Paperspace Core Reviews

Fast onboarding, very stable and reliable, reasonable prices

Monthly subscription fee on top of machine costs, multi-GPU training can be pricy

DL infrastructure as a service

I like the way all the regular headaches taken care of, such as the management of datasets. The GPUs offered are quite powerful and reasonably priced.

The integrated jupyter interface is not perfect and does not have some keyboard shortcuts. They do offer the native Jupyter Notebook interface though.

My problem of accessing more powerful machines than my own laptop, and having my models and code safe online.

I'm a data science student and first-steps researcher. I ran into paperspace because I needed decent computing power for proper pricing. I tried all cloud computing solutions, but they were expensive for small to mid-size solutions. Some of the cheapest/free ones have very restricted limitations, and I couldn't implement any DL advanced model or architecture because I ran out of RAM very quickly. With paperspace you don't have to worry about that. It offers very fair pricing for what you get in exchange, you can also deploy models, and you don't depend on building any formal API on any other hosting service. It's an all-at-once for researchers or professionals. Its gradient UI is very intuitive and informative, for it displays metrics about energy and memory consumption of your hardware resources to efficiently track your model's performance or any application you are running in this environment. Its pro version offers a great variety of GPUs, and you can have access to powerful ones for only $8 a month.

I highly recommend paperspace as a default machine learning platform for anyone wanting to deep dive into deep learning or use machine learning professionally.

Some bugs on the Chrome IU running some applications like keras for example. I also didn't like that you don't have at least two GPU's you can choose form on the free tier but I find the available one super powerful and fair for a free demo.

Chrome BUGS, I had to stop my runtime in the middle of it because bugs didn't let me see the performance of a model on the progress bar from keras. I wasted 6 hours because of that bug.

I like their GPU-based severs, making it very easy to deploy a GPU-based server in minutes.

They can enable Linux servers even cheaper than Windows servers Since Linux servers don't have license fees which windows have.

I use Paperspace to get immediate access to high GPU servers to run high consuming GPU projects.

Paperspace makes you feel that you are not just using a virtual machine, but you are using a physical device. The responsiveness and speed you get are off the roof. It provides lots of choices for users depending on their needs and requirements.

Paperspace API is also helpful when integrating it into your projects. It is well documented and readily available for developers to read and implement.

Paperspace support is amiable and fast to reply and take action.

What I dislike about paperspace core are the limited regions they offer and the load time of each machine.

Recommendation: In the future, add more regions and make loading or creation times of machines faster.

I recommend utilizing the API provided by paperspace.

Creating machines and deploying machines on our website. It helps a lot because machines are readily available and accessible.

Paperspace provides an excellent virtual windows machine that works no matter where you are and even on Macs and Chromebooks.

The resolution can be buggy as it tries to fit your browser width and has to refresh should anything be adjusted.

Paperspace provides affordable access to a Windows machine without you having to purchase a whole new computer. You can access it over the internet from any tablet, Mac or Chromebook..

Allows me to access PC applications and Office files even when I'm on a Chromebook or Mac.

Available plans, network speed, pricing, customer support

Lack of a southern US point of presence (e.g. Dallas)

The ability to work from absolutely anywhere I can access an Internet-connected computer

Free virtual machines with GPU access allow casual users and professionals to perform computationally-intensive program development and machine learning training and testing.

There are times when all of the free virtual machines are unavailable. A streamlined process for integrating Conda environments with the virtualized Jupyterlab would increase quality of life for users.

Paperspace has been a lifesaver for my work. I quite frequently have to train and test multiple machine learning algorithms, and if my company's supercomputer goes down, I know I can count on the easy-to-use and reliable Paperspace virtualization service to keep my work on track.

What I like about Paperspace is it provides Virtual Machines fast and cheap. Its UI is simple and beautiful. It offers a lot of machines from Standard GPU, CPU only, and with CUDA.

What I dislike about Paperspace is that often some machine type is not available for one region, and while it is for the other region, it's fast to run out of a machine.

I would advise others to try using Paperspace if you are planning to deploy VM in your business.

Creating VM's, be it Windows or Linux, server or desktop. We are using paper space to create a machine and deploy it on our website.