aws_sagemaker_remote.inference package

Submodules

aws_sagemaker_remote.inference.command module

class aws_sagemaker_remote.inference.command.InferenceCommand(config: aws_sagemaker_remote.inference.command.InferenceCommandConfig, help='Run inference')

Bases: aws_sagemaker_remote.commands.Command

configure(parser: argparse.ArgumentParser)
run(args)
class aws_sagemaker_remote.inference.command.InferenceCommandConfig(module, input=None, output=None, model_dir=None, input_type=None, output_type='application/json')

Bases: object

aws_sagemaker_remote.inference.command.run_inference_module(config: aws_sagemaker_remote.inference.command.InferenceCommandConfig)

aws_sagemaker_remote.inference.endpoint module

aws_sagemaker_remote.inference.endpoint.endpoint_create(config, name, client, force)
aws_sagemaker_remote.inference.endpoint.endpoint_delete(name, client)
aws_sagemaker_remote.inference.endpoint.endpoint_describe(name, client, field=None)
aws_sagemaker_remote.inference.endpoint.endpoint_exists(name, client)
aws_sagemaker_remote.inference.endpoint.endpoint_invoke(model_dir, name, model, variant, input, output, input_type, input_glob, output_type, runtime_client)

aws_sagemaker_remote.inference.endpoint_config module

aws_sagemaker_remote.inference.endpoint_config.endpoint_config_create(model, name, instance_type, force, session)
aws_sagemaker_remote.inference.endpoint_config.endpoint_config_delete(name, client)
aws_sagemaker_remote.inference.endpoint_config.endpoint_config_describe(name, client, field=None)
aws_sagemaker_remote.inference.endpoint_config.endpoint_config_exists(name, client)

aws_sagemaker_remote.inference.iam module

aws_sagemaker_remote.inference.iam.ensure_inference_role(iam, role_name)

aws_sagemaker_remote.inference.inputs module

aws_sagemaker_remote.inference.local module

aws_sagemaker_remote.inference.local.inference_handler(model_dir)
aws_sagemaker_remote.inference.local.inference_local(model_dir, tasks, input_type, output_type)
aws_sagemaker_remote.inference.local.inference_run(handler, context, input, output, input_type, output_type)

aws_sagemaker_remote.inference.mime module

aws_sagemaker_remote.inference.model module

aws_sagemaker_remote.inference.model.model_create(job, model_artifact, name, session: sagemaker.session.Session, inference_image, inference_image_path, inference_image_accounts, role, force, multimodel=False, accelerator_type=None)
aws_sagemaker_remote.inference.model.model_delete(name, client)
aws_sagemaker_remote.inference.model.model_describe(name, client, field=None)
aws_sagemaker_remote.inference.model.model_exists(name, client)

aws_sagemaker_remote.inference.outputs module

aws_sagemaker_remote.inference.outputs.output_fn_json(prediction, content_type)

aws_sagemaker_remote.inference.package module

class aws_sagemaker_remote.inference.package.ExportModelCommand(spec, help='Export package', **kwargs)

Bases: aws_sagemaker_remote.training.main.TrainingCommand

build_spec(args)
return ExportModelSpec(
requirements=args.requirements, args=args

)

todo: add args to control spec

main(args)
class aws_sagemaker_remote.inference.package.ExportModelSpec(dependencies=None, args=None)

Bases: object

aws_sagemaker_remote.inference.package.export_model_spec(model_dir, spec: aws_sagemaker_remote.inference.package.ExportModelSpec)

Module contents