Skip to content
Permalink

Comparing changes

Choose two branches to see what’s changed or to start a new pull request. If you need to, you can also or learn more about diff comparisons.

Open a pull request

Create a new pull request by comparing changes across two branches. If you need to, you can also . Learn more about diff comparisons here.
base repository: kserve/modelmesh-runtime-adapter
Failed to load repositories. Confirm that selected base ref is valid, then try again.
Loading
base: main
Choose a base ref
...
head repository: kserve/modelmesh-runtime-adapter
Failed to load repositories. Confirm that selected head ref is valid, then try again.
Loading
compare: release-0.8
Choose a head ref
  • 3 commits
  • 9 files changed
  • 2 contributors

Commits on Feb 11, 2022

  1. feat: Add GCS support to pullman (#10) 

    * feat: Add GCS support to pullman

Signed-off-by: Paul Van Eck <[email protected]>

* Add authentication capabilities

Signed-off-by: Paul Van Eck <[email protected]>

* Some cleanup and refactoring

Signed-off-by: Paul Van Eck <[email protected]>

* Enable concurrent downloading

Signed-off-by: Paul Van Eck <[email protected]>

* Minor refactoring and improvements

Signed-off-by: Paul Van Eck <[email protected]>
    @pvaneck
    pvaneck committed Feb 11, 2022
    Configuration menu
    Copy the full SHA
    e675b92 View commit details
    Browse the repository at this point in the history

  2. chore: enable gcs and http in model puller (#13) 

    Signed-off-by: Paul Van Eck <[email protected]>
    @pvaneck
    pvaneck committed Feb 11, 2022
    Configuration menu
    Copy the full SHA
    eb4f4be View commit details
    Browse the repository at this point in the history

  3. fix: Include model_type in loadModel modelKey json passed by the pull… 

    …er (#15)

* fix: Include model_type in loadModel modelKey json passed by the puller

Motivation

When using the puller with non-builtin runtimes, the modelType provided in the Predictor is passed within some json in the modelKey field because the internal modelType field is used for model placement constraints (it's historical/unfortunate and we hope to change that soon).

However, after the recent storage-related refactor this separate model_type attribute is no longer being passed through.

Also, the storage_key and storage_params attributes _are_ passed along by the puller but this doesn't really make sense because they're consumed by the puller and intended to be abstracted away from the runtime.

Modifications

- Ensure the model_type json attribute is passed through
- Remove storage_key, storage_params and bucket attributes from the transformed request
- Update unit tests to cover this behaviour

Result

Non-built-in runtimes have direct access to the predictor's modelType and are shielded from storage parameters that they shouldn't have to know about.

Signed-off-by: Nick Hill <[email protected]>

* Address @tjohnson31415's review comments

Signed-off-by: Nick Hill <[email protected]>
    @njhill @pvaneck
    njhill authored and pvaneck committed Feb 11, 2022
    Configuration menu
    Copy the full SHA
    adeeddb View commit details
    Browse the repository at this point in the history
Loading