Results for
These got released last week and the process for using them on your local machine with MATLAB is very similar to how you use the local deepseek models as I demonstrated in my February blog post How to run local DeepSeek models and use them with MATLAB » The MATLAB Blog - MATLAB & Simulink
You need Ollama and the LLMs with MATLAB package installed (Details on how to do this in the blog post above). Then you run the following in your operating systems' command line
ollama pull gpt-oss:20b
Over to MATLAB and set up a chat session
>> chat = ollamaChat("gpt-oss:20b")
chat =
ollamaChat with properties:
ModelName: "gpt-oss:20b"
Endpoint: "127.0.0.1:11434"
TopK: Inf
MinP: 0
TailFreeSamplingZ: 1
Temperature: 1
TopP: 1
StopSequences: [0×0 string]
TimeOut: 120
SystemPrompt: []
ResponseFormat: "text"
FunctionNames: []
txt = generate(chat,"Who are you?")
txt =
"I’m ChatGPT – a conversational AI developed by OpenAI. My core is the GPT‑4 language model, which has been trained on a massive mix of text from books, websites, articles and other sources to understand and generate human‑like language. I don’t have feelings, consciousness, or a personal identity; I’m a tool that can help answer questions, brainstorm ideas, explain concepts, draft text, and more. My goal is to understand the context you give me and respond in a helpful, accurate and safe way. If there’s something specific you’d like to know or do, just let me know!"
This is the smaller of the two, new open models and it is bringing my aging desktop to its knees. My GPU is too small to do the work so I think everything is happening on the CPU and its slooooow. Will try on my Mac next
Let me know if you try this out!
Long before I joined MathWorks, I was a member of the academic Research Software Engineering (RSE) community where part of my mission was to introduce basic software engineering concepts to the research community. Things like version control, testing and even simply writing code instead of using only pointy-clicky GUIs before copying and pasting the results plot into a word document. I've seen things..........*shudders*
The RSE movement is still going very strong and I am elated that MathWorks is increasingly interacting with it. One example of such interaction is a video tutorial contributed by my colleauge @Mihaela Jarema to a comminity seminar series called 'A summer of Testing' It's linked to below
The video assumes you've never run a test before and gently guides you through the principles. Along the way you'll learn about some of MATLAB's superb testing capabilities. Things like
- Unit testing Framework
- Test Browser App
- Code Coverage
- Test Fixtures (Setup and teardown)
- Test driven devellopment
- Function argument validation
- CI/CD using GitHub actions
Go check out out.
I found some beautiful computational art made by a developer called @yuruyurau who used a language called Processing. Unfortunately, I know very little about this language so I asked Claude to convert it to MATLAB for me.
Give it a try yourself and show me what you come up with.
Details here: Pair programming with Claude to produce computational art in MATLAB » The MATLAB Blog - MATLAB & Simulink
In a discussion on LInkedin about my recent blog post, Do these 3 things to increase the reach of your open source MATLAB toolbox, I was asked by "Could you elaborate on why someone might consider opening/sharing their code? Thinking of early-career researchers, what might be in it for them?"
I'll give my answer here but I'm more interested in yours. How would you have answered this?
This is what I said:
- It's the right thing to do scientifically. A computational paper is essentially just an advertisement of what you've done. The code contains vital details about how you actually did it. A computational paper is incomplete without the code.
- If you only describe your algorithm in a paper, I have to implement it before I can apply your research to my problem. If you share the code, I can get started much more quickly using your research. This means I publish faster and since I am a good scientist, this means you get cited faster.
- Other scientists start off as users of your code. This leads to citations. Over time, some of them start deeply using and modifying your code, this leads to collaborators.
- Once you decide to share code via something like GitHub, you quickly start adopting good software engineering practices without initially realizing it. This improves the quality of your research since adopting good software practices makes it more likely that your software will give the right answers.
That last point can be a little hard to get your head around sometimes. Even if all you do is use file upload to get your stuff onto GitHub (i.e. you're not using git properly yet) you will start to naturally converge towards better code.
Why? Because as soon as you share code, you have to solve the problem of getting it to run on someone else's machine.
A trivial example concerns hard coded paths, for example. If you only ever run it on your machine then having a line like datafile = "C:\Mystuff\data.csv" always works but it breaks as soon as I try to run it on my machine. You'll look at this and think "Maybe there's a better way to do that".
Similarly dependencies. Your Path may be full of stuff that isn't present on my machine. As soon as I try to run your code, it won't work and you'll have to figure out how to handle dependencies in a reproducible way.
Documentation! An empty README.md is no good if you expect me to know how to use your code. You at least have to say something like "To run this, type runme(N) into MATLAB where N is the size of the model...etc etc)
The act of sharing, and dealing with the consequences, leads to much better code than if you keep it to yourself.
In case you missed it in my overview of the MATLAB R2025a release, Markdown support has been greatly improved. This picture says it all
During the past twelve months, PIVlab, a MATLAB Community Toolbox for particle interference velocimetry (a technique for fluid flow measurement) set a new record for all-time File Exchange downloads, surpassing one hundred thousand, dating back to 2010. It also recently eclipsed the 1000 downloads/month mark on File Exchange.
Congratulations to @William Thielicke and his team for this fantastic long term achievement and head over to the File Exchange to download and use it: PIVlab - particle image velocimetry (PIV) tool with GUI - File Exchange - MATLAB Central
Me: If you have parallel code and you apply this trick that only requires changing one line then it might go faster.
Reddit user: I did and it made my code 3x faster
Not bad for just one line of code!
Which makes me wonder. Could it make your MATLAB program go faster too? If you have some MATLAB code that makes use of parallel constructs like parfor or parfeval then start up your parallel pool like this
parpool("Threads")
before running your program.
The worst that will happen is you get an error message and you'll send us a bug report....or maybe it doesn't speed up much at all....
....or maybe you'll be like the Reddit user and get 3x speed-up for 10 seconds work. It must be worth a try...after all, you're using parallel computing to make your code faster right? May as well go all the way.
In an artificial benchmark I tried, I got 10x speedup! More details in my recent blog post: Parallel computing in MATLAB: Have you tried ThreadPools yet? » The MATLAB Blog - MATLAB & Simulink
Give it a try and let me know how you get on.
There has been a lot of discussion here about the R2025a Prerelease that has really helped us get it ready for the prime time. Thank you for that!
A new update of the Prerelease has just dropped. So fresh it is still warm from the oven! In my latest blog post I discuss changes in the way MathWorks has been asking-for and processing feedback...and you have all been a part of that.
If you haven't tried the Prerelease in a while, I suggest you update and see how things are looking now.
If you have already submitted a bug report and it hasn't been fixed in this update, you don't need to submit another one. Everything is being tracked!
Have a play, discuss it here and thanks for again for being part of the process.
Imagine you are developing a new toolbox for MATLAB. You have a folder full of a few .m files defining a bunch of functions and you are thinking 'This would be useful for others, I'm going to make it available to the world'
What process would you go through? What's the first thing you'd do?
I have my own opinions but don't want to pollute the start of the conversation :)
Learn the basic of quantum computing, how to simulate quantum circuits on MATLAB and how to run them on real quantum computers using Amazon Braket. There will also be a demonstration of machine learning using quantum computers!
Details at MATLAB-AMAZON Braket Hands-on Quantum Machine Learning Workshop - MATLAB & Simulink. This will be led by MathWorker Hossein Jooya.
I kicked off my own exploration of Quantum Computing in MATLAB a year or so ago and wrote it up on The MATLAB Blog: Quantum computing in MATLAB R2023b: On the desktop and in the cloud » The MATLAB Blog - MATLAB & Simulink. This made use of the MATLAB Support Package for Quantum Computing - File Exchange - MATLAB Central
tiledlayout(4,1);
% Plot "L" (y = 1/(x+1), for x > -1)
x = linspace(-0.9, 2, 100); % Avoid x = -1 (undefined)
y =1 ./ (x+1) ;
nexttile;
plot(x, y, 'r', 'LineWidth', 2);
xlim([-10,10])
% Plot "O" (x^2 + y^2 = 9)
theta = linspace(0, 2*pi, 100);
x = 3 * cos(theta);
y = 3 * sin(theta);
nexttile;
plot(x, y, 'r', 'LineWidth', 2);
axis equal;
% Plot "V" (y = -2|x|)
x = linspace(-1, 1, 100);
y = 2 * abs(x);
nexttile;
plot(x, y, 'r', 'LineWidth', 2);
axis equal;
% Plot "E" (x = -3 |sin(y)|)
y = linspace(-pi, pi, 100);
x = -3 * abs(sin(y));
nexttile;
plot(x, y, 'r', 'LineWidth', 2);
axis equal;
On 27th February María Elena Gavilán Alfonso and I will be giving an online seminar that has been a while in the making. We'll be covering MATLAB with Jupyter, Visual Studio Code, Python, Git and GitHub, how to make your MATLAB projects available to the world (no installation required!) and much much more.
Sign up (it's free!) at MATLAB Without Borders: Connecting your Projects with Python and other Open-Source Tools - MATLAB & Simulink
For Valentine's day this year I tried to do something a little more than just the usual 'Here's some MATLAB code that draws a picture of a heart' and focus on how to share MATLAB code. TL;DR, here's my advice
- Put the code on GitHub. (Allows people to access and collaborate on your code)
- Set up 'Open in MATLAB Online' in your GitHub repo (Allows people to easily run it)
I used code by @Zhaoxu Liu / slandarer and others to demonstrate. I think that those two steps are the most impactful in that they get you from zero to one but If I were to offer some more advice for research code it would be
3. Connect the GitHub repo to File Exchange (Allows MATLAB users to easily find it in-product).
4. Get a Digitial Object Identifier (DOI) using something like Zenodo. (Allows people to more easily cite your code)
There is still a lot more you can do of course but if everyone did this for any MATLAB code relating to a research paper, we'd be in a better place I think.
Here's the article: On love and research software: Sharing code with your Valentine » The MATLAB Blog - MATLAB & Simulink
What do you think?
So you've downloaded the R2025a pre-release, tried Dark mode and are wondering what else is new. A lot! A lot is new!
One thing I am particularly happy about is the fact that Apple Accelerate is now the default BLAS on Apple Silicon machines. Check it out by doing
>> version -blas
ans =
'Apple Accelerate BLAS (ILP64)'
If you compare this to R2024b that is using OpenBLAS you'll see some dramatic speed-ups in some areas. For example, I saw up to 3.7x speed-up for matrix-matrix multiplication on my M2 Mabook Pro and 2x faster LU factorisation.
Details regarding my experiments are in this blog post Life in the fast lane: Making MATLAB even faster on Apple Silicon with Apple Accelerate » The MATLAB Blog - MATLAB & Simulink . Back then you had to to some trickery to switch to Apple Accelerate, now its the default.
Hi everyone
The R2025a pre-release is now available to licensed users. I highly encourage you to download, give it a try and give us some feedback.
The first thing I tried was switching to Dark mode. Here's the magic
>> s = settings;
>> s.matlab.appearance.MATLABTheme.PersonalValue = "Dark";
The MATLAB Online Training Suite has been updated in the areas of Deep Learning and traditional Machine Learning! These are great self-paced courses that can get you from zero to hero pretty quickly.
Deep Learning Onramp (Free to everyone!) has been updated to use the dlnetwork workflow, such as the trainnet function, which became the preferred method for creating and training deep networks in R2024a.
- Content streamlined to reduce the focus on data processing and feature extraction, and emphasize the machine learning workflow.
- Course example simplified by using a sample of the original data.
- Classification Learner used in the course where appropriate.
The rest of the updates are for subscribers to the full Online Training Suite
The Deep Learning Techniques in MATLAB for Image Applications learning path teaches skills to tackle a variety of image applications. It is made up of the following four short courses:
- Explore Convolutional Neural Networks
- Tune Deep Learning Training Options
- Regression with Deep Learning
- Object Detection with Deep Learning
Two more deep learning short courses are also available:
The Machine Learning Techniques in MATLAB learning path helps learners build their traditional machine learning skill set:
Many of my best friends at MathWorks speak Spanish as their first language and we have a large community of Spanish-speaking users. You can see good evidence of this by checking out our relatively new Spanish YouTube channel which recently crossed the 10,000 subscriber mark
I've always used MATLAB with other languages. In the early days, C and C++ via mex files were the most common ways I spliced two languages together. Other than that I've also used MATLAB with Java, Excel and even Fortran.
In more recent years, Python is the language I tend to use most alongside MATLAB and support for this combination is steadily improving. In my latest blog post, I show how easy it has become to use Python's Numpy with MATLAB.
Have you used this functionality much? If so, what for? How well did it work for you?
Next week is MATLAB EXPO week and it will be the first one that I'm presenting at! I'll be giving two presentations, both of which are related to the intersection of MATLAB and open source software.
- Open Source Software and MATLAB: Principles, Practices, and Python Along with MathWorks' Heather Gorr. We we discuss three different types of open source software with repsect to their relationship to MATLAB
- The CLASSIX Story: Developing the Same Algorithm in MATLAB and Python Simultaneously A collaboration with Prof. Stefan Guettel from University of Manchester. Developing his clustering algorithm, CLASSIX, in both Python and MATLAB simulatenously helped provide insights that made the final code better than if just one language was used.
There are a ton of other great talks too. Come join us! (It's free!) MATLAB EXPO 2024
