components mmtracking_video_multi_object_tracking_model_import

Video Multi Object Tracking MMTracking Model Import

mmtracking_video_multi_object_tracking_model_import

Overview

Import PyTorch / MLflow model

Version: 0.0.10

View in Studio: https://ml.azure.com/registries/azureml/components/mmtracking_video_multi_object_tracking_model_import/version/0.0.10

Inputs

Model family

Name	Description	Type	Default	Optional	Enum
model_family	Which framework the model belongs to.	string	MmTrackingVideo	True	['MmTrackingVideo']

Model name

Name	Description	Type	Default	Optional	Enum
model_name	Please select models from AzureML Model Assets for all supported models. For MMTracking, provide the model's config name here, same as its specified in MMTracking Model Zoo. To find the correct model name, go to https://github.com/open-mmlab/mmtracking/tree/v0.14.0/configs/mot click on the model type and you will find the model name in the metafile.yml file which is present at configs/<MODEL_TYPE>/metafile.yml location. It is the user responsibility to comply with the model's license terms.	string		True

Continual-Finetuning model path

Name	Description	Type	Default	Optional	Enum
pytorch_model	Pytorch Model registered in AzureML Asset.	custom_model		True
mlflow_model	Mlflow Model registered in AzureML Asset.	mlflow_model		True
validation_output	Validation status.	uri_file		True
download_from_source	Download model directly from MMTracking instead of system registry	boolean	False	True

Outputs

Name	Description	Type
output_dir	Folder to store model metadata.	uri_folder

Environment

azureml://registries/azureml/environments/acft-mmtracking-video-gpu/versions/33