How to Install Microsoft Phi-3 Using LM Studio

In our earlier post, we discussed extensively how to install Microsoft Phi-3 AI locally on Windows. The Phi-3 family of language models from Microsoft is now available. They are unquestionably a class apart for their size and are already outperforming other newly introduced models like the Mixtral and Llama 3 in several areas. Phi-3 is small enough to operate locally on your Windows computer. Here’s how to use LM Studio to accomplish this.

Nothing compares to LM Studio if you prefer a specialized interface for Microsoft’s Phi-3 on Windows rather than utilizing your terminal program to communicate with it. Here’s how to install Phi-3 in LM Studio and initiate a local conversation with the model.

1. Install LM Studio

• LM Studio can be downloaded using this download link.

• Once downloaded, run the installer and let the LM Studio install.

Recommended: How to Force Adjust the Brightness on Windows

2. Download the Phi-3 gguf file

Phi-3 will not be available for download or search within LM Studio. The Phi-3 guff file must be downloaded separately. Here’s how:

• Phi-3-Mini Guff file can be downloaded using this download link.
• Use the link given above and click on Files.

• Two iterations of the Phi-3 model are available here. Pick one. We are choosing the smaller version for our objectives.

• Click on download and save it on your computer.

Recommended: How to Install Microsoft Phi-3 AI Locally on Windows

3. Load the Phi-3 model

The downloaded Phi-3 model will then be loaded. To accomplish this, take the following actions:

• Open LM Studio and click on My Models on the left.

• The “Local models folder” is worth mentioning. The downloaded Phi-3 guff file has to be moved here. To open the directory, click Show in File Explorer.

• Make a new folder called Microsoft here.
• Make another folder called Phi-3 inside the Microsoft folder.
• In the Phi-3 folder, paste the Phi-3 guff file that you got.
• The Phi-3 guff file will also show up in the LM Studio after you’ve moved it.

Recommended: How to Move Large Files to PenDrives?

For LM Studio to detect Phi-3 in its directory, you might need to restart it.

• Click the AI Chat option on the left to load the Phi-3 model.

• Click on a Model to load and select the Phi-3 model.

Recommended: 15 Top Open-Source SSH Clients for Windows PC

• Hold off until it loads. After that, you can begin speaking with Phi-3. To avoid putting too much strain on your CPU, we advise offloading the model to the GPU. To accomplish this, select Advanced Configuration > Hardware Settings from the “System Prompt” menu on the right.

• Under ‘GPU Acceleration’, click on Max and click on Reload model to apply the configuration.

Recommended: Euro Truck Simulator 2 – ProMods 2.71 Map Collection Released

• You can initiate a conversation with Phi-3 as soon as the model loads.

That is all there is to it. Kindly type in a prompt. Microsoft’s Phi-3 model can now be accessed locally on your Windows computer, regardless of whether you have an internet connection.