v2 demo_part3 error

[zhjygit]

Speed is adjustable speed = 1.0

for language, text in texts.items(): model = TTS(language=language, device=device) speaker_ids = model.hps.data.spk2id

for speaker_key in speaker_ids.keys(): speaker_id = speaker_ids[speaker_key] speaker_key = speakerkey.lower().replace('', '-')

source_se = torch.load(f'checkpoints_v2/base_speakers/ses/{speaker_key}.pth', map_location=device)
model.tts_to_file(text, speaker_id, src_path, speed=speed)
save_path = f'{output_dir}/output_v2_{speaker_key}.wav'

# Run the tone color converter
encode_message = "@MyShell"
tone_color_converter.convert(
    audio_src_path=src_path, 
    src_se=source_se, 
    tgt_se=target_se, 
    output_path=save_path,
    message=encode_message)

[Zengyi-Qin]

It's the network error.

[zhjygit]

However, i can visit the file via web browser, just like visit google.com.

[zhjygit]

已经解决了。通过降级urllibs和requests包即可。后面还是会下载1-2个G的模型包，主要是语言类的，最好还是有vpn支持。