Відгуки: TFX on Cloud AI Platform Pipelines

Lab not updated, dependencies from install.sh does not work

Miguel L. · Відгук надано близько 2 років тому

It's impossible to pass last verification points. Deploy to AI Platform Pipeline is not tracked properly, service is deprecated. TFX CLI fails with http 400 on invalid region. UPD: seems like initial setup script is for us-central1 while instructions are for us-west1. TFX run fails with an error on model deploy saying that "The provided GCE region 'us-west1' is not available, or your project needs to be whitelisted to use it. Other available regions are [asia-northeast1, europe-west1, us-central1, us-east1, us-east4].". Details: "[{'@type': 'type.googleapis.com/google.rpc.BadRequest', 'fieldViolations': [{'field': 'model.regions', 'description': "The provided GCE region 'us-west1' is not available, or your project needs to be whitelisted to use it. Other available regions are [asia-northeast1, europe-west1, us-central1, us-east1, us-east4]."}]}]"> During handling of the above exception, another exception occurred:" UPD Sept 26, 2023 Lab content has been corrected so now it requests cluster creation in us-east4 region. Also I managed to get a successful job run! Unfortunately, these result are not counted by checkpoints 'Deploy the pipeline package to AI Platform Pipelines" and "Create kubeflow pipeline run using TFX CLI" despite all requirements being satisfied! It's high time to update checkpoints and ensure they works properly!!!1111

Pavel K. · Відгук надано близько 2 років тому

Fahad A. · Відгук надано близько 2 років тому

Swati M. · Відгук надано близько 2 років тому

Job failed. top of directions say to create cludster in useast1a, but the NB trains in uscentral. Is that a problem?

Marcia P. · Відгук надано близько 2 років тому

Venu P. · Відгук надано близько 2 років тому

Abraham R. · Відгук надано близько 2 років тому

Yash J. · Відгук надано близько 2 років тому

Yash J. · Відгук надано близько 2 років тому

Outdated tools, please upgrade. The grading solutions also are not able to get the runs we triggered.

Bhavesh Naresh W. · Відгук надано близько 2 років тому

Harshit P. · Відгук надано близько 2 років тому

Good Lab

Robert M. · Відгук надано близько 2 років тому

Mohammad Irshad K. · Відгук надано близько 2 років тому

PLEASE CHECK TASK Test completed tasks - Deploy the pipeline package to AI Platform Pipelines Click Check my progress to verify the objective. ===The check doesn't change, but I have deployed the pipeline=== MY CODE (SAME IN THE SOLUTION NOTEBOOK) !tfx pipeline create \ --pipeline_path=runner.py \ --endpoint={ENDPOINT} \ --build_target_image={CUSTOM_TFX_IMAGE} THE RESULT CONFIRM THE DEPLOYMENT 'id': 'aa01dbf5-5086-40b8-9e35-4e7578be609e', 'name': 'tfx_covertype_continuous_training', 'parameters': [{'name': 'pipeline-root', 'value': 'gs://qwiklabs-gcp-00-14dd53365c32-kubeflowpipelines-default/tfx_covertype_continuous_training/{{workflow.uid}}'}, {'name': 'data-root-uri', 'value': 'gs://cloud-training/OCBL203/workshop-datasets'}, {'name': 'eval-steps', 'value': '500'}, {'name': 'train-steps', 'value': '5000'}], 'resource_references': None, 'url': None} Please access the pipeline detail page at http://2b597440b79d2c46-dot-us-central1.pipelines.googleusercontent.com/#/pipelines/details/aa01dbf5-5086-40b8-9e35-4e7578be609e Pipeline "tfx_covertype_continuous_training" created successfully. But nothing change!

Sergio B. · Відгук надано близько 2 років тому

Still problems when deploying the pipeline !tfx pipeline create \ --pipeline_path=runner.py \ --endpoint={ENDPOINT} \ --build_target_image={CUSTOM_TFX_IMAGE} It create the pipeline, but cloudskills check 4 doesn't seem to detect the pipeline generated.

Sergio B. · Відгук надано близько 2 років тому

Can't complete, problem with the lab: zones/us-central1-a' does not have enough resources available to fulfill the request.

Robert M. · Відгук надано близько 2 років тому

Robert M. · Відгук надано близько 2 років тому

Could not complete the lab...

Robert M. · Відгук надано близько 2 років тому

unable to resolve errors

Chris T. · Відгук надано близько 2 років тому