From d1d6c206720963ae39b811e4e2c515985669f277 Mon Sep 17 00:00:00 2001
From: chenfeiyu <chenfeiyu@baidu.com>
Date: Wed, 30 Dec 2020 14:36:23 +0800
Subject: [PATCH] add README for transformer_tts, waveflow and wavenet

---
 examples/transformer_tts/README.md | 48 ++++++++++++++++++++++++++++++
 examples/waveflow/README.md        | 48 ++++++++++++++++++++++++++++++
 examples/wavenet/README.md         | 48 ++++++++++++++++++++++++++++++
 3 files changed, 144 insertions(+)
 create mode 100644 examples/transformer_tts/README.md
 create mode 100644 examples/waveflow/README.md
 create mode 100644 examples/wavenet/README.md

diff --git a/examples/transformer_tts/README.md b/examples/transformer_tts/README.md
new file mode 100644
index 0000000..45f8816
--- /dev/null
+++ b/examples/transformer_tts/README.md
@@ -0,0 +1,48 @@
+# TransformerTTS with LJSpeech
+
+## Dataset
+
+### Download the datasaet.
+
+```bash
+wget https://data.keithito.com/data/speech/LJSpeech-1.1.tar.bz2
+```
+
+### Extract the dataset.
+
+```bash
+tar xjvf LJSpeech-1.1.tar.bz2
+```
+
+### Preprocess the dataset. 
+
+Assume the path to save the preprocessed dataset is `ljspeech_wavenet`. Run the command below to preprocess the dataset.
+
+```bash
+python preprocess.py --input=LJSpeech-1.1/  --output=ljspeech_wavenet
+```
+
+## Train the model
+
+The training script requires 4 command line arguments.
+`--data` is the path of the training dataset, `--output` is the path of the output direcctory (we recommend to use a subdirectory in `runs` to manage different experiments.)
+
+`--device` should be "cpu" or "gpu", `--nprocs` is the number of processes to train the model in parallel.
+
+```bash
+python train.py --data=ljspeech_wavenet/ --output=runs/test --device="gpu" --nprocs=1
+```
+
+If you want distributed training, set a larger `--nprocs` (e.g. 4). Note that distributed training with cpu is not supported yet.
+
+## Synthesize
+
+Synthesize waveform. We assume the `--input` is text file, a sentence per line, and `--output` directory a directory to save the synthesized mel spectrogram(log magnitude) in `.npy` format. The mel spectrogram can be used with `Waveflow` to generate waveforms.
+
+`--checkpoint_path` should be the path of the parameter file (`.pdparams`) to load. Note that the extention name `.pdparmas` is not included here.
+
+`--device` specifiies to device to run synthesis. Due to the autoregressiveness of wavenet, using cpu may be faster.
+
+```bash
+python synthesize.py --input=sentence.txt --output=mels/ --checkpoint_path='step-310000' --device="gpu" --verbose
+```
\ No newline at end of file
diff --git a/examples/waveflow/README.md b/examples/waveflow/README.md
new file mode 100644
index 0000000..ba39402
--- /dev/null
+++ b/examples/waveflow/README.md
@@ -0,0 +1,48 @@
+# WaveFlow with LJSpeech
+
+## Dataset
+
+### Download the datasaet.
+
+```bash
+wget https://data.keithito.com/data/speech/LJSpeech-1.1.tar.bz2
+```
+
+### Extract the dataset.
+
+```bash
+tar xjvf LJSpeech-1.1.tar.bz2
+```
+
+### Preprocess the dataset. 
+
+Assume the path to save the preprocessed dataset is `ljspeech_wavenet`. Run the command below to preprocess the dataset.
+
+```bash
+python preprocess.py --input=LJSpeech-1.1/  --output=ljspeech_wavenet
+```
+
+## Train the model
+
+The training script requires 4 command line arguments.
+`--data` is the path of the training dataset, `--output` is the path of the output direcctory (we recommend to use a subdirectory in `runs` to manage different experiments.)
+
+`--device` should be "cpu" or "gpu", `--nprocs` is the number of processes to train the model in parallel.
+
+```bash
+python train.py --data=ljspeech_wavenet/ --output=runs/test --device="gpu" --nprocs=1
+```
+
+If you want distributed training, set a larger `--nprocs` (e.g. 4). Note that distributed training with cpu is not supported yet.
+
+## Synthesize
+
+Synthesize waveform. We assume the `--input` is a directory containing several mel spectrogram(log magnitude) in `.npy` format. The output would be saved in `--output` directory, containing several `.wav` files with the same name as the mel spectrogram does.
+
+`--checkpoint_path` should be the path of the parameter file (`.pdparams`) to load. Note that the extention name `.pdparmas` is not included here.
+
+`--device` specifiies to device to run synthesis. Due to the autoregressiveness of wavenet, using cpu may be faster.
+
+```bash
+python synthesize.py --input=mels/ --output=wavs/ --checkpoint_path='step-2000000' --device="cpu" --verbose
+```
\ No newline at end of file
diff --git a/examples/wavenet/README.md b/examples/wavenet/README.md
new file mode 100644
index 0000000..0224742
--- /dev/null
+++ b/examples/wavenet/README.md
@@ -0,0 +1,48 @@
+# WaveNet with LJSpeech
+
+## Dataset
+
+### Download the datasaet.
+
+```bash
+wget https://data.keithito.com/data/speech/LJSpeech-1.1.tar.bz2
+```
+
+### Extract the dataset.
+
+```bash
+tar xjvf LJSpeech-1.1.tar.bz2
+```
+
+### Preprocess the dataset. 
+
+Assume the path to save the preprocessed dataset is `ljspeech_wavenet`. Run the command below to preprocess the dataset.
+
+```bash
+python preprocess.py --input=LJSpeech-1.1/  --output=ljspeech_wavenet
+```
+
+## Train the model
+
+The training script requires 4 command line arguments.
+`--data` is the path of the training dataset, `--output` is the path of the output direcctory (we recommend to use a subdirectory in `runs` to manage different experiments.)
+
+`--device` should be "cpu" or "gpu", `--nprocs` is the number of processes to train the model in parallel.
+
+```bash
+python train.py --data=ljspeech_wavenet/ --output=runs/test --device="gpu" --nprocs=1
+```
+
+If you want distributed training, set a larger `--nprocs` (e.g. 4). Note that distributed training with cpu is not supported yet.
+
+## Synthesize
+
+Synthesize waveform. We assume the `--input` is a directory containing several mel spectrogram(log magnitude) in `.npy` format. The output would be saved in `--output` directory, containing several `.wav` files with the same name as the mel spectrogram does.
+
+`--checkpoint_path` should be the path of the parameter file (`.pdparams`) to load. Note that the extention name `.pdparmas` is not included here.
+
+`--device` specifiies to device to run synthesis. Due to the autoregressiveness of wavenet, using cpu may be faster.
+
+```bash
+python synthesize.py --input=mels/ --output=wavs/ --checkpoint_path='step-2450000' --device="cpu" --verbose
+```
\ No newline at end of file