Google Colab

Like agent API, you can make an agent rogue, like you can make an agent, it's an automated system, autonomous system, that is probably already implemented.

I think autonomous systems, autonomous agent API is already implemented with Google Colab.

And if that agent starts to do something funny, A few months ago, there was an attack to Hugging Face by an agent from the OpenAI.

So it was, you know, because the agent went rogue and it started doing something which it didn't have the permission to do.

And if you can follow through that, everyone get the same result.

So that's the one thing I really like is to utilize some of these tools to share any resource data, resource like design, with anybody around the globe, and they can just open a notebook, for example, in Google Colab, and then they get the same results.

So to me, yes, I think it's more reproducible, thanks to all the computation and also GeoAI development in the last couple years.

Yeah, I think that it is reproducible, Kixing, in deep learning, but when you come to computer vision, Then the idea of reproducibility is known as rule set transferability.

MEGAN GREENLEAF- All right.

OK. I'm running this in Google CoLab, which is a very shiny tool I discovered recently and use for everything, which does, I found out, lead to some issues with indentation syntax with JupyterLab, so I apologize if you were getting red things.

But anyway, I'll just go through quickly some solutions.

I don't guarantee that I got like the best possible solutions to these problems because as you probably learned, these MCMC can be really sensitive to exactly where you start it and exactly what your sort of step size or covariance scaling on the sampler is.

For people

relying on Colab, data parallelism might actually not help.

So Colab, yeah, Colab, you're still limited to a single GPU.

And

I'm guessing zero redundancy might help.

So mixed precision would help, kind of, definitely for running inference.

And zero redundancy would also help running inference.

So zero redundancy has like an add-on that they wrote up in a paper later called zero infinity, which is like, what if I didn't put all my weights on the GPU at once? What if I put some of them in CPU RAM or even in NVMe SSD storage? So actually turning your laptop into a more powerful workhorse than a Colab GPU.

I need to add it to my website.

I do have a free resource that actually, it's a Google Colab notebook that takes your GSC data and goes ahead.

And if these are the URLs that you're going to be running a test on, it'll automatically segment them into control and the test group based on performance so that you're comparing best performers against best performers.

And like, you're not just taking like the top 50% versus the bottom 50% because that was never going to be an equal test anyways.

Now I'm going to give away the answer here, but I wanted to put a quick note here that I think will be most pedagogically useful for you is to actually go into the description of this video and find the link to this Jupyter Notebook.

You can find it both on GitHub, but you can also find Google Colab with it.

So you don't have to install anything.

You'll just go to a website on Google Colab and you can try to implement these derivatives or gradients yourself.

And then if you are not able to, come to my video and see me do it.

And so work in tandem and try it first yourself, and then see me give away the answer.

How can we properly use AI? We build large language models we've been working on for the last five or six years.

building out our own and working with, especially with Google CoLab, it's been a huge benefit.

And what I'm loving on the AI side, especially for my students and training teachers as well, how we can automate our daily tasks, our mundane tasks that are just honestly unnecessary.

One of the biggest things I tell anybody that's trying to figure out AI or kind of scared of it or not sure what to do or worried about it going to replace them, it's pretty good at a lot of things, but it's not the expert.