huggingface accelerate出现Pytorch: RuntimeError: Expected to have finished reduction in the prior iter

最新推荐文章于 2024-11-07 21:32:28 发布

铁拳斗士

最新推荐文章于 2024-11-07 21:32:28 发布

阅读量2.3k

点赞数

文章标签： python bert

本文链接：https://blog.csdn.net/weixin_45465323/article/details/124225797

版权

在使用accelerate进行多GPU训练时遇到RuntimeError，原因是模型中有未参与计算loss的参数。通过检查发现BertModel的pooler层没有接收到梯度。解决方法是检查loss函数，确保所有参数都被使用，或者选择冻结未使用的参数。

摘要由CSDN通过智能技术生成

单卡没有问题，在使用accelerate多卡训练时出现

RuntimeError: Expected to have finished reduction in the prior iteration before starting a new one. This error indicates that your module has parameters that were not used in producing loss. You can enable unused parameter detection by (1) passing the keyword argument `find_unused_parameters=True` to `torch.nn.parallel.DistributedDataParallel`; (2) making sure all `forward` function outputs participate in calculating loss. If you already have done the above two steps, then the distributed data parallel module wasn't able to locate the output tensors in the return value of your module's `forward` function. Please include the loss function and the structure of the return value of `forward` of your module when reporting this issue (e.g. list, dict, iterable).

然后参考文章的解决方法，在梯度回传之后，参数更新之前，插入了代码

for name, param in model.named_parameters():
    if param.grad is None:
        print(name)

发现是Bert Model的pooler没有接收到梯度。这是因为我的loss没有使用Bert输出的loss。
因此，冻结对应参数即可