mlp.0.log (1298B)
1 load norm file ...... 2 norm file load 3 parameters input: 4 fea_file: ../dataset/fea_debug.pfile 5 norm_file: ../dataset/train.norm 6 targ_file: ../dataset/lab_debug.pfile 7 outwts_file: ./mlp.0.wts 8 log_file: ./mlp.0.log 9 initwts_file: ../dataset/mlp.0.wts 10 train_sent_range: 0-1000 11 cv_sent_range: 0-0 12 fea_dim: 39 13 fea_context: 11 14 bunchsize: 1024 15 train_cache: 102400 16 init_randem_seed: 27863875 17 targ_offset: 5 18 init_randem_weight_max: 0.100000 19 init_randem_weight_min: -0.100000 20 init_randem_bias_max: 0.100000 21 init_randem_bias_min: -0.100000 22 learnrate: 0.768000 23 layersizes: 429,2048,2048,2048,2048,2048,2048,8991, 24 nmod: 1 25 remainder: 0 26 discard_prob: 0.000000 27 discardLabs: 28 Please check... 29 Get pfile info over: Training data has 390729 frames, 1001 sentences. 30 Get chunk info over: Training sentences have 4 chunks, 380689 samples. 31 Get cv chunk info over: CV sentences have 1 chunks, 73 samples. 32 start training: 33 --chunk(0) : containing samples 72704 34 --chunk(1) : containing samples 102400 35 --chunk(2) : containing samples 102400 36 --chunk(3) : containing samples 102400 37 training over 38 total time cost: 2197670.000000ms