[TTS]Update VITS to support VITS and its voice cloning training on AIShell-3 (#2268)

* code for training vits voice clone on aishell3.
Co-authored-by: TianYuan <white-sky@qq.com>

README.md

@@ -613,7 +613,7 @@ PaddleSpeech supports a series of most popular models. They are summarized in [r
       </td>
     </tr>
     <tr>
-      <td rowspan="3">Voice Cloning</td>
+      <td rowspan="4">Voice Cloning</td>
       <td>GE2E</td>
       <td >Librispeech, etc.</td>
       <td>
@@ -633,13 +633,20 @@ PaddleSpeech supports a series of most popular models. They are summarized in [r
       <td>
       <a href = "./examples/aishell3/vc1">ge2e-fastspeech2-aishell3</a>
       </td>
+    </tr>
+    <tr>
+      <td>GE2E + VITS</td>
+      <td>AISHELL-3</td>
+      <td>
+      <a href = "./examples/aishell3/vits-vc">ge2e-vits-aishell3</a>
+      </td>
     </tr>
      <tr>
       <td rowspan="3">End-to-End</td>
       <td>VITS</td>
-      <td >CSMSC</td>
+      <td>CSMSC / AISHELL-3</td>
       <td>
-      <a href = "./examples/csmsc/vits">VITS-csmsc</a>
+      <a href = "./examples/csmsc/vits">VITS-csmsc</a> / <a href = "./examples/aishell3/vits">VITS-aishell3</a>
       </td>
     </tr>
   </tbody>

README_cn.md

@@ -608,7 +608,7 @@ PaddleSpeech 的 **语音合成** 主要包含三个模块：文本前端、声
       </td>
     </tr>
     <tr>
-      <td rowspan="3">声音克隆</td>
+      <td rowspan="4">声音克隆</td>
       <td>GE2E</td>
       <td >Librispeech, etc.</td>
       <td>
@@ -629,13 +629,20 @@ PaddleSpeech 的 **语音合成** 主要包含三个模块：文本前端、声
       <a href = "./examples/aishell3/vc1">ge2e-fastspeech2-aishell3</a>
       </td>
     </tr>
+    <tr>
+      <td>GE2E + VITS</td>
+      <td>AISHELL-3</td>
+      <td>
+      <a href = "./examples/aishell3/vits-vc">ge2e-vits-aishell3</a>
+      </td>
+    </tr>
     </tr>
      <tr>
       <td rowspan="3">端到端</td>
       <td>VITS</td>
-      <td >CSMSC</td>
+      <td>CSMSC / AISHELL-3</td>
       <td>
-      <a href = "./examples/csmsc/vits">VITS-csmsc</a>
+      <a href = "./examples/csmsc/vits">VITS-csmsc</a> / <a href = "./examples/aishell3/vits">VITS-aishell3</a>
       </td>
     </tr>
   </tbody>

examples/aishell3/vits-vc/README.md

+# VITS with AISHELL-3
+This example contains code used to train a [VITS](https://arxiv.org/abs/2106.06103) model with [AISHELL-3](http://www.aishelltech.com/aishell_3). The trained model can be used in Voice Cloning Task, We refer to the model structure of  [Transfer Learning from Speaker Verification to Multispeaker Text-To-Speech Synthesis](https://arxiv.org/pdf/1806.04558.pdf). The general steps are as follows:
+1. Speaker Encoder: We use Speaker Verification to train a speaker encoder. Datasets used in this task are different from those used in `VITS` because the transcriptions are not needed, we use more datasets, refer to  [ge2e](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/examples/other/ge2e).
+2. Synthesizer and Vocoder: We use the trained speaker encoder to generate speaker embedding for each sentence in AISHELL-3. This embedding is an extra input of `VITS` which will be concated with encoder outputs. The vocoder is part of `VITS` due to its special structure.
+
+## 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 phonemes for VITS, the durations of MFA are not needed here.
+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.
+
+## Pretrained GE2E Model
+We use pretrained GE2E model to generate speaker embedding for each sentence.
+
+Download pretrained GE2E model from here [ge2e_ckpt_0.3.zip](https://bj.bcebos.com/paddlespeech/Parakeet/released_models/ge2e/ge2e_ckpt_0.3.zip), and `unzip` it.
+
+## 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`.
+Assume the path to the pretrained ge2e model is `./ge2e_ckpt_0.3`.
+
+Run the command below to
+1. **source path**.
+2. preprocess the dataset.
+3. train the model.
+4. synthesize waveform from `metadata.jsonl`.
+5. start a voice cloning inference.
+
+```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
+CUDA_VISIBLE_DEVICES=${gpus} ./local/preprocess.sh ${conf_path} ${ge2e_ckpt_path}
+```
+When it is done. A `dump` folder is created in the current directory. The structure of the dump folder is listed below.
+
+```text
+dump
+├── dev
+│   ├── norm
+│   └── raw
+├── embed
+│   ├── SSB0005
+│   ├── SSB0009
+│   ├── ...
+│   └── ...
+├── phone_id_map.txt
+├── speaker_id_map.txt
+├── test
+│   ├── norm
+│   └── raw
+└── train
+    ├── feats_stats.npy
+    ├── norm
+    └── raw
+```
+The `embed` contains the generated speaker embedding for each sentence in AISHELL-3, which has the same file structure with wav files and the format is  `.npy`.
+
+The computing time of utterance embedding can be x hours.
+
+The dataset is split into 3 parts, namely `train`, `dev`, and` test`, each of which contains a `norm` and `raw` subfolder. The raw folder contains wave and linear spectrogram of each utterance, while the norm folder contains normalized ones. The statistics used to normalize features are computed from the training set, which is located in `dump/train/feats_stats.npy`.
+
+Also, there is a `metadata.jsonl` in each subfolder. It is a table-like file that contains phones, text_lengths, feats, feats_lengths, the path of linear spectrogram features, the path of raw waves, speaker, and the id of each utterance.
+
+The preprocessing step is very similar to that one of [vits](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/examples/aishell3/vits), but there is one more `ge2e/inference` step here.
+
+### Model Training
+```bash
+CUDA_VISIBLE_DEVICES=${gpus} ./local/train.sh ${conf_path} ${train_output_path}
+```
+The training step is very similar to that one of [vits](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/examples/aishell3/vits), but we should set `--voice-cloning=True` when calling `${BIN_DIR}/train.py`.
+
+### Synthesizing
+
+`./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} ${ckpt_name}
+```
+```text
+usage: synthesize.py [-h] [--config CONFIG] [--ckpt CKPT]
+                     [--phones_dict PHONES_DICT] [--speaker_dict SPEAKER_DICT]
+                     [--voice-cloning VOICE_CLONING] [--ngpu NGPU]
+                     [--test_metadata TEST_METADATA] [--output_dir OUTPUT_DIR]
+
+Synthesize with VITS
+
+optional arguments:
+  -h, --help            show this help message and exit
+  --config CONFIG       Config of VITS.
+  --ckpt CKPT           Checkpoint file of VITS.
+  --phones_dict PHONES_DICT
+                        phone vocabulary file.
+  --speaker_dict SPEAKER_DICT
+                        speaker id map file.
+  --voice-cloning VOICE_CLONING
+                        whether training voice cloning model.
+  --ngpu NGPU           if ngpu == 0, use cpu.
+  --test_metadata TEST_METADATA
+                        test metadata.
+  --output_dir OUTPUT_DIR
+                        output dir.
+```
+The synthesizing step is very similar to that one of [vits](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/examples/aishell3/vits), but we should set `--voice-cloning=True` when calling `${BIN_DIR}/../synthesize.py`.
+
+### Voice Cloning
+Assume there are some  reference audios in `./ref_audio`
+```text
+ref_audio
+├── 001238.wav
+├── LJ015-0254.wav
+└── audio_self_test.mp3
+```
+`./local/voice_cloning.sh` calls `${BIN_DIR}/voice_cloning.py`
+
+```bash
+CUDA_VISIBLE_DEVICES=${gpus} ./local/voice_cloning.sh ${conf_path} ${train_output_path} ${ckpt_name} ${ge2e_params_path} ${add_blank} ${ref_audio_dir}
+```
+
+If you want to convert a speaker audio file to refered speaker, run:
+
+```bash
+CUDA_VISIBLE_DEVICES=${gpus} ./local/voice_cloning.sh ${conf_path} ${train_output_path} ${ckpt_name} ${ge2e_params_path} ${add_blank} ${ref_audio_dir} ${src_audio_path}
+```
+
+<!-- TODO display these after we trained the model -->
+<!-- 
+## Pretrained Model
+
+The pretrained model can be downloaded here:
+
+- [vits_vc_aishell3_ckpt_1.1.0.zip](https://paddlespeech.bj.bcebos.com/Parakeet/released_models/vits/vits_vc_aishell3_ckpt_1.1.0.zip) (add_blank=true)
+
+VITS checkpoint contains files listed below.
+(There is no need for `speaker_id_map.txt` here )
+
+```text
+vits_vc_aishell3_ckpt_1.1.0
+├── default.yaml              # default config used to train vitx
+├── phone_id_map.txt          # phone vocabulary file when training vits
+└── snapshot_iter_333000.pdz  # model parameters and optimizer states
+```
+
+ps: This ckpt is not good enough, a better result is training
+
+-->

examples/aishell3/vits-vc/conf/default.yaml

+# This configuration tested on 4 GPUs (V100) with 32GB GPU
+# memory. It takes around 2 weeks to finish the training
+# but 100k iters model should generate reasonable results.
+###########################################################
+#                FEATURE EXTRACTION SETTING               #
+###########################################################
+
+fs: 22050         # sr
+n_fft: 1024        # FFT size (samples).
+n_shift: 256       # Hop size (samples). 12.5ms
+win_length: null   # Window length (samples). 50ms
+                   # If set to null, it will be the same as fft_size.
+window: "hann"     # Window function.
+
+
+##########################################################
+#                  TTS MODEL SETTING                     #
+##########################################################
+model:
+    # generator related
+    generator_type: vits_generator
+    generator_params:
+        hidden_channels: 192
+        spk_embed_dim: 256
+        global_channels: 256
+        segment_size: 32
+        text_encoder_attention_heads: 2
+        text_encoder_ffn_expand: 4
+        text_encoder_blocks: 6
+        text_encoder_positionwise_layer_type: "conv1d"
+        text_encoder_positionwise_conv_kernel_size: 3
+        text_encoder_positional_encoding_layer_type: "rel_pos"
+        text_encoder_self_attention_layer_type: "rel_selfattn"
+        text_encoder_activation_type: "swish"
+        text_encoder_normalize_before: True
+        text_encoder_dropout_rate: 0.1
+        text_encoder_positional_dropout_rate: 0.0
+        text_encoder_attention_dropout_rate: 0.1
+        use_macaron_style_in_text_encoder: True
+        use_conformer_conv_in_text_encoder: False
+        text_encoder_conformer_kernel_size: -1
+        decoder_kernel_size: 7
+        decoder_channels: 512
+        decoder_upsample_scales: [8, 8, 2, 2]
+        decoder_upsample_kernel_sizes: [16, 16, 4, 4]
+        decoder_resblock_kernel_sizes: [3, 7, 11]
+        decoder_resblock_dilations: [[1, 3, 5], [1, 3, 5], [1, 3, 5]]
+        use_weight_norm_in_decoder: True
+        posterior_encoder_kernel_size: 5
+        posterior_encoder_layers: 16
+        posterior_encoder_stacks: 1
+        posterior_encoder_base_dilation: 1
+        posterior_encoder_dropout_rate: 0.0
+        use_weight_norm_in_posterior_encoder: True
+        flow_flows: 4
+        flow_kernel_size: 5
+        flow_base_dilation: 1
+        flow_layers: 4
+        flow_dropout_rate: 0.0
+        use_weight_norm_in_flow: True
+        use_only_mean_in_flow: True
+        stochastic_duration_predictor_kernel_size: 3
+        stochastic_duration_predictor_dropout_rate: 0.5
+        stochastic_duration_predictor_flows: 4
+        stochastic_duration_predictor_dds_conv_layers: 3
+    # discriminator related
+    discriminator_type: hifigan_multi_scale_multi_period_discriminator
+    discriminator_params:
+        scales: 1
+        scale_downsample_pooling: "AvgPool1D"
+        scale_downsample_pooling_params:
+            kernel_size: 4
+            stride: 2
+            padding: 2
+        scale_discriminator_params:
+            in_channels: 1
+            out_channels: 1
+            kernel_sizes: [15, 41, 5, 3]
+            channels: 128
+            max_downsample_channels: 1024
+            max_groups: 16
+            bias: True
+            downsample_scales: [2, 2, 4, 4, 1]
+            nonlinear_activation: "leakyrelu"
+            nonlinear_activation_params:
+                negative_slope: 0.1
+            use_weight_norm: True
+            use_spectral_norm: False
+        follow_official_norm: False
+        periods: [2, 3, 5, 7, 11]
+        period_discriminator_params:
+            in_channels: 1
+            out_channels: 1
+            kernel_sizes: [5, 3]
+            channels: 32
+            downsample_scales: [3, 3, 3, 3, 1]
+            max_downsample_channels: 1024
+            bias: True
+            nonlinear_activation: "leakyrelu"
+            nonlinear_activation_params:
+                negative_slope: 0.1
+            use_weight_norm: True
+            use_spectral_norm: False
+    # others
+    sampling_rate: 22050          # needed in the inference for saving wav
+    cache_generator_outputs: True # whether to cache generator outputs in the training
+          
+###########################################################
+#                        LOSS SETTING                     #
+###########################################################
+# loss function related
+generator_adv_loss_params:
+    average_by_discriminators: False # whether to average loss value by #discriminators
+    loss_type: mse                   # loss type, "mse" or "hinge"
+discriminator_adv_loss_params:
+    average_by_discriminators: False # whether to average loss value by #discriminators
+    loss_type: mse                   # loss type, "mse" or "hinge"
+feat_match_loss_params:
+    average_by_discriminators: False # whether to average loss value by #discriminators
+    average_by_layers: False         # whether to average loss value by #layers of each discriminator
+    include_final_outputs: True      # whether to include final outputs for loss calculation
+mel_loss_params:
+    fs: 22050          # must be the same as the training data
+    fft_size: 1024        # fft points
+    hop_size: 256    # hop size
+    win_length: null   # window length
+    window: hann       # window type
+    num_mels: 80         # number of Mel basis
+    fmin: 0            # minimum frequency for Mel basis
+    fmax: null         # maximum frequency for Mel basis
+    log_base: null     # null represent natural log
+
+###########################################################
+#               ADVERSARIAL LOSS SETTING                  #
+###########################################################
+lambda_adv: 1.0        # loss scaling coefficient for adversarial loss
+lambda_mel: 45.0       # loss scaling coefficient for Mel loss
+lambda_feat_match: 2.0 # loss scaling coefficient for feat match loss
+lambda_dur: 1.0        # loss scaling coefficient for duration loss
+lambda_kl: 1.0         # loss scaling coefficient for KL divergence loss
+# others
+sampling_rate: 22050          # needed in the inference for saving wav
+cache_generator_outputs: True # whether to cache generator outputs in the training
+
+
+###########################################################
+#                  DATA LOADER SETTING                    #
+###########################################################
+batch_size: 50              # Batch size.
+num_workers: 4              # Number of workers in DataLoader.
+
+##########################################################
+#            OPTIMIZER & SCHEDULER SETTING               #
+##########################################################
+# optimizer setting for generator
+generator_optimizer_params:
+    beta1: 0.8
+    beta2: 0.99
+    epsilon: 1.0e-9
+    weight_decay: 0.0
+generator_scheduler: exponential_decay
+generator_scheduler_params:
+    learning_rate: 2.0e-4
+    gamma: 0.999875                   
+
+# optimizer setting for discriminator
+discriminator_optimizer_params:
+    beta1: 0.8
+    beta2: 0.99
+    epsilon: 1.0e-9
+    weight_decay: 0.0
+discriminator_scheduler: exponential_decay
+discriminator_scheduler_params:
+    learning_rate: 2.0e-4          
+    gamma: 0.999875
+generator_first: False # whether to start updating generator first
+
+##########################################################
+#                OTHER TRAINING SETTING                  #
+##########################################################
+num_snapshots: 10            # max number of snapshots to keep while training
+train_max_steps: 350000      # Number of training steps. == total_iters / ngpus, total_iters = 1000000
+save_interval_steps: 1000    # Interval steps to save checkpoint.
+eval_interval_steps: 250     # Interval steps to evaluate the network.
+seed: 777                    # random seed number

examples/aishell3/vits-vc/local/preprocess.sh

+#!/bin/bash
+
+stage=0
+stop_stage=100
+
+config_path=$1
+add_blank=$2
+ge2e_ckpt_path=$3
+
+# gen speaker embedding
+if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
+    python3 ${MAIN_ROOT}/paddlespeech/vector/exps/ge2e/inference.py \
+        --input=~/datasets/data_aishell3/train/wav/ \
+        --output=dump/embed \
+        --checkpoint_path=${ge2e_ckpt_path}
+fi
+
+# copy from tts3/preprocess
+if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
+    # get durations from MFA's result
+    echo "Generate durations.txt from MFA results ..."
+    python3 ${MAIN_ROOT}/utils/gen_duration_from_textgrid.py \
+        --inputdir=./aishell3_alignment_tone \
+        --output durations.txt \
+        --config=${config_path}
+fi
+
+if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
+    # extract features
+    echo "Extract features ..."
+    python3 ${BIN_DIR}/preprocess.py \
+        --dataset=aishell3 \
+        --rootdir=~/datasets/data_aishell3/ \
+        --dumpdir=dump \
+        --dur-file=durations.txt \
+        --config=${config_path} \
+        --num-cpu=20 \
+        --cut-sil=True \
+        --spk_emb_dir=dump/embed
+fi
+
+if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
+    # get features' stats(mean and std)
+    echo "Get features' stats ..."
+    python3 ${MAIN_ROOT}/utils/compute_statistics.py \
+        --metadata=dump/train/raw/metadata.jsonl \
+        --field-name="feats"
+fi
+
+if [ ${stage} -le 4 ] && [ ${stop_stage} -ge 4 ]; then
+    # normalize and covert phone/speaker to id, dev and test should use train's stats
+    echo "Normalize ..."
+    python3 ${BIN_DIR}/normalize.py \
+        --metadata=dump/train/raw/metadata.jsonl \
+        --dumpdir=dump/train/norm \
+        --feats-stats=dump/train/feats_stats.npy \
+        --phones-dict=dump/phone_id_map.txt \
+        --speaker-dict=dump/speaker_id_map.txt \
+        --add-blank=${add_blank} \
+        --skip-wav-copy
+
+    python3 ${BIN_DIR}/normalize.py \
+        --metadata=dump/dev/raw/metadata.jsonl \
+        --dumpdir=dump/dev/norm \
+        --feats-stats=dump/train/feats_stats.npy \
+        --phones-dict=dump/phone_id_map.txt \
+        --speaker-dict=dump/speaker_id_map.txt \
+        --add-blank=${add_blank} \
+        --skip-wav-copy
+
+    python3 ${BIN_DIR}/normalize.py \
+        --metadata=dump/test/raw/metadata.jsonl \
+        --dumpdir=dump/test/norm \
+        --feats-stats=dump/train/feats_stats.npy \
+        --phones-dict=dump/phone_id_map.txt \
+        --speaker-dict=dump/speaker_id_map.txt \
+        --add-blank=${add_blank} \
+        --skip-wav-copy
+fi

examples/aishell3/vits-vc/local/synthesize.sh

+#!/bin/bash
+
+config_path=$1
+train_output_path=$2
+ckpt_name=$3
+stage=0
+stop_stage=0
+
+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 ${BIN_DIR}/synthesize.py \
+        --config=${config_path} \
+        --ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+        --phones_dict=dump/phone_id_map.txt \
+        --test_metadata=dump/test/norm/metadata.jsonl \
+        --output_dir=${train_output_path}/test \
+        --voice-cloning=True
+fi

examples/aishell3/vits-vc/local/train.sh

+#!/bin/bash
+
+config_path=$1
+train_output_path=$2
+
+# install monotonic_align
+cd ${MAIN_ROOT}/paddlespeech/t2s/models/vits/monotonic_align
+python3 setup.py build_ext --inplace
+cd -
+
+python3 ${BIN_DIR}/train.py \
+    --train-metadata=dump/train/norm/metadata.jsonl \
+    --dev-metadata=dump/dev/norm/metadata.jsonl \
+    --config=${config_path} \
+    --output-dir=${train_output_path} \
+    --ngpu=4 \
+    --phones-dict=dump/phone_id_map.txt \
+    --voice-cloning=True

examples/aishell3/vits-vc/local/voice_cloning.sh

+#!/bin/bash
+
+config_path=$1
+train_output_path=$2
+ckpt_name=$3
+ge2e_params_path=$4
+add_blank=$5
+ref_audio_dir=$6
+src_audio_path=$7
+
+FLAGS_allocator_strategy=naive_best_fit \
+FLAGS_fraction_of_gpu_memory_to_use=0.01 \
+python3 ${BIN_DIR}/voice_cloning.py \
+    --config=${config_path} \
+    --ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+    --ge2e_params_path=${ge2e_params_path} \
+    --phones_dict=dump/phone_id_map.txt \
+    --text="凯莫瑞安联合体的经济崩溃迫在眉睫。" \
+    --audio-path=${src_audio_path} \
+    --input-dir=${ref_audio_dir} \
+    --output-dir=${train_output_path}/vc_syn \
+    --add-blank=${add_blank}

examples/aishell3/vits-vc/path.sh

+#!/bin/bash
+export MAIN_ROOT=`realpath ${PWD}/../../../`
+
+export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
+export LC_ALL=C
+
+export PYTHONDONTWRITEBYTECODE=1
+# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
+export PYTHONIOENCODING=UTF-8
+export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
+
+MODEL=vits
+export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/${MODEL}
\ No newline at end of file

examples/aishell3/vits-vc/run.sh

+#!/bin/bash
+
+set -e
+source path.sh
+
+gpus=0,1,2,3
+stage=0
+stop_stage=100
+
+conf_path=conf/default.yaml
+train_output_path=exp/default
+ckpt_name=snapshot_iter_153.pdz
+add_blank=true
+ref_audio_dir=ref_audio
+src_audio_path=''
+
+# not include ".pdparams" here
+ge2e_ckpt_path=./ge2e_ckpt_0.3/step-3000000
+
+# include ".pdparams" here
+ge2e_params_path=${ge2e_ckpt_path}.pdparams
+
+# with the following command, you can choose the stage range you want to run
+# such as `./run.sh --stage 0 --stop-stage 0`
+# this can not be mixed use with `$1`, `$2` ...
+source ${MAIN_ROOT}/utils/parse_options.sh || exit 1
+
+if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
+    # prepare data
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/preprocess.sh ${conf_path} ${add_blank}  ${ge2e_ckpt_path} || exit -1
+fi
+
+if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
+    # train model, all `ckpt` under `train_output_path/checkpoints/` dir
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/train.sh ${conf_path} ${train_output_path} || exit -1
+fi
+
+if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
+    CUDA_VISIBLE_DEVICES=${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/voice_cloning.sh ${conf_path} ${train_output_path} ${ckpt_name} \
+        ${ge2e_params_path} ${add_blank} ${ref_audio_dir} ${src_audio_path} || exit -1
+fi

examples/aishell3/vits/README.md

+# VITS with AISHELL-3
+This example contains code used to train a [VITS](https://arxiv.org/abs/2106.06103) model with [AISHELL-3](http://www.aishelltech.com/aishell_3).
+
+AISHELL-3 is a large-scale and high-fidelity multi-speaker Mandarin speech corpus that could be used to train multi-speaker Text-to-Speech (TTS) systems.
+
+We use AISHELL-3 to train a multi-speaker VITS model here.
+## 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 phonemes for VITS, the durations of MFA are not needed here.
+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 a 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 ${conf_path}
+```
+When it is done. A `dump` folder is created in the current directory. The structure of the dump folder is listed below.
+
+```text
+dump
+├── dev
+│   ├── norm
+│   └── raw
+├── phone_id_map.txt
+├── speaker_id_map.txt
+├── test
+│   ├── norm
+│   └── raw
+└── train
+    ├── feats_stats.npy
+    ├── norm
+    └── raw
+```
+The dataset is split into 3 parts, namely `train`, `dev`, and` test`, each of which contains a `norm` and `raw` subfolder. The raw folder contains wave and linear spectrogram of each utterance, while the norm folder contains normalized ones. The statistics used to normalize features are computed from the training set, which is located in `dump/train/feats_stats.npy`.
+
+Also, there is a `metadata.jsonl` in each subfolder. It is a table-like file that contains phones, text_lengths, feats, feats_lengths, the path of linear spectrogram features, the path of raw waves, speaker, and the id of each utterance.
+
+### Model Training
+```bash
+CUDA_VISIBLE_DEVICES=${gpus} ./local/train.sh ${conf_path} ${train_output_path}
+```
+`./local/train.sh` calls `${BIN_DIR}/train.py`.
+Here's the complete help message.
+```text
+usage: train.py [-h] [--config CONFIG] [--train-metadata TRAIN_METADATA]
+                [--dev-metadata DEV_METADATA] [--output-dir OUTPUT_DIR]
+                [--ngpu NGPU] [--phones-dict PHONES_DICT]
+                [--speaker-dict SPEAKER_DICT] [--voice-cloning VOICE_CLONING]
+
+Train a VITS model.
+
+optional arguments:
+  -h, --help            show this help message and exit
+  --config CONFIG       config file to overwrite default config.
+  --train-metadata TRAIN_METADATA
+                        training data.
+  --dev-metadata DEV_METADATA
+                        dev data.
+  --output-dir OUTPUT_DIR
+                        output dir.
+  --ngpu NGPU           if ngpu == 0, use cpu.
+  --phones-dict PHONES_DICT
+                        phone vocabulary file.
+  --speaker-dict SPEAKER_DICT
+                        speaker id map file for multiple speaker model.
+  --voice-cloning VOICE_CLONING
+                        whether training voice cloning model.
+```
+1. `--config` is a config file in yaml format to overwrite the default config, which can be found at `conf/default.yaml`.
+2. `--train-metadata` and `--dev-metadata` should be the metadata file in the normalized subfolder of `train` and `dev` in the `dump` folder.
+3. `--output-dir` is the directory to save the results of the experiment. Checkpoints are saved in `checkpoints/` inside this directory.
+4. `--ngpu` is the number of gpus to use, if ngpu == 0, use cpu.
+5. `--phones-dict` is the path of the phone vocabulary file.
+6. `--speaker-dict` is the path of the speaker id map file when training a multi-speaker VITS.
+
+### Synthesizing
+
+`./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} ${ckpt_name}
+```
+```text
+usage: synthesize.py [-h] [--config CONFIG] [--ckpt CKPT]
+                     [--phones_dict PHONES_DICT] [--speaker_dict SPEAKER_DICT]
+                     [--voice-cloning VOICE_CLONING] [--ngpu NGPU]
+                     [--test_metadata TEST_METADATA] [--output_dir OUTPUT_DIR]
+
+Synthesize with VITS
+
+optional arguments:
+  -h, --help            show this help message and exit
+  --config CONFIG       Config of VITS.
+  --ckpt CKPT           Checkpoint file of VITS.
+  --phones_dict PHONES_DICT
+                        phone vocabulary file.
+  --speaker_dict SPEAKER_DICT
+                        speaker id map file.
+  --voice-cloning VOICE_CLONING
+                        whether training voice cloning model.
+  --ngpu NGPU           if ngpu == 0, use cpu.
+  --test_metadata TEST_METADATA
+                        test metadata.
+  --output_dir OUTPUT_DIR
+                        output dir.
+```
+`./local/synthesize_e2e.sh` calls `${BIN_DIR}/synthesize_e2e.py`, which can synthesize waveform from text file.
+```bash
+CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh ${conf_path} ${train_output_path} ${ckpt_name}
+```
+```text
+usage: synthesize_e2e.py [-h] [--config CONFIG] [--ckpt CKPT]
+                         [--phones_dict PHONES_DICT]
+                         [--speaker_dict SPEAKER_DICT] [--spk_id SPK_ID]
+                         [--lang LANG]
+                         [--inference_dir INFERENCE_DIR] [--ngpu NGPU]
+                         [--text TEXT] [--output_dir OUTPUT_DIR]
+
+Synthesize with VITS
+
+optional arguments:
+  -h, --help            show this help message and exit
+  --config CONFIG       Config of VITS.
+  --ckpt CKPT           Checkpoint file of VITS.
+  --phones_dict PHONES_DICT
+                        phone vocabulary file.
+  --speaker_dict SPEAKER_DICT
+                        speaker id map file.
+  --spk_id SPK_ID       spk id for multi speaker acoustic model
+  --lang LANG           Choose model language. zh or en
+  --inference_dir INFERENCE_DIR
+                        dir to save inference models
+  --ngpu NGPU           if ngpu == 0, use cpu.
+  --text TEXT           text to synthesize, a 'utt_id sentence' pair per line.
+  --output_dir OUTPUT_DIR
+                        output dir.
+```
+1. `--config`, `--ckpt`, `--phones_dict` and `--speaker_dict` are arguments for acoustic model, which correspond to the 3 files in the VITS pretrained model.
+2. `--lang` is the model language, which can be `zh` or `en`.
+3. `--test_metadata` should be the metadata file in the normalized subfolder of `test`  in the `dump` folder.
+4. `--text` is the text file, which contains sentences to synthesize.
+5. `--output_dir` is the directory to save synthesized audio files.
+6. `--ngpu` is the number of gpus to use, if ngpu == 0, use cpu.
+
+<!-- TODO display these after we trained the model -->
+<!-- 
+## Pretrained Model
+
+The pretrained model can be downloaded here:
+
+- [vits_aishell3_ckpt_1.1.0.zip](https://paddlespeech.bj.bcebos.com/Parakeet/released_models/vits/vits_aishell3_ckpt_1.1.0.zip) (add_blank=true)
+
+VITS checkpoint contains files listed below.
+```text
+vits_aishell3_ckpt_1.1.0
+├── default.yaml              # default config used to train vitx
+├── phone_id_map.txt          # phone vocabulary file when training vits
+├── speaker_id_map.txt        # speaker id map file when training a multi-speaker vits
+└── snapshot_iter_333000.pdz  # model parameters and optimizer states
+```
+
+ps: This ckpt is not good enough, a better result is training
+
+You can use the following scripts to synthesize for `${BIN_DIR}/../sentences.txt` using pretrained VITS.
+
+```bash
+source path.sh
+add_blank=true
+
+FLAGS_allocator_strategy=naive_best_fit \
+FLAGS_fraction_of_gpu_memory_to_use=0.01 \
+python3 ${BIN_DIR}/synthesize_e2e.py \
+    --config=vits_aishell3_ckpt_1.1.0/default.yaml \
+    --ckpt=vits_aishell3_ckpt_1.1.0/snapshot_iter_333000.pdz \
+    --phones_dict=vits_aishell3_ckpt_1.1.0/phone_id_map.txt \
+    --speaker_dict=vits_aishell3_ckpt_1.1.0/speaker_id_map.txt \
+    --output_dir=exp/default/test_e2e \
+    --text=${BIN_DIR}/../sentences.txt \
+    --add-blank=${add_blank} 
+```
+-->

examples/aishell3/vits/conf/default.yaml

+# This configuration tested on 4 GPUs (V100) with 32GB GPU
+# memory. It takes around 2 weeks to finish the training
+# but 100k iters model should generate reasonable results.
+###########################################################
+#                FEATURE EXTRACTION SETTING               #
+###########################################################
+
+fs: 22050         # sr
+n_fft: 1024        # FFT size (samples).
+n_shift: 256       # Hop size (samples). 12.5ms
+win_length: null   # Window length (samples). 50ms
+                   # If set to null, it will be the same as fft_size.
+window: "hann"     # Window function.
+
+
+##########################################################
+#                  TTS MODEL SETTING                     #
+##########################################################
+model:
+    # generator related
+    generator_type: vits_generator
+    generator_params:
+        hidden_channels: 192
+        global_channels: 256
+        segment_size: 32
+        text_encoder_attention_heads: 2
+        text_encoder_ffn_expand: 4
+        text_encoder_blocks: 6
+        text_encoder_positionwise_layer_type: "conv1d"
+        text_encoder_positionwise_conv_kernel_size: 3
+        text_encoder_positional_encoding_layer_type: "rel_pos"
+        text_encoder_self_attention_layer_type: "rel_selfattn"
+        text_encoder_activation_type: "swish"
+        text_encoder_normalize_before: True
+        text_encoder_dropout_rate: 0.1
+        text_encoder_positional_dropout_rate: 0.0
+        text_encoder_attention_dropout_rate: 0.1
+        use_macaron_style_in_text_encoder: True
+        use_conformer_conv_in_text_encoder: False
+        text_encoder_conformer_kernel_size: -1
+        decoder_kernel_size: 7
+        decoder_channels: 512
+        decoder_upsample_scales: [8, 8, 2, 2]
+        decoder_upsample_kernel_sizes: [16, 16, 4, 4]
+        decoder_resblock_kernel_sizes: [3, 7, 11]
+        decoder_resblock_dilations: [[1, 3, 5], [1, 3, 5], [1, 3, 5]]
+        use_weight_norm_in_decoder: True
+        posterior_encoder_kernel_size: 5
+        posterior_encoder_layers: 16
+        posterior_encoder_stacks: 1
+        posterior_encoder_base_dilation: 1
+        posterior_encoder_dropout_rate: 0.0
+        use_weight_norm_in_posterior_encoder: True
+        flow_flows: 4
+        flow_kernel_size: 5
+        flow_base_dilation: 1
+        flow_layers: 4
+        flow_dropout_rate: 0.0
+        use_weight_norm_in_flow: True
+        use_only_mean_in_flow: True
+        stochastic_duration_predictor_kernel_size: 3
+        stochastic_duration_predictor_dropout_rate: 0.5
+        stochastic_duration_predictor_flows: 4
+        stochastic_duration_predictor_dds_conv_layers: 3
+    # discriminator related
+    discriminator_type: hifigan_multi_scale_multi_period_discriminator
+    discriminator_params:
+        scales: 1
+        scale_downsample_pooling: "AvgPool1D"
+        scale_downsample_pooling_params:
+            kernel_size: 4
+            stride: 2
+            padding: 2
+        scale_discriminator_params:
+            in_channels: 1
+            out_channels: 1
+            kernel_sizes: [15, 41, 5, 3]
+            channels: 128
+            max_downsample_channels: 1024
+            max_groups: 16
+            bias: True
+            downsample_scales: [2, 2, 4, 4, 1]
+            nonlinear_activation: "leakyrelu"
+            nonlinear_activation_params:
+                negative_slope: 0.1
+            use_weight_norm: True
+            use_spectral_norm: False
+        follow_official_norm: False
+        periods: [2, 3, 5, 7, 11]
+        period_discriminator_params:
+            in_channels: 1
+            out_channels: 1
+            kernel_sizes: [5, 3]
+            channels: 32
+            downsample_scales: [3, 3, 3, 3, 1]
+            max_downsample_channels: 1024
+            bias: True
+            nonlinear_activation: "leakyrelu"
+            nonlinear_activation_params:
+                negative_slope: 0.1
+            use_weight_norm: True
+            use_spectral_norm: False
+    # others
+    sampling_rate: 22050          # needed in the inference for saving wav
+    cache_generator_outputs: True # whether to cache generator outputs in the training
+          
+###########################################################
+#                        LOSS SETTING                     #
+###########################################################
+# loss function related
+generator_adv_loss_params:
+    average_by_discriminators: False # whether to average loss value by #discriminators
+    loss_type: mse                   # loss type, "mse" or "hinge"
+discriminator_adv_loss_params:
+    average_by_discriminators: False # whether to average loss value by #discriminators
+    loss_type: mse                   # loss type, "mse" or "hinge"
+feat_match_loss_params:
+    average_by_discriminators: False # whether to average loss value by #discriminators
+    average_by_layers: False         # whether to average loss value by #layers of each discriminator
+    include_final_outputs: True      # whether to include final outputs for loss calculation
+mel_loss_params:
+    fs: 22050          # must be the same as the training data
+    fft_size: 1024        # fft points
+    hop_size: 256    # hop size
+    win_length: null   # window length
+    window: hann       # window type
+    num_mels: 80         # number of Mel basis
+    fmin: 0            # minimum frequency for Mel basis
+    fmax: null         # maximum frequency for Mel basis
+    log_base: null     # null represent natural log
+
+###########################################################
+#               ADVERSARIAL LOSS SETTING                  #
+###########################################################
+lambda_adv: 1.0        # loss scaling coefficient for adversarial loss
+lambda_mel: 45.0       # loss scaling coefficient for Mel loss
+lambda_feat_match: 2.0 # loss scaling coefficient for feat match loss
+lambda_dur: 1.0        # loss scaling coefficient for duration loss
+lambda_kl: 1.0         # loss scaling coefficient for KL divergence loss
+# others
+sampling_rate: 22050          # needed in the inference for saving wav
+cache_generator_outputs: True # whether to cache generator outputs in the training
+
+
+###########################################################
+#                  DATA LOADER SETTING                    #
+###########################################################
+batch_size: 50              # Batch size.
+num_workers: 4              # Number of workers in DataLoader.
+
+##########################################################
+#            OPTIMIZER & SCHEDULER SETTING               #
+##########################################################
+# optimizer setting for generator
+generator_optimizer_params:
+    beta1: 0.8
+    beta2: 0.99
+    epsilon: 1.0e-9
+    weight_decay: 0.0
+generator_scheduler: exponential_decay
+generator_scheduler_params:
+    learning_rate: 2.0e-4
+    gamma: 0.999875                   
+
+# optimizer setting for discriminator
+discriminator_optimizer_params:
+    beta1: 0.8
+    beta2: 0.99
+    epsilon: 1.0e-9
+    weight_decay: 0.0
+discriminator_scheduler: exponential_decay
+discriminator_scheduler_params:
+    learning_rate: 2.0e-4          
+    gamma: 0.999875
+generator_first: False # whether to start updating generator first
+
+##########################################################
+#                OTHER TRAINING SETTING                  #
+##########################################################
+num_snapshots: 10            # max number of snapshots to keep while training
+train_max_steps: 350000      # Number of training steps. == total_iters / ngpus, total_iters = 1000000
+save_interval_steps: 1000    # Interval steps to save checkpoint.
+eval_interval_steps: 250     # Interval steps to evaluate the network.
+seed: 777                    # random seed number

examples/aishell3/vits/local/preprocess.sh

+#!/bin/bash
+
+stage=0
+stop_stage=100
+
+config_path=$1
+add_blank=$2
+
+# copy from tts3/preprocess
+if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
+    # get durations from MFA's result
+    echo "Generate durations.txt from MFA results ..."
+    python3 ${MAIN_ROOT}/utils/gen_duration_from_textgrid.py \
+        --inputdir=./aishell3_alignment_tone \
+        --output durations.txt \
+        --config=${config_path}
+fi
+
+if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
+    # extract features
+    echo "Extract features ..."
+    python3 ${BIN_DIR}/preprocess.py \
+        --dataset=aishell3 \
+        --rootdir=~/datasets/data_aishell3/ \
+        --dumpdir=dump \
+        --dur-file=durations.txt \
+        --config=${config_path} \
+        --num-cpu=20 \
+        --cut-sil=True
+fi
+
+if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
+    # get features' stats(mean and std)
+    echo "Get features' stats ..."
+    python3 ${MAIN_ROOT}/utils/compute_statistics.py \
+        --metadata=dump/train/raw/metadata.jsonl \
+        --field-name="feats"
+fi
+
+if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
+    # normalize and covert phone/speaker to id, dev and test should use train's stats
+    echo "Normalize ..."
+    python3 ${BIN_DIR}/normalize.py \
+        --metadata=dump/train/raw/metadata.jsonl \
+        --dumpdir=dump/train/norm \
+        --feats-stats=dump/train/feats_stats.npy \
+        --phones-dict=dump/phone_id_map.txt \
+        --speaker-dict=dump/speaker_id_map.txt \
+        --add-blank=${add_blank} \
+        --skip-wav-copy
+
+    python3 ${BIN_DIR}/normalize.py \
+        --metadata=dump/dev/raw/metadata.jsonl \
+        --dumpdir=dump/dev/norm \
+        --feats-stats=dump/train/feats_stats.npy \
+        --phones-dict=dump/phone_id_map.txt \
+        --speaker-dict=dump/speaker_id_map.txt \
+        --add-blank=${add_blank} \
+        --skip-wav-copy
+
+    python3 ${BIN_DIR}/normalize.py \
+        --metadata=dump/test/raw/metadata.jsonl \
+        --dumpdir=dump/test/norm \
+        --feats-stats=dump/train/feats_stats.npy \
+        --phones-dict=dump/phone_id_map.txt \
+        --speaker-dict=dump/speaker_id_map.txt \
+        --add-blank=${add_blank} \
+        --skip-wav-copy
+fi

examples/aishell3/vits/local/synthesize.sh

+#!/bin/bash
+
+config_path=$1
+train_output_path=$2
+ckpt_name=$3
+stage=0
+stop_stage=0
+
+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 ${BIN_DIR}/synthesize.py \
+        --config=${config_path} \
+        --ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+        --phones_dict=dump/phone_id_map.txt \
+        --speaker_dict=dump/speaker_id_map.txt \
+        --test_metadata=dump/test/norm/metadata.jsonl \
+        --output_dir=${train_output_path}/test
+fi

examples/aishell3/vits/local/synthesize_e2e.sh

+#!/bin/bash
+
+config_path=$1
+train_output_path=$2
+ckpt_name=$3
+add_blank=$4
+
+stage=0
+stop_stage=0
+
+
+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 ${BIN_DIR}/synthesize_e2e.py \
+        --config=${config_path} \
+        --ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+        --phones_dict=dump/phone_id_map.txt \
+        --speaker_dict=dump/speaker_id_map.txt \
+        --spk_id=0 \
+        --output_dir=${train_output_path}/test_e2e \
+        --text=${BIN_DIR}/../sentences.txt \
+        --add-blank=${add_blank}
+fi

examples/aishell3/vits/local/train.sh

+#!/bin/bash
+
+config_path=$1
+train_output_path=$2
+
+# install monotonic_align
+cd ${MAIN_ROOT}/paddlespeech/t2s/models/vits/monotonic_align
+python3 setup.py build_ext --inplace
+cd -
+
+python3 ${BIN_DIR}/train.py \
+    --train-metadata=dump/train/norm/metadata.jsonl \
+    --dev-metadata=dump/dev/norm/metadata.jsonl \
+    --config=${config_path} \
+    --output-dir=${train_output_path} \
+    --ngpu=4 \
+    --phones-dict=dump/phone_id_map.txt \
+    --speaker-dict=dump/speaker_id_map.txt

examples/aishell3/vits/path.sh

+#!/bin/bash
+export MAIN_ROOT=`realpath ${PWD}/../../../`
+
+export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
+export LC_ALL=C
+
+export PYTHONDONTWRITEBYTECODE=1
+# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
+export PYTHONIOENCODING=UTF-8
+export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
+
+MODEL=vits
+export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/${MODEL}
\ No newline at end of file

examples/aishell3/vits/run.sh

+#!/bin/bash
+
+set -e
+source path.sh
+
+gpus=0,1,2,3
+stage=0
+stop_stage=100
+
+conf_path=conf/default.yaml
+train_output_path=exp/default
+ckpt_name=snapshot_iter_153.pdz
+add_blank=true
+
+# with the following command, you can choose the stage range you want to run
+# such as `./run.sh --stage 0 --stop-stage 0`
+# this can not be mixed use with `$1`, `$2` ...
+source ${MAIN_ROOT}/utils/parse_options.sh || exit 1
+
+if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
+    # prepare data
+    ./local/preprocess.sh ${conf_path} ${add_blank}|| exit -1
+fi
+
+if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
+    # train model, all `ckpt` under `train_output_path/checkpoints/` dir
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/train.sh ${conf_path} ${train_output_path} || exit -1
+fi
+
+if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
+    CUDA_VISIBLE_DEVICES=${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} ${add_blank}|| exit -1
+fi

examples/csmsc/vits/run.sh

@@ -3,7 +3,7 @@
 set -e
 source path.sh
 
-gpus=0,1
+gpus=0,1,2,3
 stage=0
 stop_stage=100
 

paddlespeech/t2s/datasets/am_batch_fn.py

@@ -483,3 +483,58 @@ def vits_single_spk_batch_fn(examples):
         "speech": speech
     }
     return batch
+
+
+def vits_multi_spk_batch_fn(examples):
+    """
+    Returns:
+        Dict[str, Any]:
+            - text (Tensor): Text index tensor (B, T_text).
+            - text_lengths (Tensor): Text length tensor (B,).
+            - feats (Tensor): Feature tensor (B, T_feats, aux_channels).
+            - feats_lengths (Tensor): Feature length tensor (B,).
+            - speech (Tensor): Speech waveform tensor (B, T_wav).
+            - spk_id (Optional[Tensor]): Speaker index tensor (B,) or (B, 1).
+            - spk_emb (Optional[Tensor]): Speaker embedding tensor (B, spk_embed_dim).
+    """
+    # fields = ["text", "text_lengths", "feats", "feats_lengths", "speech", "spk_id"/"spk_emb"]
+    text = [np.array(item["text"], dtype=np.int64) for item in examples]
+    feats = [np.array(item["feats"], dtype=np.float32) for item in examples]
+    speech = [np.array(item["wave"], dtype=np.float32) for item in examples]
+    text_lengths = [
+        np.array(item["text_lengths"], dtype=np.int64) for item in examples
+    ]
+    feats_lengths = [
+        np.array(item["feats_lengths"], dtype=np.int64) for item in examples
+    ]
+
+    text = batch_sequences(text)
+    feats = batch_sequences(feats)
+    speech = batch_sequences(speech)
+
+    # convert each batch to paddle.Tensor
+    text = paddle.to_tensor(text)
+    feats = paddle.to_tensor(feats)
+    text_lengths = paddle.to_tensor(text_lengths)
+    feats_lengths = paddle.to_tensor(feats_lengths)
+
+    batch = {
+        "text": text,
+        "text_lengths": text_lengths,
+        "feats": feats,
+        "feats_lengths": feats_lengths,
+        "speech": speech
+    }
+    # spk_emb has a higher priority than spk_id
+    if "spk_emb" in examples[0]:
+        spk_emb = [
+            np.array(item["spk_emb"], dtype=np.float32) for item in examples
+        ]
+        spk_emb = batch_sequences(spk_emb)
+        spk_emb = paddle.to_tensor(spk_emb)
+        batch["spk_emb"] = spk_emb
+    elif "spk_id" in examples[0]:
+        spk_id = [np.array(item["spk_id"], dtype=np.int64) for item in examples]
+        spk_id = paddle.to_tensor(spk_id)
+        batch["spk_id"] = spk_id
+    return batch

paddlespeech/t2s/exps/vits/synthesize.py

@@ -15,6 +15,7 @@ import argparse
 from pathlib import Path
 
 import jsonlines
+import numpy as np
 import paddle
 import soundfile as sf
 import yaml
@@ -23,6 +24,7 @@ from yacs.config import CfgNode
 
 from paddlespeech.t2s.datasets.data_table import DataTable
 from paddlespeech.t2s.models.vits import VITS
+from paddlespeech.t2s.utils import str2bool
 
 
 def evaluate(args):
@@ -40,8 +42,26 @@ def evaluate(args):
     print(config)
 
     fields = ["utt_id", "text"]
+    converters = {}
+
+    spk_num = None
+    if args.speaker_dict is not None:
+        print("multiple speaker vits!")
+        with open(args.speaker_dict, 'rt') as f:
+            spk_id = [line.strip().split() for line in f.readlines()]
+        spk_num = len(spk_id)
+        fields += ["spk_id"]
+    elif args.voice_cloning:
+        print("Evaluating voice cloning!")
+        fields += ["spk_emb"]
+    else:
+        print("single speaker vits!")
+    print("spk_num:", spk_num)
 
-    test_dataset = DataTable(data=test_metadata, fields=fields)
+    test_dataset = DataTable(
+        data=test_metadata,
+        fields=fields,
+        converters=converters, )
 
     with open(args.phones_dict, "r") as f:
         phn_id = [line.strip().split() for line in f.readlines()]
@@ -49,6 +69,7 @@ def evaluate(args):
     print("vocab_size:", vocab_size)
 
     odim = config.n_fft // 2 + 1
+    config["model"]["generator_params"]["spks"] = spk_num
 
     vits = VITS(idim=vocab_size, odim=odim, **config["model"])
     vits.set_state_dict(paddle.load(args.ckpt)["main_params"])
@@ -65,7 +86,15 @@ def evaluate(args):
         phone_ids = paddle.to_tensor(datum["text"])
         with timer() as t:
             with paddle.no_grad():
-                out = vits.inference(text=phone_ids)
+                spk_emb = None
+                spk_id = None
+                # multi speaker
+                if args.voice_cloning and "spk_emb" in datum:
+                    spk_emb = paddle.to_tensor(np.load(datum["spk_emb"]))
+                elif "spk_id" in datum:
+                    spk_id = paddle.to_tensor(datum["spk_id"])
+                out = vits.inference(
+                    text=phone_ids, sids=spk_id, spembs=spk_emb)
             wav = out["wav"]
             wav = wav.numpy()
             N += wav.size
@@ -90,6 +119,13 @@ def parse_args():
         '--ckpt', type=str, default=None, help='Checkpoint file of VITS.')
     parser.add_argument(
         "--phones_dict", type=str, default=None, help="phone vocabulary file.")
+    parser.add_argument(
+        "--speaker_dict", type=str, default=None, help="speaker id map file.")
+    parser.add_argument(
+        "--voice-cloning",
+        type=str2bool,
+        default=False,
+        help="whether training voice cloning model.")
     # other
     parser.add_argument(
         "--ngpu", type=int, default=1, help="if ngpu == 0, use cpu.")

paddlespeech/t2s/exps/vits/synthesize_e2e.py

@@ -42,12 +42,23 @@ def evaluate(args):
     # frontend
     frontend = get_frontend(lang=args.lang, phones_dict=args.phones_dict)
 
+    spk_num = None
+    if args.speaker_dict is not None:
+        print("multiple speaker vits!")
+        with open(args.speaker_dict, 'rt') as f:
+            spk_id = [line.strip().split() for line in f.readlines()]
+        spk_num = len(spk_id)
+    else:
+        print("single speaker vits!")
+    print("spk_num:", spk_num)
+
     with open(args.phones_dict, "r") as f:
         phn_id = [line.strip().split() for line in f.readlines()]
     vocab_size = len(phn_id)
     print("vocab_size:", vocab_size)
 
     odim = config.n_fft // 2 + 1
+    config["model"]["generator_params"]["spks"] = spk_num
 
     vits = VITS(idim=vocab_size, odim=odim, **config["model"])
     vits.set_state_dict(paddle.load(args.ckpt)["main_params"])
@@ -78,7 +89,10 @@ def evaluate(args):
                 flags = 0
                 for i in range(len(phone_ids)):
                     part_phone_ids = phone_ids[i]
-                    out = vits.inference(text=part_phone_ids)
+                    spk_id = None
+                    if spk_num is not None:
+                        spk_id = paddle.to_tensor(args.spk_id)
+                    out = vits.inference(text=part_phone_ids, sids=spk_id)
                     wav = out["wav"]
                     if flags == 0:
                         wav_all = wav
@@ -109,6 +123,13 @@ def parse_args():
         '--ckpt', type=str, default=None, help='Checkpoint file of VITS.')
     parser.add_argument(
         "--phones_dict", type=str, default=None, help="phone vocabulary file.")
+    parser.add_argument(
+        "--speaker_dict", type=str, default=None, help="speaker id map file.")
+    parser.add_argument(
+        '--spk_id',
+        type=int,
+        default=0,
+        help='spk id for multi speaker acoustic model')
     # other
     parser.add_argument(
         '--lang',

paddlespeech/t2s/exps/vits/train.py

@@ -28,6 +28,7 @@ from paddle.io import DistributedBatchSampler
 from paddle.optimizer import Adam
 from yacs.config import CfgNode
 
+from paddlespeech.t2s.datasets.am_batch_fn import vits_multi_spk_batch_fn
 from paddlespeech.t2s.datasets.am_batch_fn import vits_single_spk_batch_fn
 from paddlespeech.t2s.datasets.data_table import DataTable
 from paddlespeech.t2s.models.vits import VITS
@@ -43,6 +44,7 @@ from paddlespeech.t2s.training.extensions.visualizer import VisualDL
 from paddlespeech.t2s.training.optimizer import scheduler_classes
 from paddlespeech.t2s.training.seeding import seed_everything
 from paddlespeech.t2s.training.trainer import Trainer
+from paddlespeech.t2s.utils import str2bool
 
 
 def train_sp(args, config):
@@ -72,6 +74,23 @@ def train_sp(args, config):
         "wave": np.load,
         "feats": np.load,
     }
+    spk_num = None
+    if args.speaker_dict is not None:
+        print("multiple speaker vits!")
+        collate_fn = vits_multi_spk_batch_fn
+        with open(args.speaker_dict, 'rt') as f:
+            spk_id = [line.strip().split() for line in f.readlines()]
+        spk_num = len(spk_id)
+        fields += ["spk_id"]
+    elif args.voice_cloning:
+        print("Training voice cloning!")
+        collate_fn = vits_multi_spk_batch_fn
+        fields += ["spk_emb"]
+        converters["spk_emb"] = np.load
+    else:
+        print("single speaker vits!")
+        collate_fn = vits_single_spk_batch_fn
+    print("spk_num:", spk_num)
 
     # construct dataset for training and validation
     with jsonlines.open(args.train_metadata, 'r') as reader:
@@ -100,18 +119,16 @@ def train_sp(args, config):
         drop_last=False)
     print("samplers done!")
 
-    train_batch_fn = vits_single_spk_batch_fn
-
     train_dataloader = DataLoader(
         train_dataset,
         batch_sampler=train_sampler,
-        collate_fn=train_batch_fn,
+        collate_fn=collate_fn,
         num_workers=config.num_workers)
 
     dev_dataloader = DataLoader(
         dev_dataset,
         batch_sampler=dev_sampler,
-        collate_fn=train_batch_fn,
+        collate_fn=collate_fn,
         num_workers=config.num_workers)
     print("dataloaders done!")
 
@@ -121,6 +138,7 @@ def train_sp(args, config):
     print("vocab_size:", vocab_size)
 
     odim = config.n_fft // 2 + 1
+    config["model"]["generator_params"]["spks"] = spk_num
     model = VITS(idim=vocab_size, odim=odim, **config["model"])
     gen_parameters = model.generator.parameters()
     dis_parameters = model.discriminator.parameters()
@@ -240,6 +258,17 @@ def main():
         "--ngpu", type=int, default=1, help="if ngpu == 0, use cpu.")
     parser.add_argument(
         "--phones-dict", type=str, default=None, help="phone vocabulary file.")
+    parser.add_argument(
+        "--speaker-dict",
+        type=str,
+        default=None,
+        help="speaker id map file for multiple speaker model.")
+
+    parser.add_argument(
+        "--voice-cloning",
+        type=str2bool,
+        default=False,
+        help="whether training voice cloning model.")
 
     args = parser.parse_args()
 

paddlespeech/t2s/exps/vits/voice_cloning.py

+# Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+import argparse
+import os
+from pathlib import Path
+
+import librosa
+import numpy as np
+import paddle
+import soundfile as sf
+import yaml
+from yacs.config import CfgNode
+
+from paddlespeech.t2s.datasets.get_feats import LinearSpectrogram
+from paddlespeech.t2s.exps.syn_utils import get_frontend
+from paddlespeech.t2s.models.vits import VITS
+from paddlespeech.t2s.utils import str2bool
+from paddlespeech.vector.exps.ge2e.audio_processor import SpeakerVerificationPreprocessor
+from paddlespeech.vector.models.lstm_speaker_encoder import LSTMSpeakerEncoder
+
+
+def voice_cloning(args):
+
+    # Init body.
+    with open(args.config) as f:
+        config = CfgNode(yaml.safe_load(f))
+
+    print("========Args========")
+    print(yaml.safe_dump(vars(args)))
+    print("========Config========")
+    print(config)
+
+    # speaker encoder
+    spec_extractor = LinearSpectrogram(
+        n_fft=config.n_fft,
+        hop_length=config.n_shift,
+        win_length=config.win_length,
+        window=config.window)
+    p = SpeakerVerificationPreprocessor(
+        sampling_rate=16000,
+        audio_norm_target_dBFS=-30,
+        vad_window_length=30,
+        vad_moving_average_width=8,
+        vad_max_silence_length=6,
+        mel_window_length=25,
+        mel_window_step=10,
+        n_mels=40,
+        partial_n_frames=160,
+        min_pad_coverage=0.75,
+        partial_overlap_ratio=0.5)
+    print("Audio Processor Done!")
+
+    speaker_encoder = LSTMSpeakerEncoder(
+        n_mels=40, num_layers=3, hidden_size=256, output_size=256)
+    speaker_encoder.set_state_dict(paddle.load(args.ge2e_params_path))
+    speaker_encoder.eval()
+    print("GE2E Done!")
+
+    frontend = get_frontend(lang=args.lang, phones_dict=args.phones_dict)
+    print("frontend done!")
+
+    with open(args.phones_dict, "r") as f:
+        phn_id = [line.strip().split() for line in f.readlines()]
+    vocab_size = len(phn_id)
+    print("vocab_size:", vocab_size)
+
+    odim = config.n_fft // 2 + 1
+
+    vits = VITS(idim=vocab_size, odim=odim, **config["model"])
+    vits.set_state_dict(paddle.load(args.ckpt)["main_params"])
+    vits.eval()
+
+    output_dir = Path(args.output_dir)
+    output_dir.mkdir(parents=True, exist_ok=True)
+
+    input_dir = Path(args.input_dir)
+
+    if args.audio_path == "":
+        args.audio_path = None
+    if args.audio_path is None:
+        sentence = args.text
+        merge_sentences = True
+        add_blank = args.add_blank
+
+        if args.lang == 'zh':
+            input_ids = frontend.get_input_ids(
+                sentence, merge_sentences=merge_sentences, add_blank=add_blank)
+        elif args.lang == 'en':
+            input_ids = frontend.get_input_ids(
+                sentence, merge_sentences=merge_sentences)
+        phone_ids = input_ids["phone_ids"][0]
+    else:
+        wav, _ = librosa.load(str(args.audio_path), sr=config.fs)
+        feats = paddle.to_tensor(spec_extractor.get_linear_spectrogram(wav))
+
+        mel_sequences = p.extract_mel_partials(
+            p.preprocess_wav(args.audio_path))
+        with paddle.no_grad():
+            spk_emb_src = speaker_encoder.embed_utterance(
+                paddle.to_tensor(mel_sequences))
+
+    for name in os.listdir(input_dir):
+        utt_id = name.split(".")[0]
+        ref_audio_path = input_dir / name
+        mel_sequences = p.extract_mel_partials(p.preprocess_wav(ref_audio_path))
+        # print("mel_sequences: ", mel_sequences.shape)
+        with paddle.no_grad():
+            spk_emb = speaker_encoder.embed_utterance(
+                paddle.to_tensor(mel_sequences))
+        # print("spk_emb shape: ", spk_emb.shape)
+
+        with paddle.no_grad():
+            if args.audio_path is None:
+                out = vits.inference(text=phone_ids, spembs=spk_emb)
+            else:
+                out = vits.voice_conversion(
+                    feats=feats, spembs_src=spk_emb_src, spembs_tgt=spk_emb)
+            wav = out["wav"]
+
+        sf.write(
+            str(output_dir / (utt_id + ".wav")),
+            wav.numpy(),
+            samplerate=config.fs)
+        print(f"{utt_id} done!")
+    # Randomly generate numbers of 0 ~ 0.2, 256 is the dim of spk_emb
+    random_spk_emb = np.random.rand(256) * 0.2
+    random_spk_emb = paddle.to_tensor(random_spk_emb, dtype='float32')
+    utt_id = "random_spk_emb"
+    with paddle.no_grad():
+        if args.audio_path is None:
+            out = vits.inference(text=phone_ids, spembs=random_spk_emb)
+        else:
+            out = vits.voice_conversion(
+                feats=feats, spembs_src=spk_emb_src, spembs_tgt=random_spk_emb)
+        wav = out["wav"]
+    sf.write(
+        str(output_dir / (utt_id + ".wav")), wav.numpy(), samplerate=config.fs)
+    print(f"{utt_id} done!")
+
+
+def parse_args():
+    # parse args and config
+    parser = argparse.ArgumentParser(description="")
+    parser.add_argument(
+        '--config', type=str, default=None, help='Config of VITS.')
+    parser.add_argument(
+        '--ckpt', type=str, default=None, help='Checkpoint file of VITS.')
+    parser.add_argument(
+        "--phones_dict", type=str, default=None, help="phone vocabulary file.")
+    parser.add_argument(
+        "--text",
+        type=str,
+        default="每当你觉得，想要批评什么人的时候，你切要记着，这个世界上的人，并非都具备你禀有的条件。",
+        help="text to synthesize, a line")
+    parser.add_argument(
+        '--lang',
+        type=str,
+        default='zh',
+        help='Choose model language. zh or en')
+    parser.add_argument(
+        "--audio-path",
+        type=str,
+        default=None,
+        help="audio as content to synthesize")
+
+    parser.add_argument(
+        "--ge2e_params_path", type=str, help="ge2e params path.")
+
+    parser.add_argument(
+        "--ngpu", type=int, default=1, help="if ngpu=0, use cpu.")
+
+    parser.add_argument(
+        "--input-dir",
+        type=str,
+        help="input dir of *.wav, the sample rate will be resample to 16k.")
+    parser.add_argument("--output-dir", type=str, help="output dir.")
+
+    parser.add_argument(
+        "--add-blank",
+        type=str2bool,
+        default=True,
+        help="whether to add blank between phones")
+
+    args = parser.parse_args()
+    return args
+
+
+def main():
+    args = parse_args()
+
+    if args.ngpu == 0:
+        paddle.set_device("cpu")
+    elif args.ngpu > 0:
+        paddle.set_device("gpu")
+    else:
+        print("ngpu should >= 0 !")
+
+    voice_cloning(args)
+
+
+if __name__ == "__main__":
+    main()

paddlespeech/t2s/models/vits/generator.py

@@ -522,6 +522,82 @@ class VITSGenerator(nn.Layer):
 
         return wav.squeeze(1), attn.squeeze(1), dur.squeeze(1)
 
+    def voice_conversion(
+            self,
+            feats: paddle.Tensor=None,
+            feats_lengths: paddle.Tensor=None,
+            sids_src: Optional[paddle.Tensor]=None,
+            sids_tgt: Optional[paddle.Tensor]=None,
+            spembs_src: Optional[paddle.Tensor]=None,
+            spembs_tgt: Optional[paddle.Tensor]=None,
+            lids: Optional[paddle.Tensor]=None, ) -> paddle.Tensor:
+        """Run voice conversion.
+        Args:
+            feats (Tensor): Feature tensor (B, aux_channels, T_feats,).
+            feats_lengths (Tensor): Feature length tensor (B,).
+            sids_src (Optional[Tensor]): Speaker index tensor of source feature (B,) or (B, 1).
+            sids_tgt (Optional[Tensor]): Speaker index tensor of target feature (B,) or (B, 1).
+            spembs_src (Optional[Tensor]): Speaker embedding tensor of source feature (B, spk_embed_dim).
+            spembs_tgt (Optional[Tensor]): Speaker embedding tensor of target feature (B, spk_embed_dim).
+            lids (Optional[Tensor]): Language index tensor (B,) or (B, 1).
+        Returns:
+            Tensor: Generated waveform tensor (B, T_wav).
+        """
+        # encoder
+        g_src = None
+        g_tgt = None
+        if self.spks is not None:
+            # (B, global_channels, 1)
+            g_src = self.global_emb(
+                paddle.reshape(sids_src, [-1])).unsqueeze(-1)
+            g_tgt = self.global_emb(
+                paddle.reshape(sids_tgt, [-1])).unsqueeze(-1)
+
+        if self.spk_embed_dim is not None:
+            # (B, global_channels, 1)
+            g_src_ = self.spemb_proj(
+                F.normalize(spembs_src.unsqueeze(0))).unsqueeze(-1)
+            if g_src is None:
+                g_src = g_src_
+            else:
+                g_src = g_src + g_src_
+
+            # (B, global_channels, 1)
+            g_tgt_ = self.spemb_proj(
+                F.normalize(spembs_tgt.unsqueeze(0))).unsqueeze(-1)
+            if g_tgt is None:
+                g_tgt = g_tgt_
+            else:
+                g_tgt = g_tgt + g_tgt_
+
+        if self.langs is not None:
+            # (B, global_channels, 1)
+            g_ = self.lang_emb(paddle.reshape(lids, [-1])).unsqueeze(-1)
+
+            if g_src is None:
+                g_src = g_
+            else:
+                g_src = g_src + g_
+
+            if g_tgt is None:
+                g_tgt = g_
+            else:
+                g_tgt = g_tgt + g_
+
+        # forward posterior encoder
+        z, m_q, logs_q, y_mask = self.posterior_encoder(
+            feats, feats_lengths, g=g_src)
+
+        # forward flow
+        # (B, H, T_feats)
+        z_p = self.flow(z, y_mask, g=g_src)
+
+        # decoder
+        z_hat = self.flow(z_p, y_mask, g=g_tgt, inverse=True)
+        wav = self.decoder(z_hat * y_mask, g=g_tgt)
+
+        return wav.squeeze(1)
+
     def _generate_path(self, dur: paddle.Tensor,
                        mask: paddle.Tensor) -> paddle.Tensor:
         """Generate path a.k.a. monotonic attention.

paddlespeech/t2s/models/vits/vits.py

@@ -381,7 +381,7 @@ class VITS(nn.Layer):
         if use_teacher_forcing:
             assert feats is not None
             feats = feats[None].transpose([0, 2, 1])
-            feats_lengths = paddle.to_tensor([paddle.shape(feats)[2]])
+            feats_lengths = paddle.to_tensor(paddle.shape(feats)[2])
             wav, att_w, dur = self.generator.inference(
                 text=text,
                 text_lengths=text_lengths,
@@ -406,3 +406,43 @@ class VITS(nn.Layer):
                 max_len=max_len, )
         return dict(
             wav=paddle.reshape(wav, [-1]), att_w=att_w[0], duration=dur[0])
+
+    def voice_conversion(
+            self,
+            feats: paddle.Tensor,
+            sids_src: Optional[paddle.Tensor]=None,
+            sids_tgt: Optional[paddle.Tensor]=None,
+            spembs_src: Optional[paddle.Tensor]=None,
+            spembs_tgt: Optional[paddle.Tensor]=None,
+            lids: Optional[paddle.Tensor]=None, ) -> paddle.Tensor:
+        """Run voice conversion.
+        Args:
+            feats (Tensor): Feature tensor (T_feats, aux_channels).
+            sids_src (Optional[Tensor]): Speaker index tensor of source feature (1,).
+            sids_tgt (Optional[Tensor]): Speaker index tensor of target feature (1,).
+            spembs_src (Optional[Tensor]): Speaker embedding tensor of source feature (spk_embed_dim,).
+            spembs_tgt (Optional[Tensor]): Speaker embedding tensor of target feature (spk_embed_dim,).
+            lids (Optional[Tensor]): Language index tensor (1,).
+        Returns:
+            Dict[str, Tensor]:
+                * wav (Tensor): Generated waveform tensor (T_wav,).
+        """
+        assert feats is not None
+        feats = feats[None].transpose([0, 2, 1])
+        feats_lengths = paddle.to_tensor(paddle.shape(feats)[2])
+
+        sids_none = sids_src is None and sids_tgt is None
+        spembs_none = spembs_src is None and spembs_tgt is None
+
+        assert not sids_none or not spembs_none
+
+        wav = self.generator.voice_conversion(
+            feats,
+            feats_lengths,
+            sids_src,
+            sids_tgt,
+            spembs_src,
+            spembs_tgt,
+            lids, )
+
+        return dict(wav=paddle.reshape(wav, [-1]))

paddlespeech/t2s/models/vits/vits_updater.py

@@ -111,6 +111,8 @@ class VITSUpdater(StandardUpdater):
                 text_lengths=batch["text_lengths"],
                 feats=batch["feats"],
                 feats_lengths=batch["feats_lengths"],
+                sids=batch.get("spk_id", None),
+                spembs=batch.get("spk_emb", None),
                 forward_generator=turn == "generator")
             # Generator
             if turn == "generator":
@@ -268,6 +270,8 @@ class VITSEvaluator(StandardEvaluator):
                 text_lengths=batch["text_lengths"],
                 feats=batch["feats"],
                 feats_lengths=batch["feats_lengths"],
+                sids=batch.get("spk_id", None),
+                spembs=batch.get("spk_emb", None),
                 forward_generator=turn == "generator")
             # Generator
             if turn == "generator":