C4W2: Autoscaling TensorFlow Model Deployments with TF Serving and Kubernetes

Task 5. Creating TensorFlow Serving deployment

Hello together,

I’ve got issues in adapting the file “deployment.yaml” according to the intendations as described in the 3 screenshots under the description. Line 52 is without adaptations by me: " resources:"
For me it says “error: error parsing tf-serving/deployment.yaml: error converting YAML to JSON: yaml: line 52: mapping values are not allowed in this context”.

In contrast to “C4W2 Assignment Issues as of September 2022: Autoscaling TensorFlow model deployments with TF Serving and Kubernetes” I deleted the version “tensorflow/serving:2.8.0” and changed it to “tensorflow/serving”, as in the screenshot in the task.
Discussion from “C4W2 Assignment Autoscaling TensorFlow model deployments with TF Serving and Kubernetes - #14 by chris.favila” sounds for me like a temporary solution, that should not work anymore, but I did not find any newer information, unfortunately.

Could you provide a hint for this, please?
Thanks a lot.


Your yaml file isn’t correct. For instance, there are two spec entries. It seems like you’ve copy pasted invalid content into the file. Please checkout the git repo and see the file. If that doesn’t fix things, please click my name and message your deployment.yaml file as an attachment.

Thanks for sending the file. Your file contains extra indentation for protocol entry and is therefore failing the assignment. On a related note, there’s no need for you to remove the annotations entry in the file.

Task 5 (creating the deployment) has been running for 20mins and I cant complete the assignment because I need the external IP

Your issue is different from the typo Simon has made. Please create a new topic. Here’s the community user guide to get started.