From a8252c414939dde04672ec39ee7490a82994be95 Mon Sep 17 00:00:00 2001
From: nyx-c-language <nyxchaoji123@163.com>
Date: Fri, 14 Mar 2025 20:44:22 +0800
Subject: [PATCH 01/12] =?UTF-8?q?run.sh=E4=BF=AE=E6=94=B9=EF=BC=9A?=
 =?UTF-8?q?=E4=B8=BA=20synthesize=20=E5=92=8C=20synthesize=5Fe2e=20?=
 =?UTF-8?q?=E6=B7=BB=E5=8A=A0=20--stage=20=E5=8F=82=E6=95=B0=E6=8E=A7?=
 =?UTF-8?q?=E5=88=B6=20vocoder=20=E6=A8=A1=E5=9E=8B=E9=80=89=E6=8B=A9?=
 =?UTF-8?q?=EF=BC=8CREAMDE.md=E4=BF=AE=E6=94=B9=EF=BC=9A=E8=A1=A5=E5=85=85?=
 =?UTF-8?q?=20stage=20=E5=8F=82=E6=95=B0=E8=AF=B4=E6=98=8E=EF=BC=8C?=
 =?UTF-8?q?=E6=98=8E=E7=A1=AE=20vocoder=20=E9=80=89=E6=8B=A9=E9=80=BB?=
 =?UTF-8?q?=E8=BE=91?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 examples/canton/tts3/README.md | 1 +
 examples/canton/tts3/run.sh    | 4 ++--
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/examples/canton/tts3/README.md b/examples/canton/tts3/README.md
index 87ef40907..d2c46f642 100644
--- a/examples/canton/tts3/README.md
+++ b/examples/canton/tts3/README.md
@@ -37,6 +37,7 @@ Run the command below to
 3. train the model.
 4. synthesize wavs.
     - synthesize waveform from `metadata.jsonl`.
+    - `--stage` controls the vocoder model during synthesis (0 = pwgan, 1 = hifigan).
     - synthesize waveform from text file.
 ```bash
 ./run.sh
diff --git a/examples/canton/tts3/run.sh b/examples/canton/tts3/run.sh
index acfc50223..29502a964 100755
--- a/examples/canton/tts3/run.sh
+++ b/examples/canton/tts3/run.sh
@@ -29,12 +29,12 @@ fi
 
 if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
     # synthesize, vocoder is pwgan by default
-    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
 fi
 
 if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
     # synthesize_e2e, vocoder is pwgan by default
-    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
 fi
 
 if [ ${stage} -le 4 ] && [ ${stop_stage} -ge 4 ]; then

From 74d6eaf64a62fa50a66ec85cdae38c3e56a1f1a7 Mon Sep 17 00:00:00 2001
From: nyx-c-language <nyxchaoji123@163.com>
Date: Fri, 14 Mar 2025 21:52:25 +0800
Subject: [PATCH 02/12] =?UTF-8?q?=E6=B7=BB=E5=8A=A0run.sh=E4=B8=ADstage?=
 =?UTF-8?q?=E5=8F=82=E6=95=B0=E7=9B=B8=E5=85=B3=E7=9A=84=E6=B3=A8=E9=87=8A?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 examples/canton/tts3/run.sh | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/examples/canton/tts3/run.sh b/examples/canton/tts3/run.sh
index 29502a964..0e1f52a1c 100755
--- a/examples/canton/tts3/run.sh
+++ b/examples/canton/tts3/run.sh
@@ -28,12 +28,12 @@ if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
 fi
 
 if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
-    # synthesize, vocoder is pwgan by default
+    # synthesize, vocoder is pwgan by default stage 0, stage 1 will use hifigan as vocoder
     CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
 fi
 
 if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
-    # synthesize_e2e, vocoder is pwgan by default
+    # synthesize_e2e, vocoder is pwgan by default stage 0, stage 1 will use hifigan as vocoder
     CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
 fi
 

From 741a5b0ab12aebe99e64813e8a895a813377e8fc Mon Sep 17 00:00:00 2001
From: nyx-c-language <nyxchaoji123@163.com>
Date: Sun, 16 Mar 2025 10:48:46 +0800
Subject: [PATCH 03/12] =?UTF-8?q?HiFiGAN=E6=94=B9=E4=B8=BAMultiBand=20MelG?=
 =?UTF-8?q?AN?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 examples/csmsc/voc3/README.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/examples/csmsc/voc3/README.md b/examples/csmsc/voc3/README.md
index 25da4c31e..c37099f75 100644
--- a/examples/csmsc/voc3/README.md
+++ b/examples/csmsc/voc3/README.md
@@ -86,7 +86,7 @@ Download pretrained MultiBand MelGAN model from [mb_melgan_csmsc_ckpt_0.1.1.zip]
 ```bash
 unzip mb_melgan_csmsc_ckpt_0.1.1.zip
 ```
-HiFiGAN checkpoint contains files listed below.
+MultiBand MelGAN checkpoint contains files listed below.
 ```text
 mb_melgan_csmsc_ckpt_0.1.1
 ├── default.yaml                    # default config used to train MultiBand MelGAN

From 5be93d436ade9406a5a2780458cf081ac66e388d Mon Sep 17 00:00:00 2001
From: nyx-c-language <nyxchaoji123@163.com>
Date: Sun, 16 Mar 2025 11:09:10 +0800
Subject: [PATCH 04/12] =?UTF-8?q?cmsc=E6=96=87=E4=BB=B6=E6=94=B9=E5=9B=9E?=
 =?UTF-8?q?=E5=8E=9F=E4=BD=8D=EF=BC=88No.15=E4=B8=8D=E4=BF=AE=E6=94=B9?=
 =?UTF-8?q?=EF=BC=89=EF=BC=8C=E8=BF=99=E9=87=8C=E5=8F=AA=E5=AF=B9No.6?=
 =?UTF-8?q?=E5=81=9A=E4=BF=AE=E6=94=B9?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 examples/csmsc/voc3/README.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/examples/csmsc/voc3/README.md b/examples/csmsc/voc3/README.md
index c37099f75..25da4c31e 100644
--- a/examples/csmsc/voc3/README.md
+++ b/examples/csmsc/voc3/README.md
@@ -86,7 +86,7 @@ Download pretrained MultiBand MelGAN model from [mb_melgan_csmsc_ckpt_0.1.1.zip]
 ```bash
 unzip mb_melgan_csmsc_ckpt_0.1.1.zip
 ```
-MultiBand MelGAN checkpoint contains files listed below.
+HiFiGAN checkpoint contains files listed below.
 ```text
 mb_melgan_csmsc_ckpt_0.1.1
 ├── default.yaml                    # default config used to train MultiBand MelGAN

From a11a4c6b5fb6c0b458433a4ff647e6b3d6876ac1 Mon Sep 17 00:00:00 2001
From: Echo-Nie <157974576+Echo-Nie@users.noreply.github.com>
Date: Mon, 7 Apr 2025 14:56:04 +0800
Subject: [PATCH 05/12] fix the Parameter about --voc and --am (#4047)

---
 examples/csmsc/voc1/local/synthesize_e2e.sh | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/examples/csmsc/voc1/local/synthesize_e2e.sh b/examples/csmsc/voc1/local/synthesize_e2e.sh
index 428c234ff..7e1a6e8b7 100644
--- a/examples/csmsc/voc1/local/synthesize_e2e.sh
+++ b/examples/csmsc/voc1/local/synthesize_e2e.sh
@@ -8,13 +8,13 @@ FLAGS_allocator_strategy=naive_best_fit \
 FLAGS_fraction_of_gpu_memory_to_use=0.01 \
 python3 ${BIN_DIR}/../../synthesize_e2e.py \
     --am=fastspeech2_csmsc \
-    --am_config=${config_path} \
-    --am_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
-    --am_stat=dump/train/speech_stats.npy \
+    --am_config=fastspeech2_nosil_baker_ckpt_0.4/default.yaml \
+    --am_ckpt=fastspeech2_nosil_baker_ckpt_0.4/snapshot_iter_76000.pdz \
+    --am_stat=fastspeech2_nosil_baker_ckpt_0.4/speech_stats.npy \
     --voc=pwgan_csmsc \
-    --voc_config=pwg_baker_ckpt_0.4/pwg_default.yaml \
-    --voc_ckpt=pwg_baker_ckpt_0.4/pwg_snapshot_iter_400000.pdz \
-    --voc_stat=pwg_baker_ckpt_0.4/pwg_stats.npy \
+    --voc_config=${config_path} \
+    --voc_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+    --voc_stat=dump/train/feats_stats.npy \
     --lang=zh \
     --text=${BIN_DIR}/../../assets/sentences.txt \
     --output_dir=${train_output_path}/test_e2e \

From ff61df9d68e97f08350a7da91350b8d8e07af491 Mon Sep 17 00:00:00 2001
From: Echo-Nie <157974576+Echo-Nie@users.noreply.github.com>
Date: Mon, 7 Apr 2025 15:06:46 +0800
Subject: [PATCH 06/12] fix the run.sh stage (#4049)

---
 examples/csmsc/tts3_rhy/run.sh | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/examples/csmsc/tts3_rhy/run.sh b/examples/csmsc/tts3_rhy/run.sh
index e49f43ee6..294ceded5 100755
--- a/examples/csmsc/tts3_rhy/run.sh
+++ b/examples/csmsc/tts3_rhy/run.sh
@@ -28,11 +28,13 @@ if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
 fi
 
 if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
-    # synthesize, vocoder is pwgan by default
-    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
+    # synthesize, vocoder is pwgan by default stage 0
+    # use stage 1-4 to select the vocoder to use {multi band melgan, style melgan, hifigan, wavernn}
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
 fi
 
 if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
-    # synthesize_e2e, vocoder is pwgan by default
-    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
+    # synthesize_e2e, vocoder is pwgan by default stage 0
+    # use stage 1,3,4 to select the vocoder to use {multi band melgan, hifigan, wavernn}
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
 fi

From ef29dbb1f066a5f96ba2d94c346a67dfcd47a357 Mon Sep 17 00:00:00 2001
From: Echo-Nie <157974576+Echo-Nie@users.noreply.github.com>
Date: Mon, 7 Apr 2025 15:11:00 +0800
Subject: [PATCH 07/12] fix the tts2 (#4044)

---
 examples/csmsc/tts2/README.md |  8 ++++++--
 examples/csmsc/tts2/run.sh    | 10 ++++++----
 2 files changed, 12 insertions(+), 6 deletions(-)

diff --git a/examples/csmsc/tts2/README.md b/examples/csmsc/tts2/README.md
index 478ae4127..3c6e7d96c 100644
--- a/examples/csmsc/tts2/README.md
+++ b/examples/csmsc/tts2/README.md
@@ -116,8 +116,10 @@ pwg_baker_ckpt_0.4
 ```
 `./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} ${ckpt_name}
+CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name}
 ```
+`--stage` controls the vocoder model during synthesis, which can use stage `0-4` to select the vocoder to use {`pwgan`, `multi band melgan`, `style melgan`, `hifigan`, `wavernn`}
+
 ```text
 usage: synthesize.py [-h]
                      [--am {speedyspeech_csmsc,fastspeech2_csmsc,fastspeech2_ljspeech,fastspeech2_aishell3,fastspeech2_vctk,tacotron2_csmsc,tacotron2_ljspeech,tacotron2_aishell3}]
@@ -164,8 +166,10 @@ optional arguments:
 ```
 `./local/synthesize_e2e.sh` calls `${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} ${ckpt_name}
+CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name}
 ```
+`--stage` controls the vocoder model during synthesis, which can use stage `0,1,3,4` to select the vocoder to use {`pwgan`, `multi band melgan`, `hifigan`, `wavernn`}
+
 ```text
 usage: synthesize_e2e.py [-h]
                          [--am {speedyspeech_csmsc,speedyspeech_aishell3,fastspeech2_csmsc,fastspeech2_ljspeech,fastspeech2_aishell3,fastspeech2_vctk,tacotron2_csmsc,tacotron2_ljspeech}]
diff --git a/examples/csmsc/tts2/run.sh b/examples/csmsc/tts2/run.sh
index 5732ea3c7..6f62bc95b 100755
--- a/examples/csmsc/tts2/run.sh
+++ b/examples/csmsc/tts2/run.sh
@@ -27,13 +27,15 @@ if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
 fi
 
 if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
-    # synthesize, vocoder is pwgan by default
-    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
+    # synthesize, vocoder is pwgan by default stage 0
+    # use stage 1-4 to select the vocoder to use {multi band melgan, style melgan, hifigan, wavernn}
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
 fi
 
 if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
-    # synthesize_e2e, vocoder is pwgan by default
-    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
+    # synthesize_e2e, vocoder is pwgan by default stage 0
+    # use stage 1,3,4 to select the vocoder to use {multi band melgan, hifigan, wavernn}
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
 fi
 
 if [ ${stage} -le 4 ] && [ ${stop_stage} -ge 4 ]; then

From de3fa28b1bde0ad475e770bf5f19a69caf666d72 Mon Sep 17 00:00:00 2001
From: Echo-Nie <157974576+Echo-Nie@users.noreply.github.com>
Date: Mon, 7 Apr 2025 15:12:13 +0800
Subject: [PATCH 08/12] =?UTF-8?q?=E3=80=90PaddleSpeech=20No.7=E3=80=91=20(?=
 =?UTF-8?q?#4043)?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

* run.sh修改：为 synthesize 和 synthesize_e2e 添加 --stage 参数控制 vocoder 模型选择，REAMDE.md修改：补充 stage 参数说明，明确 vocoder 选择逻辑

* 添加run.sh中stage参数相关的注释

* HiFiGAN改为MultiBand MelGAN

* cmsc文件改回原位（No.15不修改），这里只对No.6做修改

* fix the tts0
---
 examples/csmsc/tts0/README.md |  9 +++++++--
 examples/csmsc/tts0/run.sh    | 10 ++++++----
 2 files changed, 13 insertions(+), 6 deletions(-)

diff --git a/examples/csmsc/tts0/README.md b/examples/csmsc/tts0/README.md
index 374270713..6269b419a 100644
--- a/examples/csmsc/tts0/README.md
+++ b/examples/csmsc/tts0/README.md
@@ -99,8 +99,10 @@ pwg_baker_ckpt_0.4
 ```
 `./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} ${ckpt_name}
+CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name}
 ```
+`--stage` controls the vocoder model during synthesis, which can use stage `0-4` to select the vocoder to use {`pwgan`, `multi band melgan`, `style melgan`, ` hifigan`, `wavernn`}
+
 ```text
 usage: synthesize.py [-h]
                      [--am {speedyspeech_csmsc,fastspeech2_csmsc,fastspeech2_ljspeech,fastspeech2_aishell3,fastspeech2_vctk,tacotron2_csmsc,tacotron2_ljspeech,tacotron2_aishell3}]
@@ -146,9 +148,12 @@ optional arguments:
                         output dir.
 ```
 `./local/synthesize_e2e.sh` calls `${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} ${ckpt_name}
+CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name}
 ```
+`--stage` controls the vocoder model during synthesis, which can use stage `0,1,3,4` to select the vocoder to use{`pwgan`, `multi band melgan`, `hifigan`, `wavernn`}
+
 ```text
 usage: synthesize_e2e.py [-h]
                          [--am {speedyspeech_csmsc,speedyspeech_aishell3,fastspeech2_csmsc,fastspeech2_ljspeech,fastspeech2_aishell3,fastspeech2_vctk,tacotron2_csmsc,tacotron2_ljspeech}]
diff --git a/examples/csmsc/tts0/run.sh b/examples/csmsc/tts0/run.sh
index 8f06e933c..83bb02a5f 100755
--- a/examples/csmsc/tts0/run.sh
+++ b/examples/csmsc/tts0/run.sh
@@ -27,13 +27,15 @@ if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
 fi
 
 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} ${ckpt_name} || exit -1
+    # synthesize, vocoder is pwgan by default stage 0
+    # stage 1-4 to select the vocoder to use {multi band melgan, style melgan, hifigan, wavernn}
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
 fi
 
 if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
-    # synthesize_e2e, vocoder is pwgan
-    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
+    # synthesize_e2e, vocoder is pwgan by default stage 0
+    # stage 1,3,4 to select the vocoder to use {multi band melgan, hifigan, wavernn}
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
 fi
 
 if [ ${stage} -le 4 ] && [ ${stop_stage} -ge 4 ]; then

From fae44e7e78e21b9d5386c2e9c1c6b741d47363f6 Mon Sep 17 00:00:00 2001
From: Echo-Nie <157974576+Echo-Nie@users.noreply.github.com>
Date: Mon, 7 Apr 2025 15:15:40 +0800
Subject: [PATCH 09/12] =?UTF-8?q?=E3=80=90PaddleSpeech=20No.12=E3=80=91=20?=
 =?UTF-8?q?(#4037)?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

* fix the svs1

* del stage 0

* 删去中文README中多余的stage 0
---
 examples/opencpop/svs1/README.md    | 6 +++++-
 examples/opencpop/svs1/README_cn.md | 6 +++++-
 examples/opencpop/svs1/run.sh       | 4 ++--
 3 files changed, 12 insertions(+), 4 deletions(-)

diff --git a/examples/opencpop/svs1/README.md b/examples/opencpop/svs1/README.md
index 092f27644..534a741fd 100644
--- a/examples/opencpop/svs1/README.md
+++ b/examples/opencpop/svs1/README.md
@@ -118,6 +118,8 @@ pwgan_opencpop_ckpt_1.4.0.zip
 ```bash
 CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh ${conf_path} ${train_output_path} ${ckpt_name}
 ```
+use `pwgan` model as vocoder.
+
 ```text
 usage: synthesize.py [-h]
                      [--am {diffsinger_opencpop}]
@@ -170,8 +172,10 @@ optional arguments:
 `local/pinyin_to_phone.txt` comes from the readme of the opencpop dataset, indicating the mapping from pinyin to phonemes in opencpop.
 
 ```bash
-CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh ${conf_path} ${train_output_path} ${ckpt_name}
+CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name}
 ```
+`--stage` controls the vocoder model during synthesis, which can be `0` or `1`, use `pwgan` or `hifigan` model as vocoder.
+
 ```text
 usage: synthesize_e2e.py [-h]
                          [--am {speedyspeech_csmsc,speedyspeech_aishell3,fastspeech2_csmsc,fastspeech2_ljspeech,fastspeech2_aishell3,fastspeech2_vctk,tacotron2_csmsc,tacotron2_ljspeech}]
diff --git a/examples/opencpop/svs1/README_cn.md b/examples/opencpop/svs1/README_cn.md
index eac72b5aa..e35967d71 100644
--- a/examples/opencpop/svs1/README_cn.md
+++ b/examples/opencpop/svs1/README_cn.md
@@ -121,6 +121,8 @@ pwgan_opencpop_ckpt_1.4.0.zip
 ```bash
 CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh ${conf_path} ${train_output_path} ${ckpt_name}
 ```
+使用 `pwgan` 模型作为声码器。
+
 ```text
 usage: synthesize.py [-h]
                      [--am {diffsinger_opencpop}]
@@ -173,8 +175,10 @@ optional arguments:
 `local/pinyin_to_phone.txt`来源于opencpop数据集中的README，表示opencpop中拼音到音素的映射。
 
 ```bash
-CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh ${conf_path} ${train_output_path} ${ckpt_name}
+CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name}
 ```
+`--stage` 用于选择合成时使用的声码器模型，取值为 `0` 或 `1`，分别对应使用 `pwgan` 或 `hifigan` 模型作为声码器。
+
 ```text
 usage: synthesize_e2e.py [-h]
                          [--am {speedyspeech_csmsc,speedyspeech_aishell3,fastspeech2_csmsc,fastspeech2_ljspeech,fastspeech2_aishell3,fastspeech2_vctk,tacotron2_csmsc,tacotron2_ljspeech}]
diff --git a/examples/opencpop/svs1/run.sh b/examples/opencpop/svs1/run.sh
index bfe5b6594..6c6688b2f 100755
--- a/examples/opencpop/svs1/run.sh
+++ b/examples/opencpop/svs1/run.sh
@@ -32,6 +32,6 @@ if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
 fi
 
 if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
-    # synthesize_e2e, vocoder is pwgan by default
-    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
+    # synthesize_e2e, vocoder is pwgan by default, stage 1 will use hifigan as vocoder
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
 fi

From 5417f426fe1eaf87391eef8e203047c16b60fc83 Mon Sep 17 00:00:00 2001
From: Echo-Nie <157974576+Echo-Nie@users.noreply.github.com>
Date: Tue, 8 Apr 2025 11:26:58 +0800
Subject: [PATCH 10/12] fix the voc3 synthesize_e2e about --am and --voc
 (#4051)

---
 examples/csmsc/voc5/local/synthesize_e2e.sh | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/examples/csmsc/voc5/local/synthesize_e2e.sh b/examples/csmsc/voc5/local/synthesize_e2e.sh
index a068cf19a..64f5d8a3f 100644
--- a/examples/csmsc/voc5/local/synthesize_e2e.sh
+++ b/examples/csmsc/voc5/local/synthesize_e2e.sh
@@ -8,13 +8,13 @@ FLAGS_allocator_strategy=naive_best_fit \
 FLAGS_fraction_of_gpu_memory_to_use=0.01 \
 python3 ${BIN_DIR}/../../synthesize_e2e.py \
     --am=fastspeech2_csmsc \
-    --am_config=${config_path} \
-    --am_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
-    --am_stat=dump/train/speech_stats.npy \
+    --am_config=fastspeech2_nosil_baker_ckpt_0.4/default.yaml \
+    --am_ckpt=fastspeech2_nosil_baker_ckpt_0.4/snapshot_iter_76000.pdz \
+    --am_stat=fastspeech2_nosil_baker_ckpt_0.4/speech_stats.npy \
     --voc=hifigan_csmsc \
-    --voc_config=hifigan_csmsc_ckpt_0.1.1/default.yaml \
-    --voc_ckpt=hifigan_csmsc_ckpt_0.1.1/snapshot_iter_2500000.pdz \
-    --voc_stat=hifigan_csmsc_ckpt_0.1.1/feats_stats.npy \
+    --voc_config=${config_path} \
+    --voc_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+    --voc_stat=dump/train/feats_stats.npy \
     --lang=zh \
     --text=${BIN_DIR}/../../assets/sentences.txt \
     --output_dir=${train_output_path}/test_e2e \

From f2be23553d04efeb4f6b114588af73c0c23df2ed Mon Sep 17 00:00:00 2001
From: Echo-Nie <157974576+Echo-Nie@users.noreply.github.com>
Date: Tue, 8 Apr 2025 11:28:06 +0800
Subject: [PATCH 11/12] fix the voc3 about --am and --voc (#4050)

---
 examples/csmsc/voc3/local/synthesize_e2e.sh | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/examples/csmsc/voc3/local/synthesize_e2e.sh b/examples/csmsc/voc3/local/synthesize_e2e.sh
index e9b52dbec..778dc7f97 100644
--- a/examples/csmsc/voc3/local/synthesize_e2e.sh
+++ b/examples/csmsc/voc3/local/synthesize_e2e.sh
@@ -8,13 +8,13 @@ FLAGS_allocator_strategy=naive_best_fit \
 FLAGS_fraction_of_gpu_memory_to_use=0.01 \
 python3 ${BIN_DIR}/../../synthesize_e2e.py \
     --am=fastspeech2_csmsc \
-    --am_config=${config_path} \
-    --am_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
-    --am_stat=dump/train/speech_stats.npy \
+    --am_config=fastspeech2_nosil_baker_ckpt_0.4/default.yaml \
+    --am_ckpt=fastspeech2_nosil_baker_ckpt_0.4/snapshot_iter_76000.pdz \
+    --am_stat=fastspeech2_nosil_baker_ckpt_0.4/speech_stats.npy \
     --voc=mb_melgan_csmsc \
-    --voc_config=mb_melgan_csmsc_ckpt_0.1.1/default.yaml \
-    --voc_ckpt=mb_melgan_csmsc_ckpt_0.1.1/snapshot_iter_1000000.pdz\
-    --voc_stat=mb_melgan_csmsc_ckpt_0.1.1/feats_stats.npy \
+    --voc_config=${config_path} \
+    --voc_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
+    --voc_stat=dump/train/feats_stats.npy \
     --lang=zh \
     --text=${BIN_DIR}/../../assets/sentences.txt \
     --output_dir=${train_output_path}/test_e2e \

From d0bb18d7d7836fddc02f4af2da53e112e85b388d Mon Sep 17 00:00:00 2001
From: Echo-Nie <157974576+Echo-Nie@users.noreply.github.com>
Date: Tue, 8 Apr 2025 11:29:22 +0800
Subject: [PATCH 12/12] =?UTF-8?q?=E3=80=90PaddleSpeech=20No.9=E3=80=91=20(?=
 =?UTF-8?q?#4045)?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

* fix the tts3

* 中英文文档保持一致
---
 examples/csmsc/tts3/README.md    |  8 ++++++--
 examples/csmsc/tts3/README_cn.md |  8 ++++++--
 examples/csmsc/tts3/run.sh       | 10 ++++++----
 3 files changed, 18 insertions(+), 8 deletions(-)

diff --git a/examples/csmsc/tts3/README.md b/examples/csmsc/tts3/README.md
index dc55fac93..9d349426b 100644
--- a/examples/csmsc/tts3/README.md
+++ b/examples/csmsc/tts3/README.md
@@ -107,8 +107,10 @@ pwg_baker_ckpt_0.4
 ```
 `./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} ${ckpt_name}
+CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name}
 ```
+`--stage` controls the vocoder model during synthesis. The parameter values range from `0-4`, corresponding to the following five vocoder models: `pwgan`, `multi band melgan`, `style melgan`, `hifigan`, and `wavernn`.
+
 ```text
 usage: synthesize.py [-h]
                      [--am {speedyspeech_csmsc,fastspeech2_csmsc,fastspeech2_ljspeech,fastspeech2_aishell3,fastspeech2_vctk,tacotron2_csmsc,tacotron2_ljspeech,tacotron2_aishell3}]
@@ -155,8 +157,10 @@ optional arguments:
 ```
 `./local/synthesize_e2e.sh` calls `${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} ${ckpt_name}
+CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name}
 ```
+`--stage` controls the vocoder model during synthesis. The parameter values are {`0,1,3,4`}, corresponding to the following four vocoder models: `pwgan`, `multi band melgan`, `hifigan`, and `wavernn`.
+
 ```text
 usage: synthesize_e2e.py [-h]
                          [--am {speedyspeech_csmsc,speedyspeech_aishell3,fastspeech2_csmsc,fastspeech2_ljspeech,fastspeech2_aishell3,fastspeech2_vctk,tacotron2_csmsc,tacotron2_ljspeech}]
diff --git a/examples/csmsc/tts3/README_cn.md b/examples/csmsc/tts3/README_cn.md
index c6eee69d0..399c15d55 100644
--- a/examples/csmsc/tts3/README_cn.md
+++ b/examples/csmsc/tts3/README_cn.md
@@ -113,8 +113,10 @@ pwg_baker_ckpt_0.4
 `./local/synthesize.sh` 调用 `${BIN_DIR}/../synthesize.py` 即可从 `metadata.jsonl`中合成波形。
 
 ```bash
-CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh ${conf_path} ${train_output_path} ${ckpt_name}
+CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name}
 ```
+`--stage` 参数用于控制合成过程中使用的声码器模型。该参数的取值范围为 `0-4`，分别对应以下五种声码器模型：`pwgan`、`multi band melgan`、`style melgan`、`hifigan` 和 `wavernn`。
+
 ```text
 usage: synthesize.py [-h]
                      [--am {speedyspeech_csmsc,fastspeech2_csmsc,fastspeech2_ljspeech,fastspeech2_aishell3,fastspeech2_vctk,tacotron2_csmsc,tacotron2_ljspeech,tacotron2_aishell3}]
@@ -162,8 +164,10 @@ optional arguments:
 `./local/synthesize_e2e.sh` 调用 `${BIN_DIR}/../synthesize_e2e.py`，即可从文本文件中合成波形。
 
 ```bash
-CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh ${conf_path} ${train_output_path} ${ckpt_name}
+CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name}
 ```
+`--stage` 参数用于控制合成过程中使用的声码器模型。该参数的取值范围为{ `0,1,3,4`}，分别对应以下四种声码器模型：`pwgan`、`multi band melgan`、`hifigan` 和 `wavernn`。
+
 ```text
 usage: synthesize_e2e.py [-h]
                          [--am {speedyspeech_csmsc,speedyspeech_aishell3,fastspeech2_csmsc,fastspeech2_ljspeech,fastspeech2_aishell3,fastspeech2_vctk,tacotron2_csmsc,tacotron2_ljspeech}]
diff --git a/examples/csmsc/tts3/run.sh b/examples/csmsc/tts3/run.sh
index a7b4e4239..96fa84471 100755
--- a/examples/csmsc/tts3/run.sh
+++ b/examples/csmsc/tts3/run.sh
@@ -27,13 +27,15 @@ if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
 fi
 
 if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
-    # synthesize, vocoder is pwgan by default
-    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
+    # synthesize, vocoder is pwgan by default stage 0
+    # use stage 1-4 to select the vocoder to use {multi band melgan, style melgan, hifigan, wavernn}
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
 fi
 
 if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
-    # synthesize_e2e, vocoder is pwgan by default
-    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
+    # synthesize_e2e, vocoder is pwgan by default stage 0
+    # use stage 1,3,4 to select the vocoder to use {multi band melgan, hifigan, wavernn}
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
 fi
 
 if [ ${stage} -le 4 ] && [ ${stop_stage} -ge 4 ]; then