Hello and welcome to another episode of AI Buzz – thank you so much for tuning in! In this post, I want to discuss 3 reasons for when you might want to consider getting a machine learning laptop in 2021. I’ll also recommend what I think is a good machine learning laptop for those users.
The first reason is that machine learning hardware is becoming more and more accessible within a laptop form factor. Years ago, getting even somewhat comparable hardware would destroy your bank account. That is no longer the case where you can fit really beastly GPUs into these smaller laptops. The one caveat here is that if you’re planning to do HARDCORE deep learning, I would advise you to get a desktop workstation where you’ll be able to fit beefier components. You’ll be able to get better cooling, the full size GPU (not the laptop version), all at a better price. For training these large models which take a long time, I also really like the permanence of a desktop where I can just set it to train overnight or for several days straight. That being said, you can still absolutely do deep learning on laptops, especially the more recent ones. The main difference between doing deep learning on a $1200 laptop and a $3000 workstation is going to be the time that it takes to train a model. The laptop could likely train the same exact model assuming it fits within the memory, but it will take longer since it has inferior hardware.
The second reason is that companies are moving to the cloud. Laptops are great for training small to medium-sized models where training time might take a night of computation. However, in the data-centric world that we live in, real-world problems have SO MUCH DATA. Even beastly workstations can take days to crank through data. That’s why cloud training is becoming so appealing. A new approach I’ve been taking has been testing things and developing on my local machine but then pushing heavy computations to the cloud, namely AWS. You can pay for any number of machines to crank through your data. So rather than become an expert at developing local models on a desktop workstation, spend the money on a laptop to develop on but then do the serious training in the cloud. This will allow you to broaden your horizons and learn technologies that are really hot with today’s tech companies.
The third reason is that a machine learning laptop is great for students. Those who want portability to bring it to class, take notes, read ebooks, and do some local development. Once graduated, companies are going to have you start working in the cloud so there may not be a need for building an expesnive $1500-2000 desktop workstation that would get you better specs.
Alright, now onto which laptop you should get for machine learning tasks. The two specs I would look out for are a good GPU and a hearty amount of RAM. Gaming laptops will usually set you up with a good GPU, but the RAM isn’t always guaranteed. However, you can usually upgrade the RAM yourself as long as you plan for that with the specs. A CPU that is on par with the GPU is also usually included. I was scouring through what people generally thought about these machines and was looking at it through a lens that would be advantageous for amchine learning.
A lower budget laptop is the Acer Nitro. This gives you 16Gb of RAM with a 2060 Nvidia chip and an i7 processor. You get a lot for that price point with this one.
Going up to the next level, if you have a higher budget and want better specs the one that I would recommend for people wanting to spend less than $2000 on a machine learning laptop is the CUK GF75 Thin by MSI. This thing is an absolute beast with 64gb of RAM, a 3060 Nvidia chip and a core i7. This packs a ton of compute into a laptop. You’re going to be happily training models with this machine. Even deep learning is gong fly with that chip.
Once again, if you want to be training really heavy deep learning models, you’ll want to build a desktop which can bring you to that next level. That’s about where laptops are going to level out and still have their main advantage which is the portability.
That’s about all I have for you today – let me know in the comments section with any questions or if you want to reach out to me on any other platforms too. Have a good day. BYE
Disclosure: AI Buzz Media is a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for sites to earn advertising fees by advertising and linking to amazon.com