modified tf-serving.libsonnet in object_detection example to fix the error of
"FileSystemStoragePathSource encountered a file-system access error:
Could not find base path /models/model for servable model"
Change-Id: I946a0a7fbb6c80992d66fe003ca90b1c21c67cfc
Signed-off-by: Henry Wang <henry.wang@arm.com>
* create pv for pets-pv
For a lot of user k8s clusters, dynamic volume provisioning isn't
enabled. So the newcomer may be blocked since pets-pv will keep
Pending. We can guide them to create a nfs PV as an option.
* tell user how to check if a default storage class is defined
* add link about how to create PV
* add object detection grpc client
Fixes: #377
* fix kubeflow-examples-presubmit error
object_detection_grpc_client.py depends on other files in
https://github.com/tensorflow/models.git, pylint will fail
for those files need to be compiled manually.
Since mnist_DDP.py has similar dependency, here just follow
mnist_DDP.py and ignore checking this file.
In tensorflow/models/research/object_detection/, only
tensorflow/models/research/object_detection/legacy/train.py
supports kubeflow sor far (construct cluster by reading
TF_CONFIG environment var).
Fixes: #277
* Fix#272Fix#272 where the `create-pet-record-job` pod produces this error: `models/research/object_detection/data/pet_label_map.pbtxt; No such file or directory`
* Update create-pet-record-job.jsonnet
* Updated Dockerfile.traning to use latest tensorflow
and tensorflow object detetion api.
* Updated tf-training-job component and added a chief
replica spec
* Corrected some typos and updated some instructions
* adding batch-predict on GPU example
* Sync with TF-serving GPU example.
* adding visualization instructions
* change the title of readme.md
* changes according to the review comments from jlewi
* Replace the links to personal project with the one in kubeflow-example project in the yaml file
* change the procedure to build images
* polish the md file
* some minor md change
* fix a broken gs link
* fix more merge errors
* Added Ksonnet prototypes to parametrize old yaml files
* Modified instructions
* Added tf-training-job component
* Removed yaml manifest files
Modified serving instructions
* Consolidate get-data and decompression jobs
* Deleted registry and prototypes
* Added components to ks-app dir
* Modified instructions
* Fixed references to user guide page
Improved instructions
* General improvements to components and instructions
* Removed obj-detection.libsonnet file
* used specific params in export-graph and create-tf-record
instead of list params like 'args' and 'command'
* Improved instructions and removed references to yaml files
* Added tutorial for object detection distributed training
Added steps on how to leverage kubeflow tooling to
submit a distributed object detection training job
in a small kubernetes cluster (minikube, 2-4 node cluster)
* Added Jobs to prepare the training data and model
* Updated instructions
* fixed typos and added export tf graph job
* Fixed paths in jobs and instructions
* Enhanced instructions and re-arranged folder structure
* Updated links to kubeflow user guide documentation
MrXinWang
Daniel Castellanos
Hougang Liu
Sam Shi
Guang Ya Liu
Konstantinos Samaras-Tsakiris
Lun-Kai Hsu
Roy Xue
Daniel Castellanos