PengYuan
PengYuan
> Yeah I'm sensing that the versions we are using do not take into account different arch's; I'll try to have a look at this, but if you have time...
Please see step1 output log in file: `output/actor-models/1.3b/training.log`
We can add job configmap to kubernetes deployment to support custom param, then mount configmap to some path to start the job.
> Please check ci errors Ok, thanks, I'm on doing this.
@TyrantLucifer Please help me review this.
@liugddx @Hisoka-X Please help me review this, thanks!
@hailin0 Please help me review this, thanks.
@CalvinKirs Please help me review this.
> I happen to have a question that I can discuss here. Now our checkpoint persistence implementation comes in through SPI, so we need to know how to release the...
> > I happen to have a question that I can discuss here. Now our checkpoint persistence implementation comes in through SPI, so we need to know how to release...