coqui-ai
/
TTS
mirror of https://github.com/coqui-ai/TTS.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
1,440 Commits
80 Branches
98 Tags
316 MiB
Commit Graph

333 Commits (1c2dc9f73909590aed91734cc1ea0a83ee1fb97c)

Author SHA1 Message Date
Eren Golge a833a3cf95 Merge branch 'twerkmeister-drop_cached_dataset' into dev-tacotron2-cond-stopnet 2019-05-14 22:42:07 +02:00
Eren Golge 9e30f4aae8 bug fix 2019-05-14 22:35:11 +02:00
Eren Golge ba08e98bd9 revert debug changes 2019-05-14 21:33:37 +02:00
Eren Golge b94f71c56a config update xz 2019-05-14 18:10:35 +02:00
Eren Golge 82d873b35a conditional stopnet and separate stopnet training 2019-05-14 17:49:20 +02:00
Thomas Werkmeister 7b2804cc0d dropped dataset caching 2019-04-29 11:07:04 +02:00
Eren Golge add08e331b Define foldr name explicitly and mozilla preprocessor update 2019-04-24 17:36:05 +02:00
Eren Golge 9466505f27 Make eos bos chars optional 2019-04-12 16:12:15 +02:00
Eren Golge e2cf35bb10 Make loss masking optional 2019-04-10 16:41:08 +02:00
Eren Golge 961af0f5cd setup_model externally based on model selection. Make forward attention and prenet type configurable in config.json 2019-04-05 17:49:18 +02:00
Eren Golge e1573440ba if reinit option is used run runtime rrror to partially iniot the model 2019-03-29 17:03:29 +01:00
Eren Golge 6edd8bc6dd add git branch and restore_path to copied config file for each run 2019-03-29 17:01:57 +01:00
Eren Golge bc51b81aae parameter name fix 2019-03-26 00:52:47 +01:00
Eren Golge 0a92c6d5a7 Set attention norm method by config.json 2019-03-26 00:48:12 +01:00
Eren Golge d8908692c5 refactor partial reinit script as a function. Allow user to select layers to reinit in finutunning 2019-03-23 17:19:40 +01:00
Eren Golge bbd90c3b4d use grad_clip from config.json 2019-03-12 16:54:42 +01:00
Eren Golge abc7b8e366 bug fix: plot linear spec if tacotron is the model 2019-03-12 10:46:48 +01:00
Eren Golge 6595f30a1d bug fix to compute number of epoch iterations 2019-03-12 01:26:30 +01:00
Eren Golge 5cbe0f83f6 bug fix 2019-03-11 14:13:32 +01:00
Eren Golge 5754116c19 bos char addded 2019-03-06 22:06:01 +01:00
Eren Golge 08162157ee generic train.py for multiple architectures set on config.json 2019-03-06 13:11:22 +01:00
Eren Golge 1e8fdec084 Modularize functions in Tacotron 2019-03-05 13:25:50 +01:00
Eren Golge bf5f18d11e Formatting changes and distributed training 2019-02-27 09:50:52 +01:00
Eren Golge 8100360c71 bug fix for initialization of the model 2019-02-20 17:14:08 +01:00
Eren Golge 71f1f0c4c0 partial initialization fix 2019-02-20 16:46:10 +01:00
Eren Golge e2d9b0e27c Merge branch 'dev' 2019-02-18 13:08:47 +01:00
Eren Golge 4d2c374cfd Filter out shape mismatch laeyrs in restoriong model for finetunning 2019-02-16 03:19:26 +01:00
Eren Golge 6ea31e47df Constant queue size for autoregression window 2019-02-16 03:18:49 +01:00
Eren Golge 4fad1971cd set learning rate for restored model 2019-02-12 10:04:39 +01:00
Eren Golge d03865ed05 configurable loss weigting for lower frequencies 2019-02-12 10:04:18 +01:00
Eren Golge 0308c92c16 pick best model by loss function is run_eval is true 2019-02-05 11:56:44 +01:00
Eren Golge b011dafbab pass num_chars in train.py 2019-01-21 14:52:40 +01:00
Eren Golge 7e020d4084 Bug fixes 2019-01-16 16:23:04 +01:00
Eren Golge 72341ae22a phoneme usage bug fix 2019-01-16 13:09:47 +01:00
Eren Golge 5733eab90b Add configurable phoneme cache path 2019-01-16 12:33:24 +01:00
Eren Golge 656542db93 Debug purpose alignment image saving 2019-01-16 12:18:54 +01:00
Eren Golge c7ee056f21 Merge branch 'master' of github.com:mozilla/TTS 2019-01-06 18:13:54 +01:00
Yves-Noel Weweler 10524b885c
Fix NameError: name 'model_dict' is not defined 
Closes #91
 2018-12-31 13:29:39 +01:00
Eren Golge f5c972dee6 bug fixes 2018-12-28 14:22:57 +01:00
Eren Golge 17b65d5cde bug fixes on train.py 2018-12-28 14:22:40 +01:00
Eren Golge 6d8b2dad91 bug fix 2018-12-28 14:18:33 +01:00
Eren Golge a1fe867371 bug fix for partial model initialization, if model is not initialized, it is tried to init model partially with only matching layers in size 2018-12-28 14:18:31 +01:00
Eren Golge 481105ccfa logger for tensorboard plotting 2018-12-28 14:18:19 +01:00
Bajibabu Bollepalli dce5a05e80
Update train.py 2018-12-18 15:49:24 +02:00
Eren Golge 8d865629a0 partial model initialization 2018-12-12 15:43:58 +01:00
Eren Golge 619c73f0f1 Init data_loaders by function beginning of each train and eval run 2018-12-11 17:52:43 +01:00
Eren Golge 1ac037d780 Update README and comments on argsparse 2018-12-11 15:50:58 +01:00
Eren Golge 2ec55e2369 print avg spec length in training 2018-12-11 15:08:02 +01:00
Eren Golge bb2a88a984 Rename LR scheduler 2018-11-26 14:09:42 +01:00
Eren Golge c98631fe36 bug fix 2018-11-24 17:23:46 +01:00
Eren Golge 36efaa789d bug fix in testing phase 2018-11-21 15:42:10 +01:00
Eren Golge 660f8c7b66 print average text length, fix for Nancy preprocessor 2018-11-20 12:54:33 +01:00
Eren Golge 6550db5251 Formatting, fixing import statements, logging learning rate, remove optimizer restore cuda call 2018-11-05 14:05:04 +01:00
Eren Golge bb04a1c6e0 optional lr schedule 2018-11-03 19:47:28 +01:00
Eren Golge c8a552e627 Batch update after data-loss 2018-11-02 16:13:51 +01:00
Eren 1ae3b3e442 Enable CPU training and fix restore_epoch 2018-10-25 14:05:27 +02:00
Eren e00c3f5044 Fix the restore_epoch 2018-10-25 13:56:33 +02:00
Eren 4427067c0b restore lr scheduler 2018-09-29 01:46:11 +02:00
Eren 231607a546 Reduce batch group size 2018-09-26 16:52:47 +02:00
Eren c89a3098dd Merge branch 'weight-decay' into attn-smoothing-bgs-sigmoid-wd 2018-09-26 16:49:39 +02:00
Eren c36a1f7deb Make audio folder and save audio with scipy 2018-09-21 17:41:11 +02:00
Eren 34eeaee58b Make audio folder and save audio with scipy 2018-09-21 17:38:55 +02:00
Eren 30fea0b957 Batch group shuffling 2018-09-20 11:31:39 +02:00
Eren 16db5159f1 Weight decay described here: http://www.fast.ai/2018/07/02/adam-weight-decay/ 2018-09-19 15:46:25 +02:00
Eren 9b29b4e281 Weight decay 2018-09-19 15:46:25 +02:00
Eren 4b08c3a13f Save audio samples per validation 2018-09-16 13:32:59 +02:00
Eren 56c6d0cac8 Remove min max mel freq 2018-09-06 15:26:20 +02:00
Eren 7d66bdc5f4 update model size to paper 2018-09-06 14:37:19 +02:00
Eren 03e1c06f5b bug fix 2018-08-13 15:04:00 +02:00
Eren 73c3ab546d Testing update 2018-08-13 15:02:30 +02:00
Eren bb526c296f Change scheduler AnnealLR and catch audio synthesis error in eval time 2018-08-13 13:13:45 +02:00
Eren 6818e11185 Make lr scheduler configurable 2018-08-12 15:02:06 +02:00
Eren f7add3c8e5 tensorboardx plotting figures 2018-08-11 16:53:09 +02:00
Eren b65f5bd1d0 bug fix for plotting 2018-08-11 16:09:47 +02:00
Eren 96e2e3c776 Use step wise LR scheduler + adapt train.py for passing squence mask directly 2018-08-10 17:48:19 +02:00
Eren G ecd31af125 Formating and printing more about the model 2018-08-08 18:45:02 +02:00
Eren G d5febfb187 Setting up network size according to the reference paper 2018-08-08 12:34:44 +02:00
Eren G f5537dc48f pep8 format all 2018-08-02 16:34:17 +02:00
Eren G 8bc4fe8aac Printing fix with flush and spaceing 2018-07-30 13:52:39 +02:00
Eren G adedd7b1a9 Set num output units based on audio processor and set imported audio processor based on config.json 2018-07-27 16:14:10 +02:00
Eren G 355df8fe9b Keep avg step time 2018-07-27 16:13:55 +02:00
Eren G 30561d0b5f Save best model based on train loss 2018-07-27 13:51:13 +02:00
Eren G 25b6769246 Some bug fixes 2018-07-26 13:33:05 +02:00
Eren G e156738920 adding cache loader 2018-07-25 19:14:07 +02:00
Eren G f24a9e74b8 Define metafiles in conf.json 2018-07-25 12:34:42 +02:00
Eren G c72f309068 Plot specs and alignments for test sentences 2018-07-23 16:15:26 +02:00
Eren G 766e88700d Define global ap and remove preemphasis 2018-07-20 16:04:29 +02:00
Eren G 20c3085695 Perform testing on example sentences 2018-07-20 13:10:25 +02:00
Eren G dfd0bc1831 Checkpoint stop token optimizer 2018-07-20 12:23:44 +02:00
Eren G bacf5485e8 Logging bug fix for evaly 2018-07-19 16:29:37 +02:00
Eren G d4f1ccd3ed Move things into main and set thread 2018-07-17 15:59:31 +02:00
Eren G adbe603af1 Bug fixes 2018-07-13 15:24:50 +02:00
Eren G 0ef3c0ac3f Remove preemphasis from audio processing 2018-07-13 14:56:05 +02:00
Eren G dac8fdffa9 Attn masking 2018-07-13 14:50:55 +02:00
Eren G 9f52833151 Merge branch 'loc-sens-attn' into loc-sens-attn-new and attention without attention-cum 2018-07-13 14:27:51 +02:00
Eren G 7f75ff39df logging update 2018-07-12 19:18:36 +02:00
Eren G 1160ea5a19 more prunning and dataloader file encoding for win10 2018-07-12 19:03:39 +02:00
Eren G 5bf3ff345f Remove tabs from logging 2018-07-12 18:09:02 +02:00
Eren G 7fad94d8a7 More logging 2018-07-11 12:42:59 +02:00
Eren G cc26cd6c7b Logging bug fix 2018-07-09 16:11:14 +02:00
Eren G 9245992817 Print step information in train 2018-07-09 15:56:30 +02:00
Eren 934639128d Change logging for the new cluster system 2018-07-05 17:30:42 +02:00
Eren Golge f791f4e5e7 Use MSE loss instead of L1 Loss 2018-06-06 07:42:51 -07:00
Eren Golge 65ea7b0afb Revert train.py 2018-05-25 05:17:08 -07:00
Eren Golge 24644b20d4 Merge branch 'master' of https://github.com/Mozilla/TTS 
Conflicts:
	README.md
	best_model_config.json
	datasets/LJSpeech.py
	layers/tacotron.py
	notebooks/TacotronPlayGround.ipynb
	notebooks/utils.py
	tests/layers_tests.py
	tests/loader_tests.py
	tests/tacotron_tests.py
	train.py
	utils/generic_utils.py
 2018-05-25 05:14:04 -07:00
Eren Golge e6112f7b2d restore fix 2018-05-16 19:20:40 -07:00
Eren Golge 40f1a3d3a5 RNN stop-token prediction 2018-05-15 16:12:47 -07:00
Eren Golge a31e60e928 bug fix, average mel spec validation loss 2018-05-15 07:13:46 -07:00
Eren Golge b2e09f9f58 larger lr for stop token 2018-05-14 19:04:29 -07:00
Eren Golge 074d1c97d3 Use SGD for stop -token layer 2018-05-13 06:35:11 -07:00
Eren Golge dc3f909ddc Separate backward pass for stop-token prediction 2018-05-11 08:38:07 -07:00
Eren Golge 5a5b9263e2 bug fix train.py add stop token 2018-05-11 04:55:05 -07:00
Eren Golge 31cc9c5443 Update train.py for stop token prediciton 2018-05-11 04:15:53 -07:00
Eren Golge 1d8ad0968c Allow debug runand don force git hash 2018-05-11 03:49:55 -07:00
Eren Golge 3c55aed6a4 Add manual seed 2018-05-11 03:47:30 -07:00
Eren Golge b4bd713581 train.py - add with torch.no_grad(): 2018-05-10 16:44:37 -07:00
Eren Golge c8bfe731d6 train.py - replace data[0] with item() 2018-05-10 16:24:30 -07:00
Eren Golge 45bb9d9eec Add model name to experiemtn folder 2018-05-10 16:14:23 -07:00
Eren Golge 1770a9338c Remove experiment folder at error and disable signal use 2018-05-10 16:13:05 -07:00
Eren Golge a2eeec31be remove 'Variable' from train.py 2018-05-10 16:07:13 -07:00
Eren Golge 43b19ab1b3 bug fix 2018-05-05 06:55:18 -07:00
Eren Golge 1fa6068e9d bug fix 2018-05-02 06:21:29 -07:00
Eren Golge 9924ae4a1d bug fix 2018-05-02 06:07:20 -07:00
Eren Golge 6b937d6059 bug fix 2018-05-02 05:00:36 -07:00
Eren Golge 59f8d246c9 bug fix 2018-05-02 04:59:51 -07:00
Eren Golge 855dd365b5 bug fix 2018-05-02 04:59:23 -07:00
Eren Golge 91d97686c0 bug fix 2018-05-02 04:58:48 -07:00
Eren Golge ccd91a67d0 bug fix 2018-05-02 04:56:52 -07:00
Eren Golge a4561c5096 config 2018-05-02 04:56:35 -07:00
Eren Golge 8319f22278 bug fix 2018-04-30 06:12:56 -07:00
Eren Golge fda7e7f6c9 carry to cuda 2018-04-30 06:12:12 -07:00
Eren Golge 40b479b3b9 bug fix 2018-04-30 06:11:22 -07:00
Eren Golge cc9bfe96af add stop loss 2018-04-30 05:47:14 -07:00
Eren Golge d68346eeba stop targets 2018-04-30 05:07:41 -07:00
Eren Golge 5fb184e278 resolve optimizer restore problem 2018-04-28 13:29:50 -07:00
Eren Golge 41430d1062 bug fix 2018-04-28 13:26:14 -07:00
Eren Golge 784996dc4e a 2018-04-28 13:16:02 -07:00
Eren Golge 1b59d8110c fix optimizer for restored model 2018-04-28 13:13:01 -07:00
Eren Golge f3bb9dd0f9 add model name to config 2018-04-27 04:48:21 -07:00
Eren Golge aa32c76ce0 remove folder in error 2018-04-26 05:53:54 -07:00
Eren Golge 19e6f4a8b7 bug fix 2018-04-26 05:47:15 -07:00
Eren Golge 6b178d5436 bug fix 2018-04-26 05:46:49 -07:00
Eren Golge 961d240534 bug fix 2018-04-26 05:46:24 -07:00
Eren Golge d143616ada remove folder if error 2018-04-26 05:44:18 -07:00
Eren Golge c894919120 optional proirity_freq 2018-04-26 05:38:52 -07:00
Eren Golge 7bfdc32b7b remove requires_grad_() 2018-04-26 05:27:08 -07:00
Eren Golge 0482a66a98 manual seeding 2018-04-26 05:13:51 -07:00
Eren Golge b1361375da bug fix 
Q
~
'
 2018-04-25 08:42:21 -07:00
Eren Golge 83d7756856 against blocking stduot 2018-04-25 08:25:55 -07:00
Eren Golge 8ef8ddb915 bug fix 2018-04-25 08:12:14 -07:00