Hey everyone,
The buzz around DeepSeek R1 is real! It’s a powerful Chinese AI model, and while you can use it for free on their website, some of us (myself included) are a bit wary of the privacy implications. So, what’s the solution? Run it locally!
There are several versions of DeepSeek R1, ranging from the distilled models (1.5B, 7B, 14B, 32B, and 70B parameters) to the full-fledged 671B parameter model. Keep in mind that the bigger the model, the more horsepower you’ll need. We’re talking serious memory, a fast CPU/GPU, and potentially even an NPU for those top-tier models.
Even for the smaller models (1.5B and 7B), you’ll need at least 8GB of RAM to avoid issues.
Method 1: LM Studio (PC - Windows/Mac/Linux)
LM Studio makes running DeepSeek R1 locally a breeze on your PC. No need to mess around with the terminal!
LM Studio is compatible with:
- Macs: Apple Silicon Chips (M1, M2, M3, M4) running macOS 13.4 or later
- Windows: Both x64 and ARM (Snapdragon X Elite) systems are supported
- Linux: Ubuntu 20.04 or newer (x64 required)
While 16 GB of RAM is the recommended sweet spot, you can get away with 8 GB for the smaller models. And if you’re rocking over 192 GB of RAM? You can even run the massive 671B DeepSeek R1 model!
Here’s how to get started:
-
Head over to lmstudio.ai and grab the installer for your operating system.
-
Follow the installation prompts for your system. Once installed, launch LM Studio and choose “Skip onboarding”.
-
Click the “Search” icon in the left-hand navigation.
Note: If the “Search” icon is missing, either switch to “Power User” mode at the bottom or go to “App Settings” (bottom-right) and enable “Model Search” in the Mission Control window.
-
Search for the DeepSeek R1 model you want. For example, “DeepSeek R1 Distill (Qwen 7B)” for the 7B parameter model, or “DeepSeek R1 Distill Qwen 1.5B” for the even smaller one. All available DeepSeek R1 models will show up, displaying the disk space each requires. Hit “Download” to grab the model.
-
After the download finishes, click “Load Model”. This will open the chat interface.
-
Start typing your messages in the chat window! You can monitor resource usage in the bottom-right corner.
-
To tweak settings, click the “Setting” icon to the left of the model name at the top.
Method 2: PocketPal AI (Phone - iOS/Android)
Want to run DeepSeek R1 on your phone? PocketPal AI is a solid option for both iOS and Android.
-
Download and install the app from either the App Store or the Play Store.
-
Open PocketPal AI and tap on “Go to Models”.
-
Tap the “+” icon in the bottom right corner.
-
Tap “Add from Hugging Face”.
-
Search for “DeepSeek R1”, then tap the model you want.
-
Tap the download icon next to one of the available GGUF files.
-
Once the download completes, go back to the main screen and tap “Load” on your downloaded model.
-
The model will load, and you can start chatting!
-
Access your chat history and model settings from the left-hand menu.
That’s it! You should be able to get up and running with DeepSeek R1 on your PC or phone by following the steps above.