Unsloth
Fine-tune LLMs 2x faster with 80% less memory
About
Unsloth is a library for faster LLM fine-tuning and inference that patches Hugging Face training internals to use less memory and run on consumer GPUs. It supports LoRA, QLoRA, and full fine-tuning across Llama, Mistral, Qwen, Gemma, and other families, on Linux, Windows, WSL, and macOS via Unsloth Studio or the code-based Unsloth Core. The project also publishes pre-quantized GGUF model exports on Hugging Face.
Reviews (0)
Leave a Review
No reviews yet. Be the first to review!
Details
- Category
- Datasets & Training
- Price
- Free
- Platform
- Local/Desktop
- Difficulty
- Intermediate (3/5)
- License
- Apache-2.0
- Minimum VRAM
- 8 GB
- Added
- Jan 29, 2026
Related Tools
Tool for fine-tuning LLMs with various configurations
Open source data labeling platform for ML projects
Hugging Face library for parameter-efficient fine-tuning
Python library for synthetic data generation and curation pipelines with batch LLM inference.
Detects label errors, outliers, and duplicates in machine learning datasets using confident learning.
Data processing system for foundation models with over 200 composable operators for multimodal data.
Mentioned in
Running LLMs Without Nvidia in 2026: ROCm and Apple Silicon
The 2026 state of local LLMs off Nvidia: which runtimes actually work on AMD ROCm and Apple Silicon, real...
Billy C
Fine-Tuning vs RAG vs Long Context in 2026: A Decision Guide
How to choose between fine-tuning, RAG, and long context in 2026, with real costs, the failure modes of each...
Max P
Fine-Tuning Llama 3.3 with Unsloth on a 16GB GPU, Step-by-Step
A practical, end-to-end fine-tuning walkthrough with Unsloth: dataset prep, LoRA config, 4-bit quantization,...
Billy C