Instructions to use MontrealCorpusTools/czech_mfa with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Montreal Forced Aligner
How to use MontrealCorpusTools/czech_mfa with Montreal Forced Aligner:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
Upload folder using huggingface_hub
Browse files- README.md +15 -19
- acoustic/final.alimdl +2 -2
- acoustic/final.mdl +2 -2
- acoustic/lda.mat +0 -0
- acoustic/meta.json +1 -1
- acoustic/tree +2 -2
- dictionary/czech_mfa.dict +0 -0
- dictionary/rules.yaml +26 -26
- g2p/czech_mfa/graphemes.txt +5 -5
- g2p/czech_mfa/meta.json +1 -1
- g2p/czech_mfa/model.fst +2 -2
README.md
CHANGED
|
@@ -109,9 +109,9 @@ To get started, follow the instructions for [installing MFA](https://montreal-fo
|
|
| 109 |
| **Open** | | | a aː | | |
|
| 110 |
|
| 111 |
##### Diphthongs
|
| 112 |
-
* aw
|
| 113 |
* ow
|
| 114 |
* ew
|
|
|
|
| 115 |
|
| 116 |
|
| 117 |
## Training Details
|
|
@@ -125,53 +125,49 @@ To get started, follow the instructions for [installing MFA](https://montreal-fo
|
|
| 125 |
- **Source:** https://voice.mozilla.org/en/datasets
|
| 126 |
- **License:** [CC-0](https://creativecommons.org/publicdomain/zero/1.0/)
|
| 127 |
- **Dialects:** N/A
|
| 128 |
-
- **Number of hours:**
|
| 129 |
-
- **Number of utterances:**
|
| 130 |
-
- **Number of speakers:**
|
| 131 |
-
- **Female speakers:**
|
| 132 |
-
- **Male speakers:**
|
| 133 |
-
- **Unknown speakers:**
|
| 134 |
-
|
| 135 |
|
| 136 |
#### Czech Parliament Meetings
|
| 137 |
|
| 138 |
- **Source:** https://lindat.mff.cuni.cz/repository/xmlui/handle/11858/00-097C-0000-0005-CF9C-4
|
| 139 |
- **License:** [CC BY-NC-ND 3.0](https://creativecommons.org/licenses/by-nc-nd/3.0/)
|
| 140 |
- **Dialects:** N/A
|
| 141 |
-
- **Number of hours:**
|
| 142 |
-
- **Number of utterances:** 51,
|
| 143 |
-
- **Number of speakers:**
|
| 144 |
- **Female speakers:** 0
|
| 145 |
- **Male speakers:** 0
|
| 146 |
- **Unknown speakers:** 67
|
| 147 |
|
| 148 |
-
|
| 149 |
#### GlobalPhone Czech
|
| 150 |
|
| 151 |
- **Source:** https://catalogue.elra.info/en-us/repository/browse/ELRA-S0196/
|
| 152 |
- **License:** [ELRA](https://www.elra.info/en/services-around-lrs/distribution/licensing/)
|
| 153 |
- **Dialects:** N/A
|
| 154 |
-
- **Number of hours:** 31.
|
| 155 |
-
- **Number of utterances:** 12,
|
| 156 |
- **Number of speakers:** 101
|
| 157 |
- **Female speakers:** 45
|
| 158 |
- **Male speakers:** 57
|
| 159 |
- **Unknown speakers:** 0
|
| 160 |
|
| 161 |
-
|
| 162 |
#### Large Corpus of Czech Parliament Plenary Hearings
|
| 163 |
|
| 164 |
- **Source:** https://lindat.mff.cuni.cz/repository/xmlui/handle/11234/1-3126
|
| 165 |
- **License:** [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)
|
| 166 |
- **Dialects:** N/A
|
| 167 |
-
- **Number of hours:**
|
| 168 |
-
- **Number of utterances:**
|
| 169 |
-
- **Number of speakers:**
|
| 170 |
- **Female speakers:** 0
|
| 171 |
- **Male speakers:** 0
|
| 172 |
- **Unknown speakers:** 84
|
| 173 |
|
| 174 |
-
|
| 175 |
### Training Procedure
|
| 176 |
|
| 177 |
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
|
|
|
| 109 |
| **Open** | | | a aː | | |
|
| 110 |
|
| 111 |
##### Diphthongs
|
|
|
|
| 112 |
* ow
|
| 113 |
* ew
|
| 114 |
+
* aw
|
| 115 |
|
| 116 |
|
| 117 |
## Training Details
|
|
|
|
| 125 |
- **Source:** https://voice.mozilla.org/en/datasets
|
| 126 |
- **License:** [CC-0](https://creativecommons.org/publicdomain/zero/1.0/)
|
| 127 |
- **Dialects:** N/A
|
| 128 |
+
- **Number of hours:** 71.36
|
| 129 |
+
- **Number of utterances:** 60,276
|
| 130 |
+
- **Number of speakers:** 999
|
| 131 |
+
- **Female speakers:** 35
|
| 132 |
+
- **Male speakers:** 304
|
| 133 |
+
- **Unknown speakers:** 670
|
|
|
|
| 134 |
|
| 135 |
#### Czech Parliament Meetings
|
| 136 |
|
| 137 |
- **Source:** https://lindat.mff.cuni.cz/repository/xmlui/handle/11858/00-097C-0000-0005-CF9C-4
|
| 138 |
- **License:** [CC BY-NC-ND 3.0](https://creativecommons.org/licenses/by-nc-nd/3.0/)
|
| 139 |
- **Dialects:** N/A
|
| 140 |
+
- **Number of hours:** 80.59
|
| 141 |
+
- **Number of utterances:** 51,445
|
| 142 |
+
- **Number of speakers:** 198
|
| 143 |
- **Female speakers:** 0
|
| 144 |
- **Male speakers:** 0
|
| 145 |
- **Unknown speakers:** 67
|
| 146 |
|
|
|
|
| 147 |
#### GlobalPhone Czech
|
| 148 |
|
| 149 |
- **Source:** https://catalogue.elra.info/en-us/repository/browse/ELRA-S0196/
|
| 150 |
- **License:** [ELRA](https://www.elra.info/en/services-around-lrs/distribution/licensing/)
|
| 151 |
- **Dialects:** N/A
|
| 152 |
+
- **Number of hours:** 31.65
|
| 153 |
+
- **Number of utterances:** 12,352
|
| 154 |
- **Number of speakers:** 101
|
| 155 |
- **Female speakers:** 45
|
| 156 |
- **Male speakers:** 57
|
| 157 |
- **Unknown speakers:** 0
|
| 158 |
|
|
|
|
| 159 |
#### Large Corpus of Czech Parliament Plenary Hearings
|
| 160 |
|
| 161 |
- **Source:** https://lindat.mff.cuni.cz/repository/xmlui/handle/11234/1-3126
|
| 162 |
- **License:** [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)
|
| 163 |
- **Dialects:** N/A
|
| 164 |
+
- **Number of hours:** 438.35
|
| 165 |
+
- **Number of utterances:** 204,411
|
| 166 |
+
- **Number of speakers:** 344
|
| 167 |
- **Female speakers:** 0
|
| 168 |
- **Male speakers:** 0
|
| 169 |
- **Unknown speakers:** 84
|
| 170 |
|
|
|
|
| 171 |
### Training Procedure
|
| 172 |
|
| 173 |
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
acoustic/final.alimdl
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:13102f8708a911d4befa4759272f40042e2d1215bea95b36364e41130aadac91
|
| 3 |
+
size 50080498
|
acoustic/final.mdl
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:04418dd045f1f5cf65c5122a82b7664add311c24641cc196c49cb4d15d9a6a64
|
| 3 |
+
size 50080498
|
acoustic/lda.mat
CHANGED
|
Binary files a/acoustic/lda.mat and b/acoustic/lda.mat differ
|
|
|
acoustic/meta.json
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"phones": ["a", "aw", "aː", "b", "c", "d", "dz", "dʒ", "ew", "f", "iː", "j", "k", "l", "l̩", "m", "m̩", "n", "n̩", "o", "ow", "oː", "p", "r", "r̝", "r̝̊", "r̩", "s", "t", "ts", "tʃ", "u", "uː", "v", "x", "z", "ŋ", "ə", "ɛ", "ɛː", "ɟ", "ɡ", "ɦ", "ɪ", "ɲ", "ʃ", "ʒ", "ʔ"], "phone_mapping": {"<eps>": 0, "sil": 1, "spn": 2, "a": 3, "aw": 4, "aː": 5, "b": 6, "c": 7, "d": 8, "dz": 9, "dʒ": 10, "ew": 11, "f": 12, "iː": 13, "j": 14, "k": 15, "l": 16, "l̩": 17, "m": 18, "m̩": 19, "n": 20, "n̩": 21, "o": 22, "ow": 23, "oː": 24, "p": 25, "r": 26, "r̝": 27, "r̝̊": 28, "r̩": 29, "s": 30, "t": 31, "ts": 32, "tʃ": 33, "u": 34, "uː": 35, "v": 36, "x": 37, "z": 38, "ŋ": 39, "ə": 40, "ɛ": 41, "ɛː": 42, "ɟ": 43, "ɡ": 44, "ɦ": 45, "ɪ": 46, "ɲ": 47, "ʃ": 48, "ʒ": 49, "ʔ": 50}, "phone_groups": {"0": ["a", "aː"], "1": ["aw"], "2": ["b"], "3": ["c"], "4": ["d"], "5": ["dz"], "6": ["dʒ"], "7": ["ew"], "8": ["f"], "9": ["iː"], "10": ["j"], "11": ["k"], "12": ["l", "l̩"], "13": ["m", "m̩"], "14": ["n", "n̩"], "15": ["o", "oː"], "16": ["ow"], "17": ["p"], "18": ["r", "r̩"], "19": ["r̝"], "20": ["r̝̊"], "21": ["s"], "22": ["t"], "23": ["ts"], "24": ["tʃ"], "25": ["u", "uː"], "26": ["v"], "27": ["x"], "28": ["z"], "29": ["ŋ"], "30": ["ə"], "31": ["ɛ", "ɛː"], "32": ["ɟ"], "33": ["ɡ"], "34": ["ɦ"], "35": ["ɪ"], "36": ["ɲ"], "37": ["ʃ"], "38": ["ʒ"], "39": ["ʔ"]}, "version": "3.
|
|
|
|
| 1 |
+
{"phones": ["a", "aw", "aː", "b", "c", "d", "dz", "dʒ", "ew", "f", "iː", "j", "k", "l", "l̩", "m", "m̩", "n", "n̩", "o", "ow", "oː", "p", "r", "r̝", "r̝̊", "r̩", "s", "t", "ts", "tʃ", "u", "uː", "v", "x", "z", "ŋ", "ə", "ɛ", "ɛː", "ɟ", "ɡ", "ɦ", "ɪ", "ɲ", "ʃ", "ʒ", "ʔ"], "phone_mapping": {"<eps>": 0, "sil": 1, "spn": 2, "a": 3, "aw": 4, "aː": 5, "b": 6, "c": 7, "d": 8, "dz": 9, "dʒ": 10, "ew": 11, "f": 12, "iː": 13, "j": 14, "k": 15, "l": 16, "l̩": 17, "m": 18, "m̩": 19, "n": 20, "n̩": 21, "o": 22, "ow": 23, "oː": 24, "p": 25, "r": 26, "r̝": 27, "r̝̊": 28, "r̩": 29, "s": 30, "t": 31, "ts": 32, "tʃ": 33, "u": 34, "uː": 35, "v": 36, "x": 37, "z": 38, "ŋ": 39, "ə": 40, "ɛ": 41, "ɛː": 42, "ɟ": 43, "ɡ": 44, "ɦ": 45, "ɪ": 46, "ɲ": 47, "ʃ": 48, "ʒ": 49, "ʔ": 50}, "phone_groups": {"0": ["a", "aː"], "1": ["aw"], "2": ["b"], "3": ["c"], "4": ["d"], "5": ["dz"], "6": ["dʒ"], "7": ["ew"], "8": ["f"], "9": ["iː"], "10": ["j"], "11": ["k"], "12": ["l", "l̩"], "13": ["m", "m̩"], "14": ["n", "n̩"], "15": ["o", "oː"], "16": ["ow"], "17": ["p"], "18": ["r", "r̩"], "19": ["r̝"], "20": ["r̝̊"], "21": ["s"], "22": ["t"], "23": ["ts"], "24": ["tʃ"], "25": ["u", "uː"], "26": ["v"], "27": ["x"], "28": ["z"], "29": ["ŋ"], "30": ["ə"], "31": ["ɛ", "ɛː"], "32": ["ɟ"], "33": ["ɡ"], "34": ["ɦ"], "35": ["ɪ"], "36": ["ɲ"], "37": ["ʃ"], "38": ["ʒ"], "39": ["ʔ"]}, "version": "3.4.3", "architecture": "gmm-hmm", "train_date": "2026-10-09 05:27:16.660822", "training": {"audio_duration": 2232620.159937846, "num_speakers": 1597, "num_utterances": 327084, "num_oovs": 0, "average_log_likelihood": -52.340404924621424}, "dictionaries": {"names": ["czech_mfa"], "default": "czech_mfa", "silence_word": "<eps>", "oov_word": "<unk>", "bracketed_word": "[bracketed]", "laughter_word": "[laughter]", "clitic_marker": "'", "position_dependent_phones": false}, "language": "unknown", "tokenization": "simple", "features": {"type": "mfcc", "use_energy": false, "frame_shift": 10, "frame_length": 25, "snip_edges": false, "low_frequency": 20, "high_frequency": 7800, "sample_frequency": 16000, "dither": 0.0, "energy_floor": 0.0, "num_coefficients": 13, "num_mel_bins": 23, "cepstral_lifter": 22, "preemphasis_coefficient": 0.97, "uses_cmvn": true, "uses_deltas": true, "uses_voiced": false, "uses_splices": false, "uses_speaker_adaptation": true, "use_pitch": false, "use_voicing": false, "min_f0": 50, "max_f0": 800, "delta_pitch": 0.005, "penalty_factor": 0.1, "silence_weight": 0.0, "splice_left_context": 3, "splice_right_context": 3}, "oov_phone": "spn", "optional_silence_phone": "sil", "phone_set_type": "UNKNOWN", "silence_probability": 0.21, "initial_silence_probability": 0.21, "final_silence_correction": 0.99, "final_non_silence_correction": 0.57, "duration_information": {"a": [-2.8193848092964555, 0.5030860793221705], "aw": [-1.9058687368665967, 0.30679030426667175], "aː": [-2.170866517208258, 0.41171918130414886], "b": [-2.719864410037368, 0.3488078109263064], "c": [-2.5125444022552776, 0.36491438122606135], "d": [-2.939650138180784, 0.37527712721736417], "dz": [-2.545117439052222, 0.3908877854148026], "dʒ": [-2.379862936646152, 0.4529812485439085], "ew": [-1.8710607814891076, 0.33721105104151344], "f": [-2.602810283982693, 0.45552669253969624], "iː": [-2.6852324217770027, 0.5552634189965184], "j": [-2.9039420793237025, 0.5445582372904737], "k": [-2.5569698629959836, 0.38002587386248654], "l": [-2.8691713117037314, 0.3817738986898634], "l̩": [-2.5633676354776638, 0.5318775187333405], "m": [-2.703556572322657, 0.41198653138197394], "m̩": [-3.0506420626928556, 0.5444193618777741], "n": [-2.8584203843072564, 0.37788264511056086], "n̩": [-2.6292156092412124, 0.5056738861305238], "o": [-2.821809760509923, 0.4515621452218157], "ow": [-2.2429398328670245, 0.4794269535971944], "oː": [-2.181069490745467, 0.443336787562333], "p": [-2.6660011033942594, 0.3849627029955284], "r": [-2.9963133556199337, 0.3053609733647267], "r̝": [-2.746309590928616, 0.35066777161336543], "r̝̊": [-2.656614088432661, 0.322339482944551], "r̩": [-2.7589162336096997, 0.37456803282149603], "s": [-2.401645195909443, 0.39272196003277665], "t": [-2.673056634210402, 0.41810493108872904], "ts": [-2.269784852333614, 0.3824491297251477], "tʃ": [-2.2375056589913163, 0.31025737221157573], "u": [-2.792663012447272, 0.54684768017601], "uː": [-2.420589257706202, 0.5671872382062705], "v": [-2.8798890147087053, 0.35689339994228714], "x": [-2.559154525732009, 0.532053560772595], "z": [-2.6462112492267904, 0.371123443966969], "ŋ": [-2.666233283105511, 0.36316722377911553], "ə": [-2.505329106841459, 1.043065246322122], "ɛ": [-2.8754446436341907, 0.48337744401028837], "ɛː": [-2.3828202599435806, 0.5241173443652304], "ɟ": [-2.6679767883588554, 0.3771272999765758], "ɡ": [-2.8671022969800677, 0.41379466286098404], "ɦ": [-2.8312384712562846, 0.47796966191609497], "ɪ": [-2.828907300441165, 0.5273825642367318], "ɲ": [-2.8619399392243463, 0.443434695526496], "ʃ": [-2.346895293423586, 0.4387221396602243], "ʒ": [-2.5961585738333355, 0.3822744645603496], "ʔ": [-3.0639211183013515, 0.3957495430994786]}}
|
acoustic/tree
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4ceb91a89c17c80d3fba8fd70d7e675bdac31252e5f6cc62433224a9b687c9e0
|
| 3 |
+
size 303085
|
dictionary/czech_mfa.dict
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
dictionary/rules.yaml
CHANGED
|
@@ -1,75 +1,75 @@
|
|
| 1 |
dialects:
|
| 2 |
us:
|
| 3 |
- following_context: '[vn]'
|
| 4 |
-
non_silence_before_correction: -0.
|
| 5 |
preceding_context: ^
|
| 6 |
probability: 0.01
|
| 7 |
replacement: d ə
|
| 8 |
segment: d
|
| 9 |
-
silence_after_probability: 0.
|
| 10 |
-
silence_before_correction: 0.
|
| 11 |
- following_context: '[d]'
|
| 12 |
-
non_silence_before_correction: -0.
|
| 13 |
preceding_context: ^
|
| 14 |
probability: 0.04
|
| 15 |
replacement: ɡ ə
|
| 16 |
segment: ɡ
|
| 17 |
-
silence_after_probability: 0.
|
| 18 |
-
silence_before_correction: 0.
|
| 19 |
- following_context: '[smd]'
|
| 20 |
-
non_silence_before_correction: 0.
|
| 21 |
preceding_context: ^
|
| 22 |
probability: 0.03
|
| 23 |
replacement: ''
|
| 24 |
segment: j
|
| 25 |
silence_after_probability: 0.22
|
| 26 |
-
silence_before_correction: -0.
|
| 27 |
- following_context: $
|
| 28 |
-
non_silence_before_correction: 0.
|
| 29 |
preceding_context: ''
|
| 30 |
probability: 0.01
|
| 31 |
replacement: u
|
| 32 |
segment: u j ɪ
|
| 33 |
-
silence_after_probability: 0.
|
| 34 |
-
silence_before_correction: -0.
|
| 35 |
- following_context: n
|
| 36 |
-
non_silence_before_correction: 0.
|
| 37 |
preceding_context: s
|
| 38 |
probability: 0.01
|
| 39 |
replacement: ''
|
| 40 |
segment: t
|
| 41 |
-
silence_after_probability: 0.
|
| 42 |
-
silence_before_correction: -0.
|
| 43 |
- following_context: $
|
| 44 |
-
non_silence_before_correction: 0.
|
| 45 |
preceding_context: ''
|
| 46 |
probability: 0.01
|
| 47 |
replacement: ''
|
| 48 |
segment: t
|
| 49 |
-
silence_after_probability: 0.
|
| 50 |
-
silence_before_correction: -0.
|
| 51 |
- following_context: $
|
| 52 |
-
non_silence_before_correction: 0.
|
| 53 |
preceding_context: ''
|
| 54 |
probability: 0.01
|
| 55 |
replacement: u j u
|
| 56 |
segment: u j ɪ
|
| 57 |
-
silence_after_probability:
|
| 58 |
-
silence_before_correction: -0.
|
| 59 |
- following_context: $
|
| 60 |
-
non_silence_before_correction: 0.
|
| 61 |
preceding_context: ''
|
| 62 |
probability: 0.01
|
| 63 |
replacement: j iː
|
| 64 |
segment: j ɛ j iː
|
| 65 |
-
silence_after_probability: 0.
|
| 66 |
-
silence_before_correction: -0.
|
| 67 |
- following_context: ''
|
| 68 |
-
non_silence_before_correction: 0.
|
| 69 |
preceding_context: ''
|
| 70 |
probability: 0.01
|
| 71 |
replacement: j ɛː
|
| 72 |
segment: j ɛ j ɛ
|
| 73 |
-
silence_after_probability: 0.
|
| 74 |
-
silence_before_correction: -0.
|
| 75 |
rules: []
|
|
|
|
| 1 |
dialects:
|
| 2 |
us:
|
| 3 |
- following_context: '[vn]'
|
| 4 |
+
non_silence_before_correction: -0.04
|
| 5 |
preceding_context: ^
|
| 6 |
probability: 0.01
|
| 7 |
replacement: d ə
|
| 8 |
segment: d
|
| 9 |
+
silence_after_probability: 0.59
|
| 10 |
+
silence_before_correction: 0.06
|
| 11 |
- following_context: '[d]'
|
| 12 |
+
non_silence_before_correction: -0.22
|
| 13 |
preceding_context: ^
|
| 14 |
probability: 0.04
|
| 15 |
replacement: ɡ ə
|
| 16 |
segment: ɡ
|
| 17 |
+
silence_after_probability: 0.35
|
| 18 |
+
silence_before_correction: 0.45
|
| 19 |
- following_context: '[smd]'
|
| 20 |
+
non_silence_before_correction: 0.06
|
| 21 |
preceding_context: ^
|
| 22 |
probability: 0.03
|
| 23 |
replacement: ''
|
| 24 |
segment: j
|
| 25 |
silence_after_probability: 0.22
|
| 26 |
+
silence_before_correction: -0.21
|
| 27 |
- following_context: $
|
| 28 |
+
non_silence_before_correction: 0.03
|
| 29 |
preceding_context: ''
|
| 30 |
probability: 0.01
|
| 31 |
replacement: u
|
| 32 |
segment: u j ɪ
|
| 33 |
+
silence_after_probability: 0.17
|
| 34 |
+
silence_before_correction: -0.1
|
| 35 |
- following_context: n
|
| 36 |
+
non_silence_before_correction: 0.04
|
| 37 |
preceding_context: s
|
| 38 |
probability: 0.01
|
| 39 |
replacement: ''
|
| 40 |
segment: t
|
| 41 |
+
silence_after_probability: 0.78
|
| 42 |
+
silence_before_correction: -0.13
|
| 43 |
- following_context: $
|
| 44 |
+
non_silence_before_correction: 0.06
|
| 45 |
preceding_context: ''
|
| 46 |
probability: 0.01
|
| 47 |
replacement: ''
|
| 48 |
segment: t
|
| 49 |
+
silence_after_probability: 0.24
|
| 50 |
+
silence_before_correction: -0.19
|
| 51 |
- following_context: $
|
| 52 |
+
non_silence_before_correction: 0.01
|
| 53 |
preceding_context: ''
|
| 54 |
probability: 0.01
|
| 55 |
replacement: u j u
|
| 56 |
segment: u j ɪ
|
| 57 |
+
silence_after_probability: 0.92
|
| 58 |
+
silence_before_correction: -0.07
|
| 59 |
- following_context: $
|
| 60 |
+
non_silence_before_correction: 0.18
|
| 61 |
preceding_context: ''
|
| 62 |
probability: 0.01
|
| 63 |
replacement: j iː
|
| 64 |
segment: j ɛ j iː
|
| 65 |
+
silence_after_probability: 0.15
|
| 66 |
+
silence_before_correction: -0.46
|
| 67 |
- following_context: ''
|
| 68 |
+
non_silence_before_correction: 0.29
|
| 69 |
preceding_context: ''
|
| 70 |
probability: 0.01
|
| 71 |
replacement: j ɛː
|
| 72 |
segment: j ɛ j ɛ
|
| 73 |
+
silence_after_probability: 0.52
|
| 74 |
+
silence_before_correction: -0.54
|
| 75 |
rules: []
|
g2p/czech_mfa/graphemes.txt
CHANGED
|
@@ -508,11 +508,11 @@ y|n 506
|
|
| 508 |
ó|d 507
|
| 509 |
j|s 508
|
| 510 |
ü 509
|
| 511 |
-
|
| 512 |
-
|
| 513 |
-
|
| 514 |
-
|
| 515 |
-
|
| 516 |
k|ř 515
|
| 517 |
e|p 516
|
| 518 |
y|z 517
|
|
|
|
| 508 |
ó|d 507
|
| 509 |
j|s 508
|
| 510 |
ü 509
|
| 511 |
+
p|s 510
|
| 512 |
+
é|z 511
|
| 513 |
+
a|š 512
|
| 514 |
+
l|n 513
|
| 515 |
+
o|j 514
|
| 516 |
k|ř 515
|
| 517 |
e|p 516
|
| 518 |
y|z 517
|
g2p/czech_mfa/meta.json
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"version": "3.
|
|
|
|
| 1 |
+
{"version": "3.4.3", "architecture": "phonetisaurus", "train_date": "2026-10-09 05:50:38.788649", "phones": ["a", "aw", "a\u02d0", "b", "c", "d", "dz", "d\u0292", "ew", "f", "i\u02d0", "j", "k", "l", "l\u0329", "m", "m\u0329", "n", "n\u0329", "o", "ow", "o\u02d0", "p", "r", "r\u031d", "r\u031d\u030a", "r\u0329", "s", "t", "ts", "t\u0283", "u", "u\u02d0", "v", "x", "z", "\u014b", "\u0259", "\u025b", "\u025b\u02d0", "\u025f", "\u0261", "\u0266", "\u026a", "\u0272", "\u0283", "\u0292", "\u0294"], "graphemes": ["x", "z", "c", "m", "t", "j", "\u011b", "\u00fd", "-", "n", "k", "v", "'", "y", "\u00e1", "g", "b", "\u016f", "p", "\u0161", "\u00fc", "\u010f", "\u0159", "\u00fa", "\u00f3", "\u017e", "h", "w", "r", "o", "q", "s", "d", "\u00e9", "\u00ed", "e", "a", "l", "\u00f6", "f", "\u010d", "\u0148", "\u0165", "u", "i"], "grapheme_order": 2, "phone_order": 2, "sequence_separator": "|", "unicode_decomposition": false, "evaluation": {"num_words": 12477, "word_error_rate": null, "phone_error_rate": null}, "training": {"num_words": 112295, "num_graphemes": 45, "num_phones": 48}}
|
g2p/czech_mfa/model.fst
CHANGED
|
|
Git LFS Details
|
|
|
Git LFS Details
|