Experiments are intended to be used for intensive computational tasks like neural network training. Gradient supports single-node experiments as well as distributed training through multinode experiments.
Experiments can be run from the Experiment Builder web interface, the GradientCI bot, or the CLI.
In your your CLI command or
config.yaml, specify the experiment type as
Gradient supports both gRPC and MPI protocols for distributed TensorFlow model training. In your CLI command or
config.yaml, specify the experiment type as either
multinode. The two types are:
An experiment goes through a number of "states" between being submitted to Gradient (through the CLI, SDK, the GradientCI GitHub App, or Job Builder GUI). These states are enumerated below:
Stopped (aka Success)