wassname/TTS
1
0
Fork 0
mirror of https://github.com/wassname/TTS.git synced 2026-09-20 12:30:22 +08:00
Code Issues Packages Projects Releases Wiki Activity
3,039 Commits 1 Branch 0 Tags
aba840b4e676c9d01202d2ba72505e62fff0560b
Commit Graph
3039 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Eren Gölge aba840b4e6 Fix loading the amp scaler from a checkpoint 🛠️ 2021-06-23 13:49:42 +02:00
Eren Gölge 18e5393f16 Add 🐍 python 3.9 to CI 2021-06-23 13:49:36 +02:00
Eren Gölge 0ff2d2336a Fix wrong argument name 🛠️ 2021-06-22 16:21:11 +02:00
Eren Gölge af12925142 Add AlignTTS recipe 👩‍🍳 2021-06-22 16:21:11 +02:00
Eren Gölge 61c3cb871f Docstring edit in TTSDataset.py ✍️ 2021-06-22 16:21:11 +02:00
Eren Gölge 6f739ea07a Fix eval_log for gan.py 🛠️ 2021-06-22 16:21:11 +02:00
Eren Gölge ebb91c0fbb Move TorchSTFT to utils.audio 2021-06-22 16:21:11 +02:00
Eren Gölge 01c4b22a2f Fixup trainer.py 🛠️ 2021-06-22 16:21:11 +02:00
Eren Gölge 9cb1062736 Create LJSpeech recipes for all the models 2021-06-22 16:21:11 +02:00
Eren Gölge 787f6177c0 Update gitignore 2021-06-22 16:21:11 +02:00
Eren Gölge d2d914cbc0 Update Pylint configuration 2021-06-22 16:21:11 +02:00
Eren Gölge 7de2756fc4 Enable support for 🐍 python 3.10 
Bump up versions numpy 1.19.5 and TF 2.5.0
 2021-06-22 16:21:11 +02:00
Eren Gölge 220e184f66 Apply small fixes for API compatibility 2021-06-22 16:21:11 +02:00
Eren Gölge 77d57dd301 Print max_decoder_steps when model reaches the limit 2021-06-22 16:21:11 +02:00
Eren Gölge 7dc2177df4 Update synthesizer for speaker and model init 2021-06-22 16:21:11 +02:00
Eren Gölge c3a0bc702e fixup configs 2021-06-22 16:21:11 +02:00
Eren Gölge 0e01c2594f Update speaker_manager 2021-06-22 16:21:11 +02:00
Eren Gölge 8182f5168f Fixup utils for the trainer 2021-06-22 16:21:11 +02:00
Eren Gölge b4bb567e04 Update vocoder utils 2021-06-22 16:21:11 +02:00
Eren Gölge f3ff5b1971 Update TTS.bin scripts for the new API 2021-06-22 16:21:11 +02:00
Eren Gölge aed919cf1c Update vocoder datasets and setup_dataset 2021-06-22 16:21:11 +02:00
Eren Gölge 59abf490a1 Implement setup_model for vocoder models 2021-06-22 16:21:11 +02:00
Eren Gölge 420820caf4 Update vocoder models 2021-06-22 16:21:11 +02:00
Eren Gölge d10f9c5676 Update tts.models.setup_model 2021-06-22 16:21:11 +02:00
Eren Gölge cae702980f Create base 🐸TTS model abstraction for tts models 2021-06-22 16:21:11 +02:00
Eren Gölge 70d968b169 Update vocoder model configs 2021-06-22 16:21:11 +02:00
Eren Gölge f8a3460818 Update tts model configs 2021-06-22 16:21:11 +02:00
Eren Gölge 20081c5396 Update tests for the new trainer API 2021-06-22 16:21:11 +02:00
Eren Gölge 76e590f33c Update model test configs 2021-06-22 16:21:11 +02:00
Eren Gölge acd96a4940 Implement unified IO utils 2021-06-22 16:21:10 +02:00
Eren Gölge 6b907554f8 Implement unified trainer 2021-06-22 16:21:10 +02:00
Eren Gölge 20c4a8c8e1 tts model abstraction with TTSModel 2021-06-22 16:21:10 +02:00
Eren Gölge b934665fc0 fix calculation of loader_start_time 2021-06-22 16:21:10 +02:00
Eren Gölge 64f0f57757 TrainerAbstract and related updates for TrainerTTS 2021-06-22 16:21:10 +02:00
Eren Gölge f077a356e0 rename to 2021-06-22 16:21:10 +02:00
Eren Gölge 4575b70826 merge if branches with the same implementation 2021-06-22 16:21:10 +02:00
Eren Gölge 59be1b9af1 adjust distribute.py for the train_tts.py 2021-06-22 16:21:10 +02:00
Eren Gölge 614738cc85 downsize melgan test model size 2021-06-22 13:12:52 +02:00
Eren Gölge 4f29725eb6 fix glow-tts inference() 2021-06-22 13:12:52 +02:00
Eren Gölge 7ea71c7586 use one testing sentence in tts tests 2021-06-22 13:12:52 +02:00
Eren Gölge a87c886497 refactor and fix multi-speaker training in Trainer and Tacotron models 2021-06-22 13:12:52 +02:00
Eren Gölge 0206bb847b add max_decoder_steps argument to tacotron models 2021-06-22 13:12:52 +02:00
Eren Gölge cbb52b3d83 fix speaker_manager init 2021-06-22 13:12:52 +02:00
Eren Gölge 3f96491168 reduce multiband melgan test model size 2021-06-22 13:12:52 +02:00
Eren Gölge d2fd6a34a1 use get_speaker_manager in Trainer and save speakers.json file when 
needed
 2021-06-22 13:12:52 +02:00
Eren Gölge 147550c65f make style and linter fixes 2021-06-22 13:12:52 +02:00
Eren Gölge a605dd3d08 Compute d_vectors and speaker_ids separately in TTSDataset 2021-06-22 13:12:52 +02:00
Eren Gölge f00ef90ce6 rename external speaker embedding arguments as d_vectors 2021-06-22 13:12:52 +02:00
Eren Gölge e7b7268c43 use to_cuda() for moving data in format_batch() 2021-06-22 13:12:52 +02:00
Eren Gölge fdce9eb234 reduce size of the metadata.csv used at testing 2021-06-22 13:12:52 +02:00