unslothai / unsloth-zoo

Unsloth Zoo - Utils for Unsloth! Finetune gpt-oss, Gemma 3n, Qwen3, Llama 4, & Mistral 2x faster with 80% less VRAM! ✨ Finetune for Free Notebooks are beginner friendly. Read our guide. Add your dataset, click "Run All", and export your finetuned model to GGUF, Ollama, vLLM or Hugging Face. | Unsloth supports | Free Notebooks | Performance | Memory use | |-----------|---------|--------|----------| | **gpt-oss (20B)** | ▶️ Start for free-Fine-tuning.ipynb) | 1.5x faster | 70% less | | **Gemma 3n (4B)** | ▶️ Start for free-Conversational.ipynb) | 1.5x faster | 50% less | | **Qwen3 (14B)** | ▶️ Start for free-Reasoning-Conversational.ipynb) | 2x faster | 70% less | | **Qwen3 (4B): GRPO** | ▶️ Start for free-GRPO.ipynb) | 2x faster | 80% less | | **Gemma 3 (4B)** | ▶️ Start for free.ipynb) | 1.6x faster | 60% less | | **Phi-4 (14B)** | ▶️ Start for free | 2x faster | 70% less | | **Llama 3.2 Vision (11B)** | ▶️ Start for free-Vision.ipynb) | 2x faster | 50% less | | **Llama 3.1 (8B)** | ▶️ Start for free-Alpaca.ipynb) | 2x faster | 70% less | | **Mistral v0.3 (7B)** | ▶️ Start for free-Conversational.ipynb) | 2.2x faster | 75% less | | **Orpheus-TTS (3B)** | ▶️ Start for free-TTS.ipynb) | 1.5x faster | 50% less | • See all our notebooks for: Kaggle, GRPO, **TTS** & Vision • See all our models and all our notebooks • See detailed documentation for Unsloth here ⚡ Quickstart Linux or WSL Windows For Windows, works only if you have Pytorch installed. For more info, read our Windows Guide. Docker Use our official Unsloth Docker image container. Read our Docker Guide. Blackwell For RTX 50x, B200, 6000 GPUs, simply do . Read our Blackwell Guide for more details. 🦥 Unsloth.ai News • 📣 **Memory-efficient RL** We're introducing even better RL. Our new kernels & algos allows faster RL with 50% less VRAM & 10× more context. Read blog • 📣 **gpt-oss** by OpenAI: For details on Unsloth Flex Attention, long-context training, bug fixes, Read our Guide. 20B works on a 14GB GPU and 120B on 65GB VRAM. gpt-oss uploads. • 📣 **Gemma 3n** by Google: Read Blog. We uploaded GGUFs, 4-bit models. • 📣 **Text-to-Speech (TTS)** is now supported, including and STT . • 📣 **Qwen3** is now supported. Qwen3-30B-A3B fits on 17.5GB VRAM. • 📣 Introducing **Dynamic 2.0** quants that set new benchmarks on 5-shot MMLU & KL Divergence. • 📣 **EVERYTHING** is now supported - all models (BERT, diffusion, Cohere, Mamba), FFT, etc. MultiGPU coming soon. Enable FFT with , 8-bit with . Click for more news • 📣 DeepSeek-R1 - run or fine-tune them with our guide. All model uploads: here. • 📣 Introducing Long-context Reasoning (GRPO) in Unsloth. Train your own reasoning model with just 5GB VRAM. Transform Llama, Phi, Mistral etc. into reasoning LLMs! • 📣 Introducing Unsloth Dynamic 4-bit Quantization! We dynamically opt not to quantize certain parameters and this greatly increases accuracy while only using 🔗 Links and Resources | Type | Links | | ------------------------------- | --------------------------------------- | | 📚 **Documentation & Wiki** | Read Our Docs | |   **Twitter (aka X)** | Follow us on X| | 💾 **Installation** | Pip install| | 🔮 **Our Models** | Unsloth Releases| | ✍️ **Blog** | Read our Blogs| |   **Reddit** | Join our Reddit| ⭐ Key Features • Supports **full-finetuning**, pretraining, 4b-bit, 16-bit and **8-bit** training • Supports **all transformer-style models** including TTS, STT, multimodal, diffusion, BERT and more! • All kernels written in OpenAI's Triton language. **Manual backprop engine**. • **0% loss in accuracy** - no approximation methods - all exact. • No change of hardware. Supports NVIDIA GPUs since 2018+. Minimum CUDA Capability 7.0 (V100, T4, Titan V, RTX 20, 30, 40x, A100, H100, L40 etc) Check your GPU! GTX 1070, 1080 works, but is slow. • Works on **Linux** and **Windows** • If you trained a model with 🦥Unsloth, you can use this cool sticker!   💾 Install Unsloth > [!warning] > Python 3.14 is supported. You can also see our documentation for more detailed installation and updating instructions here. Pip Installation **Install with pip (recommended) for Linux devices:** **To update Unsloth:** See here for advanced pip install instructions. Windows Installation • **Install NVIDIA Video Driver:** You should install the latest version of your GPUs driver. Download drivers here: NVIDIA GPU Drive. • **Install Visual Studio C++:** You will need Visual Studio, with C++ installed. By default, C++ is not installed with Visual Studio, so make sure you select all of the C++ options. Also select options for Windows 10/11 SDK. For detailed instructions with options, see here. • **Install CUDA Toolkit:** Follow the instructions to install CUDA Toolkit. • **Install PyTorch:** You will need the correct version of PyTorch that is compatible with your CUDA drivers, so make sure to select them carefully. Install PyTorch. • **Install Unsloth:** Notes To run Unsloth directly on Windows: • Install Triton from this Windows fork and follow the instructions here (be aware that the Windows fork requires PyTorch >= 2.4 and CUDA 12) • In the , set to avoid a crashing issue: Advanced/Troubleshooting For **advanced installation instructions** or if you see weird errors during installations: First try using an isolated environment via then • Install and . Go to https://pytorch.org to install it. For example • Confirm if CUDA is installed correctly. Try . If that fails, you need to install or CUDA drivers. • Install manually via: Check if succeeded with Go to https://github.com/facebookresearch/xformers. Another option is to install for Ampere GPUs and ignore • For GRPO runs, you can try installing and seeing if succeeds. • Double check that your versions of Python, CUDA, CUDNN, , , and are compatible with one another. The PyTorch Compatibility Matrix may be useful. • Finally, install and check it with Conda Installation (Optional) . Select either for CUDA 11.8 or CUDA 12.1. We…