Created by: witness99

请问在按照deepFM中所说的步骤走的情况下，为什么最后在local_train阶段epoch的时间会这么短？

以下为得到的结果： args.embedding_size: 10 args.num_field: 39 args.num_feat: 1086460 args.layer_sizes: [400, 400, 400] args.act: relu args.reg: 0.0001

epoch1 is finished and takes 0.069958 s epoch2 is finished and takes 0.030759 s epoch3 is finished and takes 0.031983 s epoch4 is finished and takes 0.039973 s epoch5 is finished and takes 0.032980 s epoch6 is finished and takes 0.033335 s epoch7 is finished and takes 0.029980 s epoch8 is finished and takes 0.029713 s epoch9 is finished and takes 0.029981 s epoch10 is finished and takes 0.038977 s epoch11 is finished and takes 0.044971 s epoch12 is finished and takes 0.028981 s epoch13 is finished and takes 0.040977 s epoch14 is finished and takes 0.037975 s epoch15 is finished and takes 0.032979 s epoch16 is finished and takes 0.032980 s epoch17 is finished and takes 0.030980 s epoch18 is finished and takes 0.031979 s epoch19 is finished and takes 0.033981 s epoch20 is finished and takes 0.032983 s epoch21 is finished and takes 0.033988 s epoch22 is finished and takes 0.031253 s epoch23 is finished and takes 0.033978 s epoch24 is finished and takes 0.031979 s epoch25 is finished and takes 0.032640 s epoch26 is finished and takes 0.032981 s epoch27 is finished and takes 0.032980 s epoch28 is finished and takes 0.036979 s epoch29 is finished and takes 0.031981 s epoch30 is finished and takes 0.030977 s

Process finished with exit code 0