Clash seems like a good option unless you want to write your DAGs more from scratch with the Compute Engine Operators.
https://github.com/ottogroup/clash https://airflow.apache.org/docs/apache-airflow-providers-google/stable/operators/cloud/compute.html I'd also suggest you pop into the Airflow Slack and follow the discussions in #airflow-gcp. GPU:s go brrrrrr, Anton Den ons 26 jan. 2022 kl 00:18 skrev Shih-gian Lee <[email protected]>: > Hello, > > I am using Airflow 1.10.10. I would like to submit a model training job to > a Google Cloud Platform GPU Compute Engine. Is there any example or > documentation on how to do that in Airflow? If this is not possible, can > you help to suggest an alternative? > > Thank you in advance for your help! > >
