Skip to content

[Op] Implement GPU version of KvResourceSparseApplyAdam. - #535

Open
fuhailin wants to merge 1 commit into
DeepRec-AI:mainfrom
fuhailin:adam_gpu
Open

[Op] Implement GPU version of KvResourceSparseApplyAdam.#535
fuhailin wants to merge 1 commit into
DeepRec-AI:mainfrom
fuhailin:adam_gpu

Conversation

@fuhailin

@fuhailin fuhailin commented Nov 9, 2022

Copy link
Copy Markdown
Collaborator

No description provided.


const Device& device = ctx->eigen_device<Device>();
OP_REQUIRES_OK(ctx,
functor::KvSparseApplyAdamAsync<Device, T, Tindex, Tstep>()(

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

这里直接调用AdamAsync的functor吗?

Copy link
Copy Markdown
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

是的,我注意到AdamAsync的functor中逻辑代码与Adam相同,只是AdamAsync多了apply_sparse_rmsprop参数,于是考虑复用functor逻辑代码,不同在于将Adam 中调用的functor 中apply_sparse_rmsprop 置为了false

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

嗯,我建议新写一个functor::KvSparseApplyAdam,而不是复用这个,避免别人改AdamAsync的functor而影响到这里的逻辑正确性

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

2 participants