[XPU] AdamW support multi_precision #61694
Merged
Conversation
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
|
你的PR提交成功，感谢你对开源项目的贡献! |
Galaxy1458
previously approved these changes
Feb 22, 2024
HarperCy
approved these changes
Feb 22, 2024
lj970926
approved these changes
Feb 22, 2024
test/xpu/test_adamw_op_xpu.py
Outdated
| def test_main(self): | ||
| xpu_version = core.get_xpu_device_version(0) | ||
| if xpu_version != core.XPUVersion.XPU3: | ||
| return |
Contributor
There was a problem hiding this comment.
这里为什么只针对KL3呀，我看KL2好像也做了支持
Contributor
Author
There was a problem hiding this comment.
因为新写的adamw_v2在XDNN里面只有KL3的实现，当时想着是以KL3为准，KL2不维护，所以这里写的只针对KL3。
另外用算子拼接实现的版本可能也支持KL2，没有特地测试过，稍后我跑一把看看情况。
test/xpu/test_adamw_op_xpu.py
Outdated
| param = paddle.randn(shape).astype(paddle.bfloat16) | ||
| master_weight = param.astype(paddle.float32) | ||
| grad = paddle.randn(shape).astype(paddle.bfloat16) | ||
| main_grad = grad.astype(paddle.bfloat16) |
Contributor
There was a problem hiding this comment.
这里的main_grad是干什么用的？我看好像和grad数据和类型都完全一致
Contributor
Author
There was a problem hiding this comment.
感谢指出，稍后我查一下，目前看是为了debug而有些修改，忘记改回去了。
chenwhql
approved these changes
Feb 23, 2024
Contributor
chenwhql
left a comment
chenwhql
left a comment
There was a problem hiding this comment.
LGTM for PADDLE_ENFORCE
Galaxy1458
approved these changes
Feb 23, 2024
QingshuChen
approved these changes
Feb 23, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
7 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Outdated suggestions cannot be applied.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 types
New features
PR changes
OPs
Description
adamw_v2函数。AdamW不支持multi_precision为true的情况，现在支持了。master_param，并转换grad的数据类型，并将结果写到master_param_outs里面。paddle/phi/kernels/gpu/adamw_kernel.cu。TestAdamWOpMultiPrecisonWithMainGrad、TestAdamWOpMultiPrecison这两个类，同步到XPU上。reduce_mean_grad和reduce_mean_grad的类型注册，追加了float16类型。