File size: 5,140 Bytes
38a17ab
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148

import os
import json
import torch
import torchaudio
from tqdm import tqdm

from src.chatterbox_.tts_turbo import ChatterboxTurboTTS
from src.chatterbox_.tts import ChatterboxTTS, punc_norm
from src.chatterbox_.models.s3tokenizer import S3_SR
from src.utils import setup_logger
from src.config import TrainConfig

logger = setup_logger(__name__)


def preprocess_dataset_json_based(config, tts_engine: ChatterboxTTS):
    
    """
    Reads metadata from JSON file, processes audio-text pairs, and saves them as .pt.
    Structure:
    - JSON contains: id, text, formatted_text, etc.
    - Audio files: {wav_dir}/{id}.wav
    """
    
    os.makedirs(config.preprocessed_dir, exist_ok=True)
    
    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
    tts_engine.ve.to(device)
    tts_engine.s3gen.to(device)
    tts_engine.ve.eval()
    tts_engine.s3gen.eval()
    

    if not os.path.exists(config.metadata_path):
        logger.error(f"ERROR: Metadata file not found: '{config.metadata_path}'!")
        return
    
    with open(config.metadata_path, 'r', encoding='utf-8') as f:
        metadata = json.load(f)
    
    if len(metadata) == 0:
        logger.error(f"ERROR: No items found in metadata file!")
        return
    
    logger.info(f"Processing dataset... Found items in JSON: {len(metadata)}")
    
    success_count = 0
    
    SPEECH_STOP_ID = getattr(tts_engine.t3.hp, 'stop_speech_token', 6562)
    for item in tqdm(metadata, desc="Preprocessing"):
        try:

            file_id = item.get("id")
            raw_text = item.get("text", "")
            
            if not file_id or not raw_text:
                logger.warning(f"Skipping item with missing id or text")
                continue
            

            wav_path = os.path.join(config.wav_dir, f"{file_id}.wav")
            
            if not os.path.exists(wav_path):
                logger.warning(f"Audio file not found, skipping: {file_id}")
                continue
            
            wav, sr = torchaudio.load(wav_path)
            
            if wav.shape[0] > 1:
                wav = wav.mean(dim=0, keepdim=True)
            
            if sr != S3_SR:
                resampler = torchaudio.transforms.Resample(sr, S3_SR)
                wav = resampler(wav)
            
            wav = wav.to(device)
            
            with torch.no_grad():

                wav_np = wav.cpu().squeeze().numpy()
                spk_emb_np = tts_engine.ve.embeds_from_wavs([wav_np], sample_rate=S3_SR)
                speaker_emb = torch.from_numpy(spk_emb_np[0]).cpu()
                
                s_tokens, _ = tts_engine.s3gen.tokenizer(wav.unsqueeze(0))
                raw_speech_tokens = s_tokens.squeeze().cpu()
                
                stop_speech_tensor = torch.tensor([SPEECH_STOP_ID], dtype=raw_speech_tokens.dtype)
                speech_tokens = torch.cat([raw_speech_tokens, stop_speech_tensor], dim=0)
                
                
                prompt_samples = int(config.prompt_duration * S3_SR)
                if wav.shape[1] < prompt_samples:
                    prompt_wav = torch.nn.functional.pad(wav, (0, prompt_samples - wav.shape[1]))
                else:
                    prompt_wav = wav[:, :prompt_samples]
                
                p_tokens, _ = tts_engine.s3gen.tokenizer(prompt_wav.unsqueeze(0))
                prompt_tokens = p_tokens.squeeze().cpu()
            
            clean_text = punc_norm(raw_text)
            
            if config.is_turbo:
                token_output = tts_engine.tokenizer(clean_text, return_tensors="pt")
                raw_text_tokens = token_output.input_ids[0].cpu()
                
                if tts_engine.tokenizer.eos_token_id is not None:
                    text_eos = torch.tensor([tts_engine.tokenizer.eos_token_id], dtype=raw_text_tokens.dtype)
                    text_tokens = torch.cat([raw_text_tokens, text_eos], dim=0)
                else:
                    text_tokens = raw_text_tokens
            
            else:
                text_tokens = tts_engine.tokenizer.text_to_tokens(clean_text).squeeze(0).cpu()
            
            save_path = os.path.join(config.preprocessed_dir, f"{file_id}.pt")
            
            torch.save({
                "speech_tokens": speech_tokens,
                "speaker_emb": speaker_emb,
                "prompt_tokens": prompt_tokens,
                "text_tokens": text_tokens,
            }, save_path)
            
            success_count += 1

            
        except Exception as e:
            logger.error(f"Error ({item.get('id', 'unknown')}): {e}")
            continue
    
    logger.info(f"Preprocessing completed! Success: {success_count}/{len(metadata)}")
    
    

if __name__ == "__main__":

    cfg = TrainConfig()
    
    if cfg.is_turbo:
        EngineClass = ChatterboxTurboTTS
    else:
        EngineClass = ChatterboxTTS
    
    logger.info(f"{EngineClass} engine starting...")
    tts_engine = EngineClass.from_local(cfg.model_dir, device="cpu")
    
    preprocess_dataset_json_based(cfg, tts_engine)