[tts][ernie_sat]add ernie sat synthesize_e2e

examples/aishell3/README.md

@@ -10,3 +10,4 @@
 * voc3 - MultiBand MelGAN
 * vc0 - Tactron2 Voice Cloning with GE2E
 * vc1 - FastSpeech2 Voice Cloning with GE2E
+* ernie_sat - ERNIE-SAT

examples/aishell3/ernie_sat/README.md

@@ -1 +1,151 @@
-# ERNIE SAT with AISHELL3 dataset
+# ERNIE-SAT with AISHELL3 dataset
+
+ERNIE-SAT 是可以同时处理中英文的跨语言的语音-语言跨模态大模型，其在语音编辑、个性化语音合成以及跨语言的语音合成等多个任务取得了领先效果。可以应用于语音编辑、个性化合成、语音克隆、同传翻译等一系列场景，该项目供研究使用。
+
+## 模型框架
+ERNIE-SAT 中我们提出了两项创新：
+- 在预训练过程中将中英双语对应的音素作为输入，实现了跨语言、个性化的软音素映射
+- 采用语言和语音的联合掩码学习实现了语言和语音的对齐
+
+<p align="center">
+    <img src="https://user-images.githubusercontent.com/24568452/186110814-1b9c6618-a0ab-4c0c-bb3d-3d860b0e8cc2.png" />
+</p>
+
+## Dataset
+### Download and Extract
+Download AISHELL-3 from it's [Official Website](http://www.aishelltech.com/aishell_3) and extract it to `~/datasets`. Then the dataset is in the directory `~/datasets/data_aishell3`.
+
+### Get MFA Result and Extract
+We use [MFA2.x](https://github.com/MontrealCorpusTools/Montreal-Forced-Aligner) to get durations for aishell3_fastspeech2.
+You can download from here [aishell3_alignment_tone.tar.gz](https://paddlespeech.bj.bcebos.com/MFA/AISHELL-3/with_tone/aishell3_alignment_tone.tar.gz), or train your MFA model reference to [mfa example](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/examples/other/mfa) (use MFA1.x now) of our repo.
+
+## Get Started
+Assume the path to the dataset is `~/datasets/data_aishell3`.
+Assume the path to the MFA result of AISHELL-3 is `./aishell3_alignment_tone`.
+Run the command below to
+1. **source path**.
+2. preprocess the dataset.
+3. train the model.
+4. synthesize wavs.
+    - synthesize waveform from `metadata.jsonl`.
+    - synthesize waveform from text file.
+
+```bash
+./run.sh
+```
+You can choose a range of stages you want to run, or set `stage` equal to `stop-stage` to use only one stage, for example, running the following command will only preprocess the dataset.
+```bash
+./run.sh --stage 0 --stop-stage 0
+```
+### Data Preprocessing
+```bash
+./local/preprocess.sh ${conf_path}
+```
+When it is done. A `dump` folder is created in the current directory. The structure of the dump folder is listed below.
+
+```text
+dump
+├── dev
+│   ├── norm
+│   └── raw
+├── phone_id_map.txt
+├── speaker_id_map.txt
+├── test
+│   ├── norm
+│   └── raw
+└── train
+    ├── norm
+    ├── raw
+    └── speech_stats.npy
+```
+The dataset is split into 3 parts, namely `train`, `dev`, and` test`, each of which contains a `norm` and `raw` subfolder. The raw folder contains speech features of each utterance, while the norm folder contains normalized ones. The statistics used to normalize features are computed from the training set, which is located in `dump/train/*_stats.npy`.
+
+Also, there is a `metadata.jsonl` in each subfolder. It is a table-like file that contains phones, text_lengths, speech_lengths, durations, the path of speech features, speaker, and id of each utterance.
+
+### Model Training
+```bash
+CUDA_VISIBLE_DEVICES=${gpus} ./local/train.sh ${conf_path} ${train_output_path}
+```
+`./local/train.sh` calls `${BIN_DIR}/train.py`.
+
+### Synthesizing
+We use [HiFiGAN](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/examples/aishell3/voc5) as the neural vocoder.
+
+Download pretrained HiFiGAN model from [hifigan_aishell3_ckpt_0.2.0.zip](https://paddlespeech.bj.bcebos.com/Parakeet/released_models/hifigan/hifigan_aishell3_ckpt_0.2.0.zip) and unzip it.
+```bash
+unzip hifigan_aishell3_ckpt_0.2.0.zip
+```
+HiFiGAN checkpoint contains files listed below.
+```text
+hifigan_aishell3_ckpt_0.2.0
+├── default.yaml                    # default config used to train HiFiGAN
+├── feats_stats.npy                 # statistics used to normalize spectrogram when training HiFiGAN
+└── snapshot_iter_2500000.pdz       # generator parameters of HiFiGAN
+```
+`./local/synthesize.sh` calls `${BIN_DIR}/../synthesize.py`, which can synthesize waveform from `metadata.jsonl`.
+```bash
+CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh ${conf_path} ${train_output_path} ${ckpt_name}
+```
+##  Speech Synthesis and Speech Editing
+### Prepare
+**prepare aligner**
+```bash
+mkdir -p tools/aligner
+cd tools
+# download MFA
+wget https://github.com/MontrealCorpusTools/Montreal-Forced-Aligner/releases/download/v1.0.1/montreal-forced-aligner_linux.tar.gz
+# extract MFA
+tar xvf montreal-forced-aligner_linux.tar.gz
+# fix .so of MFA
+cd montreal-forced-aligner/lib
+ln -snf libpython3.6m.so.1.0 libpython3.6m.so
+cd -
+# download align models and dicts
+cd aligner
+wget https://paddlespeech.bj.bcebos.com/MFA/ernie_sat/aishell3_model.zip
+wget https://paddlespeech.bj.bcebos.com/MFA/AISHELL-3/with_tone/simple.lexicon
+wget https://paddlespeech.bj.bcebos.com/MFA/ernie_sat/vctk_model.zip
+wget https://paddlespeech.bj.bcebos.com/MFA/LJSpeech-1.1/cmudict-0.7b
+cd ../../
+```
+**prepare pretrained FastSpeech2 models**
+
+ERNIE-SAT use FastSpeech2 as phoneme duration predictor:
+```bash
+mkdir download
+cd download
+wget https://paddlespeech.bj.bcebos.com/Parakeet/released_models/fastspeech2/fastspeech2_conformer_baker_ckpt_0.5.zip
+wget https://paddlespeech.bj.bcebos.com/Parakeet/released_models/fastspeech2/fastspeech2_nosil_ljspeech_ckpt_0.5.zip
+unzip fastspeech2_conformer_baker_ckpt_0.5.zip
+unzip fastspeech2_nosil_ljspeech_ckpt_0.5.zip
+cd ../
+```
+**prepare source data**
+```bash
+mkdir source
+cd source
+wget https://paddlespeech.bj.bcebos.com/Parakeet/released_models/ernie_sat/source/SSB03540307.wav
+wget https://paddlespeech.bj.bcebos.com/Parakeet/released_models/ernie_sat/source/SSB03540428.wav
+wget https://paddlespeech.bj.bcebos.com/Parakeet/released_models/ernie_sat/source/LJ050-0278.wav
+wget https://paddlespeech.bj.bcebos.com/Parakeet/released_models/ernie_sat/source/p243_313.wav
+wget https://paddlespeech.bj.bcebos.com/Parakeet/released_models/ernie_sat/source/p299_096.wav
+wget https://paddlespeech.bj.bcebos.com/Parakeet/released_models/ernie_sat/source/this_was_not_the_show_for_me.wav
+wget https://paddlespeech.bj.bcebos.com/Parakeet/released_models/ernie_sat/source/README.md
+cd ../
+```
+
+You can check the text of downloaded wavs in `source/README.md`.
+### Speech Synthesis and Speech Editing
+```bash
+./run.sh --stage 3 --stop-stage 3 --gpus 0
+```
+`stage 3` of `run.sh` calls `local/synthesize_e2e.sh`, `stage 0` of it is **Speech Synthesis** and  `stage 1` of it is **Speech Editing**.
+
+You can modify `--wav_path`、`--old_str` and `--new_str` yourself, `--old_str`  should be the text corresponding to the audio of  `--wav_path`, `--new_str` should be designed according to `--task_name`, both `--source_lang` and `--target_lang` should be `zh` for model trained with AISHELL3 dataset.
+## Pretrained Model
+Pretrained ErnieSAT model:
+- [erniesat_aishell3_ckpt_1.2.0.zip](https://paddlespeech.bj.bcebos.com/Parakeet/released_models/ernie_sat/erniesat_aishell3_ckpt_1.2.0.zip)
+
+Model | Step | eval/mlm_loss | eval/loss
+:-------------:| :------------:| :-----: | :-----:
+default| 8(gpu) x 289500|51.723782|51.723782

examples/aishell3/ernie_sat/conf/default.yaml

@@ -1,3 +1,6 @@
+# This configuration tested on 8 GPUs (A100) with 80GB GPU memory.
+# It takes around 3 days to finish the training,You can adjust
+# batch_size、num_workers here and ngpu in local/train.sh for your machine
 ###########################################################
 #                FEATURE EXTRACTION SETTING               #
 ###########################################################
@@ -21,8 +24,8 @@ mlm_prob: 0.8
 ###########################################################
 #                       DATA SETTING                      #
 ###########################################################
-batch_size: 20
-num_workers: 2
+batch_size: 40
+num_workers: 8
 
 ###########################################################
 #                       MODEL SETTING                     #
@@ -280,4 +283,4 @@ token_list:
 - o3
 - iang5
 - ei5
-- <sos/eos>
+- <sos/eos>

examples/aishell3/ernie_sat/local/synthesize.sh

@@ -4,28 +4,11 @@ config_path=$1
 train_output_path=$2
 ckpt_name=$3
 
-stage=1
-stop_stage=1
-
-# pwgan
-if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
-    FLAGS_allocator_strategy=naive_best_fit \
-    FLAGS_fraction_of_gpu_memory_to_use=0.01 \
-    python3 ${BIN_DIR}/synthesize.py \
-        --erniesat_config=${config_path} \
-        --erniesat_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
-        --erniesat_stat=dump/train/speech_stats.npy \
-        --voc=pwgan_aishell3 \
-        --voc_config=pwg_aishell3_ckpt_0.5/default.yaml \
-        --voc_ckpt=pwg_aishell3_ckpt_0.5/snapshot_iter_1000000.pdz \
-        --voc_stat=pwg_aishell3_ckpt_0.5/feats_stats.npy \
-        --test_metadata=dump/test/norm/metadata.jsonl \
-        --output_dir=${train_output_path}/test \
-        --phones_dict=dump/phone_id_map.txt
-fi
+stage=0
+stop_stage=0
 
 # hifigan
-if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
+if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
     FLAGS_allocator_strategy=naive_best_fit \
     FLAGS_fraction_of_gpu_memory_to_use=0.01 \
     python3 ${BIN_DIR}/synthesize.py \

examples/aishell3/ernie_sat/local/synthesize_e2e.sh

@@ -0,0 +1,52 @@
+#!/bin/bash
+
+config_path=$1
+train_output_path=$2
+ckpt_name=$3
+
+stage=0
+stop_stage=1
+
+if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
+    echo 'speech synthesize !'
+    FLAGS_allocator_strategy=naive_best_fit \
+    FLAGS_fraction_of_gpu_memory_to_use=0.01 \
+    python3 ${BIN_DIR}/synthesize_e2e.py \
+        --task_name=synthesize \
+        --wav_path=source/SSB03540307.wav\
+        --old_str='请播放歌曲小苹果。' \
+        --new_str='歌曲真好听。' \
+        --source_lang=zh \
+        --target_lang=zh \
+        --erniesat_config=${config_path} \
+        --phones_dict=dump/phone_id_map.txt \
+        --erniesat_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+        --erniesat_stat=dump/train/speech_stats.npy \
+        --voc=hifigan_aishell3 \
+        --voc_config=hifigan_aishell3_ckpt_0.2.0/default.yaml \
+        --voc_ckpt=hifigan_aishell3_ckpt_0.2.0/snapshot_iter_2500000.pdz \
+        --voc_stat=hifigan_aishell3_ckpt_0.2.0/feats_stats.npy \
+        --output_name=exp/pred_gen.wav
+fi
+
+if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
+    echo 'speech edit !'
+    FLAGS_allocator_strategy=naive_best_fit \
+    FLAGS_fraction_of_gpu_memory_to_use=0.01 \
+    python3 ${BIN_DIR}/synthesize_e2e.py \
+        --task_name=edit \
+        --wav_path=source/SSB03540428.wav \
+        --old_str='今天天气很好' \
+        --new_str='今天心情很好' \
+        --source_lang=zh \
+        --target_lang=zh \
+        --erniesat_config=${config_path} \
+        --phones_dict=dump/phone_id_map.txt \
+        --erniesat_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+        --erniesat_stat=dump/train/speech_stats.npy \
+        --voc=hifigan_aishell3 \
+        --voc_config=hifigan_aishell3_ckpt_0.2.0/default.yaml \
+        --voc_ckpt=hifigan_aishell3_ckpt_0.2.0/snapshot_iter_2500000.pdz \
+        --voc_stat=hifigan_aishell3_ckpt_0.2.0/feats_stats.npy \
+        --output_name=exp/pred_edit.wav
+fi

examples/aishell3/ernie_sat/local/train.sh

@@ -8,5 +8,5 @@ python3 ${BIN_DIR}/train.py \
     --dev-metadata=dump/dev/norm/metadata.jsonl \
     --config=${config_path} \
     --output-dir=${train_output_path} \
-    --ngpu=2 \
-    --phones-dict=dump/phone_id_map.txt
+    --ngpu=8 \
+    --phones-dict=dump/phone_id_map.txt

examples/aishell3/ernie_sat/run.sh

@@ -9,7 +9,7 @@ stop_stage=100
 
 conf_path=conf/default.yaml
 train_output_path=exp/default
-ckpt_name=snapshot_iter_153.pdz
+ckpt_name=snapshot_iter_289500.pdz
 
 # with the following command, you can choose the stage range you want to run
 # such as `./run.sh --stage 0 --stop-stage 0`
@@ -30,3 +30,7 @@ if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
     # synthesize, vocoder is pwgan
     CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
 fi
+
+if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
+fi

examples/aishell3_vctk/README.md

@@ -1 +1,2 @@
 # Mixed Chinese and English TTS with AISHELL3 and VCTK datasets
+* ernie_sat - ERNIE-SAT