BatchContainerOverrides¶
-
class
aws_cdk.aws_stepfunctions_tasks.
BatchContainerOverrides
(*, command=None, environment=None, gpu_count=None, instance_type=None, memory=None, vcpus=None)¶ Bases:
object
The overrides that should be sent to a container.
- Parameters
command (
Optional
[List
[str
]]) – The command to send to the container that overrides the default command from the Docker image or the job definition. Default: - No command overridesenvironment (
Optional
[Mapping
[str
,str
]]) – The environment variables to send to the container. You can add new environment variables, which are added to the container at launch, or you can override the existing environment variables from the Docker image or the job definition. Default: - No environment overridesgpu_count (
Union
[int
,float
,None
]) – The number of physical GPUs to reserve for the container. The number of GPUs reserved for all containers in a job should not exceed the number of available GPUs on the compute resource that the job is launched on. Default: - No GPU reservationinstance_type (
Optional
[InstanceType
]) – The instance type to use for a multi-node parallel job. This parameter is not valid for single-node container jobs. Default: - No instance type overridesmemory (
Optional
[Size
]) – Memory reserved for the job. Default: - No memory overrides. The memory supplied in the job definition will be used.vcpus (
Union
[int
,float
,None
]) – The number of vCPUs to reserve for the container. This value overrides the value set in the job definition. Default: - No vCPUs overrides
Attributes
-
command
¶ The command to send to the container that overrides the default command from the Docker image or the job definition.
- Default
No command overrides
- Return type
Optional
[List
[str
]]
-
environment
¶ The environment variables to send to the container.
You can add new environment variables, which are added to the container at launch, or you can override the existing environment variables from the Docker image or the job definition.
- Default
No environment overrides
- Return type
Optional
[Mapping
[str
,str
]]
-
gpu_count
¶ The number of physical GPUs to reserve for the container.
The number of GPUs reserved for all containers in a job should not exceed the number of available GPUs on the compute resource that the job is launched on.
- Default
No GPU reservation
- Return type
Union
[int
,float
,None
]
-
instance_type
¶ The instance type to use for a multi-node parallel job.
This parameter is not valid for single-node container jobs.
- Default
No instance type overrides
- Return type
Optional
[InstanceType
]
-
memory
¶ Memory reserved for the job.
- Default
No memory overrides. The memory supplied in the job definition will be used.
- Return type
Optional
[Size
]
-
vcpus
¶ The number of vCPUs to reserve for the container.
This value overrides the value set in the job definition.
- Default
No vCPUs overrides
- Return type
Union
[int
,float
,None
]