Error while deploying a model on an endpoint- Vertex AI

user_9731134 0
Viewed 235

I'm working on a model that I need to deploy on a Vertex AI endpoint. The model is a DNN developed in Tensorflow. I've saved the model locally, loaded to GCS and imported it in the Vertex AI Model section without problems. When I'm trying to deploy it to a new endpoint Vertex responses is the following: enter image description here

I've tried to search the error List of found errors: 1.Field: deployed_model.prediction_resources; Message: Required field is not set. but found nothing. What could be the cause?

Version of tensorflow and python are the same of the pre-built environment given by Google, the model is saved through tf.keras.Model.save(model, model_path).

google-cloud-vertex-ai gcloud tensorflow google-cloud-platform
1 Answers

Sorry this is not an answer but just want to add more context to this issue. I think this is a bug because the script I had used to upload the model and deploy it to endpoint just failed yesterday because of this error. The weirdest thing is that the error appears when I try to gcloud ai model list after successfully uploading the model. More info here: https://issuetracker.google.com/issues/237421435

user_74567510
Related