Gradient Community Notebooks let you train ML models on FREE Cloud GPUs ⚡
Discussion
Would you recommend this product?
2 Reviews5.0/5
Super happy to introduce our newest product — Gradient Community Notebooks. There is still a ton to build but we think that making it possible for anyone to run complex ML and AI application on free cloud GPUs is a game changer. Let us know what you think!
Love the project showcase 🔥
Such an amazing concept. I know it's an early version, but design is clean and straightforward. Keep up the good work!
I've been using Paperspace's Gradient Notebooks (the private kind) for a while now. These new Community Notebooks will take things to the next level. Let me say why I believe this is better than existing options (Google Colab and Kaggle Kernels to name a few): - Persistent environment. Currently you need to know how to build dockerfiles and images in order to make a custom container for your notebooks. However, if you take a base images and simply add a new package in it via conda or pip it doesn't go away! Colab and Kaggle are essentially useless if you are using anything other than vanilla packages as you have to install them manually *every time you open the notebook*. - Native Notbook/Lab interface. JupyterLab is significantly better than Colab or Kernels for various reasons. Gradient Notebooks gives you a native coding environment where I don't need to relearn keyboard shortcuts or figure out how markdown will behave. - Persistent data storage. These Community Notebooks come with 5GB of persistent storage. This storage is more performant than Google Drive mounting in Colab (which is painfully slow). Plus I can run git commands in an actual terminal on the free machine to sync my data up instead of relying on Jupyter magics to hack it into working. In short, this is pretty exciting.