[TTS]Canton CLI, test=tts (#2977)

paddlespeech/cli/tts/infer.py

@@ -82,6 +82,7 @@ class TTSExecutor(BaseExecutor):
                 'tacotron2_csmsc',
                 'tacotron2_ljspeech',
                 'fastspeech2_male',
+                'fastspeech2_canton',
             ],
             help='Choose acoustic model type of tts task.')
         self.parser.add_argument(
@@ -273,7 +274,7 @@ class TTSExecutor(BaseExecutor):
             use_pretrained_voc = False
         voc_lang = lang
         # When speaker is 174 (csmsc), use csmsc's vocoder is better than aishell3's
-        if lang == 'mix':
+        if lang == 'mix' or lang == 'canton':
             voc_dataset = voc[voc.rindex('_') + 1:]
             if voc_dataset in {"ljspeech", "vctk"}:
                 voc_lang = 'en'
@@ -487,7 +488,7 @@ class TTSExecutor(BaseExecutor):
             # fastspeech2
             else:
                 # multi speaker
-                if am_dataset in {'aishell3', 'vctk', 'mix'}:
+                if am_dataset in {'aishell3', 'vctk', 'mix', 'canton'}:
                     mel = self.am_inference(
                         part_phone_ids, spk_id=paddle.to_tensor(spk_id))
                 else:

paddlespeech/resource/pretrained_models.py

@@ -889,6 +889,24 @@ tts_dynamic_pretrained_models = {
             'phone_id_map.txt',
         },
     },
+    "fastspeech2_canton-canton": {
+        '1.0': {
+            'url':
+            'https://paddlespeech.bj.bcebos.com/Parakeet/released_models/fastspeech2/fastspeech2_canton_ckpt_1.4.0.zip',
+            'md5':
+            '504560c082deba82120927627c900374',
+            'config':
+            'default.yaml',
+            'ckpt':
+            'snapshot_iter_140000.pdz',
+            'speech_stats':
+            'speech_stats.npy',
+            'phones_dict':
+            'phone_id_map.txt',
+            'speaker_dict':
+            'speaker_id_map.txt',
+        },
+    },
     "fastspeech2_ljspeech-en": {
         '1.0': {
             'url':

tests/unit/cli/test_cli.sh

@@ -61,6 +61,7 @@ paddlespeech tts --am tacotron2_csmsc --voc wavernn_csmsc --input "你好，欢
 paddlespeech tts --am tacotron2_ljspeech --voc pwgan_ljspeech --lang en --input "Life was like a box of chocolates, you never know what you're gonna get."
 paddlespeech tts --am fastspeech2_male --voc pwgan_male --lang zh --input "你好，欢迎使用百度飞桨深度学习框架！"
 paddlespeech tts --am fastspeech2_male --voc pwgan_male --lang en --input "Life was like a box of chocolates, you never know what you're gonna get."
+paddlespeech tts --am fastspeech2_canton --voc pwgan_aishell3 --input "各个国家有各个国家嘅国歌" --lang canton --spk_id 10
 
 # mix tts
 # The `am` must be `fastspeech2_mix`!