Brainsteam

Search

SearchSearch
      • AI is Causing Chaos
      • AI Metrics
      • Anthropic Claude
      • Artificial Intelligence and Machine Learning Team Responsibilities
      • Best Practices in ML
      • Best-Worst (MaxDiff) Ranking
      • ColBERT
      • Commercial AI APIs
      • Copilot Lag
      • evaluate
      • Gemma3
      • GPT-3
      • GPT-4
      • GPT-4o
      • List of LLM and VLM models
      • LLaMa
      • llama-cpp-server
      • Llama3
      • LLaVa
      • LLM Memory and Compute Requirements
      • llms
      • LLMs as judges
      • Local LLMs
      • LoRA
      • Machine Learning
      • Mistral 7b
      • Model Context Protocol
      • My GPTs
      • OLMOE
      • Pegasus
      • QLoRa
      • Qwen
      • Qwen 2.5
      • Qwen 3
      • Qwen-VL
      • Speech to Text (STT)
            • End of an Era
        • Anthropic
        • Google
        • Huggingface
        • OpenAI
          • If your websites full of assholes, its your fault - Anil Dash
          • How We Became the McWorld - Global Culture is Getting More Boring
          • This is My Most Expensive Habit - RyanHoliday.net
          • Creative Doing
          • Hunt the Stars
          • Malevolent Seven
          • They Knew: How a Culture of Conspiracy Keeps America Complacent
          • The Fall Guy
          • Wicked Little Letters
          • Black Doves
        • Evaluating LLMs
        • LIME
        • SHAP
        • Dr Rangan Chatterjee
        • Ed Zitron
        • Marc Andreessen
        • joplin-hypothesis
        • PenParse
        • Arrange Act Assert
        • asyncio
        • Flask HTTP 415 with get_json
        • GCP
        • Google Artifact Registry
        • Google BigQuery
        • Google GCP
        • Google GCP Instance Schedules
        • Google GCP Startup Scripts
        • hatch
        • Hoisting in javascript
        • html2canvas
        • HTML5 Canvas
        • pyright
        • python
        • Python AttributeError cython_sources
        • Python Packaging
        • Python Testing
        • accelerate
        • Airbyte
        • Aphrodite Engine
        • Argilla
        • AWS API Gateway
        • AWS Lambda Functions
        • AWS Neptune
        • AWS SAM
        • axolotl
        • BERTopic
        • Bookstack
        • BotKit
        • Boto3
        • brew
        • Caddy
        • Calibre
        • CapRover
        • Cline
        • Coolify
        • django
        • Docker
        • docker-compose
        • Doku (LLM Monitoring)
        • DVC
        • Eleventy
        • evaluate
        • fasthtml
        • fasthx
        • Firefox
        • flask
        • Forgejo
        • Github CoPilot
        • Gitlab
        • Gnome Shell
        • GoAccess
        • Grype
        • gunicorn
        • Hoarder
        • Huggingface Optimum
        • Huggingface SFT
        • hypothes.is
        • Ideogram
        • Joplin
        • LiteLLM
        • llama.cpp
        • llm
        • LMDeploy
        • memos
        • MLFlow
        • mongodb
        • Neo4j
        • NeoVIM
        • Nomic Atlas
        • NotebookLM
        • Nvidia Triton
        • ollama
        • ONNXRuntime
        • Open Web UI
        • Pandas
        • Paperless-ngx
        • PDM
        • PEFT
        • Podman
        • Poetry
        • PostgreSQL
        • pydantic
        • pyenv
        • pylance
        • Reclaim-The-Stack
        • Redis
        • Requests
        • Restic
        • Scikit-learn
        • Silverbullet
        • Small Text
        • spaCy
        • SQLAlchemy
        • SQLite
        • SQLModel
        • Syncthing
        • text-generation-ui
        • Transformers
        • Unraid
        • Unsloth
        • uv
        • vllm
        • Volta
        • VSCode
        • wallabag
        • Zed
      • /ai
      • /using - my tech stack
      • Anxiety
      • Artificial Intelligence
      • AWS Sagemaker Endpoints
      • Best-Worst Scaling More Reliable than Rating Scales: A Case Study on Sentiment Intensity Annotation
      • Blame Fuse
      • Colophon
      • continue-dev
      • Contrastive Training
      • Deep Java Learning
      • digital garden
      • flow state
      • foss
      • GeoJSON
      • GGUF
      • GLiNER
      • Google Gemini
      • Google Gemini 1.5 Flash
      • Gremlin
      • Ikigai
      • Jupyter
      • Kobo Elipsa
      • Langfuse
      • Large Multi-Modal Model
      • Learning in Public
      • LilyGo T-Deck Plus
      • Measuring LLM Summarisation performance
      • mental health
      • meshtastic
      • Morning Person
      • mosh
      • MTEB
      • music
      • My Publication Workflow
      • Named Entity Recognition
      • Natural Language Processing
      • NextJS
      • Nomic Embed
      • NVIDIA GPUs
      • Obsidian
      • OCR for handwriting
      • OpenCypher
      • OpenVino
      • ORM
      • Personal Knowledge Management
      • Phudge
      • pip
      • PostGIS
      • Prompt-Based Modelling
      • puppeteer-video-recorder
      • pwdlib
      • pyllmcore
      • Regression Analysis
      • Retrieval Augmented Generation
      • Roland AE-10
      • Samsung Tizen
      • SeaQL
      • Set up a new Docker Compose Deployment Pipeline in Github CI
      • SetFit
      • Sling
      • Small Language Models
      • Software Engineering
      • SPARQL
      • ssl and python
      • Structured Output
      • The Bitter Lesson
      • The Toolbox Fallacy
      • Time Series Processing
      • Vitest
      • WizardCoder
      • WKT
      • writing practice
    Home

    ❯

    GGUF

    GGUF

    Apr 25, 20251 min read

    GGUF is a model serialization format for Machine Learning models with built in quantization.

    How Does GGUF Work?

    There is an excellent guide to the format by Vicki Boykiss (mirror) that explains different file formats and the need for GGUF models.

    Converting to GGUF

    llama.cpp


    Graph View

    • How Does GGUF Work?
    • Converting to GGUF

    Backlinks

    • OCR for handwriting
    • Aphrodite Engine
    • llama.cpp
    • ollama
    • text-generation-ui
    • WizardCoder

    Created with Quartz v4.2.4 © 2025

    • Blog
    • Mastodon
    • Bluesky