Deploy with ease
OctoML uses the latest optimization techniques to shrink model size, reduce latency, and maintain accuracy, making it easier and faster to deploy cutting edge models to production.
Compatible across deep learning frameworks
Create your model in any framework — including TensorFlow, PyTorch, Keras, MXNet, CoreML and ONNX — and switch between frameworks to maximize your productivity.
Deploy on the cloud, hardware, or edge
Run your model across diverse hardware targets from server-class GPUs and CPUs to specialized accelerators (FPGAs, ASICs), mobile phones, IoT and edge devices.
We simplify the hardest parts of ML deployment