Skip to content
Switch branches/tags
Go to file
* Change e2e test job to point to kserve repo

* Fix prow config

* Remove dir dep

* fix image-transformer built image tag (#44)

* fix image-transformer built image tag

Signed-off-by: Theofilos Papapanagiotou <>

* rename ecr registry directory for kserve

Signed-off-by: Theofilos Papapanagiotou <>

* merging kserve branch to master (#35)

* KServe Go module and API group change

Updated go source files with new go module.
Changed apigroup in yaml files.
Made changes in Makefile to represent the repository and apigroup names.

Verified `make test` is passing.

Removed reference for old names.

* Pointed correct repo reference for ko images (#23)

* Rename changes python - New (#19)

* python changes-initial-commit

* updated api gen.
Renamed package references in python.

* Renamed python/kfserving to python/kserve

* renamed python folder according to package kserve

* Ran client-gen. Updated swagger.json

* renamed constants from KFSERVING_ to KSERVE_.
Yet to change namespace from kfserving-system to kserve and hence not changed in python side.

* Updated kfserving reference to kserve
Updated version for Python Package Index to 0.7.0rc0
Renamed KFServingSampleModel to KServeSampleModel

* Python package and corresponding directory change has to be reflected in test and dockerfiles

* updated a specific version in requirments to address issue with pip resolver.
Added more specific versions in requirements,txt and updated paddleserver with kserve dependency.

* Changed namespace from kfserving-system to kserve (#31)

* Update local dev scripts (#33)

* initial commit of dev-scripts update

* Updated quick-install

* Updated python Makefile

* Generated install for new release

* Updated generate-install script with new release

* Reverting controller name in makefile.
Will be updated in separate pr.

* Github action changes (#34)

* Github action changes

Migrated images publish to kserve repo.
Added github workflows for the ones that was being pushed to from release/triggers.

* Removed batcher publisher since batcher.Dockerfile does not exist
Updated tf2openapi.Dockerfile  to correct directory and package.

* Fixed issue with working directory of tf2openapi workflow

* Removed the deptrecated logger docker

* Excluded tests from linting. (#36)

They were pointing to older directory and package names and
hence were not being excluded while linting.

* Update license (#20)

* Fixed errors thrown by flake8 linter (#37)

* Kfserving kserve manifests changes and code reference changes (#39)

* kfserving -> kserve migartion initial commit

* kfserving -> kserve config changes

* Changed constants.
Updated image reference from kserve to andyarok

* Reverted kserve models web app  image repo.

* go changes for kfserving-kserve

* Updated KFServing refernce in python code

* Updated KFServing to KServe in python doc comments.
Updated references of v1alpha2 from duplicate aliases to just use pkg name.
Updated docs samples references with proper name.
Made changes to use Constant for container name instead of literal "kserve-container".

* Updated gitignore for generated files that came in after folder rename.
Removed the generated files.
Fixed lint warnings for go.

* Removed travis yml file.

* KServe doc update (#38)

* Update main and pythonserver readme

* Update python docs

* Fix role binding

* Update kserve sdk doc

* Fix e2e test script

* Fix import

* Kserve error fix (#40)

* Fixed linting errors in python

* Fixed make test.
Issue with storage initializer name/path.

* Fixed failing python test.
Updated ray version.
Updated KFServingClient reference to KServeClient in kserve init.

* Updated ray version from 1.4.0 to 1.5.0

* Fixed error with e2e due to kfserving reference in test overlay (#42)

Updated quick install and added 0.7.0-rc0 install manifests.
Updated e2e namespace.

* Controller repo update (#45)

Made changes in workflow to update image repo to kserve instead of kfserving

* Controller repo update (#45)

Made changes in workflow to update image repo to kserve instead of kfserving

* update pytorch image

* Fix transformer dockerfile

Co-authored-by: Andrews Arokiam <>
Co-authored-by: andyi2it <>

* Generate 0.7.0 release candidate and update roadmap/release process (#46)

* Fix apiVersion for examples and docs (#55)

Fix apiVersion for examples and docs change from

* add default deployment and fix startup without knative to master branch (#57)

* add default deployment and fix startup without knative

* fix suite_test error

* fix suit_test error

* add constant vb for RawDeployment and Serverless

* update predictor string RawDeployment to constants.RawDeployment

* modify getDeploymentMode return type

* Update kserve layer diagram (#58)

* update diag (#59)

* Remove kubeflow copyright (#56)

* Adding kfs_architect.drawio and changing kfs_architect.png's "KFserving" name to "KServe" (#60)

* Delete kfs_architect.png

* Add files via upload

Adding kfs_architect.drawio and changing kfs_architect.png's "KFserving" name to "KServe"

* Delete kfs_architect.png

* Delete kfs_architect.drawio

* added Metric scrapping title

* Delete kfs_architect.png

* Delete kfs_architect.drawio

* fixed typo

* add Nick Hill as a reviewer (#66)

also removing rkelkar

* Separate out web-app repo (#68)

* Sepate out web-app repo

* Remove models web-app github action

* Fix feast example

Co-authored-by: Theofilos Papapanagiotou <>
Co-authored-by: Andrews Arokiam <>
Co-authored-by: andyi2it <>
Co-authored-by: Chin Huang <>
Co-authored-by: Qingshan Chen <>
Co-authored-by: Animesh Singh <>
Co-authored-by: js-ts <>
9 contributors

Users who have contributed to this file

@animeshsingh @ellistarn @yuzisun @andyi2it @theofpa @gaocegege @chinhuang007 @js-ts @Iamlovingit

KServe Roadmap

2021 Q4/2022 Q1

Kubernetes Deployment

Objective: "Enable raw kubernetes deployment as alternative mode"

  • Support existing ML frameworks, transformer/explainer, logger and batching
  • Make Istio/KNative optional and unlock KNative limitations
    • Allow multiple volumes mounted
    • Allow TCP/UDP

Inference Graph

Objective: "Enable model serving pipelines with flexible routing graph"


Objective: "Unifying interface for SingleModel and ModelMesh deployment"

  • Ability to perform inference using Predict v2 API with REST/gRPC
  • Unify the storage support for single and ModelMesh
  • InferenceService controller to utilize ServingRuntime
  • Single install for KServe which includes SingleModel and ModelMesh Serving