1NGRAMPERPLEXITY(1)               User Commands              NGRAMPERPLEXITY(1)
2
3
4

NAME

6       ngramperplexity - manual page for ngramperplexity 1.3.14
7

DESCRIPTION

9       Apply n-gram model to input FST archive.
10
11              Usage: ngramperplexity [--options] ngram.fst [in.far [out.txt]]
12
13       PROGRAM FLAGS:
14
15       --OOV_class_size: type = double, default = 10000
16
17              Number of members of OOV class --OOV_probability: type = double,
18              default = 0 Unigram probability for OOVs  --OOV_symbol:  type  =
19              std::string,  default  = "" Existing symbol for OOV class --con‐
20              text_pattern: type = std::string, default = "" Restrict perplex‐
21              ity  computation to contexts defined by pattern (default: no re‐
22              striction) --use_phimatcher: type = bool, default  =  false  Use
23              phi matcher and composition
24
25       LIBRARY FLAGS:
26
27       Flags from: compile-strings.cc
28
29       --far_field_separator: type = std::string, default = "
30              "
31
32              Set of characters used as a separator between printed fields
33
34       Flags from: flags.cc
35
36       --help: type = bool, default = false
37
38              show usage information --helpshort: type = bool, default = false
39              show brief usage information --tmpdir: type =  std::string,  de‐
40              fault  = "/tmp" temporary directory --v: type = int32_t, default
41              = 0 verbosity level
42
43       Flags from: fst.cc
44
45       --fst_align: type = bool, default = false
46
47              Write FST data aligned where appropriate --fst_default_cache_gc:
48              type  =  bool, default = true Enable garbage collection of cache
49              --fst_default_cache_gc_limit: type = int64_t, default =  1048576
50              Cache    byte    size    that    triggers   garbage   collection
51              --fst_read_mode: type = std::string, default  =  "read"  Default
52              file  reading  mode  for mappable files --fst_verify_properties:
53              type = bool, default = false Verify FST  properties  queried  by
54              TestProperties  --save_relabel_ipairs:  type  = std::string, de‐
55              fault =  ""  Save  input  relabel  pairs  to  file  --save_rela‐
56              bel_opairs: type = std::string, default = "" Save output relabel
57              pairs to file
58
59       Flags from: ngram-output.cc
60
61       --end_symbol: type = std::string, default = "</s>"
62
63              Class label for sentence end --start_symbol: type = std::string,
64              default = "<s>" Class label for sentence start
65
66       Flags from: symbol-table.cc
67
68       --fst_compat_symbols: type = bool, default = true
69
70              Require symbol tables to match when appropriate --fst_field_sep‐
71              arator: type = std::string, default = "         " Set of charac‐
72              ters used as a separator between printed fields
73
74       Flags from: util.cc
75
76       --fst_error_fatal: type = bool, default = true
77
78              FST  errors are fatal; o.w. return objects flagged as bad: e.g.,
79              FSTs:  kError  property  set,  FST  weights:  not   a   Member()
80              --ngram_error_fatal:  type  =  bool, default = true NGram errors
81              are fatal if true; otherwise returns  objects  flagged  as  bad:
82              e.g., NGramModel::Error() is true
83
84       Flags from: weight.cc
85
86       --fst_weight_parentheses: type = std::string, default = ""
87
88              Characters  enclosing  the  first  weight of a printed composite
89              weight (e.g., pair weight, tuple weight and derived classes)  to
90              ensure  proper I/O of nested composite weights; must have size 0
91              (none) or 2 (open and close parenthesis) --fst_weight_separator:
92              type  =  std::string,  default = "," Character separator between
93              printed composite weights; must be a single character
94
95
96
97ngramperplexity 1.3.14           February 2022              NGRAMPERPLEXITY(1)
Impressum