After many weeks of developing a web-based platform, that we hope will change the way we’ll share and explore the Machine Learning field, we’re excited to celebrate our Private Beta Pre-Release. Read on to learn what that means to you.
In the past few months we’ve been working on implementing features that we think are crucial to provide a platform for the web on which everyone and can share, explore and execute Machine Learning code. This includes things like configuring labs, sharing and embedding them, or even having multiple and parallel executions.
Let’s take a closer look.
Sharing is one of the most important things you can do with MachineLabs. Instead of linking to Git repositories that people have to clone and install to try them out, in MachineLabs, you can simply hit the share button and choose your favourite social platform to demonstrate your latest experiments. Of course, as with every website, you can also just go ahead and copy the link to your lab from the browser address bar.
Sharing your experiments has never been easier!
Yes that’s right! MachineLabs comes with an optimized UI so you can embed all of your labs in other blog articles and websites. So instead of showing just code snippets when writing about your latest learnings in AI, your readers can play with your labs from right within your blog.
Embedding a lab is as easy as hitting the embed button and copy pasting the generated HTML code into your blog or website.
Below is an embedded lab in action that demonstrates a simple implementation of a neural net that learns to recognize hand-written digits using the MNIST dataset.
It’s very natural in the Machine Learning field to try out different things in order to get the best results possible. Sometimes you want to train your neural net with more layers, different activation functions or different optimizers. You’re experimenting after all!
MachineLabs is built with experimentation in mind. Every time you execute your lab, it’ll show up in the execution list, resulting naturally in having multiple executions per lab. Of course, you don’t have to wait until your execution is done until you can try the next experiment. Running multiple executions at the same time comes as batteries included.
Check out our MNIST Lab, which has different executions for different configurations. You can compare them to see which setup performs best!
As mentioned earlier, today we’re celebrating our Private Beta Pre-Release. Now you might wonder, why a Pre-Release? And what does that even mean? Private Beta Pre-Release means that MachineLabs is up online and ready to be used as in, everyone can view labs with existing executions. However, to run a lab yourself you need to get a private beta invite. We haven’t let anyone in yet as we’re fleshing out the last bits and pieces that are needed - this is really just fine-tuning from now on.
If you’re seeing MachineLabs for the first time today or simply haven’t applied for our private beta yet, there’s still a chance to get an invite! Head over to get.machinelabs.ai and let us know why you want to use MachineLabs and why you’d be a good candidate.