components medimage_insight_ft_pipeline - Azure/azureml-assets GitHub Wiki
Pipeline Component to finetune MedImageInsight Model.
Version: 0.0.2
View in Studio: https://ml.azure.com/registries/azureml/components/medimage_insight_ft_pipeline/version/0.0.2
Name | Description | Type | Default | Optional | Enum |
---|---|---|---|---|---|
mlflow_embedding_model_path | Path to the MLflow model to be imported. | uri_folder | False | ||
eval_image_tsv | Path to the evaluation image TSV file. | uri_file | False | ||
eval_text_tsv | Path to the evaluation text TSV file. | uri_file | False | ||
image_tsv | Path to the image TSV file. | uri_file | False | ||
text_tsv | Path to the text TSV file. | uri_file | False | ||
label_file | Path to the label file. | uri_file | False | ||
conf_files | Path to the configuration files. | uri_file | False | ||
instance_type_finetune | Instance type to be used for finetune component in case of serverless compute, eg. standard_nc24rs_v3. The parameter compute_finetune must be set to 'serverless' for instance_type to be used | string | Standard_nc24rs_v3 | True | |
compute_finetune | compute to be used for finetune eg. provide 'FT-Cluster' if your compute is named 'FT-Cluster'. Special characters like \ and ' are invalid in the parameter value. If compute cluster name is provided, instance_type field will be ignored and the respective cluster will be used | string | serverless | True | |
process_count_per_instance | Number of processes to run per instance. This is used to set the number of GPUs to use for training. | integer | 1 | True | |
instance_count | Number of instances to use for training. | integer | 1 | True |
Name | Description | Type |
---|---|---|
save_dir | Directory to save the model and checkpoints, used for pipeline's internal operations. | uri_folder |
embedding_mlflow_model | Directory to save the MLflow model. | mlflow_model |
classification_mlflow_model | Path to save the output model configured with labels. | mlflow_model |