🦝 Raccoon Mission: Silero TTS integration complete with 140 voices
✅ Integrated Silero TTS as tts-1-silero model - Fixed omegaconf dependency - Fixed Silero API integration (torch.hub.load returns 2 values) - Fixed model.to(device) returning None bug - Mapped all 140 Silero voices across 5 languages: * English (en): 118 speakers (en_0 to en_117) + random * Russian (ru): 5 speakers (aidar, baya, kseniya, xenia, eugene) + random * German (de): 5 speakers (bernd_ungerer, eva_k, friedrich, hokuspokus, karlsson) + random * Spanish (es): 3 speakers (es_0, es_1, es_2) + random * French (fr): 6 speakers (fr_0 to fr_5) + random 📝 Configuration changes: - requirements.txt: Added omegaconf for Silero - voice_to_speaker.default.yaml: All 140 Silero voices mapped - speech.py: Silero wrapper class with proper API handling 🎯 Working TTS engines: 3 - Piper TTS (tts-1) - Fast, lightweight - XTTS v2 (tts-1-hd) - High quality, voice cloning - Silero TTS (tts-1-silero) - CPU-friendly, 5 languages, actively maintained 🦝 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
parent
4deedb9539
commit
01e51b08b5
3 changed files with 620 additions and 21 deletions
|
|
@ -10,6 +10,7 @@ coqui-tts[languages]
|
|||
# Silero TTS - actively maintained, small efficient models
|
||||
# Note: Silero models are loaded via torch.hub, no package install needed
|
||||
# Models: ~50-100MB each, CPU-friendly, real-time capable
|
||||
omegaconf # Required by Silero TTS
|
||||
# Chatterbox - emotion control, 23 languages (Resemble AI)
|
||||
# Install from git since no PyPI package exists yet
|
||||
# 🦝 RACCOON NOTE: Disabled due to dependency conflict with Coqui TTS
|
||||
|
|
|
|||
46
speech.py
46
speech.py
|
|
@ -113,35 +113,59 @@ class xtts_wrapper():
|
|||
self.last_used = time.time()
|
||||
|
||||
class silero_wrapper():
|
||||
"""Wrapper for Silero TTS models"""
|
||||
def __init__(self, language='en', speaker='v4_en', device='cpu'):
|
||||
"""Wrapper for Silero TTS models
|
||||
|
||||
Silero torch.hub.load returns: (model, example_text)
|
||||
The model has a method apply_tts(text, speaker, sample_rate)
|
||||
"""
|
||||
def __init__(self, language='en', speaker='v3_en', device='cpu'):
|
||||
self.language = language
|
||||
self.speaker = speaker
|
||||
self.device = device
|
||||
|
||||
logger.info(f"Loading Silero model for {language} on {device}")
|
||||
logger.info(f"Loading Silero model for {language} with speaker {speaker} on {device}")
|
||||
|
||||
import torch
|
||||
self.model, self.symbols, self.sample_rate, self.example_text, self.apply_tts = torch.hub.load(
|
||||
try:
|
||||
# torch.hub.load returns (model, example_text)
|
||||
self.model, example_text = torch.hub.load(
|
||||
repo_or_dir='snakers4/silero-models',
|
||||
model='silero_tts',
|
||||
language=language,
|
||||
speaker=speaker
|
||||
speaker=speaker,
|
||||
verbose=False
|
||||
)
|
||||
self.model = self.model.to(device)
|
||||
|
||||
def tts(self, text, speaker_id='en_0'):
|
||||
logger.info(f"Model loaded, type: {type(self.model)}")
|
||||
self.model.to(device) # Move to device (in-place for Silero)
|
||||
self.sample_rate = 48000 # Silero uses 48kHz
|
||||
logger.info(f"Successfully loaded Silero {language}/{speaker}, example: {example_text}")
|
||||
except Exception as e:
|
||||
logger.error(f"Failed to load Silero model: {e}")
|
||||
raise
|
||||
|
||||
def tts(self, text, speaker_id='lj_16khz'):
|
||||
"""Generate speech from text"""
|
||||
import torch
|
||||
|
||||
logger.info(f"Silero tts() called: model={self.model}, speaker_id={speaker_id}")
|
||||
|
||||
if self.model is None:
|
||||
raise RuntimeError("Silero model is None - model failed to load")
|
||||
|
||||
if not hasattr(self.model, 'apply_tts'):
|
||||
logger.error(f"Model has no apply_tts method. Model type: {type(self.model)}, dir: {dir(self.model)}")
|
||||
raise AttributeError(f"Silero model {type(self.model)} has no apply_tts method")
|
||||
|
||||
with torch.no_grad():
|
||||
audio = self.apply_tts(
|
||||
# Use model's apply_tts method
|
||||
audio = self.model.apply_tts(
|
||||
text=text,
|
||||
speaker=speaker_id,
|
||||
sample_rate=self.sample_rate
|
||||
)
|
||||
# Convert to float32 PCM
|
||||
audio_np = audio.cpu().numpy()
|
||||
return audio_np.tobytes()
|
||||
# audio is a tensor, convert to numpy float32
|
||||
return audio.cpu().numpy().tobytes()
|
||||
|
||||
def default_exists(filename: str):
|
||||
if not os.path.exists(filename):
|
||||
|
|
|
|||
|
|
@ -58,27 +58,601 @@ tts-1-hd:
|
|||
top_p: 0.85
|
||||
comment: You can add a comment here also, which will be persistent and otherwise ignored.
|
||||
tts-1-silero:
|
||||
# OpenAI-compatible voice aliases (for compatibility)
|
||||
alloy:
|
||||
language: en
|
||||
speaker: en_0
|
||||
silero_speaker: v4_en
|
||||
silero_speaker: v3_en
|
||||
echo:
|
||||
language: en
|
||||
speaker: en_1
|
||||
silero_speaker: v4_en
|
||||
silero_speaker: v3_en
|
||||
fable:
|
||||
language: en
|
||||
speaker: en_2
|
||||
silero_speaker: v4_en
|
||||
silero_speaker: v3_en
|
||||
onyx:
|
||||
language: en
|
||||
speaker: en_3
|
||||
silero_speaker: v4_en
|
||||
silero_speaker: v3_en
|
||||
nova:
|
||||
language: en
|
||||
speaker: en_4
|
||||
silero_speaker: v4_en
|
||||
silero_speaker: v3_en
|
||||
shimmer:
|
||||
language: en
|
||||
speaker: en_5
|
||||
silero_speaker: v4_en
|
||||
silero_speaker: v3_en
|
||||
# All 118 Silero v3_en voices with proper names
|
||||
en_0:
|
||||
language: en
|
||||
speaker: en_0
|
||||
silero_speaker: v3_en
|
||||
en_1:
|
||||
language: en
|
||||
speaker: en_1
|
||||
silero_speaker: v3_en
|
||||
en_2:
|
||||
language: en
|
||||
speaker: en_2
|
||||
silero_speaker: v3_en
|
||||
en_3:
|
||||
language: en
|
||||
speaker: en_3
|
||||
silero_speaker: v3_en
|
||||
en_4:
|
||||
language: en
|
||||
speaker: en_4
|
||||
silero_speaker: v3_en
|
||||
en_5:
|
||||
language: en
|
||||
speaker: en_5
|
||||
silero_speaker: v3_en
|
||||
en_6:
|
||||
language: en
|
||||
speaker: en_6
|
||||
silero_speaker: v3_en
|
||||
en_7:
|
||||
language: en
|
||||
speaker: en_7
|
||||
silero_speaker: v3_en
|
||||
en_8:
|
||||
language: en
|
||||
speaker: en_8
|
||||
silero_speaker: v3_en
|
||||
en_9:
|
||||
language: en
|
||||
speaker: en_9
|
||||
silero_speaker: v3_en
|
||||
en_10:
|
||||
language: en
|
||||
speaker: en_10
|
||||
silero_speaker: v3_en
|
||||
en_11:
|
||||
language: en
|
||||
speaker: en_11
|
||||
silero_speaker: v3_en
|
||||
en_12:
|
||||
language: en
|
||||
speaker: en_12
|
||||
silero_speaker: v3_en
|
||||
en_13:
|
||||
language: en
|
||||
speaker: en_13
|
||||
silero_speaker: v3_en
|
||||
en_14:
|
||||
language: en
|
||||
speaker: en_14
|
||||
silero_speaker: v3_en
|
||||
en_15:
|
||||
language: en
|
||||
speaker: en_15
|
||||
silero_speaker: v3_en
|
||||
en_16:
|
||||
language: en
|
||||
speaker: en_16
|
||||
silero_speaker: v3_en
|
||||
en_17:
|
||||
language: en
|
||||
speaker: en_17
|
||||
silero_speaker: v3_en
|
||||
en_18:
|
||||
language: en
|
||||
speaker: en_18
|
||||
silero_speaker: v3_en
|
||||
en_19:
|
||||
language: en
|
||||
speaker: en_19
|
||||
silero_speaker: v3_en
|
||||
en_20:
|
||||
language: en
|
||||
speaker: en_20
|
||||
silero_speaker: v3_en
|
||||
en_21:
|
||||
language: en
|
||||
speaker: en_21
|
||||
silero_speaker: v3_en
|
||||
en_22:
|
||||
language: en
|
||||
speaker: en_22
|
||||
silero_speaker: v3_en
|
||||
en_23:
|
||||
language: en
|
||||
speaker: en_23
|
||||
silero_speaker: v3_en
|
||||
en_24:
|
||||
language: en
|
||||
speaker: en_24
|
||||
silero_speaker: v3_en
|
||||
en_25:
|
||||
language: en
|
||||
speaker: en_25
|
||||
silero_speaker: v3_en
|
||||
en_26:
|
||||
language: en
|
||||
speaker: en_26
|
||||
silero_speaker: v3_en
|
||||
en_27:
|
||||
language: en
|
||||
speaker: en_27
|
||||
silero_speaker: v3_en
|
||||
en_28:
|
||||
language: en
|
||||
speaker: en_28
|
||||
silero_speaker: v3_en
|
||||
en_29:
|
||||
language: en
|
||||
speaker: en_29
|
||||
silero_speaker: v3_en
|
||||
en_30:
|
||||
language: en
|
||||
speaker: en_30
|
||||
silero_speaker: v3_en
|
||||
en_31:
|
||||
language: en
|
||||
speaker: en_31
|
||||
silero_speaker: v3_en
|
||||
en_32:
|
||||
language: en
|
||||
speaker: en_32
|
||||
silero_speaker: v3_en
|
||||
en_33:
|
||||
language: en
|
||||
speaker: en_33
|
||||
silero_speaker: v3_en
|
||||
en_34:
|
||||
language: en
|
||||
speaker: en_34
|
||||
silero_speaker: v3_en
|
||||
en_35:
|
||||
language: en
|
||||
speaker: en_35
|
||||
silero_speaker: v3_en
|
||||
en_36:
|
||||
language: en
|
||||
speaker: en_36
|
||||
silero_speaker: v3_en
|
||||
en_37:
|
||||
language: en
|
||||
speaker: en_37
|
||||
silero_speaker: v3_en
|
||||
en_38:
|
||||
language: en
|
||||
speaker: en_38
|
||||
silero_speaker: v3_en
|
||||
en_39:
|
||||
language: en
|
||||
speaker: en_39
|
||||
silero_speaker: v3_en
|
||||
en_40:
|
||||
language: en
|
||||
speaker: en_40
|
||||
silero_speaker: v3_en
|
||||
en_41:
|
||||
language: en
|
||||
speaker: en_41
|
||||
silero_speaker: v3_en
|
||||
en_42:
|
||||
language: en
|
||||
speaker: en_42
|
||||
silero_speaker: v3_en
|
||||
en_43:
|
||||
language: en
|
||||
speaker: en_43
|
||||
silero_speaker: v3_en
|
||||
en_44:
|
||||
language: en
|
||||
speaker: en_44
|
||||
silero_speaker: v3_en
|
||||
en_45:
|
||||
language: en
|
||||
speaker: en_45
|
||||
silero_speaker: v3_en
|
||||
en_46:
|
||||
language: en
|
||||
speaker: en_46
|
||||
silero_speaker: v3_en
|
||||
en_47:
|
||||
language: en
|
||||
speaker: en_47
|
||||
silero_speaker: v3_en
|
||||
en_48:
|
||||
language: en
|
||||
speaker: en_48
|
||||
silero_speaker: v3_en
|
||||
en_49:
|
||||
language: en
|
||||
speaker: en_49
|
||||
silero_speaker: v3_en
|
||||
en_50:
|
||||
language: en
|
||||
speaker: en_50
|
||||
silero_speaker: v3_en
|
||||
en_51:
|
||||
language: en
|
||||
speaker: en_51
|
||||
silero_speaker: v3_en
|
||||
en_52:
|
||||
language: en
|
||||
speaker: en_52
|
||||
silero_speaker: v3_en
|
||||
en_53:
|
||||
language: en
|
||||
speaker: en_53
|
||||
silero_speaker: v3_en
|
||||
en_54:
|
||||
language: en
|
||||
speaker: en_54
|
||||
silero_speaker: v3_en
|
||||
en_55:
|
||||
language: en
|
||||
speaker: en_55
|
||||
silero_speaker: v3_en
|
||||
en_56:
|
||||
language: en
|
||||
speaker: en_56
|
||||
silero_speaker: v3_en
|
||||
en_57:
|
||||
language: en
|
||||
speaker: en_57
|
||||
silero_speaker: v3_en
|
||||
en_58:
|
||||
language: en
|
||||
speaker: en_58
|
||||
silero_speaker: v3_en
|
||||
en_59:
|
||||
language: en
|
||||
speaker: en_59
|
||||
silero_speaker: v3_en
|
||||
en_60:
|
||||
language: en
|
||||
speaker: en_60
|
||||
silero_speaker: v3_en
|
||||
en_61:
|
||||
language: en
|
||||
speaker: en_61
|
||||
silero_speaker: v3_en
|
||||
en_62:
|
||||
language: en
|
||||
speaker: en_62
|
||||
silero_speaker: v3_en
|
||||
en_63:
|
||||
language: en
|
||||
speaker: en_63
|
||||
silero_speaker: v3_en
|
||||
en_64:
|
||||
language: en
|
||||
speaker: en_64
|
||||
silero_speaker: v3_en
|
||||
en_65:
|
||||
language: en
|
||||
speaker: en_65
|
||||
silero_speaker: v3_en
|
||||
en_66:
|
||||
language: en
|
||||
speaker: en_66
|
||||
silero_speaker: v3_en
|
||||
en_67:
|
||||
language: en
|
||||
speaker: en_67
|
||||
silero_speaker: v3_en
|
||||
en_68:
|
||||
language: en
|
||||
speaker: en_68
|
||||
silero_speaker: v3_en
|
||||
en_69:
|
||||
language: en
|
||||
speaker: en_69
|
||||
silero_speaker: v3_en
|
||||
en_70:
|
||||
language: en
|
||||
speaker: en_70
|
||||
silero_speaker: v3_en
|
||||
en_71:
|
||||
language: en
|
||||
speaker: en_71
|
||||
silero_speaker: v3_en
|
||||
en_72:
|
||||
language: en
|
||||
speaker: en_72
|
||||
silero_speaker: v3_en
|
||||
en_73:
|
||||
language: en
|
||||
speaker: en_73
|
||||
silero_speaker: v3_en
|
||||
en_74:
|
||||
language: en
|
||||
speaker: en_74
|
||||
silero_speaker: v3_en
|
||||
en_75:
|
||||
language: en
|
||||
speaker: en_75
|
||||
silero_speaker: v3_en
|
||||
en_76:
|
||||
language: en
|
||||
speaker: en_76
|
||||
silero_speaker: v3_en
|
||||
en_77:
|
||||
language: en
|
||||
speaker: en_77
|
||||
silero_speaker: v3_en
|
||||
en_78:
|
||||
language: en
|
||||
speaker: en_78
|
||||
silero_speaker: v3_en
|
||||
en_79:
|
||||
language: en
|
||||
speaker: en_79
|
||||
silero_speaker: v3_en
|
||||
en_80:
|
||||
language: en
|
||||
speaker: en_80
|
||||
silero_speaker: v3_en
|
||||
en_81:
|
||||
language: en
|
||||
speaker: en_81
|
||||
silero_speaker: v3_en
|
||||
en_82:
|
||||
language: en
|
||||
speaker: en_82
|
||||
silero_speaker: v3_en
|
||||
en_83:
|
||||
language: en
|
||||
speaker: en_83
|
||||
silero_speaker: v3_en
|
||||
en_84:
|
||||
language: en
|
||||
speaker: en_84
|
||||
silero_speaker: v3_en
|
||||
en_85:
|
||||
language: en
|
||||
speaker: en_85
|
||||
silero_speaker: v3_en
|
||||
en_86:
|
||||
language: en
|
||||
speaker: en_86
|
||||
silero_speaker: v3_en
|
||||
en_87:
|
||||
language: en
|
||||
speaker: en_87
|
||||
silero_speaker: v3_en
|
||||
en_88:
|
||||
language: en
|
||||
speaker: en_88
|
||||
silero_speaker: v3_en
|
||||
en_89:
|
||||
language: en
|
||||
speaker: en_89
|
||||
silero_speaker: v3_en
|
||||
en_90:
|
||||
language: en
|
||||
speaker: en_90
|
||||
silero_speaker: v3_en
|
||||
en_91:
|
||||
language: en
|
||||
speaker: en_91
|
||||
silero_speaker: v3_en
|
||||
en_92:
|
||||
language: en
|
||||
speaker: en_92
|
||||
silero_speaker: v3_en
|
||||
en_93:
|
||||
language: en
|
||||
speaker: en_93
|
||||
silero_speaker: v3_en
|
||||
en_94:
|
||||
language: en
|
||||
speaker: en_94
|
||||
silero_speaker: v3_en
|
||||
en_95:
|
||||
language: en
|
||||
speaker: en_95
|
||||
silero_speaker: v3_en
|
||||
en_96:
|
||||
language: en
|
||||
speaker: en_96
|
||||
silero_speaker: v3_en
|
||||
en_97:
|
||||
language: en
|
||||
speaker: en_97
|
||||
silero_speaker: v3_en
|
||||
en_98:
|
||||
language: en
|
||||
speaker: en_98
|
||||
silero_speaker: v3_en
|
||||
en_99:
|
||||
language: en
|
||||
speaker: en_99
|
||||
silero_speaker: v3_en
|
||||
en_100:
|
||||
language: en
|
||||
speaker: en_100
|
||||
silero_speaker: v3_en
|
||||
en_101:
|
||||
language: en
|
||||
speaker: en_101
|
||||
silero_speaker: v3_en
|
||||
en_102:
|
||||
language: en
|
||||
speaker: en_102
|
||||
silero_speaker: v3_en
|
||||
en_103:
|
||||
language: en
|
||||
speaker: en_103
|
||||
silero_speaker: v3_en
|
||||
en_104:
|
||||
language: en
|
||||
speaker: en_104
|
||||
silero_speaker: v3_en
|
||||
en_105:
|
||||
language: en
|
||||
speaker: en_105
|
||||
silero_speaker: v3_en
|
||||
en_106:
|
||||
language: en
|
||||
speaker: en_106
|
||||
silero_speaker: v3_en
|
||||
en_107:
|
||||
language: en
|
||||
speaker: en_107
|
||||
silero_speaker: v3_en
|
||||
en_108:
|
||||
language: en
|
||||
speaker: en_108
|
||||
silero_speaker: v3_en
|
||||
en_109:
|
||||
language: en
|
||||
speaker: en_109
|
||||
silero_speaker: v3_en
|
||||
en_110:
|
||||
language: en
|
||||
speaker: en_110
|
||||
silero_speaker: v3_en
|
||||
en_111:
|
||||
language: en
|
||||
speaker: en_111
|
||||
silero_speaker: v3_en
|
||||
en_112:
|
||||
language: en
|
||||
speaker: en_112
|
||||
silero_speaker: v3_en
|
||||
en_113:
|
||||
language: en
|
||||
speaker: en_113
|
||||
silero_speaker: v3_en
|
||||
en_114:
|
||||
language: en
|
||||
speaker: en_114
|
||||
silero_speaker: v3_en
|
||||
en_115:
|
||||
language: en
|
||||
speaker: en_115
|
||||
silero_speaker: v3_en
|
||||
en_116:
|
||||
language: en
|
||||
speaker: en_116
|
||||
silero_speaker: v3_en
|
||||
en_117:
|
||||
language: en
|
||||
speaker: en_117
|
||||
silero_speaker: v3_en
|
||||
random:
|
||||
language: en
|
||||
speaker: random
|
||||
silero_speaker: v3_en
|
||||
# Russian voices (v4_ru model)
|
||||
ru_aidar:
|
||||
language: ru
|
||||
speaker: aidar
|
||||
silero_speaker: v4_ru
|
||||
ru_baya:
|
||||
language: ru
|
||||
speaker: baya
|
||||
silero_speaker: v4_ru
|
||||
ru_kseniya:
|
||||
language: ru
|
||||
speaker: kseniya
|
||||
silero_speaker: v4_ru
|
||||
ru_xenia:
|
||||
language: ru
|
||||
speaker: xenia
|
||||
silero_speaker: v4_ru
|
||||
ru_eugene:
|
||||
language: ru
|
||||
speaker: eugene
|
||||
silero_speaker: v4_ru
|
||||
ru_random:
|
||||
language: ru
|
||||
speaker: random
|
||||
silero_speaker: v4_ru
|
||||
# German voices (v3_de model)
|
||||
de_bernd_ungerer:
|
||||
language: de
|
||||
speaker: bernd_ungerer
|
||||
silero_speaker: v3_de
|
||||
de_eva_k:
|
||||
language: de
|
||||
speaker: eva_k
|
||||
silero_speaker: v3_de
|
||||
de_friedrich:
|
||||
language: de
|
||||
speaker: friedrich
|
||||
silero_speaker: v3_de
|
||||
de_hokuspokus:
|
||||
language: de
|
||||
speaker: hokuspokus
|
||||
silero_speaker: v3_de
|
||||
de_karlsson:
|
||||
language: de
|
||||
speaker: karlsson
|
||||
silero_speaker: v3_de
|
||||
de_random:
|
||||
language: de
|
||||
speaker: random
|
||||
silero_speaker: v3_de
|
||||
# Spanish voices (v1_es model)
|
||||
es_0:
|
||||
language: es
|
||||
speaker: es_0
|
||||
silero_speaker: v1_es
|
||||
es_1:
|
||||
language: es
|
||||
speaker: es_1
|
||||
silero_speaker: v1_es
|
||||
es_2:
|
||||
language: es
|
||||
speaker: es_2
|
||||
silero_speaker: v1_es
|
||||
es_random:
|
||||
language: es
|
||||
speaker: random
|
||||
silero_speaker: v1_es
|
||||
# French voices (v3_fr model)
|
||||
fr_0:
|
||||
language: fr
|
||||
speaker: fr_0
|
||||
silero_speaker: v3_fr
|
||||
fr_1:
|
||||
language: fr
|
||||
speaker: fr_1
|
||||
silero_speaker: v3_fr
|
||||
fr_2:
|
||||
language: fr
|
||||
speaker: fr_2
|
||||
silero_speaker: v3_fr
|
||||
fr_3:
|
||||
language: fr
|
||||
speaker: fr_3
|
||||
silero_speaker: v3_fr
|
||||
fr_4:
|
||||
language: fr
|
||||
speaker: fr_4
|
||||
silero_speaker: v3_fr
|
||||
fr_5:
|
||||
language: fr
|
||||
speaker: fr_5
|
||||
silero_speaker: v3_fr
|
||||
fr_random:
|
||||
language: fr
|
||||
speaker: random
|
||||
silero_speaker: v3_fr
|
||||
Loading…
Add table
Add a link
Reference in a new issue