chore(components): Change #AutoML #Vision data converter output artifact type
PiperOrigin-RevId: 580344061
This commit is contained in:
parent
033bfa264d
commit
79a916d067
|
|
@ -1,6 +1,7 @@
|
|||
## Upcoming release
|
||||
* Bump supported KFP versions to kfp>=2.0.0b10,<=2.4.0
|
||||
* Add LLM Eval pipeline parameter for customizing eval dataset reference ground truth field
|
||||
* Fix the output parameter `output_dir` of `preview.automl.vision.DataConverterJobOp`.
|
||||
|
||||
## Release 2.5.0
|
||||
* Upload tensorboard metrics from `preview.llm.rlhf_pipeline` if a `tensorboard_resource_id` is provided at runtime.
|
||||
|
|
|
|||
|
|
@ -29,7 +29,7 @@ def data_converter(
|
|||
input_file_path: str,
|
||||
input_file_type: str,
|
||||
objective: str,
|
||||
output_dir: dsl.OutputPath(str),
|
||||
output_dir: dsl.Output[dsl.Artifact],
|
||||
gcp_resources: dsl.OutputPath(str),
|
||||
location: str = 'us-central1',
|
||||
timeout: str = '604800s',
|
||||
|
|
@ -59,7 +59,7 @@ def data_converter(
|
|||
encryption_spec_key_name: Customer-managed encryption key options for the CustomJob. If this is set, then all resources created by the CustomJob will be encrypted with the provided encryption key.
|
||||
project: Project to create the custom training job in. Defaults to the project in which the PipelineJob is run.
|
||||
Returns:
|
||||
output_dir: Cloud Storage directory storing converted data and pipeline information.
|
||||
output_dir: Cloud Storage directory storing converted data.
|
||||
gcp_resources: Serialized JSON of `gcp_resources` [proto](https://github.com/kubeflow/pipelines/tree/master/components/google-cloud/google_cloud_pipeline_components/proto) which tracks the CustomJob.
|
||||
"""
|
||||
# fmt: on
|
||||
|
|
@ -83,7 +83,7 @@ def data_converter(
|
|||
'--objective',
|
||||
objective,
|
||||
'--output_dir',
|
||||
output_dir,
|
||||
output_dir.path,
|
||||
IfPresent(
|
||||
input_name='output_shape',
|
||||
then=Concat('--output_shape', output_shape),
|
||||
|
|
|
|||
Loading…
Reference in New Issue