Local UI to run and train LLMs and diffusion models. Supports GGUF, MLX, Qwen3.8, DeepSeek-V4.1, MiniMax-H3, Gemma 4, FLUX and more.
Python agentaichatgptdeepseekfine-tuninggemmaimage-generationllamallmllmsopenaipythonqwenreinforcement-learningself-hostedstable-diffusiontext-to-speechttsuiunsloth
docs: clarify the Docker quickstart shell, Windows setup and logs command (#11036)
Lee Jackson7e8c8faa5350
diff --git a/README.md b/README.mdindex 4e18517a2bd..6318ff357a0 100644--- a/README.md+++ b/README.md@@ -151,7 +151,9 @@ unsloth studio --secure```#### Docker-Use our [Docker image](https://hub.docker.com/r/unsloth/unsloth) ```unsloth/unsloth```. On Linux, set up GPU access once with `curl -fsSL https://raw.githubusercontent.com/unslothai/unsloth/main/docker/install_nvidia_toolkit.sh -o install_nvidia_toolkit.sh && sudo -E bash install_nvidia_toolkit.sh` (Windows: Docker Desktop with WSL 2). Run:+Use our [Docker image](https://hub.docker.com/r/unsloth/unsloth) ```unsloth/unsloth```. On Linux, set up GPU access once with `curl -fsSL https://raw.githubusercontent.com/unslothai/unsloth/main/docker/install_nvidia_toolkit.sh -o install_nvidia_toolkit.sh && sudo -E bash install_nvidia_toolkit.sh` (Windows: [Docker Desktop with WSL 2](https://unsloth.ai/docs/get-started/install/docker)).++**Linux / WSL (Bash):**```bashdocker run -d --gpus all --ipc=host \-p 8000:8000 -p 8888:8888 \@@ -160,7 +162,7 @@ docker run -d --gpus all --ipc=host \-v unsloth-studio:/opt/unsloth-studio \unsloth/unsloth```-Follow startup with `docker logs -f`. Studio is at `http://localhost:8000` (user `unsloth`), JupyterLab at `http://localhost:8888`. The `unsloth-studio` volume keeps your accounts, chats and trained models across `docker rm`; each image brings its own Studio code, and a volume from an older image is migrated on the first start (its old code is kept under `.unsloth-studio-legacy/`). Tags (`unsloth/unsloth:core` for notebooks only), GPU support and options: [Docker Hub](https://hub.docker.com/r/unsloth/unsloth).+Follow startup with `docker logs -f <container>`, where `<container>` is the ID printed by `docker run -d` or a name from `docker ps`. Studio is at `http://localhost:8000` (user `unsloth`), JupyterLab at `http://localhost:8888`. The `unsloth-studio` volume keeps your accounts, chats and trained models across `docker rm`; each image brings its own Studio code, and a volume from an older image is migrated on the first start (its old code is kept under `.unsloth-studio-legacy/`). Tags (`unsloth/unsloth:core` for notebooks only), GPU support and options: [Docker Hub](https://hub.docker.com/r/unsloth/unsloth).#### Remote HTTPS & LAN AccessServer-side tools are on by default - so **be careful**! Keep your password safe, or use `--disable-tools` when exposing Unsloth.