Modèle Français 0.3.4
Pre-release
Pre-release
·
102 commits
to master
since this release
Jeux de données :
- Lingua Libre (~20h)
- Common Voice FR (v2) (~120h, en autorisant des duplicatas)
- Training Speech (~180h)
- African Accented French (~15h)
- M-AILABS French (~315h)
Total : ~650h
Paramètres :
- LEARNING_RATE=0.0001
- DROPOUT=0.3
- BATCH_SIZE=96
- LM_ALPHA=0.65
- LM_BETA=1.45
Language Model : dump wikipedia + dump débats assemblée nationale.
Fonctionne avec DeepSpeech v0.6.0. Ré-export de 0.3.3 pour corriger un bug dans TFLite
Résultats test set:
Testing model on /mnt/extracted/data/lingualibre/lingua_libre_Q21-fra-French_test.csv
Test epoch | Steps: 75 | Elapsed Time: 0:01:44
Test on /mnt/extracted/data/lingualibre/lingua_libre_Q21-fra-French_test.csv - WER: 0.467659, CER: 0.138508, loss: 6.800947
--------------------------------------------------------------------------------
WER: 4.000000, CER: 2.200000, loss: 39.604939
- wav: file:///mnt/extracted/data/lingualibre/lingua_libre/Q21-fra-French/Vahidmasrour/abhal.wav
- src: "abhal"
- res: "le panel a bal"
--------------------------------------------------------------------------------
WER: 3.000000, CER: 0.600000, loss: 2.462182
- wav: file:///mnt/extracted/data/lingualibre/lingua_libre/Q21-fra-French/WikiLucas00/irato.wav
- src: "irato"
- res: "il a to"
--------------------------------------------------------------------------------
WER: 3.000000, CER: 0.111111, loss: 3.428576
- wav: file:///mnt/extracted/data/lingualibre/lingua_libre/Q21-fra-French/WikiLucas00/ultratrifoliophile.wav
- src: "ultratrifoliophile"
- res: "ultra trifolio phile"
--------------------------------------------------------------------------------
WER: 3.000000, CER: 0.333333, loss: 5.036440
- wav: file:///mnt/extracted/data/lingualibre/lingua_libre/Q21-fra-French/WikiLucas00/cuthomiurophile.wav
- src: "cuthomiurophile"
- res: "culto miro phile"
--------------------------------------------------------------------------------
WER: 3.000000, CER: 0.333333, loss: 5.090287
- wav: file:///mnt/extracted/data/lingualibre/lingua_libre/Q21-fra-French/WikiLucas00/remiauler.wav
- src: "remiauler"
- res: "remi au le"
--------------------------------------------------------------------------------
WER: 3.000000, CER: 0.285714, loss: 6.972348
- wav: file:///mnt/extracted/data/lingualibre/lingua_libre/Q21-fra-French/Lyokoï/indoeuropéiste.wav
- src: "indoeuropéiste"
- res: "in doro péiste"
--------------------------------------------------------------------------------
WER: 3.000000, CER: 0.454545, loss: 7.742430
- wav: file:///mnt/extracted/data/lingualibre/lingua_libre/Q21-fra-French/Jules78120/Antarctique.wav
- src: "antarctique"
- res: "en parti que"
--------------------------------------------------------------------------------
WER: 3.000000, CER: 0.833333, loss: 8.499911
- wav: file:///mnt/extracted/data/lingualibre/lingua_libre/Q21-fra-French/Guilhelma (Ives)/padena.wav
- src: "padena"
- res: "pas de nom"
--------------------------------------------------------------------------------
WER: 3.000000, CER: 0.307692, loss: 8.974085
- wav: file:///mnt/extracted/data/lingualibre/lingua_libre/Q21-fra-French/Lyokoï/pleurogynique.wav
- src: "pleurogynique"
- res: "pleu rogi mique"
--------------------------------------------------------------------------------
WER: 3.000000, CER: 0.230769, loss: 9.156916
- wav: file:///mnt/extracted/data/lingualibre/lingua_libre/Q21-fra-French/WikiLucas00/bonhomisation.wav
- src: "bonhomisation"
- res: "bon ami sation"
--------------------------------------------------------------------------------
Testing model on /mnt/extracted/data/trainingspeech/ts_2019-04-11_fr_FR_test.csv
Test epoch | Steps: 129 | Elapsed Time: 0:10:36
Test on /mnt/extracted/data/trainingspeech/ts_2019-04-11_fr_FR_test.csv - WER: 0.185852, CER: 0.061034, loss: 21.406639
--------------------------------------------------------------------------------
WER: 4.000000, CER: 1.222222, loss: 28.107866
- wav: file:///mnt/extracted/data/trainingspeech/ts_2019-04-11_fr_FR/MonsieurLecoqP1C16_0188.converted.wav
- src: "continuez"
- res: "quand il ne est"
--------------------------------------------------------------------------------
WER: 2.333333, CER: 0.818182, loss: 123.491005
- wav: file:///mnt/extracted/data/trainingspeech/ts_2019-04-11_fr_FR/LesMysteresDeParisT1P1C5_0129.converted.wav
- src: "diminution de fourloir"
- res: "des minutions de de fournoue a sa fin"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.142857, loss: 0.981466
- wav: file:///mnt/extracted/data/trainingspeech/ts_2019-04-11_fr_FR/LaGloireDuComacchio_0097.converted.wav
- src: "pardieu"
- res: "par dieu"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.500000, loss: 5.709780
- wav: file:///mnt/extracted/data/trainingspeech/ts_2019-04-11_fr_FR/LeComteDeMonteCristoT1Chap3_0240.converted.wav
- src: "hola"
- res: "a la"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.750000, loss: 6.360806
- wav: file:///mnt/extracted/data/trainingspeech/ts_2019-04-11_fr_FR/madamebovaryC24_0123.converted.wav
- src: "leon"
- res: "et on"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.750000, loss: 8.123431
- wav: file:///mnt/extracted/data/trainingspeech/ts_2019-04-11_fr_FR/LesMysteresDeParisT3P5C12_0281.converted.wav
- src: "cici"
- res: "si si"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.750000, loss: 8.998843
- wav: file:///mnt/extracted/data/trainingspeech/ts_2019-04-11_fr_FR/MonsieurLecoqT2P04_0012.converted.wav
- src: "hola"
- res: "ou la"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.444444, loss: 9.201403
- wav: file:///mnt/extracted/data/trainingspeech/ts_2019-04-11_fr_FR/MonsieurLecoqT2P29_0219.converted.wav
- src: "jarnibieu"
- res: "jami bien"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.300000, loss: 11.241019
- wav: file:///mnt/extracted/data/trainingspeech/ts_2019-04-11_fr_FR/LesMysteresDeParisT2P3C2_0146.converted.wav
- src: "infortunee"
- res: "un fortune"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.833333, loss: 12.523339
- wav: file:///mnt/extracted/data/trainingspeech/ts_2019-04-11_fr_FR/LaGloireDuComacchio_0841.converted.wav
- src: "infame"
- res: "un charme"
--------------------------------------------------------------------------------
Testing model on /mnt/extracted/data/cv-fr/clips/test.csv
Test epoch | Steps: 150 | Elapsed Time: 0:06:26
Test on /mnt/extracted/data/cv-fr/clips/test.csv - WER: 0.372634, CER: 0.177308, loss: 38.467953
--------------------------------------------------------------------------------
WER: 3.000000, CER: 0.700000, loss: 18.267256
- wav: file:///mnt/extracted/data/cv-fr/clips/3506b8983d26cdaf9ac27c5099824bf9192a5a21e464b3e8bc18f82176ddffe8e50eac8826f41f4f96055d8a441e627ca40560db2c424bf75cdc987004359f7a.wav
- src: "lesquelles"
- res: "il est le"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.200000, loss: 10.854227
- wav: file:///mnt/extracted/data/cv-fr/clips/49bf7986cddf86ef7f319b4fa0a1deb7225e813dd9918d8c54326d42257a7e5d3e7eb2270f692a8836f8977d3e678b4f4e418e7a73ea64ec686cf980e96f6577.wav
- src: "lesquelles"
- res: "les celles"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.333333, loss: 11.127397
- wav: file:///mnt/extracted/data/cv-fr/clips/890d482adb285fdfcdf1ad5e877d175be48cfe9544834136b7ef094f7252083211f6293c49b49bdc5fcbebeccc9b3eab11f0883358648441015127d2b05e902f.wav
- src: "bienvenue"
- res: "bien menu"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.285714, loss: 11.260811
- wav: file:///mnt/extracted/data/cv-fr/clips/27a4b648313e1daa05708c74af8d0f68d010e34b949735a1ea85a58ebe1057d416a22a1f21f9e4913a8f556f57534833c39658d3fd87a554c003903ae07552e2.wav
- src: "dommage"
- res: "de mage"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.250000, loss: 13.728657
- wav: file:///mnt/extracted/data/cv-fr/clips/556f15d193c06afdaa99f58a9ee0e90dbce11653ef858d7644e1142e958c1fdc95331f5eb85255fc757d6183969bd13c64a79ab89456b3391bc2277b7f603cb9.wav
- src: "mensonge"
- res: "en songe"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.181818, loss: 17.680149
- wav: file:///mnt/extracted/data/cv-fr/clips/84d82e3417e9ec95af3b0d84f438261e78c99345203ac775be63d8b4ec76066ae6da226fdf8d0d5449e9788439edcd2e329d6cf24bc9078bd2851fdf2754a997.wav
- src: "malédiction"
- res: "male diction"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 1.000000, loss: 18.397339
- wav: file:///mnt/extracted/data/cv-fr/clips/0b2f4c148fd3d74a0dbd68f9b52ace48b880fec44de078045313e4058cf1a8fc565cf3bb9eca0cfee85616f0abe9b7212a39815ffa810f6e9d21adf7b38f72ac.wav
- src: "écoutez"
- res: "le côté"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.454545, loss: 21.874004
- wav: file:///mnt/extracted/data/cv-fr/clips/e7cfa56b14f04aa3ef3199fb21e9500e257a8c99784de8f785143007bacf3c17f1c64325235afe1112775b1385d0a15647d69594b6c012107266f8637b794cf8.wav
- src: "défavorable"
- res: "des adorables"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.533333, loss: 30.569332
- wav: file:///mnt/extracted/data/cv-fr/clips/6adb27aca9f87c8c97d3c37107a7f2ae8121b9ed005ec1a93ea1daaa317f082866bcf615e4d776080292cc3f32ee2a5476e2bef75c1487d9dfb2b37eabd12e8f.wav
- src: "en substitution"
- res: "en su que tu tiens"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.578947, loss: 40.773918
- wav: file:///mnt/extracted/data/cv-fr/clips/7fb3c5c2e998d2d51cbefab33e230b804257bc0628658c7f4dc5afed21eb8a6419d7cd63de4398055a418e812705dc672ba8a5c3e3dc5cfdfbe0fb37c8af4acf.wav
- src: "complètement trempé"
- res: "on peut en trente"
--------------------------------------------------------------------------------
Testing model on /mnt/extracted/data/M-AILABS/fr_FR/fr_FR_test.csv [346/1944]
Test epoch | Steps: 148 | Elapsed Time: 0:18:58
Test on /mnt/extracted/data/M-AILABS/fr_FR/fr_FR_test.csv - WER: 0.077349, CER: 0.023200, loss: 13.334200
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.250000, loss: 5.419570
- wav: file:///mnt/extracted/data/M-AILABS/fr_FR/male/bernard/le_pays_des_fourrures/wavs/le_pays_des_fourrures_2_21_f000013.wav
- src: "attendre"
- res: "à tendre"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.300000, loss: 10.197102
- wav: file:///mnt/extracted/data/M-AILABS/fr_FR/female/ezwa/monsieur_lecoq/wavs/monsieur_lecoq_1_08_f000011.wav
- src: "hhh tables"
- res: "h h h table"
--------------------------------------------------------------------------------
WER: 2.000000, CER: 5.000000, loss: 12.230239
- wav: file:///mnt/extracted/data/M-AILABS/fr_FR/female/nadine_eckert_boulet/les_mysteres_de_paris/wavs/les_mysteres_de_paris_4_13_f000027.wav
- src: "m"
- res: "on ne"
--------------------------------------------------------------------------------
WER: 1.125000, CER: 0.562500, loss: 111.289154
- wav: file:///mnt/extracted/data/M-AILABS/fr_FR/female/nadine_eckert_boulet/les_mysteres_de_paris/wavs/les_mysteres_de_paris_2_09_f000184.wav
- src: "m césar bradamanti qui l'a guéri d'un rhumatisme"
- res: "pauvre mon cesar batenti qu'il a gris d'au matise"
--------------------------------------------------------------------------------
WER: 1.000000, CER: 0.333333, loss: 0.632192
- wav: file:///mnt/extracted/data/M-AILABS/fr_FR/male/gilles_g_le_blanc/lupin_contre_holmes/wavs/lupin_contre_holmes_01_f000270.wav
- src: "non"
- res: "on"
--------------------------------------------------------------------------------
WER: 1.000000, CER: 0.250000, loss: 3.804301
- wav: file:///mnt/extracted/data/M-AILABS/fr_FR/female/ezwa/monsieur_lecoq/wavs/monsieur_lecoq_1_18_f000067.wav
- src: "lesquels"
- res: "lesquelles"
--------------------------------------------------------------------------------
WER: 1.000000, CER: 0.125000, loss: 4.740473
- wav: file:///mnt/extracted/data/M-AILABS/fr_FR/male/gilles_g_le_blanc/lupin_contre_holmes/wavs/lupin_contre_holmes_12_f000206.wav
- src: "superbes"
- res: "superbe"
--------------------------------------------------------------------------------
WER: 1.000000, CER: 0.666667, loss: 5.663177
- wav: file:///mnt/extracted/data/M-AILABS/fr_FR/female/nadine_eckert_boulet/madame_bovary/wavs/madame_bovary_3_06_f000037.wav
- src: "yes"
- res: "le"
--------------------------------------------------------------------------------
WER: 1.000000, CER: 0.666667, loss: 5.752508
- wav: file:///mnt/extracted/data/M-AILABS/fr_FR/male/gilles_g_le_blanc/lupin_contre_holmes/wavs/lupin_contre_holmes_01_f000248.wav
- src: "certes"
- res: "ce"
--------------------------------------------------------------------------------
WER: 1.000000, CER: 0.666667, loss: 7.209431
- wav: file:///mnt/extracted/data/M-AILABS/fr_FR/male/gilles_g_le_blanc/les_miserables_tome_5/wavs/les_miserables_tome_5_18_f000067.wav
- src: "certes"
- res: "ce"
--------------------------------------------------------------------------------
Testing model on /mnt/extracted/data/African_Accented_French/African_Accented_French/African_Accented_French_test.csv
Test epoch | Steps: 13 | Elapsed Time: 0:00:25
Test on /mnt/extracted/data/African_Accented_French/African_Accented_French/African_Accented_French_test.csv - WER: 0.482247, CER: 0.276829, loss: 46.125565
--------------------------------------------------------------------------------
WER: 2.000000, CER: 0.142857, loss: 5.044052
- wav: file:///mnt/extracted/data/African_Accented_French/African_Accented_French/speech/train/yaounde/answers/ctell5-73/ctell5-73-130.wav
- src: "bonsoir"
- res: "bon soir"
--------------------------------------------------------------------------------
WER: 1.500000, CER: 0.600000, loss: 23.252029
- wav: file:///mnt/extracted/data/African_Accented_French/African_Accented_French/speech/train/yaounde/answers/ctell4-57/ctell4-57-131.wav
- src: "bonne nuit"
- res: "bon ni meci"
--------------------------------------------------------------------------------
WER: 1.500000, CER: 0.933333, loss: 203.758545
- wav: file:///mnt/extracted/data/African_Accented_French/African_Accented_French/speech/train/yaounde/answers/ctell3-51/ctell3-51-084.wav
- src: "de quelle couleur est sa barbe"
- res: "il n'y a pas de bas monsieur la nave"
--------------------------------------------------------------------------------
WER: 1.333333, CER: 0.631579, loss: 52.995483
- wav: file:///mnt/extracted/data/African_Accented_French/African_Accented_French/speech/train/yaounde/read/ctell3-36/ctell3-36-0031.wav
- src: "adresse passe force"
- res: "la gare a fort"
--------------------------------------------------------------------------------
WER: 1.333333, CER: 1.750000, loss: 68.027031
- wav: file:///mnt/extracted/data/African_Accented_French/African_Accented_French/speech/train/ca16/052/afc-gabon_16.06.15_052_conv_0137.wav
- src: "ah ça va"
- res: "ah je suis moi cela"
--------------------------------------------------------------------------------
WER: 1.250000, CER: 0.840000, loss: 77.566818
- wav: file:///mnt/extracted/data/African_Accented_French/African_Accented_French/speech/train/yaounde/answers/ctell2-26/ctell2-26-192.wav
- src: "souffrez vous de vertiges"
- res: "ce que vous avez souffert de pise"
--------------------------------------------------------------------------------
WER: 1.250000, CER: 1.285714, loss: 100.534805
- wav: file:///mnt/extracted/data/African_Accented_French/African_Accented_French/speech/train/yaounde/answers/ctell5-78/ctell5-78-253.wav
- src: "où fait il mal"
- res: "au niveau de la vendre"
--------------------------------------------------------------------------------
WER: 1.250000, CER: 1.000000, loss: 149.877106
- wav: file:///mnt/extracted/data/African_Accented_French/African_Accented_French/speech/train/yaounde/answers/ctell3-39/ctell3-39-099.wav
- src: "quelle est sa religion"
- res: "je crois qu'il est peu distant"
--------------------------------------------------------------------------------
WER: 1.200000, CER: 0.375000, loss: 46.732121
- wav: file:///mnt/extracted/data/African_Accented_French/African_Accented_French/speech/train/yaounde/read/ctell1-02/ctell1-02-0089.wav
- src: "pain défendu réveille l' appétit"
- res: "pin de fendu levine la pitit"
--------------------------------------------------------------------------------
WER: 1.200000, CER: 0.566667, loss: 67.323540
- wav: file:///mnt/extracted/data/African_Accented_French/African_Accented_French/speech/train/yaounde/read/ctell1-03/ctell1-03-0161.wav
- src: "le bernois feint faux courtois"
- res: "le ben moins prompte pro pour toi"
--------------------------------------------------------------------------------