update vctk voc1, test=tts (#1294)

examples/vctk/tts3/README.md

@@ -95,16 +95,16 @@ optional arguments:
 ### Synthesizing
 We use [parallel wavegan](https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/examples/vctk/voc1) as the neural vocoder.
 
-Download pretrained parallel wavegan model from [pwg_vctk_ckpt_0.5.zip](https://paddlespeech.bj.bcebos.com/Parakeet/released_models/pwgan/pwg_vctk_ckpt_0.5.zip)and unzip it.
+Download pretrained parallel wavegan model from [pwg_vctk_ckpt_0.1.1.zip](https://paddlespeech.bj.bcebos.com/Parakeet/released_models/pwgan/pwg_vctk_ckpt_0.1.1.zip) and unzip it.
 ```bash
-unzip pwg_vctk_ckpt_0.5.zip
+unzip pwg_vctk_ckpt_0.1.1.zip
 ```
 Parallel WaveGAN checkpoint contains files listed below.
 ```text
-pwg_vctk_ckpt_0.5
-├── pwg_default.yaml               # default config used to train parallel wavegan
-├── pwg_snapshot_iter_1000000.pdz  # generator parameters of parallel wavegan
-└── pwg_stats.npy                  # statistics used to normalize spectrogram when training parallel wavegan
+pwg_vctk_ckpt_0.1.1
+├── default.yaml                   # default config used to train parallel wavegan
+├── snapshot_iter_1500000.pdz      # generator parameters of parallel wavegan
+└── feats_stats.npy                # statistics used to normalize spectrogram when training parallel wavegan
 ```
 `./local/synthesize.sh` calls `${BIN_DIR}/../synthesize.py`, which can synthesize waveform from `metadata.jsonl`.
 ```bash

examples/vctk/tts3/local/synthesize.sh

@@ -12,9 +12,9 @@ python3 ${BIN_DIR}/../synthesize.py \
     --am_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
     --am_stat=dump/train/speech_stats.npy \
     --voc=pwgan_vctk \
-    --voc_config=pwg_vctk_ckpt_0.5/pwg_default.yaml \
-    --voc_ckpt=pwg_vctk_ckpt_0.5/pwg_snapshot_iter_1000000.pdz \
-    --voc_stat=pwg_vctk_ckpt_0.5/pwg_stats.npy \
+    --voc_config=pwg_vctk_ckpt_0.1.1/default.yaml  \
+    --voc_ckpt=pwg_vctk_ckpt_0.1.1/snapshot_iter_1500000.pdz \
+    --voc_stat=pwg_vctk_ckpt_0.1.1/feats_stats.npy \
     --test_metadata=dump/test/norm/metadata.jsonl \
     --output_dir=${train_output_path}/test \
     --phones_dict=dump/phone_id_map.txt \

examples/vctk/tts3/local/synthesize_e2e.sh

@@ -12,9 +12,9 @@ python3 ${BIN_DIR}/../synthesize_e2e.py \
     --am_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
     --am_stat=dump/train/speech_stats.npy \
     --voc=pwgan_vctk \
-    --voc_config=pwg_vctk_ckpt_0.5/pwg_default.yaml  \
-    --voc_ckpt=pwg_vctk_ckpt_0.5/pwg_snapshot_iter_1000000.pdz \
-    --voc_stat=pwg_vctk_ckpt_0.5/pwg_stats.npy \
+    --voc_config=pwg_vctk_ckpt_0.1.1/default.yaml  \
+    --voc_ckpt=pwg_vctk_ckpt_0.1.1/snapshot_iter_1500000.pdz \
+    --voc_stat=pwg_vctk_ckpt_0.1.1/feats_stats.npy \
     --lang=en \
     --text=${BIN_DIR}/../sentences_en.txt \
     --output_dir=${train_output_path}/test_e2e \

examples/vctk/voc1/README.md

@@ -132,15 +132,15 @@ optional arguments:
 5. `--ngpu` is the number of gpus to use, if ngpu == 0, use cpu.
 
 ## Pretrained Model
-Pretrained models can be downloaded here [pwg_vctk_ckpt_0.5.zip](https://paddlespeech.bj.bcebos.com/Parakeet/released_models/pwgan/pwg_vctk_ckpt_0.5.zip).
+Pretrained models can be downloaded here [pwg_vctk_ckpt_0.1.1.zip](https://paddlespeech.bj.bcebos.com/Parakeet/released_models/pwgan/pwg_vctk_ckpt_0.1.1.zip).
 
 Parallel WaveGAN checkpoint contains files listed below.
 
 ```text
-pwg_vctk_ckpt_0.5
-├── pwg_default.yaml               # default config used to train parallel wavegan
-├── pwg_snapshot_iter_1000000.pdz  # generator parameters of parallel wavegan
-└── pwg_stats.npy                  # statistics used to normalize spectrogram when training parallel wavegan
+pwg_vctk_ckpt_0.1.1
+├── default.yaml                   # default config used to train parallel wavegan
+├── snapshot_iter_1500000.pdz      # generator parameters of parallel wavegan
+└── feats_stats.npy                # statistics used to normalize spectrogram when training parallel wavegan
 ```
 ## Acknowledgement
 We adapted some code from https://github.com/kan-bayashi/ParallelWaveGAN.

examples/vctk/voc1/conf/default.yaml

@@ -70,7 +70,7 @@ lambda_adv: 4.0  # Loss balancing coefficient.
 ###########################################################
 #                  DATA LOADER SETTING                    #
 ###########################################################
-batch_size: 8              # Batch size.
+batch_size: 6              # Batch size.
 batch_max_steps: 24000     # Length of each audio in batch. Make sure dividable by n_shift.
 num_workers: 2             # Number of workers in DataLoader.
 
@@ -100,7 +100,7 @@ discriminator_grad_norm: 1 # Discriminator's gradient norm.
 #                    INTERVAL SETTING                     #
 ###########################################################
 discriminator_train_start_steps: 100000 # Number of steps to start to train discriminator.
-train_max_steps: 1000000                # Number of training steps.
+train_max_steps: 1500000                # Number of training steps.
 save_interval_steps: 5000               # Interval steps to save checkpoint.
 eval_interval_steps: 1000               # Interval steps to evaluate the network.
 

paddlespeech/cli/tts/infer.py

@@ -156,15 +156,15 @@ pretrained_models = {
     },
     "pwgan_vctk-en": {
         'url':
-        'https://paddlespeech.bj.bcebos.com/Parakeet/released_models/pwgan/pwg_vctk_ckpt_0.5.zip',
+        'https://paddlespeech.bj.bcebos.com/Parakeet/released_models/pwgan/pwg_vctk_ckpt_0.1.1.zip',
         'md5':
-        '322ca688aec9b127cec2788b65aa3d52',
+        'b3da1defcde3e578be71eb284cb89f2c',
         'config':
-        'pwg_default.yaml',
+        'default.yaml',
         'ckpt':
-        'pwg_snapshot_iter_1000000.pdz',
+        'snapshot_iter_1500000.pdz',
         'speech_stats':
-        'pwg_stats.npy',
+        'feats_stats.npy',
     },
     # mb_melgan
     "mb_melgan_csmsc-zh": {