Skip to main content

Running LLMs locally using LM Studio

As I like to experiment a lot with AI, I always have to be careful and keep my token usage under control. And although the token cost has decreased over time for most models, the expenses can go up quite fast.

That is one of the reasons I like to use  (Large) Language Models locally.  There are multiple ways to run a model locally but my preferred way so far was Ollama (together with OpenWebUI). I also experimented with Podman AI Lab but I always returned to Ollama in the end. 

Recently a colleague introduced me to LM Studio, another tool to run and test LLM’s locally. With LM Studio, you can:

  • Run LLMs offline on your local machine
  • Download and run models from Hugging Face
  • Integrate your own application with a local model using the LM Studio SDK or through the OpenAI endpoints
  • Use the built-in RAG support to chat with your local documents

More than enough reasons to give it a try…

Getting started

  • I downloaded the installer from the website and followed the installation wizard. Once the installation has completed click on Finish:

 

  • If you selected the Run LM Studio checkbox, LM Studio will be started immediately and you will be welcomed by an onboarding wizard:

 

  • Click on the Get your first LLM button. The next onboarding step appears where it suggests me to download a first model(in this case DeepSeek R1).
    • Notice the Enable local LLM service on login option. By checking this option, you can use the LM Studio local LLM server without having the LM Studio application open.

 

  • Click on Download to download the suggested model.

 

  • Once the download has completed, a Start New Chat button appears. Click on it to continue.

  • We finally get the LM Studio UI in front of us. Let us first load the downloaded model by clicking on the Load Model button in the notification popup.

  • Once the model is loaded, we can start asking questions by typing a message in the chat window and hitting Send.
 

Exploring some features

Now that we have the basics runnng, let’s explore some of the nice features that LM Studio has to offer.

If you click on the settings icon in the top right corner, you get a lot of options to easily tweak the used model. 

For example you can easily enable structured output(if the model supports it):

 

Another feature I like is that you can switch between ‘User’ and ‘Assistant’ mode in the chat. This is useful for few-shot prompting and other scenario’s:

I already mentioned the RAG integration:

Don’t forget to check out the Developer tab as well where you can run the LLM server and expose OpenAI compatible endpoints to your applications:

 

It’s too soon to tell if I will use this instead of Ollama but it is certainly worth to give it a try.

More information

Tokenizer - OpenAI API

Run LLMs locally using Podman AI Lab

Running large language models locally using Ollama

Explore and test local modals using Ollama and OpenWebUI

LM Studio - Discover, download, and run local LLMs

Labels:

Popular posts from this blog

Podman– Command execution failed with exit code 125

After updating WSL on one of the developer machines, Podman failed to work. When we took a look through Podman Desktop, we noticed that Podman had stopped running and returned the following error message: Error: Command execution failed with exit code 125 Here are the steps we tried to fix the issue: We started by running podman info to get some extra details on what could be wrong: >podman info OS: windows/amd64 provider: wsl version: 5.3.1 Cannot connect to Podman. Please verify your connection to the Linux system using `podman system connection list`, or try `podman machine init` and `podman machine start` to manage a new Linux VM Error: unable to connect to Podman socket: failed to connect: dial tcp 127.0.0.1:2655: connectex: No connection could be made because the target machine actively refused it. That makes sense as the podman VM was not running. Let’s check the VM: >podman machine list NAME         ...
Read more

Cache stampede: when our cache turned against us

While investigating some performance issues, we ran into an ASP.NET Core API that cached a fairly expensive aggregation query for 60 seconds. Under normal load, that was fine: one request rebuilds the cache, everyone else reads from it. Under peak load, dozens of requests would arrive in that same expiry window, all see a cache miss, and all fire the same expensive query in parallel. The database didn't like that. That was the moment when our caching layer stopped helping and started hurting. A burst of requests comes in at the same time, all miss the cache, and all go hammer the database or the downstream API at once. That's a cache stampede . The cache was supposed to protect our backend, and for a few hundred milliseconds it did the opposite. Why this happens IMemoryCache.GetOrCreate (and its async sibling) looks like it protects you, but it doesn't add any locking on its own. Look at the naive version: public async Task<Report> GetReportAsync(string key) ...
Read more

A complex system designed from scratch never works

A few years ago, I worked as an architect on a big mainframe rewrite. I still count it as one of my failures. Not because the technology was wrong, but because I couldn't convince the management team to simplify the approach. Years later, the organization is still struggling to get the new system up and running. I left the project at the time, because I couldn't put my name behind an approach that would take very long and cost a lot of money without a working system to show for it along the way. Gall’s Law That memory keeps coming back to me, because it's a textbook case of Gall's Law playing out in real life. Gall's Law , from John Gall's Systemantics , states it plainly: A complex system that works is invariably found to have evolved from a simple system that worked. A complex system designed from scratch never works, and it cannot be patched to make it work. You have to start over with a simple system that works. What does that mean in practice,...
Read more