Yep! It complained about passing data between the scheduler and the workers
Matt Alhonte
05/02/2022, 10:15 PM
So basically I'm trying to do Hyperparameter optimization - I wanna load the data in 1 task and then map the training task over the hyperparam combinations.
Matt Alhonte
05/02/2022, 10:15 PM
Lemme look through the CloudWatch logs for the exact error...
k
Kevin Kho
05/02/2022, 11:58 PM
^ DM’ed Matt. He will just read from the workers rather than passing the dataframe through the scheduler. Reading from workers directly is more efficient anyway
Bring your towel and join one of the fastest growing data communities. Welcome to our second-generation open source orchestration platform, a completely rethought approach to dataflow automation.