHub Copilot with a code comment, the regular expression is generated swiftly and accurately.
3. Structuring data-related notes: A support engineer at GitHub uses GitHub Copilot Chat to structure and compile their notes into Markdown tables, enabling efficient documentation of a complex troubleshooting process.
4. Exploring and learning: A colleague uses GitHub Copilot to develop a program in Rust, a programming language he was not previously familiar with. In just 23 minutes, he successfully created a functional program that converts numerical input into written English equivalents.
A collection of Raspberry Pi-sourced system configuration files and associated scripts.
ShellCheck is a tool that helps you find bugs in your shell scripts. It's free, open-source, and available on various platforms. Paste a script to try it out, or apply fixes, report bugs, or load random examples. The tool also provides output and supports integration in major editors and services like CodeClimate, Codacy, and CodeFactor.
Rundeck is an open source automation service that allows users to easily run automation tasks across a set of nodes. This repository contains the source code for Rundeck, with features such as a web console, command line tools, and a WebAPI. It is built with Gradle and requires Java 1.8 and NodeJs 16. Documentation, development guides, and issue tracker are available.
Bookmarks on this page are managed by an admin user. 