Commit
·
4584388
1
Parent(s):
c2f4320
add japanese
Browse files
app.py
CHANGED
|
@@ -44,6 +44,11 @@ tts = TTS("tts_models/multilingual/multi-dataset/xtts_v1.1")
|
|
| 44 |
model_path = os.path.join(get_user_data_dir("tts"), "tts_models--multilingual--multi-dataset--xtts_v1.1")
|
| 45 |
config = XttsConfig()
|
| 46 |
config.load_json(os.path.join(model_path, "config.json"))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 47 |
model = Xtts.init_from_config(config)
|
| 48 |
model.load_checkpoint(
|
| 49 |
config,
|
|
@@ -61,7 +66,7 @@ DEVICE_ASSERT_LANG=None
|
|
| 61 |
|
| 62 |
def predict(prompt, language, audio_file_pth, mic_file_path, use_mic, voice_cleanup, no_lang_auto_detect, agree,):
|
| 63 |
if agree == True:
|
| 64 |
-
supported_languages=["en","es","fr","de","it","pt","pl","tr","ru","nl","cs","ar","zh-cn"]
|
| 65 |
|
| 66 |
if language not in supported_languages:
|
| 67 |
gr.Warning(f"Language you put {language} in is not in is not in our Supported Languages, please choose from dropdown")
|
|
@@ -79,6 +84,13 @@ def predict(prompt, language, audio_file_pth, mic_file_path, use_mic, voice_clea
|
|
| 79 |
if language_predicted == "zh":
|
| 80 |
#we use zh-cn
|
| 81 |
language_predicted = "zh-cn"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 82 |
print(f"Detected language:{language_predicted}, Chosen language:{language}")
|
| 83 |
|
| 84 |
# After text character length 15 trigger language detection
|
|
@@ -390,6 +402,16 @@ examples = [
|
|
| 390 |
False,
|
| 391 |
True,
|
| 392 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 393 |
]
|
| 394 |
|
| 395 |
|
|
|
|
| 44 |
model_path = os.path.join(get_user_data_dir("tts"), "tts_models--multilingual--multi-dataset--xtts_v1.1")
|
| 45 |
config = XttsConfig()
|
| 46 |
config.load_json(os.path.join(model_path, "config.json"))
|
| 47 |
+
|
| 48 |
+
if "ja-jp" not in config.languages:
|
| 49 |
+
#fix to have JP before next TTS update
|
| 50 |
+
config.languages.append("ja-jp")
|
| 51 |
+
|
| 52 |
model = Xtts.init_from_config(config)
|
| 53 |
model.load_checkpoint(
|
| 54 |
config,
|
|
|
|
| 66 |
|
| 67 |
def predict(prompt, language, audio_file_pth, mic_file_path, use_mic, voice_cleanup, no_lang_auto_detect, agree,):
|
| 68 |
if agree == True:
|
| 69 |
+
supported_languages=["en","es","fr","de","it","pt","pl","tr","ru","nl","cs","ar","zh-cn","ja-jp"]
|
| 70 |
|
| 71 |
if language not in supported_languages:
|
| 72 |
gr.Warning(f"Language you put {language} in is not in is not in our Supported Languages, please choose from dropdown")
|
|
|
|
| 84 |
if language_predicted == "zh":
|
| 85 |
#we use zh-cn
|
| 86 |
language_predicted = "zh-cn"
|
| 87 |
+
|
| 88 |
+
if language_predicted == "ja":
|
| 89 |
+
#we use zh-cn
|
| 90 |
+
language_predicted = "ja-jp"
|
| 91 |
+
if language == "ja":
|
| 92 |
+
language = "ja-jp"
|
| 93 |
+
|
| 94 |
print(f"Detected language:{language_predicted}, Chosen language:{language}")
|
| 95 |
|
| 96 |
# After text character length 15 trigger language detection
|
|
|
|
| 402 |
False,
|
| 403 |
True,
|
| 404 |
],
|
| 405 |
+
[
|
| 406 |
+
"かつて6歳のとき、素晴らしい絵を見ました",
|
| 407 |
+
"ja-jp",
|
| 408 |
+
"examples/female.wav",
|
| 409 |
+
None,
|
| 410 |
+
False,
|
| 411 |
+
True,
|
| 412 |
+
False,
|
| 413 |
+
True,
|
| 414 |
+
],
|
| 415 |
]
|
| 416 |
|
| 417 |
|