It has always been a concern for many people who want to step into the field of machine learning or molecular modelling, that they worry about their laptop is not good enough interms of CPU or GPU capibilities. The good news is in most cases, especially for beginners, you don't really need perfect local hardware since there are numberous online choices avaliable there, one of them is the padlepadle platform, developed by Chinese tech giant Baidu.
It is very similar to google's Colab, but much more generous to provide you more GPU power, as good as 8 hours per day to use Tesla V100, for free, while Google only gives you P100 or other GPUs much weaker than Tesla V100. Basically, what you need to do is just log in and start running any GPU related project, you then get the free 8 hours every day. To use it, you can either create a new project in the ipynb format to run your job interactively which is very similar to Colab again, or you can create a script project to submit your job in a batch mode, and run the job after having closed the browser, for free, as long as you could handle the coding process. If you want to get the same thing from Google, you have to pay around 50 US dollars to avoid your job not being killed after you close your browser. Ultimately, PadelPadel provide you with a Unix-like system on the cloud, so you could use it to do any job that would require a local Unix-like terminal would allow you to do, except the visualization function is missing. For me, I mostly use it for molecular modelling related tasks, such as installing some molecular dynamics packages like Gromacs Gromacs compile on AI Studio PadelPadel and Gromacs tutorial on AI Studio PadelPadel
Have a look at my AI Studio PadelPadel space if you are interested Chinese language warning if you can't understand it! You can always switch the web language to English to minimise the language barrier
