极度视界

训练神经网络片段： { # Train the MLP train_ali=exp/tri4b_train_ali cv_ali=exp/tri4b_dev_ali $cuda_cmd $dir/log/train_nnet.log \ steps/train_nnet.sh --copy-feats false \ --apply-cmvn true --norm-vars tru

通过src/util/parse-options.h 中定义的ParseOptions类来解析参数 Shell脚本的使用： src/nnetbin/nnet-train-frmshuff.cc 用来训练神经网络 即上图 train_tool=nnet-train-frmshuff NnetTrainOptions trn_opts; trn_o

IBM系： 【1】B. Kingsbury, “Lattice-Based Optimization of Sequence Classification Criteria for Neural-Network Acoustic Modeling,” in Proc. ICASSP,2009. 【2】Kingsbury B, Sainath T N, Soltau H. Scalable Mi