Hi, first time here 🙂 We are looking forward to deploy a fairly simple xgboost based regression model in the cloud. We have developed the model in Google Colab with GPU and it is working. Now we want to deploy it in a way that a we can send prediction queries to a REST endpoint and get results. No GPU needed for this inference part, and it needs to be up always. And also for the training part, once daily a GPU based machine that will need to be spin up for couple of hours to train and save the model. Now, after looking into ClearML documentation, looks like for the training and deployment there are no ClearML cloud instances? Do we have to train/deploy externally to ClearML with the ClearML Agent? Am I missing something? Early days for me, any pointers to the right direction will be highly helpful. Thanks a lot in advance.
Looks like you need the https://clear.ml/docs/latest/docs/clearml_serving/clearml_serving and https://clear.ml/docs/latest/docs/pipelines/pipelines features with a https://clear.ml/pricing/ in SaaS deployment so you can use the https://clear.ml/docs/latest/docs/webapp/applications/apps_gcp_autoscaler to manage the workers for you
I see, many thanks for the prompt reply. So, ClearML themselves does not host any machines for training and serving.
Nope but it can manage your cloud vms for you
