shikajiro
|
3660a63508
|
feat(components): Add json format for BigQuery output to GCS (#4994)
* Support for json output
* update to_gcs component.yaml
* add output json format test.
* add command-line param.
|
2021-01-14 21:25:51 -08:00 |
Chen Sun
|
8463992aaa
|
fix(components): Update GCP component container to Python 3.7. Fixes #4959 (#4960)
* Update GCP component container to Python 3.7, fix Dataflow client
* fix test
* add back the http patch
|
2021-01-06 17:09:46 -08:00 |
|
Chen Sun
|
d115d467ae
|
fix(components): fix ml_engine/train interface (#4593)
* Fix missing `job_dir_output_path`
* fix test
|
2020-10-09 14:45:39 -07:00 |
Alexey Volkov
|
2beddd054a
|
fix(components): Fixed the GCP - ML Engine - Batch predict component. Fixes #4430 (#4432)
* Components - Fixed the GCP - ML Engine - Batch predict component
Fixes https://github.com/kubeflow/pipelines/issues/4430
The issue was introduced in https://github.com/kubeflow/pipelines/pull/3850. That PR has added a new parameter in the middle of the the create_job function signature which can cause breaking changes as the parameter ordering changes.
* Fixed test
|
2020-09-02 22:27:41 -07:00 |
|
Alexey Volkov
|
fe6a10a446
|
chore(components): Components - De-hardcoded outputs paths in the GCP components (#4123)
|
2020-08-04 03:16:19 -07:00 |
Andrew Leach
|
58f1d13214
|
Adds job_id parameter. (#3850)
* Adds job_id parameter to ml_engine train component, which takes precedence over job_id generated from job_id_prefix.
* Restores ipynb config.
Co-authored-by: andrewleach <andrewleach@google.com>
|
2020-06-01 15:04:15 -07:00 |
Jiaxiao Zheng
|
1914fa2f34
|
[Component] Fix the image version specification in create cluster component. (#2213)
* Fix
* Fix test
|
2019-09-24 17:11:59 -07:00 |
Alexey Volkov
|
0e0f84c697
|
Moved component_sdk to components/gcp/ (#1698)
|
2019-07-31 19:56:13 -07:00 |