PeriFlow is a reliable, speedy, and efficient service for training and serving your own large-scale AI models on any data of your choice. It removes the overhead of ML operations and (cloud or on-premise) resource management. With just one click, PeriFlow runs massive-scale AI workloads with our highly-optimized training and serving systems. PeriFlow can run on our self-managed cloud infrastructure or on the infrastructure that you own.

With datasets and models becoming ever larger, the time of large-scale AI has arrived.

Size comes with its own cost. Now every company wants to use large-scale models, but utilizing them is no simple task. The technical difficulty of developing and managing large-scale AI is hard to take for many companies. And the cost of using large-scale ML is often beyond one’s practice.

This is where PeriFlow comes to the rescue.

PeriFlow provides an end-to-end machine learning pipeline for large-scale AI, from developing models to serving inference requests. It provides an integrated interface for utilizing diverse cloud or on-premise resources and native support for large-scale distributed execution as well. It takes charge of both the training and serving process, all the while automatically resolving faults and performance problems, thereby removing the overhead for any manual monitoring.

Experience seamless workflow with PeriFlow.

You can conduct the whole process of AI development from training to serving on PeriFlow. Upload the data, and train your original model with our optimized training system. When the model is ready to be deployed, you can also serve your model with our efficient serving system. PeriFlow automatically handles faults and problems that arise during the training/serving process. You can also choose to train/serve your model either on our managed cloud system or on your resources.

PeriFlowFriendliAI

PeriFlow

FriendliAI

PeriFlow

FriendliAI

PeriFlow
FriendliAI