[Question]: 多卡时运行run_pretrain.py存在报错，单卡时可以正常运行

[Buddingpopp]

请提出你的问题

单卡时训练可以正常完成，在相同配置下，单机双GPU运行 python -u -m paddle.distributed.launch --gpus "0,1" run_pretrain.py ./config/qwen/pretrain_argument_0p5b.json
会有如下报错：

OSError: [Errno 101] Network is unreachable

urllib3.exceptions.NewConnectionError: HTTPSConnection(host='bj.bcebos.com', port=443): Failed to establish a new connection: [Errno 101] Network is unreachable

urllib3.exceptions.MaxRetryError: HTTPSConnectionPool(host='bj.bcebos.com', port=443): Max retries exceeded with url: /paddlenlp/models/community/Qwen/Qwen2.5-0.5B/added_tokens.json (Caused by NewConnectionError("HTTPSConnection(host='bj.bcebos.com', port=443): Failed to establish a new connection: [Errno 101] Network is unreachable"))

requests.exceptions.ConnectionError: HTTPSConnectionPool(host='bj.bcebos.com', port=443): Max retries exceeded with url: /paddlenlp/models/community/Qwen/Qwen2.5-0.5B/added_tokens.json (Caused by NewConnectionError("HTTPSConnection(host='bj.bcebos.com', port=443): Failed to establish a new connection: [Errno 101] Network is unreachable"))

OSError: Can't load the model for 'Qwen/Qwen2.5-0.5B'. If you were trying to load it from 'BOS', make sure you don't have a local directory with the same name. Otherwise, make sure 'Qwen/Qwen2.5-0.5B' is the correct path to a directory containing one of the ['added_tokens.json']

requests.exceptions.ConnectionError: HTTPSConnectionPool(host='bj.bcebos.com', port=443): Max retries exceeded with url: /paddlenlp/models/community/Qwen/Qwen2.5-0.5B/added_tokens.json (Caused by NewConnectionError("HTTPSConnection(host='bj.bcebos.com', port=443): Failed to establish a new connection: [Errno 101] Network is unreachable"))

OSError: Can't load the model for 'Qwen/Qwen2.5-0.5B'. If you were trying to load it from 'BOS', make sure you don't have a local directory with the same name. Otherwise, make sure 'Qwen/Qwen2.5-0.5B' is the correct path to a directory containing one of the ['added_tokens.json']

[nepeplwu]

@Buddingpopp 感谢反馈，后续PaddleNLP中的大模型训练相关功能迁移到PaddleFormers支持，如果有大模型训练相关问题，请移步：https://github.com/PaddlePaddle/PaddleFormers/issues 提问

[github-actions]

This issue is stale because it has been open for 60 days with no activity. 当前issue 60天内无活动，被标记为stale。