respark / epoch9 /BATCH_INFERENCE_README.md
yueyulin's picture
Upload folder using huggingface_hub
03fc8c0 verified

ζ‰Ήι‡ζŽ¨η†εŠŸθƒ½θ―΄ζ˜Ž

ζœ¬ζ–‡ζ‘£δ»‹η»δΊ† ReSpark TTS ζ¨‘εž‹ηš„ζ‰Ήι‡ζŽ¨η†εŠŸθƒ½οΌŒθ―₯εŠŸθƒ½ε―δ»₯ζ˜Ύθ‘—ζι«˜ε€šδΈͺζ–‡ζœ¬ηš„θ―­ιŸ³εˆζˆζ•ˆηŽ‡γ€‚

使用方法

εŸΊζœ¬ζ‰Ήι‡ζŽ¨η†

from utilities import generate_embeddings_batch
from tts_batch_infer import generate_speech_batch

# ε‡†ε€‡ζ–‡ζœ¬εˆ—θ‘¨
texts = [
    "第一δΈͺθ¦εˆζˆηš„ζ–‡ζœ¬γ€‚",
    "第二δΈͺθ¦εˆζˆηš„ζ–‡ζœ¬γ€‚",
    "第三δΈͺθ¦εˆζˆηš„ζ–‡ζœ¬γ€‚"
]

# ζ‰Ήι‡η”Ÿζˆθ―­ιŸ³
wavs = generate_speech_batch(
    model, tokenizer, texts, audio_tokenizer,
    prompt_text="ζη€Ίζ–‡ζœ¬",
    prompt_audio=prompt_audio,
    device=device
)

# δΏε­˜ιŸ³ι’‘ζ–‡δ»Ά
for i, wav in enumerate(wavs):
    sf.write(f'output_{i}.wav', wav, sample_rate)