NvidiaGPU

class verta.endpoint.resources.NvidiaGPU(number, model=None)

Nvidia GPU resources.

New in version 0.24.1.

The JSON equivalent for this is:

{
    "resources": {"nvidia_gpu": {"number": 1, "model": "V100"}}
}

Parameters:

• number (int > 0) – Number of GPUs requested to run this endpoint’s model.

• model (NvidiaGPUModel, optional) – Model of GPU requested to run this endpoint’s model. Currently available models are specified in the NvidiaGPUModel enum.

Examples

from verta.endpoint.resources import Resources, NvidiaGPU, NvidiaGPUModel
resources = Resources(nvidia_gpu=NvidiaGPU(1, NvidiaGPUModel.V100)