Merge pull request #2087 from yt605155624/add_blank

[TTS]install CPython version monotonic_align before training

README.md

@@ -1,3 +1,4 @@
+
 ([简体中文](./README_cn.md)|English)
 <p align="center">
   <img src="./docs/images/PaddleSpeech_logo.png" />
@@ -494,6 +495,14 @@ PaddleSpeech supports a series of most popular models. They are summarized in [r
       <a href = "./examples/aishell3/vc1">ge2e-fastspeech2-aishell3</a>
       </td>
     </tr>
+     <tr>
+      <td rowspan="3">End-to-End</td>
+      <td>VITS</td>
+      <td >CSMSC</td>
+      <td>
+      <a href = "./examples/csmsc/vits">VITS-csmsc</a>
+      </td>
+    </tr>
   </tbody>
 </table>
 

README_cn.md

@@ -1,3 +1,4 @@
+
 (简体中文|[English](./README.md))
 <p align="center">
   <img src="./docs/images/PaddleSpeech_logo.png" />
@@ -481,6 +482,15 @@ 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>
+     <tr>
+      <td rowspan="3">端到端</td>
+      <td>VITS</td>
+      <td >CSMSC</td>
+      <td>
+      <a href = "./examples/csmsc/vits">VITS-csmsc</a>
+      </td>
+    </tr>
   </tbody>
 </table>
 

examples/csmsc/vits/README.md

@@ -144,3 +144,34 @@ optional arguments:
 6. `--ngpu` is the number of gpus to use, if ngpu == 0, use cpu.
 
 ## Pretrained Model
+
+The pretrained model can be downloaded here:
+
+- [vits_csmsc_ckpt_1.1.0.zip](https://paddlespeech.bj.bcebos.com/Parakeet/released_models/vits/vits_csmsc_ckpt_1.1.0.zip) (add_blank=true)
+
+VITS checkpoint contains files listed below.
+```text
+vits_csmsc_ckpt_1.1.0
+├── default.yaml              # default config used to train vitx
+├── phone_id_map.txt          # phone vocabulary file when training vits
+└── snapshot_iter_350000.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 ${BIN_DIR}/synthesize_e2e.py \
+    --config=vits_csmsc_ckpt_1.1.0/default.yaml \
+    --ckpt=vits_csmsc_ckpt_1.1.0/snapshot_iter_350000.pdz \
+    --phones_dict=vits_csmsc_ckpt_1.1.0/phone_id_map.txt \
+    --output_dir=exp/default/test_e2e \
+    --text=${BIN_DIR}/../sentences.txt \
+    --add-blank=${add_blank} 
+```

examples/csmsc/vits/local/synthesize.sh

@@ -15,4 +15,4 @@ if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
         --phones_dict=dump/phone_id_map.txt \
         --test_metadata=dump/test/norm/metadata.jsonl \
         --output_dir=${train_output_path}/test
-fi
+fi

examples/csmsc/vits/local/train.sh

@@ -3,6 +3,11 @@
 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 \