Neu.ro puts data manipulation, model management, and training at your fingertips. The platform allows you to focus on model development tasks at hand by managing critical aspects of infrastructure and system integration, including resource allocation and storage, image management and sharing, secure web and terminal access to running jobs.
The key components of Neu.ro are:
- Environment. A Docker image that can be launched on the platform. For your convenience, we provide a base image that is built on the
deepoDocker image and contains the most popular ML tools (Tensorflow, Keras, PyTorch, TensorBoard, Jupyter Notebooks, and JupyterLab).
Environments, storage volumes, and jobs can be published and shared among users.
Each time you start a job, Neu.ro will:
- Wait for requested resources to become available.
- Pull an environment Docker image and launch it.
- Attach storage volumes to the container's local mount points.
All your interactions with the platform are happening either through the command-line interface (CLI) tool or through the web interface.