Built-in Types

ProjectMetadata

Project metadata.

Bases

dataclass

FieldTypeDescription
namestrThe name of the project, for example: "My Cool Project"

TagMetadata

Tag metadata.

Bases

dataclass

FieldTypeDescription
namestrThe name of the tag
colorstrThe color code of the tag, for example: "#73aaff"
descriptionstrThe description of the tag.

Pagination

Pagination parameters.

Bases

dataclass

FieldTypeDescription
pagestrA string that represents the page of results
limitintA limit on the number of objects to be returned, the Limit can range between 1 and 100, and the default is 100

AssetFilter

Asset list filter.

Bases

dataclass

FieldTypeDescription
groupsList[str]The filter groups.
filenamestrThe query filename.
statusstrThe status of the asset, enum: [Annotated, Review, Completed, Tofix, None]
metadata_querystrThe metadata query string. Please refer Here.

AssetMetadata

Asset Metadata.

πŸ“˜

The custom metadata size that can be attached depends on your price plan. Please reference here for more information!

Bases

dataclass

FieldTypeDescription
statusstrThe status of the asset, enum: [Annotated, Review, Completed, Tofix, None]
custom_metadatadictThis metadata dictionary should contain the metadata information you want to preserve or add for use on the Nexus platform.

AnnotationFilter

Annotation list filter.

Bases

dataclass

FieldTypeDescription
asset_idsList[str]The list of asset ids.

AnnotationMetadata

Annotation metadata.

Bases

dataclass

FieldTypeDescription
asset_idstrThe id of the asset, example: "asset_9f61bab8-ec60-476f-a1cf-63eca5c313dd"
tagstrThe name of the tag
bound_typestrThe bound type of the annotation, Enum: [Polygon, Rectangle,Classification,Keypoints]
boundlistThe bound of the annotation, example: [[0.0325, 0.8436],[0.0325, 0.9094],[0.1275, 0.9094],[0.1275, 0.8431]]

AnnotationExportOptions

Annotation export options.

Bases

dataclass

FieldTypeDescription
split_ratiointThe ratio used to split the data into training and validation sets
seedintThe number used to initialize a pseudorandom number generator
normalizedboolThe boolean indication on whether the exported annotations should be shuffled

AnnotationExportMetadata

Annotation Export Metadata.

Bases

dataclass

FieldTypeDescription
formatAnnotationExportFormatThe format to export the annotations.
optionsAnnotationExportOptionsThe options to export the annotations

AssetMetadata

Asset Metadata.

πŸ“˜

The custom metadata size that can be attached depends on your price plan. Please reference here for more information!

Bases

dataclass

FieldTypeDescription
statusstrThe status of the asset, enum: [Annotated, Review, Completed, Tofix, None]
custom_metadatadictThis metadata dictionary should contain the metadata information you want to preserve or add for use on the Nexus platform.

AnnotationExportFormat

Annotation export formats.

Bases

Enum

TypeDescription
strThe format of the annotation.

Classification: [csv_classification, classification_tfrecord]
Bounding Box: [csv_fourcorner, csv_widthheight, coco, pascal_voc, yolo_keras_pytorch, yolo_darknet, createml, tfrecord]
Segmentation: [polygon_single, polygon_coco]
Keypoint: [keypoints_coco]

FlowMetadata

Workflow Metadata.

Bases

dataclass

FieldTypeDescription
titlestrThe title of the Workflow, for example: "New workflow title"

RunSetupMetadata

Training settings.

Bases

dataclass

FieldTypeDescription
acceleratorAcceleratorThe hardware accelerator to use for the training run
checkpointCheckpointThe checkpoint strategy name for the training run.
limitLimitThe limit configuration for the training
previewboolA boolean value indicating whether to enable preview for the training
matrixboolA boolean value indicating whether to enable the matrix for the training

Accelerator

Run Accelerator Metadata.

Bases

dataclass

FieldTypeDescription
namestrThe GPU type to use for the training run, enum: [GPU_T4, GPU_L4,GPU_P100,GPU_V100,GPU_A100_40GB]
countintThe number of GPUs to use for the training run

Checkpoint

Run Checkpoint Metadata.

Bases

dataclass

FieldTypeDescription
strategystrThe checkpointing strategy name for the training, enum: [ STRAT_EVERY_N_EPOCH, STRAT_ALWAYS_SAVE_LATEST, STRAT_LOWEST_VALIDATION_LOSS, STRAT_HIGHEST_ACCURACY]
metricstrThe checkpointing metric for the training, enum: [Loss/total_loss, Loss/regularization_loss, Loss/localization_loss, Loss/classification_loss, DetectionBoxes_Precision/mAP, DetectionBoxes_Precision/[email protected], DetectionBoxes_Precision/mAP (small), DetectionBoxes_Precision/mAP (medium), DetectionBoxes_Precision/mAP (large), DetectionBoxes_Recall/AR@1, DetectionBoxes_Recall/AR@10, DetectionBoxes_Recall/AR@100, DetectionBoxes_Recall/AR@100 (small), DetectionBoxes_Recall/AR@100 (medium), DetectionBoxes_Recall/AR@100 (large)]
evaluation_intervalintThe checkpoint evaluation interval value for the training

Limit

Run Limit Metadata.

Bases

dataclass

FieldTypeDescription
metricstrThe limit metric for the training, enum: [LIM_MINUTE, LIM_EPOCH, LIM_NONE]
valueintThe limit metric value for the training

ArtifactFilters

Artifact filters data class.

Bases

dataclass

FieldTypeDescription
run_idsList[str]The list of run ids.

ArtifactExportMetadata

Artifact exporting settings.

Bases

dataclass

FieldTypeDescription
formatstrThe format of the artifact exports, enum: [ONNX, TensorFlow, PyTorch, TFLite]
quantizationstrThe quantization of the artifact exports, optional.

DeploymentMetadata

Deployment settings.

Bases

dataclass

FieldTypeDescription
namestrThe name of the inference API instance
artifact_idintThe unique identifier of the artifact to deploy
num_of_instancesintThe number of the inference API instances to deploy
version_tagstrThe current version tag of the inference instance.
resourcesDeploymentResourcesThe resource allocation for the deployment instance, optional.
optionsDeploymentOptionsThe configuration options for the deployment instance, optional.

DeploymentResources

The resource allocation for the deployment instance, optional.

Bases

dataclass

FieldTypeDescription
GPU_T4intThe number of NVIDIA Tesla T4 GPUs to allocate to each Inference API instance, optional.

DeploymentOptions

The configuration options for creating each deployment API instance, optional.

Bases

dataclass

FieldTypeDescription
evaluation_strategystrThe evaluation strategy to use of each Inference API, default entropy_score.
evaluation_thresholdintThe evaluation threshold to use to trigger post-evaluation actions, default 0.5.
evaluation_groupstrThe asset group to which assets triggered by the active learning route will be uploaded, is a comma-separated list.