#4326 onnxruntime.capi.onnxruntime_pybind11_state.RuntimeException: [ONNXRuntimeError] : 6 : RUNTIME_EXCEPTION : Non-zero stat

2408:8256* Posted at: 2 hours ago 👁11

multiprocessing.pool.RemoteTraceback:
"""
Traceback (most recent call last):
File "multiprocessing\pool.py", line 125, in worker
File "videotrans\task\vad.py", line 45, in get_speech_timestamp_silero
File "faster_whisper\vad.py", line 98, in get_speech_timestamps
File "faster_whisper\vad.py", line 377, in call
File "onnxruntime\capi\onnxruntime_inference_collection.py", line 287, in run

self._validate_input(list(input_dict_ort_values.keys()))

onnxruntime.capi.onnxruntime_pybind11_state.RuntimeException: [ONNXRuntimeError] : 6 : RUNTIME_EXCEPTION : Non-zero status code returned while running LSTM node. Name:'/decoder/rnn/LSTM' Status Message: bad allocation
"""

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
File "videotrans\task\_only_one.py", line 44, in run
File "videotrans\task\trans_create.py", line 380, in recogn
File "videotrans\recognition\__init__.py", line 293, in run
File "videotrans\recognition\_base.py", line 138, in run
File "videotrans\recognition\_overall.py", line 46, in _download
File "videotrans\recognition\_base.py", line 121, in _vad_split
File "videotrans\configure\_base.py", line 285, in _new_process
File "videotrans\process\signelobj.py", line 13, in result
File "multiprocessing\pool.py", line 774, in get
onnxruntime.capi.onnxruntime_pybind11_state.RuntimeException: [ONNXRuntimeError] : 6 : RUNTIME_EXCEPTION : Non-zero status code returned while running LSTM node. Name:'/decoder/rnn/LSTM' Status Message: bad allocation


=
system:Windows-10-10.0.19045-SP0
version:v3.99
frozen:True
language:zh
root_dir:E:/win-pyvideotrans-v3.99-420

1 Replies

问题源于 Silero VAD 模型运行时,ONNX Runtime 内存分配失败 (bad allocation)。这通常意味着系统可用内存或显存资源不足。

可尝试以下步骤解决:

  1. 尝试切换 VAD 模型:打开 菜单 - 工具/选项 - 高级选项,在【语音识别参数】区域找到 选择VAD 选项,将其从 silero 切换为 ten-vad
  2. 降低任务并行度:在主界面关闭 二次识别 功能,避免多任务同时消耗大量资源。
  3. 调整识别参数以减少资源占用:在【语音识别参数】区域进行如下修改:

    • 识别准确度beam_size:改为 1
    • 识别准确度best_of:改为 1
    • CUDA数据类型:尝试更换为 float16int8;如果你正在使用 int8,则尝试切换回 defaultfloat16
  4. 启用音频预分割:在【语音识别参数】中,勾选 Whisper预分割音频,有助于节省处理长音频时的资源。

如果以上方法均无效,则可能是模型文件下载不完整,可尝试手动下载,方法参考下方文档。

请查阅相关文档:

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.

Related resource