#7196 :{'text': '你好啊，我亲爱的朋友，希望你的每一天都是美好愉快的！', 'rate': '+0%', 'role': 'en_female_2.wav', 'filename': 'D:/win-pyvideotrans-v4.12

The underlying cause is that the ONNX model file for the MOSS-TTS-Nano (tts_type 5) dubbing channel was not downloaded or is incomplete. The error is explicit:

Load model ...\models\MOSS-TTS-Nano-100M-ONNX\moss_tts_prefill.onnx failed: File doesn't exist

This is a model download failure, not a code or configuration problem.

Solution: manually download the required models

1. Download the MOSS-TTS-Nano voice model from:
   https://huggingface.co/OpenMOSS-Team/MOSS-TTS-Nano-100M/tree/main
   Place all files into:
   软件目录/models/MOSS-TTS-Nano-100M

2. Download the MOSS audio tokenizer (the ONNX one your error is complaining about) from:
   https://huggingface.co/OpenMOSS-Team/MOSS-Audio-Tokenizer-Nano-ONNX/tree/main
   Place all files into:
   软件目录/models/MOSS-TTS-Nano-100M-ONNX

   Note the exact missing file is moss_tts_prefill.onnx inside that folder — make sure it exists there after downloading.

3. If you are in mainland China, huggingface.co is blocked. Either enable a stable VPN/proxy, or use the mirror https://hf-mirror.com with the same paths.
4. After copying, restart the software and retry the dubbing. If it still fails, delete the models/MOSS-TTS-Nano-100M-ONNX folder entirely and re-download (partial downloads are common and produce exactly this error).

Tips: make sure filenames are unchanged (browsers may rename duplicates to xxx(2).onnx). Each model folder here is large; a flaky connection frequently results in missing files.

请查阅相关文档：

• 手动下载模型