Effortless infrastructure for Machine Learning and Data Science
Welcome to Paperspace’s latest offering, designed to eliminate infrastructure bottlenecks for Deep Learning practitioners. Today, the product consists of two primary components:
- Jupyter notebook integration
- GPU job runner
This offering gives you instant access to a Jupyter notebook instance that is backed by our powerful GPU infrastructure. You no longer need to create a full VM instance to access a notebook and you also get access to powerful features like artifact downloading, versioning, collaboration, and notebook cloning.
The Paperspace Job Runner is designed for users who want to execute code (such as training a deep neural network) on a cluster of GPUs easily without managing infrastructure. The Job Runner enables you to work on your local machine and submit “jobs” to the cloud to be processed. A job consists of:
- collection of files (code, resources, etc)
- docker container (with code dependencies and packages pre-installed)
- command to execute (i.e.
The Job Runner includes a Python module, compatible with Python 2 and 3. Use pip, pipenv, or conda to install the paperspace-python package, e.g.
pip install paperspace
A few things to note
- Gradient° is still evolving quickly and is rough around the edges.
- We have a Help Center section dedicated to Gradient°
- Our support team is here to help with technical issues as much as they can but given that this product and workflow is relatively new we have created this community portal where we encourage you to share your issues/feature requests / etc.
As always, your feedback is invaluable to us – here’s a brief survey where you can submit your ideas, complaints, feature requests etc.
We’re excited to see what you will build!