From 7b7f2a9c29e156d3d514012f8f4545e17c74b49a Mon Sep 17 00:00:00 2001 From: ming024 Date: Thu, 25 Jun 2020 23:13:07 +0800 Subject: [PATCH] minor revision --- .gitignore | 2 +- LICENSE | 21 +++++++++++++++++++++ README.md | 4 ++-- train.py | 3 +-- 4 files changed, 25 insertions(+), 5 deletions(-) create mode 100644 LICENSE diff --git a/.gitignore b/.gitignore index 909089f8cd..e24925fbb6 100644 --- a/.gitignore +++ b/.gitignore @@ -113,7 +113,7 @@ montreal-forced-aligner/ # data, checkpoint, and models preprocessed/ ckpt/ -result/ +results/ synth/ log/ eval/ diff --git a/LICENSE b/LICENSE new file mode 100644 index 0000000000..c4d469548b --- /dev/null +++ b/LICENSE @@ -0,0 +1,21 @@ +MIT License + +Copyright (c) 2020 Chung-Ming Chien + +Permission is hereby granted, free of charge, to any person obtaining a copy +of this software and associated documentation files (the "Software"), to deal +in the Software without restriction, including without limitation the rights +to use, copy, modify, merge, publish, distribute, sublicense, and/or sell +copies of the Software, and to permit persons to whom the Software is +furnished to do so, subject to the following conditions: + +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +SOFTWARE. diff --git a/README.md b/README.md index f29edc4c74..547b3d1311 100644 --- a/README.md +++ b/README.md @@ -6,7 +6,7 @@ This is a Pytorch implementation of Microsoft's text-to-speech system [**FastSpe # Audio Samples Audio samples generated by this implementation can be found [here](https://ming024.github.io/FastSpeech2/). -- The model used to generate these samples is trained for 30k steps on [LJSpeech](https://keithito.com/LJ-Speech-Dataset/) dataset. +- The model used to generate these samples is trained for 300k steps on [LJSpeech](https://keithito.com/LJ-Speech-Dataset/) dataset. - Audio samples are converted from mel-spectrogram to raw waveform via [NVIDIA's pretrained WaveGlow](https://github.com/NVIDIA/waveglow). # Quickstart @@ -89,7 +89,7 @@ Train your model with python3 train.py ``` -The model takes less than 10000 steps (less than 1 hour on my GTX1080 GPU) of training to generate audio samples with acceptable quality, which is much more efficient than the autoregressive models such as Tacotron2. +The model takes less than 10k steps (less than 1 hour on my GTX1080 GPU) of training to generate audio samples with acceptable quality, which is much more efficient than the autoregressive models such as Tacotron2. There might be some room for improvement for this repository. For example, I just simply add up the duration loss, f0 loss, energy loss and mel loss without any weighting. Please inform me if you find any useful tip for training the FastSpeech2 model. diff --git a/train.py b/train.py index b1ac88e6d7..201911ede0 100644 --- a/train.py +++ b/train.py @@ -101,8 +101,7 @@ def main(args): # Cal Loss mel_loss, mel_postnet_loss, d_loss, f_loss, e_loss = Loss( - duration_output, D, f0_output, f0, energy_output, energy, mel_output, mel_postnet_output, mel_target, mel_len) - + duration_output, D, f0_output, f0, energy_output, energy, mel_output, mel_postnet_output, mel_target) total_loss = mel_loss + mel_postnet_loss + d_loss + f_loss + e_loss # Logger