This is a repository for all finite-state machines that are compatible with the Carmel toolkit (http://www.isi.edu/licensed-sw/carmel).
| FSA | Description |
|---|---|
| none yet | none yet |
| WFSA | Description |
|---|---|
| wfsa001 | English letter 2-gram model, P(e) |
| wfsa002 | English phoneme 3-gram model, P(epron); plus generic n-gram WFSA builder |
| wfsa003 | Letter 3-gram models for dozens of languages, built from UNDHR data |
| wfsa004 | Letter 3-gram model for English, built from Penn Treebank plaintext |
| wfsa005 | Pronounceable grapheme segment 3-gram model for English, built from the aligned CMU Pronunciation Dictionary |
| wfsa006 | Accepts any word of alphabet [A-Z] except for the word "FILTER." Includes code for building any such filter. |
| FST | Description |
|---|---|
| fst001 | English-to-Cipher letter substitutor, fully connected |
| fst002 | English-to-Phonemes, built from CMU Pron Dict |
| fst002a | English-to-Phonemes, with fixes |
| fst003 | Pinyin-if-to-pinyin, bi-directional |
| WFST | Description |
|---|---|
| wfst001 | English-to-French word-to-word translator, P(f|e) |
| wfst002 | French-to-English word-to-word translator, P(e|f) |
| wfst003 | Pinyin-to-Chinese bi-directional P(p|c) and P(c|p) |
| wfst004 | Pinyin-if-to-Epron, Pinyin initial/final to IPA (English Pronounciation) |
| wfst005 | English-to-Syllable-Stress, includes original CMU pronunciation lexicon |
| wfst006 | English phonemes to grapheme (spell out), useful for OOV |
| wfst007 | English graphemes to pronounceable grapheme segment sequences |
| wfst008 | Adds start and end tags to English pronounceable grapheme segment sequences |
| wfst009 | English pronounceable grapheme segment sequences to phoneme (pronouncer), useful for OOV |
| Cascade | Description |
|---|---|
| casc001 | English-to-English word-to-word paraphraser, P(e|e) |
| casc002 | English letter-substitution cipher solver |
| casc003 | English-to-Chinglish |
| casc004 | English word to pronounciation |
| casc005 | Symbol clustering using EM (any # of clusters, any n-gram order) |