pytorch system_error ai_generated partial

运行时错误：DataLoader工作进程（pid 12345）收到信号11（段错误）。可能原因：共享内存耗尽或数据集内存损坏

RuntimeError: DataLoader worker (pid 12345) received signal 11 (Segmentation fault). Possible causes: shared memory exhausted or dataset memory corruption

ID: pytorch/dataloader-worker-segfault

其他格式: JSON · Markdown 中文 · English
85%修复率
82%置信度
1证据数
2023-06-20首次发现

版本兼容性

版本状态引入弃用备注
pytorch>=1.13 active — — —
torchvision>=0.14 active — — —
cuda>=11.7 active — — —
python>=3.8 active — — —

根因分析

DataLoader工作进程因段错误崩溃，通常由于共享内存不足（Linux上/dev/shm太小）或数据集存在内存损坏错误（例如使用不可pickle对象的不安全多进程）。

English

A DataLoader worker process crashed with a segmentation fault, typically due to insufficient shared memory (on Linux, /dev/shm is too small) or a buggy dataset that corrupts memory (e.g., unsafe multiprocessing with non-picklable objects).

generic

官方文档

https://pytorch.org/docs/stable/data.html#multi-process-data-loading

解决方案

  1. Increase shared memory size: In Docker, add --shm-size=8g; on bare metal, remount /dev/shm with larger size: sudo mount -o remount,size=8G /dev/shm
  2. Reduce num_workers to 0 or 1 to avoid multiprocessing issues: DataLoader(..., num_workers=0)
  3. Switch multiprocessing context to 'spawn' (instead of 'fork'): torch.multiprocessing.set_start_method('spawn', force=True)

无效尝试

常见但无效的做法:

  1. Increasing num_workers to speed up data loading 90% 失败

    More workers consume more shared memory, making the segfault more likely.

  2. Using pin_memory=True without increasing shared memory 85% 失败

    Pin memory also uses shared memory, exacerbating the exhaustion.