dcd recog

dcd-recog

SYNOPSIS

dcd-decode [options] transition-rspecifier fsts-rspecifier features-rspecifier response-wspecifier

DESCRIPTION

dcd-recog is the main decoding command. It accepts a transition models, component search fsts and acoustic features or pre-computed acoustic scores and outputs the utterance hypotheses.

OPTIONS

--acoustic_lookahead : (float, default = 0)

--acoustic_scale : (float, default = 0.1)

--band : (int, default = 2147483647)

--beam : (float, default = 2.14748e+09)

--colorize : (bool, default = true)

--early_mission : (bool, default = false)

--fst_reset_period : (int, default = 32)

--gc_period : (int, default = 25)

--gen_lattice : (bool, default = false)

--insertion_penalty : (float, default = 0)

--lattice_beam : (float, default = 2.14748e+09)

--lingware : (string, default = "")

--logfile : /dev/stderr (string, default = "/dev/stderr")

--nbest : (int, default = 0)

--prune_eps : (bool, default = true)

--tmtype : Type of transition model to use (string, default = "generic")

--trans_scale : (float, default = 0.1)

--use_lattice_pool : (bool, default = false)

--use_search_pool : (bool, default = false)

--word_symbols_table : (string, default = "")

DIAGNOSTICS

--cache_dest_states : (bool, default = true)

--dump_traceback : (bool, default = false)

--gc_check : (bool, default = false)