Skip to content
repository radar
Voice, Vision & Multimodal

Unsloth

ACTIVE STEADY

unslothai/unsloth · homepage ↗

Local UI to run and train LLMs and diffusion models, including Qwen3.8, Kimi K3, MiniMax-H3, Gemma 4, DeepSeek-V4, FLUX and more.

⭐ Hugely popular: 72k stars, gaining about 1k a week

View on GitHub ↗

repo profile

vintage 2023 3 years old
delivery product
language Python
license Apache-2.0

momentum

total stars 72k
stars added last week +1k
HN peak 385 pts 2y 8mo ago
weekly downloads 655k /wk · PyPI
binary downloads 23k total · GitHub releases
commits / week 244 accelerating · +152% vs prior mo
issues closed 74% ~4d to close, median
contributors 274 (+16 in 17d)
release cadence weekly
last activity yesterday

durability

backing VC-backed Y Combinator
openness permissive
security score 4.5/10 OpenSSF Scorecard · 2026-08-14
bus factor 2 concentrated
top-author share 48% 6 mo

bus factor = how many people it takes to cover more than half the commits (6 months). 1 is a solo project; higher means the work is spread across a team. top-author share is the single busiest author's slice of those commits.

since we covered it

monthly average + 2k/mo (+8%/mo) · + 42k total since PR#2

why it's a big deal

  • Reduces GPU memory usage, making large-scale model tuning accessible on consumer hardware.
  • Supports 4-bit and 16-bit QLoRA fine-tuning, optimizing efficiency without compromising accuracy.
  • Allows for exporting models to GGUF, Ollama, and Hugging Face with seamless integration.

under the hood

  • Written in OpenAI's Triton, ensuring optimized backpropagation and fast training loops.
  • Implements advanced tensor techniques for memory efficiency in fine-tuning tasks.
  • Benchmarked against Hugging Face’s standard implementations, showing significant speed and context-length improvements.

our take from PR#2, 2025-02-19

star history

PR#2 · 30k72k now Nov 2023Aug 2026
  1. PR#2 30k 2025-02-19
  2. now 72k + 42k since first covered

curve is sampled from GitHub's star history, plus our own daily readings since we covered it; the dashed stretch is before we first covered it, the solid line since. figures at coverage are the numbers we printed then (approx.), current count is live.

understory

Quietly building: more output than attention, for now.

+22 understory score output 99 · clout 77
Aug 2025 Jul 2026
  • output, commits & releases
  • clout, star velocity

output = commits & releases; clout = star velocity, both 0 to 100 monthly indices; the gap where output runs above clout is the understory. The understory →

covered in

  • PR#2 2025-02-19 on the radar

    Faster, Memory-Efficient Fine-Tuning for LLMs

similar projects

compare these →
  • 🦙 LLaMA-Factory

    Unified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)

    74k ACTIVE
  • 🎬 OpenMontage

    AGPL-3.0 licensed

    World's first open-source, agentic video production system. 12 production pipelines, 100+ tools, 700+ agent skill and production-knowledge files. Turn your AI coding assistant into a full video production studio.

    48k ACTIVE
  • 🖥️ Open WebUI

    2.1× the stars

    User-friendly AI Interface (Supports Ollama, OpenAI API,...)

    149k ACTIVE

comments

Sign in with GitHub to add your blip on Unsloth.

loading comments…