Issue with the GPUs are in almost all cases related to improper Dockefile. I put together a notebook that shows two ways (primarily two different dockerfiles) that properly utilize the GPUs with PyTorch. Full code see here:https://colab.research.goog...

As kvandre said Yes you can use Vertex AI Experiments without TensorBoard instance (as long as you don't need Time Series Metrics).The TensorBoard Instance is not automatically created. To create it, you have to call explicitly:experiment_tensorboard...

Hi NityaTo deploy your model you have to pack it into a serving container this can be either done with a custom prediction container or with custom prediction routine. I have written a deep dive article and videohttps://medium.com/google-cloud/servin...

Hello Mota,This is a known issue, and Google will release a fix starting next week. Best regards

Can you share the code or gcloud command you're using to submit the job. With a custom container you don't need to define a trainer.