Локальные LLM
Ollama
The simplest way to run open language models on your own machine.
★ 182,1 тыс. звёздMIT
Смотреть руководство →Unsloth — проект с открытым исходным кодом из категории «Локальные LLM» от unslothai (unslothai/unsloth на GitHub). У него 77 168 звёзд на GitHub и 7 096 форков, он распространяется по лицензии Apache-2.0 и последний раз обновлялся на GitHub 3 октября 2026.
Run and fine-tune LLMs and diffusion models locally — faster training, smaller memory footprint.
Grab Unsloth Desktop for Windows, macOS, or Linux; open Unsloth Studio in the browser; or pip-install Unsloth Core for code-first work.
Choose from the latest open families — Qwen, DeepSeek-V4, Gemma 4, MiniMax-H3, FLUX, and many more.
Run models locally for private chat, or fine-tune on your own data with Unsloth's memory-efficient kernels.
Save your tuned model to GGUF or MLX and run it in any compatible app or on-device runtime.
Custom kernels train sooner and leaner — fine-tune bigger models on smaller GPUs.
One workbench for language models and image models, including FLUX.
Take your tuned models to llama.cpp, Ollama-style runners, or Apple Silicon apps.
Desktop app, web-based Studio, or Core for scripts and notebooks.
Tool calling, RAG, and integrations with coding agents and MCP-style tooling.
Day-one-ish support for new open releases across the ecosystem.
Download Unsloth Desktop for your OS from the releases page and install it like any app.
Prefer the browser? Open Unsloth Studio from the docs site — no install at all.
For notebooks and scripts, install the code package and follow the docs' training notebooks.
pip install unslothOllama is the simplest way to run models. Unsloth is for running plus training — fine-tuning, LoRAs, and efficient experimentation. They pair well together.
The open-source project is Apache-2.0 licensed. Check the official site for anything around hosted or premium offerings.
Efficiency is the whole point — Unsloth is designed to train bigger models on smaller cards than stock tooling allows.
The Desktop app and Studio are the friendly on-ramps; Core and notebooks serve the code-first crowd.
The tooling is Apache-2.0. Also check the license of the base model you fine-tune, since that travels with your derivative.
The simplest way to run open language models on your own machine.
Alibaba's open voice models: expressive, streaming-capable speech with free-form voice design.
Turn the beloved Kokoro voice model into an OpenAI-compatible speech API with one Docker command.
GPT4All — run local LLMs on any device: a free desktop chat client and Python SDK that run open models entirely on your CPU, no GPU or cloud required
Check if your GPU can fine-tune an LLM — memory estimates and LoRA recipes.
Run controllable pre-trained AI models locally, written in Rust.
Другие проекты из категории «Локальные LLM» с наибольшим числом звёзд в каталоге LibreAI — реалистичный список альтернатив, если Unsloth вам не подходит.
The simplest way to run open language models on your own machine.
User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
LLM inference in C/C++
A high-throughput and memory-efficient inference and serving engine for LLMs
🤯 LobeHub is your Chief Agent Operator, organizing your agents into 7×24 operations by hiring, scheduling, and reporting on your entire AI team.
Соседние проекты из категории «Локальные LLM» в каталоге, ближайшие по числу звёзд к Unsloth — продолжайте изучать полку по обе стороны от него.
GPT4All — run local LLMs on any device: a free desktop chat client and Python SDK that run open models entirely on your CPU, no GPU or cloud required
The fastest, litest AI Gateway. Rust core with Python SDK. Call 100+ LLM APIs in OpenAI (or native) format with cost tracking, guardrails, load balancing, an…
🤯 LobeHub is your Chief Agent Operator, organizing your agents into 7×24 operations by hiring, scheduling, and reporting on your entire AI team.
Run frontier AI locally.
A high-throughput and memory-efficient inference and serving engine for LLMs
Open-source desktop app for local LLMs. Text, vision, tool-calling, OpenAI/Anthropic-compatible API. 100% private
Бесплатно и открыто — забирайте исходный код, ставьте звезду репозиторию и начинайте создавать.