yolov8 ultralytics库实现多机多卡DDP训练_yolo8多机多卡

参考：
https://github.com/ultralytics/ultralytics/issues/6286

ddp训练报错，问题修改：
https://blog.csdn.net/weixin_41012399/article/details/134379417

RuntimeError: CUDA error: invalid device ordinal CUDA kernel errors
might be asynchronously reported at some other API call, so the
stacktrace below might be incorrect. For debugging consider passing
CUDA_LAUNCH_BLOCKING=1. Compile with TORCH_USE_CUDA_DSA to enable
device-side assertions.

主要修改机器二上代码即可（主节点不用修改）：
1）/root/miniconda3/envs/up/lib/python3.9/site-packages/ultralytics/engine/trainer.py

from ultralytics.utils import (DEFAULT_CFG, LOGGER, RANK