4th-batch-50-分布式训练初始化一致性问题#75784
Merged
luotao1 merged 1 commit intoPaddlePaddle:developfrom Oct 14, 2025
Merged
Conversation
|
你的PR提交成功，感谢你对开源项目的贡献! |
luotao1
approved these changes
Oct 14, 2025
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
PR Category
Execute Infrastructure
PR Types
Bug fixes
Description
第四批-编号50（共1个）
代码虽然设置了固定的
seed，但在多进程环境下，若np.random.seed(seed)没有被全局统一且原子地设置，不同进程可能因执行顺序、延迟或其他干扰导致实际使用的随机状态不一致。此外，w是在每个进程中独立生成的，尽管种子相同，但如果底层 NumPy 实现受环境影响（如多线程干扰、动态加载等），仍可能产生差异。最关键的是，正确的做法应是在一个进程中生成完整权重后广播或切分，而不是每个进程都独立采样。修复方案确保权重矩阵
w只在rank == 0的进程中生成，并通过paddle.distributed.broadcast将其广播至所有其他进程，避免各进程独立调用np.random.normal导致潜在的不一致性。这样可以保证所有设备上的初始权重切分基于完全相同的原始矩阵，满足分布式训练对初始化一致性的要求。