RuntimeError: The shape of the mask [1, 9476] at index 1 does not match the shape of the indexed tensor [1, 9477, 3584] at index 1
![RuntimeError: The shape of the mask [1, 9476] at index 1 does not match the shape of the indexed tensor [1, 9477, 3584] at index 1](https://www.chat-gpts.plus/wp-content/uploads/2026/09/48835-104181fb-768x403.jpg)
这个报错通常出现在用 VibeVoice-ASR 转写较长音频(24 kHz 采样下超过约 11.65 分钟、即样本数超过 2**24)时,processor 与 model 对音频占位 token 数量的浮点计算精度不一致,导致 mask 与特征张量形状对不上;优先排查输入音频长度是否是 hop(








