focus on experiments spell removes the time it takes to set up and manage machine learning infrastructure. use pre-built environments with the most popular ml tools and libraries to deploy in seconds. jupyter notebooks gpu-powered jupyter workspaces get to work on collaborative gpu-powered jupyter workspaces. instantly spin up notebooks up for data preparation, cleaning, transformation, and experimentation. simple command line run code, software, and binaries with the spell cli. simply type ‘spell run’ in front of commands to get started. seamless movement from writing code to deploying models. open frameworks access the framework of your choice instantly. spell supports the best machine learning and deep learning frameworks: tensorflow, pytorch, caffe, fastai, and more. community support connect with a growing community of data scientists and developers. network, learn, and share tutorials, industry insights, and best practices in machine learning. save time on mlops don’t be held back by infrastructure. access the fastest cpu and gpu machine types and frameworks in seconds. accelerate training and optimization easily distribute your code to run projects in parallel. built-in hyperparameter optimization tools, and integrations with tensorboard and weights & biases help improve models up to 10x faster. cloud management and integration deploy easily to your private cloud and quickly start machine learning projects. bring your own aws or gcp credits, keep data in your s3 or google cloud buckets, and deploy models within your private cloud infrastructure. model serving and management deploy models in one-click on industrial-grade, auto-scaling, kubernetes-based infrastructure. easily manage the model life cycle, model versions, and performance results of inference. focus on business value spell allows teams and business units to focus on getting to results faster. governance and security spell integrates with single sign-on systems, internal data stores, and governance systems. spell provides secure user management. real-time support white glove service that features and one-on-one premium support. a private slack channel for real-time access to support from spell. priority requests we are constantly iterating on spell and we prioritize enterprise client needs. all product feature requests are sent directly to our product team.
There is no how it works explanation for this product
Making developer tools is a huge passion of mine. When I first saw Spell, I was blown away by how simple and elegant it was. Having wrestled with provisioning and configuring ML environments a lot, I immediately understood the power of Spell. It really feels like magic. I'm thrilled to be partnering with Spell and can't wait to see what our users create. CHRIS VAN PELT, CO-FOUNDER AT WEIGHTS & BIASES
One of the big advantages of Spell was our team could get up and running right away without slowing down ongoing projects. Spell plugged in seamlessly to our existing infrastructure and tools and we could migrate workloads as we saw more and more value. CHRIS ACKERSON, DIRECTOR OF PRODUCT AT ALPHASENSE
At this point, I've trained 100's of models on Spell. It's amazing to never have to think about infrastructure - I feel like I can go directly from idea to training a model. There's a lot of talk about the democratization of machine learning, but in practice it's quite difficult to set up the necessary building blocks to train much more than a toy model. Spell seems like a real step forward in making machine learning approachable for more people. DANNY MEYER, ML ENGINEER AT ORIGINATE
I had the opportunity to work with Spell on a challenge for Omdena to preprocessing our datasets. This was a great experience because on my laptop this process took many hours while on Spell it was much faster and my laptop didn't freeze. Additionally, Spell had logs where I can monitor my process, kill the run if it is necessary, and I can see how long the run is so I can monitor performance in my code. ANA LOPEZ, ML ENGINEER AT OMDENA
Spell is an infrastructure platform for AI and Deep Learning. We provide simple tools that are flexible enough to meet any workflow and robust enough to scale with you as you grow. Our pre-built environments ensure your engineers can start coding immediately using the tools and frameworks they’re already familiar with.View Company