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

NAME

6       ngramrandtest - manual page for ngramrandtest 1.3.14
7

DESCRIPTION

9       Generates random data/models.
10
11              Usage: ngramrandtest [--options]
12
13       PROGRAM FLAGS:
14
15       --directory: type = std::string, default = "."
16
17              directory  where  files  will  be  placed  --max_length:  type =
18              int32_t, default = 1000 Maximum sentence  length  --mean_length:
19              type  =  int32_t,  default  =  100  maximum  mean  string length
20              --ngram_max: type = int32_t, default = 3  maximum  n-gram  order
21              size  --sample_max: type = int32_t, default = 10000 maximum sam‐
22              ple corpus size --seed: type =  int32_t,  default  =  1653672809
23              Randomization seed --thresh_max: type = double, default = 3 max‐
24              imum threshold size --vars: type =  std::string,  default  =  ""
25              file  name for outputting variable values --vocabulary_max: type
26              = int32_t, default = 5000 maximum vocabulary size
27
28       LIBRARY FLAGS:
29
30       Flags from: flags.cc
31
32       --help: type = bool, default = false
33
34              show usage information --helpshort: type = bool, default = false
35              show  brief  usage information --tmpdir: type = std::string, de‐
36              fault = "/tmp" temporary directory --v: type = int32_t,  default
37              = 0 verbosity level
38
39       Flags from: fst.cc
40
41       --fst_align: type = bool, default = false
42
43              Write FST data aligned where appropriate --fst_default_cache_gc:
44              type = bool, default = true Enable garbage collection  of  cache
45              --fst_default_cache_gc_limit:  type = int64_t, default = 1048576
46              Cache   byte   size    that    triggers    garbage    collection
47              --fst_read_mode:  type  =  std::string, default = "read" Default
48              file reading mode for  mappable  files  --fst_verify_properties:
49              type  =  bool,  default = false Verify FST properties queried by
50              TestProperties --save_relabel_ipairs: type  =  std::string,  de‐
51              fault  =  ""  Save  input  relabel  pairs  to  file --save_rela‐
52              bel_opairs: type = std::string, default = "" Save output relabel
53              pairs to file
54
55       Flags from: ngram-output.cc
56
57       --end_symbol: type = std::string, default = "</s>"
58
59              Class label for sentence end --start_symbol: type = std::string,
60              default = "<s>" Class label for sentence start
61
62       Flags from: symbol-table.cc
63
64       --fst_compat_symbols: type = bool, default = true
65
66              Require symbol tables to match when appropriate --fst_field_sep‐
67              arator: type = std::string, default = "         " Set of charac‐
68              ters used as a separator between printed fields
69
70       Flags from: util.cc
71
72       --fst_error_fatal: type = bool, default = true
73
74              FST errors are fatal; o.w. return objects flagged as bad:  e.g.,
75              FSTs:   kError   property  set,  FST  weights:  not  a  Member()
76              --ngram_error_fatal: type = bool, default =  true  NGram  errors
77              are  fatal  if  true;  otherwise returns objects flagged as bad:
78              e.g., NGramModel::Error() is true
79
80       Flags from: weight.cc
81
82       --fst_weight_parentheses: type = std::string, default = ""
83
84              Characters enclosing the first weight  of  a  printed  composite
85              weight  (e.g., pair weight, tuple weight and derived classes) to
86              ensure proper I/O of nested composite weights; must have size  0
87              (none) or 2 (open and close parenthesis) --fst_weight_separator:
88              type = std::string, default = ","  Character  separator  between
89              printed composite weights; must be a single character
90
91
92
93ngramrandtest 1.3.14             February 2022                NGRAMRANDTEST(1)
Impressum