Deep Learning as a Service

IBM’s experiment-centric deep learning service within Watson Studio allows data scientists to visually design their neural networks and scale out their training runs while auto-allocation means paying only for the resources utilized. Optimized for production environments, scale up your training using the NVIDIA® Tesla® V100 GPU with your preferred deep learning framework then easily deploy to the cloud or at the edge. -Experiment assistant Initiate and monitor batch training experiments then compare cross-model performance in real-time without worrying about log transfers and scripts to visualize results. -Open and flexible Use your preferred deep learning framework: Tensorflow, Keras, PyTorch, Caffe and more. -Elastic GPU compute Train neural networks in parallel using market-leading NVIDIA® Tesla® GPUs - K80, P100, and V100. -Hyperparameter optimization