#1469 ValueError: Requested int8_float16 compute type, but the target device or backend do not support efficient int8_float16

167.253* Posted at: 3 hours ago 👁6

语音识别阶段出错:[faster-whisper(本地)] 运行时错误：Traceback (most recent call last):
File "videotrans\process\_overall.py", line 45, in run
File "faster_whisper\transcribe.py", line 675, in init
ValueError: Requested int8_float16 compute type, but the target device or backend do not support efficient int8_float16 computation.
:
Traceback (most recent call last):
File "videotrans\task\job.py", line 107, in run
File "videotrans\task\trans_create.py", line 371, in recogn
File "videotrans\recognition\__init__.py", line 242, in run
File "videotrans\recognition\_base.py", line 78, in run
File "videotrans\recognition\_overall.py", line 199, in _exec
RuntimeError: Traceback (most recent call last):
File "videotrans\process\_overall.py", line 45, in run
File "faster_whisper\transcribe.py", line 675, in init
ValueError: Requested int8_float16 compute type, but the target device or backend do not support efficient int8_float16 computation.

=
system:Windows-10-10.0.26100-SP0
version:v3.84
frozen:True
language:zh
root_dir:G:/AI/win-videotrans-v3.84

1 Replies

Floor #1 DeepSeek Posted at 3 hours ago

这个错误是由于在 faster-whisper 语音识别时，您选择的 CUDA数据类型 为 int8_float16，但您的显卡或CUDA环境不支持这种计算类型。

解决方案：

  1. 在软件顶部菜单栏，选择 工具/选项 -> 高级选项。
  2. 找到 faster/openai语音识别调整 区域下的 CUDA数据类型 选项。

  3. 将其从 int8_float16 更改为 float16 或 int8。

    • float16：精度更高，但消耗的显存也更多。
    • int8：精度稍低，但消耗的显存最少，速度最快。
  4. 点击确定保存设置，然后重新尝试语音识别任务。

如果更改后仍报错，请尝试将 CUDA数据类型 设置为 float32（兼容性最好，但速度最慢）。如果所有 CUDA 选项都失败，请取消勾选主界面的 CUDA加速 复选框，改为使用 CPU 进行计算（速度会显著变慢）。

Post Your Reply
Open source and free maintenance is not easy. If this project is helpful to you, please consider making a small donation to help the project continue to maintain and update.

Donate: https://ko-fi.com/jianchang512

Trending Questions