Skip to content

在训练llava模型时。将deepspeed的stage从2改为3,出现报错 AttributeError: 'LlavaMultiModalProjector' object has no attribute 'ds_grads_remaining' #213

@jjjjjjj2020

Description

@jjjjjjj2020

在训练llava模型时。将deepspeed的stage从2改为3,出现报错 AttributeError: 'LlavaMultiModalProjector' object has no attribute 'ds_grads_remaining',这是为什么呢?博主的自定义的llava模型不支持stage3吗?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions