move ssl into t2s.frontend; fix spk_id for 0-D tensor;

paddlespeech/t2s/__init__.py

@@ -18,6 +18,5 @@ from . import exps
 from . import frontend
 from . import models
 from . import modules
-from . import ssml
 from . import training
 from . import utils

paddlespeech/t2s/exps/synthesize_e2e.py

@@ -148,7 +148,7 @@ def evaluate(args):
                         # multi speaker
                         if am_dataset in {"aishell3", "vctk", "mix", "canton"}:
                             # multi-speaker
-                            spk_id = paddle.to_tensor(args.spk_id)
+                            spk_id = paddle.to_tensor([args.spk_id])
                             mel = am_inference(part_phone_ids, spk_id)
                         else:
                             # single-speaker
@@ -157,7 +157,7 @@ def evaluate(args):
                         part_tone_ids = frontend_dict['tone_ids'][i]
                         if am_dataset in {"aishell3", "vctk", "mix"}:
                             # multi-speaker
-                            spk_id = paddle.to_tensor(args.spk_id)
+                            spk_id = paddle.to_tensor([args.spk_id])
                             mel = am_inference(part_phone_ids, part_tone_ids,
                                                spk_id)
                         else:

paddlespeech/t2s/frontend/__init__.py

@@ -15,6 +15,7 @@ from .generate_lexicon import *
 from .normalizer import *
 from .phonectic import *
 from .punctuation import *
+from .ssml import *
 from .tone_sandhi import *
 from .vocab import *
 from .zh_normalization import *

paddlespeech/t2s/frontend/mix_frontend.py

@@ -19,8 +19,8 @@ import numpy as np
 import paddle
 
 from paddlespeech.t2s.frontend import English
+from paddlespeech.t2s.frontend.ssml.xml_processor import MixTextProcessor
 from paddlespeech.t2s.frontend.zh_frontend import Frontend
-from paddlespeech.t2s.ssml.xml_processor import MixTextProcessor
 
 
 class MixFrontend():