With thousands of customers served, Clarifai offers a ton of practical know-how about what it takes to deliver production grade AI and we offer platform architecture that significantly speeds time to market. Read on to learn about important platform updates that we announced at Perceive 2020.
Data labeling with Scribe gives you the power to tailor your AI solution to your data and your domain. Build efficient, accurate and targeted models that help your business do things that it could never do before.
With Scribe, machine learning is at the core of your data labeling pipeline. You can label data automatically, or send it to your (or our) human workforce for review. Working with a labeling workforce has never been easier: user rolls, task delegation and quality management tools make it easy to get up and running quickly.
You can even use Mesh (see below) to filter and sort your data before labeling.
Spacetime helps you search through your own image, video and text data as easily as an Internet search. You can rank, sort and filter your data on the same platform that you label it, learn from it, and process it. Spacetime can also power customer-facing solutions like similar product searches or loyalty check-ins.
Save searches to record specific subsets of your data for sharing and model building. Or use localized search to identify and search for people and objects in your data. Video search finds people or objects when and where they occur in your video streams and archival data. Spacetime can even search for custom trained concepts without re-indexing your data.
Train and deploy models easily with Enlight. Train new models quickly by leveraging context, or learn new domains with deep training. Use prebuilt models from Clarifai or upload your own third party models.
Enlight speeds up the training process for quicker iterations. Collaborate as a team to train models and track research across your organization. Track model versions and evaluate performance metrics. Scale up your training infrastructure for parallelized experimentation.
Armada delivers 24/7 AI with no DevOps. All you have to do is train your model and call it when you need it. Armada monitors in-flight requests to each model within the API and automatically assigns models to predict servers according to load demand.
Dynamic loading of models minimizes your infrastructure and DevOps costs as you scale. Armada can add or remove model instances on the fly, based on the number of messages and the capacity for each server instance. Armada delivers optimized runtime to maximize GPU utilization of each model replica.
Armada helps your models scale efficiently for unlimited traffic. Each Armada server can run Clarifai pre-trained models, Clarifai-powered models, or 3rd party models.
Mesh provides a modular architecture for your inference pipeline. You can easily express business logic and gain more insights. Mesh uses your AI models and other fixed-function models as composable building blocks that handle input/output types flexibly. You can leverage AI models along with filters, trackers and alerts. Build complex inference pipelines and even understand data across multiple modalities.
Mesh makes it easy to improve AI performance and prevent data drift with collectors that gather data from model predictions. Mesh helps you reduce time and costs associated with labelling this data. You can auto-annotate data, or provide AI-powered labeling recommendations to your human workforce.
Your data is automatically indexed for search at upload time, making your data easy to work with right away. You can take advantage of highly optimized parallel processing of data and workflow nodes, or even customize a serverless block for your own processing and APIs.
Go where the data is. Flare lets you deploy AI anywhere. Access the information you need with integrated IOT solutions and lightweight devices in the field. You can even build on our cloud platform and then deploy on the edge.
Enable trillions of low-power and battery operated devices and sensors with the power of AI. Common use cases include predictive maintenance to help lower operational costs, quality control systems used to improve efficiencies and quality through the supply chain, improving employee safety from physical elements and hazards, and AI-powered customer experiences like alerts and analytics.