Saturn Cloud offers a seamless integration with Jupyter notebooks, allowing users to easily run, track, and debug RL models. Another valuable feature is the ability to scale up and down resources like GPUs and CPUs, providing flexibility and cost-effectiveness.
One of the users shared with us: "The availability of pre-configured environments saves time and hassle, particularly when working with complex setups involving packages like CUDA and PyTorch".
Saturn Cloud supports GPUs, which are essential for many machine learning tasks, and allows users to edit the environment before starting the cloud resources. Additionally, Saturn Cloud provides lightning-fast CPUs and offers lots of free computing resources upfront, allowing for easy prototyping and fine-tuning of models. This can give users a head start in deployment and iterate quickly, regardless of whether they are in academia or industry.
- "They provide a centralized space for data, code, and results."
- "There is plenty of computational resources (both GPU, CPU and disk space)."
- "It offered an excellent development environment while not touching our production cloud resources."
Saturn Cloud could improve the user experience by making the process of setting up custom environments more beginner-friendly. They should provide more detailed and beginner-friendly documentation, especially for advanced features.
It would be beneficial to include prebuilt images for advanced data science packages, such as LightGBM and a Kaggle image. "The option to choose specific data science subpackages in the environment would be useful." Usage reporting should be more precise and quantify use in minutes instead of just hours.
It would be helpful if Saturn Cloud offered a wider range of libraries for installation, including support for C++. Providing a way to easily and freely host API/web apps would be advantageous, as it is a common need for data scientists to deploy web apps for their machine learning models.
Finally, the pricing should be more transparent.
- "My main suggestion for improvement centers on pricing. Introducing a tier modelled after AWS spot instances would be a game-changer."
- "It would be nice to have more hardware category options, like TPU coprocessors or ARM64 CPUs."
- "Public Clouds integration and sandbox environments would be a true game changer."
