#!/bin/bash
  
  # Copyright 2012-2014  Johns Hopkins University (Author: Daniel Povey).  Apache 2.0.
  
  # This script trains a fairly vanilla network with tanh nonlinearities.
  
  # train_tanh_fast.sh is a new, improved version of train_tanh.sh, which uses
  # the 'online' preconditioning method.  For GPUs it's about two times faster
  # than before (although that's partly due to optimizations that will also help
  # the old recipe), and for CPUs it gives better performance than the old method
  # (I believe); also, the difference in optimization performance between CPU and
  # GPU is almost gone.  The old train_tanh.sh script is now deprecated.
  # We made this a separate script because not all of the options that the
  # old script accepted, are still accepted.
  
  
  # Begin configuration section.
  cmd=run.pl
  num_epochs=15      # Number of epochs during which we reduce
                     # the learning rate; number of iteration is worked out from this.
  num_epochs_extra=5 # Number of epochs after we stop reducing
                     # the learning rate.
  num_iters_final=20 # Maximum number of final iterations to give to the
                     # optimization over the validation set.
  initial_learning_rate=0.04
  final_learning_rate=0.004
  bias_stddev=0.5
  shrink_interval=5 # shrink every $shrink_interval iters except while we are
                    # still adding layers, when we do it every iter.
  shrink=true
  num_frames_shrink=2000 # note: must be <= --num-frames-diagnostic option to get_egs.sh, if
                         # given.
  final_learning_rate_factor=0.5 # Train the two last layers of parameters half as
                                 # fast as the other layers, by default.
  
  hidden_layer_dim=300 #  You may want this larger, e.g. 1024 or 2048.
  
  minibatch_size=128 # by default use a smallish minibatch size for neural net
                     # training; this controls instability which would otherwise
                     # be a problem with multi-threaded update.
  
  samples_per_iter=200000 # each iteration of training, see this many samples
                          # per job.  This option is passed to get_egs.sh.
  num_jobs_nnet=8    # Number of neural net jobs to run in parallel.  This option
                     # is passed to get_egs.sh.
  get_egs_stage=0
  
  shuffle_buffer_size=5000 # This "buffer_size" variable controls randomization of
                  # the samples on each iter.  You could set it to 0 or to a large
                  # value for complete randomization, but this would both consume
                  # memory and cause spikes in disk I/O.  Smaller is easier on
                  # disk and memory but less random.  It's not a huge deal though,
                  # as samples are anyway randomized right at the start.
  
  add_layers_period=2 # by default, add new layers every 2 iterations.
  num_hidden_layers=3 # This is an important configuration value that you might
                      # want to tune.
  stage=-5
  
  io_opts="--max-jobs-run 5" # for jobs with a lot of I/O, limits the number running at one time.   These don't
  splice_width=4 # meaning +- 4 frames on each side for second LDA
  randprune=4.0 # speeds up LDA.
  alpha=4.0 # relates to preconditioning.
  update_period=4 # relates to online preconditioning: says how often we update the subspace.
  num_samples_history=2000 # relates to online preconditioning
  max_change_per_sample=0.075
  # we make the [input, output] ranks less different for the tanh setup than for
  # the pnorm setup, as we don't have the difference in dimensions to deal with.
  precondition_rank_in=30  # relates to online preconditioning
  precondition_rank_out=60 # relates to online preconditioning
  mix_up=0 # Number of components to mix up to (should be > #tree leaves, if
           # specified.)
  num_threads=16
  parallel_opts="--num-threads 16 --mem 1G" # by default we use 16 threads; this lets the queue know.
    # note: parallel_opts doesn't automatically get adjusted if you adjust num-threads.
  combine_parallel_opts="--num-threads 8"  # queue options for the "combine" stage.
  combine_num_threads=8
  cleanup=true
  egs_dir=
  lda_opts=
  egs_opts=
  transform_dir=
  cmvn_opts=  # will be passed to get_lda.sh and get_egs.sh, if supplied.
              # only relevant for "raw" features, not lda.
  feat_type=  # Can be used to force "raw" features.
  prior_subset_size=10000 # 10k samples per job, for computing priors.  Should be
                          # more than enough.
  # End configuration section.
  
  
  echo "$0 $@"  # Print the command line for logging
  
  if [ -f path.sh ]; then . ./path.sh; fi
  . parse_options.sh || exit 1;
  
  
  if [ $# != 4 ]; then
    echo "Usage: $0 [opts] <data> <lang> <ali-dir> <exp-dir>"
    echo " e.g.: $0 data/train data/lang exp/tri3_ali exp/tri4_nnet"
    echo ""
    echo "Main options (for others, see top of script file)"
    echo "  --config <config-file>                           # config file containing options"
    echo "  --cmd (utils/run.pl|utils/queue.pl <queue opts>) # how to run jobs."
    echo "  --num-epochs <#epochs|15>                        # Number of epochs of main training"
    echo "                                                   # while reducing learning rate (determines #iterations, together"
    echo "                                                   # with --samples-per-iter and --num-jobs-nnet)"
    echo "  --num-epochs-extra <#epochs-extra|5>             # Number of extra epochs of training"
    echo "                                                   # after learning rate fully reduced"
    echo "  --initial-learning-rate <initial-learning-rate|0.02> # Learning rate at start of training, e.g. 0.02 for small"
    echo "                                                       # data, 0.01 for large data"
    echo "  --final-learning-rate  <final-learning-rate|0.004>   # Learning rate at end of training, e.g. 0.004 for small"
    echo "                                                   # data, 0.001 for large data"
    echo "  --num-hidden-layers <#hidden-layers|2>           # Number of hidden layers, e.g. 2 for 3 hours of data, 4 for 100hrs"
    echo "  --initial-num-hidden-layers <#hidden-layers|1>   # Number of hidden layers to start with."
    echo "  --add-layers-period <#iters|2>                   # Number of iterations between adding hidden layers"
    echo "  --mix-up <#pseudo-gaussians|0>                   # Can be used to have multiple targets in final output layer,"
    echo "                                                   # per context-dependent state.  Try a number several times #states."
    echo "  --num-jobs-nnet <num-jobs|8>                     # Number of parallel jobs to use for main neural net"
    echo "                                                   # training (will affect results as well as speed; try 8, 16)"
    echo "                                                   # Note: if you increase this, you may want to also increase"
    echo "                                                   # the learning rate."
    echo "  --num-threads <num-threads|16>                   # Number of parallel threads per job (will affect results"
    echo "                                                   # as well as speed; may interact with batch size; if you increase"
    echo "                                                   # this, you may want to decrease the batch size."
    echo "  --parallel-opts <opts|\"--num-threads 16 --mem 1G\">      # extra options to pass to e.g. queue.pl for processes that"
    echo "                                                   # use multiple threads... "
    echo "  --io-opts <opts|\"--max-jobs-run 10\">                      # Options given to e.g. queue.pl for jobs that do a lot of I/O."
    echo "  --minibatch-size <minibatch-size|128>            # Size of minibatch to process (note: product with --num-threads"
    echo "                                                   # should not get too large, e.g. >2k)."
    echo "  --samples-per-iter <#samples|200000>             # Number of samples of data to process per iteration, per"
    echo "                                                   # process."
    echo "  --splice-width <width|4>                         # Number of frames on each side to append for feature input"
    echo "                                                   # (note: we splice processed, typically 40-dimensional frames"
    echo "  --lda-dim <dim|250>                              # Dimension to reduce spliced features to with LDA"
    echo "  --num-iters-final <#iters|20>                    # Number of final iterations to give to nnet-combine-fast to "
    echo "                                                   # interpolate parameters (the weights are learned with a validation set)"
    echo "  --stage <stage|-9>                               # Used to run a partially-completed training process from somewhere in"
    echo "                                                   # the middle."
  
    exit 1;
  fi
  
  data=$1
  lang=$2
  alidir=$3
  dir=$4
  
  # Check some files.
  for f in $data/feats.scp $lang/L.fst $alidir/ali.1.gz $alidir/final.mdl $alidir/tree; do
    [ ! -f $f ] && echo "$0: no such file $f" && exit 1;
  done
  
  
  # Set some variables.
  num_leaves=`am-info $alidir/final.mdl 2>/dev/null | awk '/number of pdfs/{print $NF}'` || exit 1;
  
  nj=`cat $alidir/num_jobs` || exit 1;  # number of jobs in alignment dir...
  # in this dir we'll have just one job.
  sdata=$data/split$nj
  utils/split_data.sh $data $nj
  
  mkdir -p $dir/log
  cp $alidir/tree $dir
  
  utils/lang/check_phones_compatible.sh $lang/phones.txt $alidir/phones.txt || exit 1;
  cp $lang/phones.txt $dir || exit 1;
  
  extra_opts=()
  [ ! -z "$cmvn_opts" ] && extra_opts+=(--cmvn-opts "$cmvn_opts")
  [ ! -z "$feat_type" ] && extra_opts+=(--feat-type $feat_type)
  [ ! -z "$online_ivector_dir" ] && extra_opts+=(--online-ivector-dir $online_ivector_dir)
  [ -z "$transform_dir" ] && transform_dir=$alidir
  extra_opts+=(--transform-dir $transform_dir)
  extra_opts+=(--splice-width $splice_width)
  
  if [ $stage -le -4 ]; then
    echo "$0: calling get_lda.sh"
    steps/nnet2/get_lda.sh $lda_opts "${extra_opts[@]}" --cmd "$cmd" $data $lang $alidir $dir || exit 1;
  fi
  
  # these files will have been written by get_lda.sh
  feat_dim=$(cat $dir/feat_dim) || exit 1;
  ivector_dim=$(cat $dir/ivector_dim) || exit 1;
  lda_dim=$(cat $dir/lda_dim) || exit 1;
  
  if [ $stage -le -3 ] && [ -z "$egs_dir" ]; then
    echo "$0: calling get_egs.sh"
    steps/nnet2/get_egs.sh $egs_opts "${extra_opts[@]}" \
        --samples-per-iter $samples_per_iter \
        --num-jobs-nnet $num_jobs_nnet --stage $get_egs_stage \
        --cmd "$cmd" $egs_opts --io-opts "$io_opts" \
        $data $lang $alidir $dir || exit 1;
  fi
  if [ -z $egs_dir ]; then
    egs_dir=$dir/egs
  fi
  
  iters_per_epoch=`cat $egs_dir/iters_per_epoch`  || exit 1;
  ! [ $num_jobs_nnet -eq `cat $egs_dir/num_jobs_nnet` ] && \
    echo "$0: Warning: using --num-jobs-nnet=`cat $egs_dir/num_jobs_nnet` from $egs_dir"
  num_jobs_nnet=`cat $egs_dir/num_jobs_nnet` || exit 1;
  
  
  if ! [ $num_hidden_layers -ge 1 ]; then
    echo "Invalid num-hidden-layers $num_hidden_layers"
    exit 1
  fi
  
  if [ $stage -le -2 ]; then
    echo "$0: initializing neural net";
  
    # Get spk-vec dim (in case we're using them).
    lda_mat=$dir/lda.mat
    tot_input_dim=$[$feat_dim+$ivector_dim]
  
    online_preconditioning_opts="alpha=$alpha num-samples-history=$num_samples_history update-period=$update_period rank-in=$precondition_rank_in rank-out=$precondition_rank_out max-change-per-sample=$max_change_per_sample"
  
    stddev=`perl -e "print 1.0/sqrt($hidden_layer_dim);"`
    cat >$dir/nnet.config <<EOF
  SpliceComponent input-dim=$tot_input_dim left-context=$splice_width right-context=$splice_width const-component-dim=$ivector_dim
  FixedAffineComponent matrix=$lda_mat
  AffineComponentPreconditionedOnline input-dim=$lda_dim output-dim=$hidden_layer_dim $online_preconditioning_opts learning-rate=$initial_learning_rate param-stddev=$stddev bias-stddev=$bias_stddev
  TanhComponent dim=$hidden_layer_dim
  AffineComponentPreconditionedOnline input-dim=$hidden_layer_dim output-dim=$num_leaves $online_preconditioning_opts learning-rate=$initial_learning_rate param-stddev=0 bias-stddev=0
  SoftmaxComponent dim=$num_leaves
  EOF
  
    # to hidden.config it will write the part of the config corresponding to a
    # single hidden layer; we need this to add new layers.
    cat >$dir/hidden.config <<EOF
  AffineComponentPreconditionedOnline input-dim=$hidden_layer_dim output-dim=$hidden_layer_dim $online_preconditioning_opts learning-rate=$initial_learning_rate param-stddev=$stddev bias-stddev=$bias_stddev
  TanhComponent dim=$hidden_layer_dim
  EOF
    $cmd $dir/log/nnet_init.log \
      nnet-am-init $alidir/tree $lang/topo "nnet-init $dir/nnet.config -|" \
      $dir/0.mdl || exit 1;
  fi
  
  if [ $stage -le -1 ]; then
    echo "Training transition probabilities and setting priors"
    $cmd $dir/log/train_trans.log \
      nnet-train-transitions $dir/0.mdl "ark:gunzip -c $alidir/ali.*.gz|" $dir/0.mdl \
      || exit 1;
  fi
  
  num_iters_reduce=$[$num_epochs * $iters_per_epoch];
  num_iters_extra=$[$num_epochs_extra * $iters_per_epoch];
  num_iters=$[$num_iters_reduce+$num_iters_extra]
  
  echo "$0: Will train for $num_epochs + $num_epochs_extra epochs, equalling "
  echo "$0: $num_iters_reduce + $num_iters_extra = $num_iters iterations, "
  echo "$0: (while reducing learning rate) + (with constant learning rate)."
  
  # This is when we decide to mix up from: halfway between when we've finished
  # adding the hidden layers and the end of training.
  finish_add_layers_iter=$[$num_hidden_layers * $add_layers_period]
  mix_up_iter=$[($num_iters + $finish_add_layers_iter)/2]
  
  if [ $num_threads -eq 1 ]; then
    train_suffix="-simple" # this enables us to use GPU code if
                           # we have just one thread.
    if ! cuda-compiled; then
      echo "$0: WARNING: you are running with one thread but you have not compiled"
      echo "   for CUDA.  You may be running a setup optimized for GPUs.  If you have"
      echo "   GPUs and have nvcc installed, go to src/ and do ./configure; make"
    fi
  else
    train_suffix="-parallel --num-threads=$num_threads"
  fi
  
  
  
  x=0
  while [ $x -lt $num_iters ]; do
    if [ $x -ge 0 ] && [ $stage -le $x ]; then
      # Set off jobs doing some diagnostics, in the background.
      $cmd $dir/log/compute_prob_valid.$x.log \
        nnet-compute-prob $dir/$x.mdl ark:$egs_dir/valid_diagnostic.egs &
      $cmd $dir/log/compute_prob_train.$x.log \
        nnet-compute-prob $dir/$x.mdl ark:$egs_dir/train_diagnostic.egs &
      if [ $x -gt 0 ] && [ ! -f $dir/log/mix_up.$[$x-1].log ]; then
        $cmd $dir/log/progress.$x.log \
          nnet-show-progress --use-gpu=no $dir/$[$x-1].mdl $dir/$x.mdl \
           ark:$egs_dir/train_diagnostic.egs '&&' \
           nnet-am-info $dir/$x.mdl &
      fi
  
      echo "Training neural net (pass $x)"
      if [ $x -gt 0 ] && \
        [ $x -le $[($num_hidden_layers-1)*$add_layers_period] ] && \
        [ $[($x-1) % $add_layers_period] -eq 0 ]; then
        mdl="nnet-init --srand=$x $dir/hidden.config - | nnet-insert $dir/$x.mdl - - |"
      else
        mdl=$dir/$x.mdl
      fi
  
      if [ $x -eq 0 ] || [ "$mdl" != "$dir/$x.mdl" ]; then
        # on iteration zero or when we just added a layer, use a smaller minibatch
        # size and just one job: the model-averaging doesn't seem to be helpful
        # when the model is changing too fast (i.e. it worsens the objective
        # function), and the smaller minibatch size will help to keep
        # the update stable.
        this_minibatch_size=$[$minibatch_size/2];
        do_average=false
      else
        this_minibatch_size=$minibatch_size
        do_average=true
      fi
  
      $cmd $parallel_opts JOB=1:$num_jobs_nnet $dir/log/train.$x.JOB.log \
        nnet-shuffle-egs --buffer-size=$shuffle_buffer_size --srand=$x \
        ark:$egs_dir/egs.JOB.$[$x%$iters_per_epoch].ark ark:- \| \
        nnet-train$train_suffix \
           --minibatch-size=$this_minibatch_size --srand=$x "$mdl" \
          ark:- $dir/$[$x+1].JOB.mdl \
        || exit 1;
  
      nnets_list=
      for n in `seq 1 $num_jobs_nnet`; do
        nnets_list="$nnets_list $dir/$[$x+1].$n.mdl"
      done
  
      learning_rate=`perl -e '($x,$n,$i,$f)=@ARGV; print ($x >= $n ? $f : $i*exp($x*log($f/$i)/$n));' $[$x+1] $num_iters_reduce $initial_learning_rate $final_learning_rate`;
      last_layer_learning_rate=`perl -e "print $learning_rate * $final_learning_rate_factor;"`;
      nnet-am-info $dir/$[$x+1].1.mdl > $dir/foo  2>/dev/null || exit 1
      nu=`cat $dir/foo | grep num-updatable-components | awk '{print $2}'`
      na=`cat $dir/foo | grep -v Fixed | grep AffineComponent | wc -l`
      # na is number of last updatable AffineComponent layer [one-based, counting only
      # updatable components.]
      # The last two layers will get this (usually lower) learning rate.
      lr_string="$learning_rate"
      for n in `seq 2 $nu`; do
        if [ $n -eq $na ] || [ $n -eq $[$na-1] ]; then lr=$last_layer_learning_rate;
        else lr=$learning_rate; fi
        lr_string="$lr_string:$lr"
      done
  
      if $do_average; then
        $cmd $dir/log/average.$x.log \
          nnet-am-average $nnets_list - \| \
          nnet-am-copy --learning-rates=$lr_string - $dir/$[$x+1].mdl || exit 1;
      else
        n=$(perl -e '($nj,$pat)=@ARGV; $best_n=1; $best_logprob=-1.0e+10; for ($n=1;$n<=$nj;$n++) {
            $fn = sprintf($pat,$n); open(F, "<$fn") || die "Error opening log file $fn";
            undef $logprob; while (<F>) { if (m/log-prob-per-frame=(\S+)/) { $logprob=$1; } }
            close(F); if (defined $logprob && $logprob > $best_logprob) { $best_logprob=$logprob;
            $best_n=$n; } } print "$best_n
  "; ' $num_jobs_nnet $dir/log/train.$x.%d.log) || exit 1;
        [ -z "$n" ] && echo "Error getting best model" && exit 1;
        $cmd $dir/log/select.$x.log \
          nnet-am-copy --learning-rates=$lr_string $dir/$[$x+1].$n.mdl $dir/$[$x+1].mdl || exit 1;
      fi
  
      if $shrink && [ $[$x % $shrink_interval] -eq 0 ]; then
        mb=$[($num_frames_shrink+$num_threads-1)/$num_threads]
        $cmd $combine_parallel_opts $dir/log/shrink.$x.log \
          nnet-subset-egs --n=$num_frames_shrink --randomize-order=true --srand=$x \
            ark:$egs_dir/train_diagnostic.egs ark:-  \| \
          nnet-combine-fast --use-gpu=no --num-threads=$combine_num_threads \
            --verbose=3 --minibatch-size=$mb \
            $dir/$[$x+1].mdl ark:- $dir/$[$x+1].mdl || exit 1;
      else
        # On other iters, do nnet-am-fix which is much faster and has roughly
        # the same effect.
        nnet-am-fix $dir/$[$x+1].mdl $dir/$[$x+1].mdl 2>$dir/log/fix.$x.log
      fi
  
      if [ "$mix_up" -gt 0 ] && [ $x -eq $mix_up_iter ]; then
        # mix up.
        echo Mixing up from $num_leaves to $mix_up components
        $cmd $dir/log/mix_up.$x.log \
          nnet-am-mixup --min-count=10 --num-mixtures=$mix_up \
          $dir/$[$x+1].mdl $dir/$[$x+1].mdl || exit 1;
      fi
      rm $nnets_list
    fi
    x=$[$x+1]
  done
  
  # Now do combination.
  # At the end, final.mdl will be a combination of the last e.g. 10 models.
  nnets_list=()
  if [ $num_iters_final -gt $num_iters_extra ]; then
    echo "Setting num_iters_final=$num_iters_extra"
  fi
  start=$[$num_iters-$num_iters_final+1]
  for x in `seq $start $num_iters`; do
    idx=$[$x-$start]
    if [ $x -gt $mix_up_iter ]; then
      nnets_list[$idx]=$dir/$x.mdl # "nnet-am-copy --remove-dropout=true $dir/$x.mdl - |"
    fi
  done
  
  if [ $stage -le $num_iters ]; then
    echo "Doing final combination to produce final.mdl"
    # Below, use --use-gpu=no to disable nnet-combine-fast from using a GPU, as if
    # there are many models it can give out-of-memory error on the GPU; set
    # num-threads to 8 to speed it up (this isn't ideal...)
    num_egs=`nnet-copy-egs ark:$egs_dir/combine.egs ark:/dev/null 2>&1 | tail -n 1 | awk '{print $NF}'`
    mb=$[($num_egs+$combine_num_threads-1)/$combine_num_threads]
    [ $mb -gt 512 ] && mb=512
    $cmd $combine_parallel_opts $dir/log/combine.log \
      nnet-combine-fast --use-gpu=no --num-threads=$combine_num_threads \
        --verbose=3 --minibatch-size=$mb "${nnets_list[@]}" ark:$egs_dir/combine.egs \
        $dir/final.mdl || exit 1;
  
    # Compute the probability of the final, combined model with
    # the same subset we used for the previous compute_probs, as the
    # different subsets will lead to different probs.
    $cmd $dir/log/compute_prob_valid.final.log \
      nnet-compute-prob $dir/final.mdl ark:$egs_dir/valid_diagnostic.egs &
    $cmd $dir/log/compute_prob_train.final.log \
      nnet-compute-prob $dir/final.mdl ark:$egs_dir/train_diagnostic.egs &
  fi
  
  if [ $stage -le $[$num_iters+1] ]; then
    echo "Getting average posterior for purposes of adjusting the priors."
    # Note: this just uses CPUs, using a smallish subset of data.
    rm $dir/post.*.vec 2>/dev/null
    $cmd JOB=1:$num_jobs_nnet $dir/log/get_post.JOB.log \
      nnet-subset-egs --n=$prior_subset_size ark:$egs_dir/egs.JOB.0.ark ark:- \| \
      nnet-compute-from-egs "nnet-to-raw-nnet $dir/final.mdl -|" ark:- ark:- \| \
      matrix-sum-rows ark:- ark:- \| vector-sum ark:- $dir/post.JOB.vec || exit 1;
  
    sleep 3;  # make sure there is time for $dir/post.*.vec to appear.
  
    $cmd $dir/log/vector_sum.log \
     vector-sum $dir/post.*.vec $dir/post.vec || exit 1;
  
    rm $dir/post.*.vec;
  
    echo "Re-adjusting priors based on computed posteriors"
    $cmd $dir/log/adjust_priors.log \
      nnet-adjust-priors $dir/final.mdl $dir/post.vec $dir/final.mdl || exit 1;
  fi
  
  sleep 2
  
  echo Done
  
  if $cleanup; then
    echo Cleaning up data
    if [ $egs_dir == "$dir/egs" ]; then
      steps/nnet2/remove_egs.sh $dir/egs
    fi
    echo Removing most of the models
    for x in `seq 0 $num_iters`; do
      if [ $[$x%100] -ne 0 ] && [ $x -lt $[$num_iters-$num_iters_final+1] ]; then
         # delete all but every 10th model; don't delete the ones which combine to form the final model.
        rm $dir/$x.mdl
      fi
    done
  fi