search

PaddlePaddle / Paddle

PArallel Distributed Deep LEarning: Machine Learning Framework from Industrial Practice （『飞桨』核心框架，深度学习&机器学习高性能单机、分布式训练和跨平台部署）
http://www.paddlepaddle.org/
Apache License 2.0
22.24k stars 5.59k forks source link

train.sh in semantic_role_labeling cannot exit at the end #286

Closed byzhang closed 8 years ago

byzhang commented 8 years ago

It hangs when the last pass is trained, for the following settings:

  1. use_gpu = false, trainer_count=4
  2. use_gpu = true, trainer_count=4
  3. use_gpu = true, trainer_count=1

It can exit normally only when

  1. use_gpu = false, trainer_count=1
byzhang commented 8 years ago

I'm at @9c5c38f

pineking commented 8 years ago

+1

wangkuiyi commented 8 years ago

这个问题我也碰到过。请 @reyoung 看看谁来解决合适。

此外，我们的值班表 https://github.com/baidu/Paddle/wiki/Paddle-Members-on-Duty 里没有安排周末值班。这会导致我们对周末提出的需求的相应变慢，也会导致周一的值班员的工作任务太重。

我可以来担任周末值班员，但是我对项目情况还不够熟悉，周末可能会Hi大家。

reyoung commented 8 years ago

Thank you for response. I'm trying to fix it.