yahoo / CaffeOnSpark

Distributed deep learning on Hadoop and Spark clusters.
Apache License 2.0
1.27k stars 357 forks source link

An issue when run in GPU mode: waiting for AM container to be allocated. #283

Closed GoodJoey closed 6 years ago

GoodJoey commented 7 years ago

It works smoothly when I use CPU mode with CaffeOnSpark. but when i switch to GPU mode, it always shows the information: INFO yarn.Client: Application report for application_1507515571247_0006 (state: ACCEPTED), then i check the spark log, seems it's waiting for AM container to be allocated. and there is no problem when using GPU mode in cafe

Does anyone encounter the some issue or know how to fix this?