ai-agents-bark-processor-devic-when-using-barkprocessor-with-a-voice-preset-the-h-6f32cd95

THEN Manually move the voice_preset tensors to the target device before calling the processor, then move the entire inputs dict to the device. For example, after constructing the voice_preset dict from an npz file, set inputs['history_prompt'] = inputs['history_prompt'].to(device) before inputs.to(device).

import numpy as np
with np.load('bark/assets/prompts/en_speaker_6.npz') as data:
    voice_preset = {
        'fine_prompt': data['fine_prompt'],
        'coarse_prompt': data['coarse_prompt'],
        'semantic_prompt': data['semantic_prompt']
    }
inputs = processor(text_prompt, voice_preset=voice_preset)
inputs['history_prompt'] = inputs['history_prompt'].to(device)
inputs.to(device)