malarinv
/
tacotron2
mirror of https://github.com/malarinv/tacotron2
1
0
Fork 0
Code Issues Releases Wiki Activity

Commit Graph

  • 6d3679d760 relax scipy version master Malar 2021-07-02 23:21:45 +0530
  • a851e80db2 1. added rpyc server optional package 2. updated path variable Malar Kannan 2020-03-05 15:59:04 +0530
  • cb0c8ddd06 1. make pyaudio as extra requirement 2. warn if pyaudio not installed if player_gen is used Malar Kannan 2020-01-22 13:52:59 +0530
  • 42a85d177e added glow as module Malar 2019-11-29 00:20:01 +0530
  • 5efb1e2758 1. fix synth_speech_fast interface 2. rename private methods Malar 2019-11-28 18:07:47 +0530
  • ea11c5199e tested gl/wavglow working Malar Kannan 2019-11-28 17:52:05 +0530
  • 78eed2d295 bugfix for vocoder arg Malar 2019-11-28 17:03:46 +0530
  • 009b87e716 update dependency version and add speed/sample rate and vocoder args Malar 2019-11-28 16:57:36 +0530
  • ac5ffcf6d5 enable gpu support if available Malar Kannan 2019-11-27 21:42:10 +0530
  • 5a30069f0a update tempo and output sample rate Malar Kannan 2019-10-09 17:25:51 +0530
  • dcc9ab3625 add tts cli args Malar Kannan 2019-10-09 16:03:29 +0530
  • dde35048b7 1. implemented corpus wav generator 2. refactor Malar Kannan 2019-10-07 16:00:35 +0530
  • 36f229449f 1. remove unneeded files 2. add Makefile Malar Kannan 2019-10-04 16:22:30 +0530
  • edbb1bd57d 1. clean-up 2. update readme and release info Malar Kannan 2019-10-04 16:15:30 +0530
  • 36c731cad0 1. update waveglow 2. add gl option and hyperparams to TTSModel Malar Kannan 2019-10-04 15:24:42 +0530
  • d0d273a698 add placeholder path to models Malar 2019-10-02 00:44:03 +0530
  • acfcadac47 update setup.py artifact and add dunder files Malar 2019-10-02 00:42:02 +0530
  • 0682eddfdc 1. add griffin lim support 2. add denoiser 3. add support to handle old and new waveglow models Malar Kannan 2019-09-26 10:53:09 +0530
  • 6d3788d858 1. using SentencePiece, Pretrained BPEmb 2. Using 40 Mel channels with 4000Hz upperbound experiments Malar Kannan 2019-09-24 10:29:34 +0530
  • f449105b79 1. updated requirements 2. spm params explicit 3. gitignore and script Malar Kannan 2019-09-23 16:26:54 +0530
  • 4d5001bdf0 1. using sentencepiece for text_to_seq 2. using 40 mel channels 3. added makefile and .gitignore Malar Kannan 2019-09-23 15:30:43 +0530
  • a10a6d517e packaged taco2 Malar Kannan 2019-09-21 01:19:30 +0530
  • 1f83e8687c implemented a grpc server for entity api Malar Kannan 2019-07-09 13:51:48 +0530
  • 755362518a update comments Malar Kannan 2019-07-05 14:33:04 +0530
  • 9413fb73b9 move codestyle to black Malar Kannan 2019-07-03 18:10:16 +0530
  • 108ce2b079 cleanup unused code and fix packaging issues Malar Kannan 2019-07-03 17:26:12 +0530
  • 5f75aa0a0d integrate tacotron2/waveglow based tts server Malar Kannan 2019-07-03 15:08:00 +0530
  • 4be2475cc1 load waveglow model from statedict Malar Kannan 2019-07-03 14:05:10 +0530
  • 08ad9ce16e 1. clean-up redundant code 2. remove ffmpeg dependency using librosa api 3. remove waveglow submodule Malar Kannan 2019-07-02 16:16:44 +0530
  • 505af768a0 1. create a class for the tts api 2. implement a grpc server for tts 3. add a demo client for the grpc service 4. update gitignore and requriements 5. cleanup Malar Kannan 2019-07-01 14:47:55 +0530
  • ccd8ab42e7 add a function to cache all responses from corpus.txt Malar Kannan 2019-06-28 18:44:24 +0530
  • 102c424eac add WORKFLOW.md and update final.py Malar Kannan 2019-06-28 14:24:36 +0530
  • 81d15abb4d return audio stream from speech Malar Kannan 2019-06-28 10:25:37 +0530
  • acf1b444a9 clean-up for cpu-only inference Malar Kannan 2019-06-28 09:46:46 +0530
  • 131c1465b4
         Merge pull request #188 from jybaek/fixed-waveglow-link Rafael Valle 2019-04-22 16:49:14 -0700
  • d5321ff0ca Fixed link to download waveglow from inference.py jybaek 2019-04-19 15:21:09 +0900
  • c76ac3b211 README.md: clarifying terminology rafaelvalle 2019-04-03 14:59:20 -0700
  • e3d2d0a5ef README.md: using proper nomenclature rafaelvalle 2019-04-03 14:56:06 -0700
  • a992aea070 README.md: updating terminology rafaelvalle 2019-04-03 14:54:45 -0700
  • eb2a171690 Merge branch 'master' of https://github.com/NVIDIA/tacotron2 rafaelvalle 2019-04-03 13:51:59 -0800
  • 821bfeba5d README.md: adding instructions to install apex rafaelvalle 2019-04-03 13:51:36 -0800
  • d6670c8ed7 Dockerfile: updating to use latest pytorch and apex rafaelvalle 2019-04-03 13:51:22 -0800
  • 0274619e45 train.py: using amp for mixed precision training rafaelvalle 2019-04-03 13:42:00 -0800
  • bb20035586 inference.ipynb: adding fp16 inference rafaelvalle 2019-04-03 13:41:11 -0800
  • 1480f82908 model.py: renaming variables, removing dropout from lstm cell state, removing conversions now handled by amp rafaelvalle 2019-04-03 13:36:35 -0800
  • 087c86755f logger.py: using new pytorch api rafaelvalle 2019-04-03 13:35:04 -0800
  • ece7d3f568 train.py: changing dataloder params given sampler Rafael Valle 2019-03-19 13:47:01 -0700
  • f37998c59d train.py: shuffling at every epoch rafaelvalle 2019-03-15 17:49:27 -0700
  • bff304f432 README.md: adding explanation on training from pre-trained model rafaelvalle 2019-03-15 17:38:40 -0700
  • 3869781877 train.py: adding routine to warm start and ignore layers, e.g. embedding.weight rafaelvalle 2019-03-15 17:34:27 -0700
  • bb67613493 hparams.py: adding ignore_layers argument to ignore text embedding layers when warm_starting rafaelvalle 2019-03-15 17:28:50 -0700
  • af1f71a975 inference.ipynb: adding code to remove waveglows bias rafaelvalle 2019-03-15 16:54:54 -0700
  • fc0d34cfce stft.py: moving window_sum to cuda if magnitude is cuda rafaelvalle 2019-03-15 14:36:56 -0700
  • f2c94d94fd
         Merge pull request #136 from GrzegorzKarchNV/master Rafael Valle 2019-02-01 12:10:42 -0800
  • df4a466af2 Fixing concatenation error for fp16 ditributed training gkarch 2019-02-01 09:55:59 +0100
  • 825ffa47d1 inference.ipynb: reverting fp16 inference for now rafaelvalle 2018-12-08 21:26:01 -0800
  • 4d7b04120a inference.ipynb: changing waverglow inference fo fp16 rafaelvalle 2018-12-05 22:14:35 -0800
  • 6e430556bd train.py: val logger on gpu 0 only rafaelvalle 2018-11-27 22:03:11 -0800
  • 3973b3e495 hparams.py: distributed using tcp rafaelvalle 2018-11-27 22:02:43 -0800
  • 52a30bb7b6 distributed.py: replacing to avoid distributed error rafaelvalle 2018-11-27 21:01:26 -0800
  • 0ad65cc053 train.py: renaming variable to n_gpus rafaelvalle 2018-11-27 21:00:05 -0800
  • 8300844fa7 hparams.py: removing 22khz rafaelvalle 2018-11-27 20:56:52 -0800
  • f06063f746 train.py: renaming function, removing dataparallel rafaelvalle 2018-11-27 18:04:12 -0800
  • 3045ba125b inference.ipynb: cleanup rafaelvalle 2018-11-27 12:04:36 -0800
  • 4c4aca3662 README.md: layout rafaelvalle 2018-11-27 11:59:05 -0800
  • 05dd8f91d2 README.md: adding submodule init to README rafaelvalle 2018-11-27 11:55:40 -0800
  • 5d66c3deab adding waveglow submodule rafaelvalle 2018-11-27 11:53:20 -0800
  • f02704f338
         Merge pull request #96 from NVIDIA/clean_slate Rafael Valle 2018-11-27 08:06:00 -0800
  • ba8cf36198 requirements.txt: removing pytorch 0.4 from requirements. upgrading to 1.0 clean_slate rafaelvalle 2018-11-27 08:04:21 -0800
  • b5e0a93946 inference.ipynb: updating inference file with relative paths rafaelvalle 2018-11-27 08:04:04 -0800
  • 58b0ec61bd README.md: updating requirements and inference demo rafaelvalle 2018-11-27 08:03:34 -0800
  • 1ad939df1a inference.ipynb: setting relative model paths rafaelvalle 2018-11-27 07:45:09 -0800
  • 8128e0e98f using waveglow, pillow, text changes latest_model rafaelvalle 2018-11-26 16:42:07 -0800
  • e0c7e0e928 adding changes to text rafaelvalle 2018-11-26 16:41:21 -0800
  • f03692c73d cleanup, new model and waveglow rafaelvalle 2018-11-26 16:37:44 -0800
  • 32b9a135d0 utils.py: updating rafaelvalle 2018-11-25 22:34:38 -0800
  • ce29e13959 train.py: updating rafaelvalle 2018-11-25 22:34:34 -0800
  • 1ea6ed5861 text/symbols.py: updating symbols rafaelvalle 2018-11-25 22:34:26 -0800
  • cdfde985e5 text/__init__.py: remove stop token rafaelvalle 2018-11-25 22:34:11 -0800
  • e314bb4cd0 stft.py: fix filter winlength error rafaelvalle 2018-11-25 22:33:52 -0800
  • 4af4ccb135 model.py: rewrite rafaelvalle 2018-11-25 22:33:38 -0800
  • 1ec0e5e8cd layers.py: rewrite rafaelvalle 2018-11-25 22:33:32 -0800
  • 249afd8043 inference.ipynb: import taco2model to be public rafaelvalle 2018-11-25 22:33:16 -0800
  • 1b243d5d5a hparams.py:rewrite rafaelvalle 2018-11-25 22:33:05 -0800
  • d0aa9e7d32 distributed.py: rewrite rafaelvalle 2018-11-25 22:32:54 -0800
  • 1683a57ae5 data_utils.py: rewrite rafaelvalle 2018-11-25 22:32:47 -0800
  • fc0cf6a89a
         Merge pull request #53 from cobr123/patch-1 Rafael Valle 2018-07-02 14:32:21 -0700
  • 8de38495be
         add pillow cobr123 2018-07-02 19:35:51 +0300
  • 7eb045206c README.md: updating readme to explicitly mention that mel representation of WaveNet and Tacotron2 must be the same rafaelvalle 2018-06-14 11:25:42 -0700
  • c67005f1be Dockerfile: adding jupyter to dockerfile rafaelvalle 2018-06-14 10:30:01 -0700
  • cb3794796f Dockerfile: removing return from Dockerfile rafaelvalle 2018-06-12 21:38:03 -0700
  • a8de973923
         Merge pull request #37 from yoks/master Rafael Valle 2018-06-12 09:00:14 -0700
  • a0ae2da05f
         `used_saved_learning_rate` fix yoks 2018-06-12 16:53:12 +0300
  • 34066ac4fc requirements.txt: setting torch to 0.4.0 rafaelvalle 2018-06-11 08:16:31 -0700
  • 12ab5ba89c model.py: setting weight initialization to xavier uniform rafaelvalle 2018-06-07 20:28:52 -0700
  • d10da5f41e hparams.py: commenting n_frames_per_step to indicate that currently only 1 frame per step is supported now rafaelvalle 2018-06-07 13:02:23 -0700
  • 5f0ea06c41 hparams.py: adding use saved learning rate param rafaelvalle 2018-06-05 08:12:49 -0700
  • 22bcff1155 hparams.py: adding use saved learning rate param rafaelvalle 2018-06-05 08:12:35 -0700
  • 2e934478a4 README.md: being explicit about action rafaelvalle 2018-06-04 17:12:32 -0700
  • 8ae231b10b README.md: more explicit about demo audio rafaelvalle 2018-06-04 16:56:22 -0700