Where Are the Models in Ollama Stored?

So, here’s the deal with finding your Ollama models—they’re tucked away in a spot that’s both convenient and well-organized, but not always super obvious if you’re new to this. The default storage location for Ollama models is in your home directory, specifically under ~/.ollama/models. If you’re not familiar with Unix-style directories, don’t sweat it; I’ll break it down in a way that makes sense, even if you’ve never cracked open a terminal before.

I remember the first time I went looking for these files. I thought, “Why not just keep them in the main project folder where I’m working?” Turns out, having a dedicated directory like ~/.ollama/models is way better for long-term organization. It keeps everything centralized and avoids cluttering your workspaces with random files. Plus, if you’re working across multiple projects, all your downloaded or custom-trained models are in one spot—no hunting around.

How to Find and Navigate the Models Directory

If you’re on a Mac or Linux system, you can access this directory pretty easily. Just open up your terminal (don’t be intimidated—it’s simpler than it looks) and type:

cd ~/.ollama/models

This will take you straight to the folder where the models are stored. If you want to see what’s inside, use the ls command:

ls

Windows users, don’t worry. The process is a bit different, but the concept is the same. On Windows, these directories are stored under something like C:\Users\<YourUsername>\.ollama\models. You’ll need to enable viewing hidden files to see the .ollama folder—Windows likes to hide anything starting with a dot.

Go to File Explorer, click on “View,” and check the box for “Hidden items.”

A Few Tips to Keep It Simple

I’ve learned a few tricks that can save you headaches when working with Ollama models. Here are my top three:

1. Create Shortcuts: If you find yourself frequently accessing the models folder, create a shortcut to it. On Linux or Mac, you can add an alias in your shell configuration file (like .bashrc or .zshrc):alias ollama_models='cd ~/.ollama/models' Next time, just type ollama_models in the terminal, and you’re there.
2. Back It Up: These models can be hefty, and if you’ve spent time customizing or training them, losing them would be painful. Set up an automated backup to a cloud service like Google Drive or Dropbox. Personally, I use rsync with a cron job, but even a manual copy works fine.
3. Watch Storage Space: Models can take up a lot of room, especially as you download or create new ones. Check your storage regularly with tools like du in the terminal:du -h ~/.ollama/models It’ll give you a quick look at how much space these models are eating up.

Lastly, don’t be afraid to clean house every now and then. If you’ve downloaded a bunch of models you don’t use anymore, delete them! Keeping only what’s necessary not only saves space but makes it easier to find what you need when you need it. Oh, and don’t forget to document any tweaks or training you’ve done—future you will thank you.