AI models
Models downloaded by local AI tools such as Ollama and LM Studio, and the ones other apps keep inside them.
13 entries in 2 groups.
- Safe
- Its owner recreates it by itself when it needs it, and nothing you made is lost.
- Usually safe
- It can be recreated, but it takes a download, a reinstall or the tool's own command.
- Review first
- It's your own data, or hard to recreate. Decide item by item.
- Managed by macOS
- macOS manages it. It's never moved to the Trash.
Local AI models
- DiffusionBee modelsStable Diffusion models that DiffusionBee downloaded or that you imported into it.Review first~/
.diffusionbee/ {downloaded_assets,downloads,imported_models} - Docker Model Runner modelsAI models pulled or packaged with Docker Model Runner, kept on your Mac outside Docker's disk image.Review first~/
.docker/ models - Draw Things modelsImage and video generation models downloaded or imported in Draw Things.Review first~/
Library/ Containers/ com.liuliu.draw-things/ Data/ Documents/ Models - GPT4All modelsLanguage model files downloaded in GPT4All, each stored as a .gguf file in its download folder.Review first~/
Library/ Application Support/ nomic.ai/ GPT4All/ *.gguf - Hugging Face cacheModels and datasets downloaded from Hugging Face, shared by every library and tool that uses them.Review first~/
.cache/ huggingface/ hub - Hugging Face Xet cacheCaches that Hugging Face's Xet storage uses to avoid transferring the same data twice.Safe~/
.cache/ huggingface/ xet - Jan modelsLanguage models downloaded or imported in Jan, grouped by the engine that runs them.Review first~/
Library/ Application Support/ Jan/ data/ {llamacpp,mlx}/ models - llama.cpp model cacheModel files llama.cpp downloaded from a web address or Docker Hub, and Hugging Face models saved by older versions.Usually safe~/
Library/ Caches/ llama.cpp - LM Studio modelsLanguage models downloaded or imported in LM Studio, in a folder for each publisher and model.Review first~/
.lmstudio/ models - Ollama modelsLanguage models you've pulled or created with Ollama, stored as shared layers.Review first~/
.ollama/ models - OpenAI Whisper modelsSpeech recognition models downloaded by OpenAI's open-source Whisper package, one file per model size.Usually safe~/
.cache/ whisper - PyTorch Hub cacheModel code and pretrained weights downloaded by PyTorch Hub and libraries built on it, such as torchvision.Usually safe~/
.cache/ torch/ hub