Add Compressedbackend for Onebit optimizers #5473
Open
Conversation
This file contains 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
Liangliang-Ma
requested review from
conglongli,
awan-10,
mrwyattii,
tjruwase,
loadams and
arashb
as code owners
delock
reviewed
May 4, 2024
csrc/xpu/packbits/packing.cpp
Outdated
|
|
||
| at::Tensor packbits(at::Tensor tensor, int input_size, int rank) | ||
| { | ||
| /* |
There was a problem hiding this comment.
@Liangliang-Ma the function documentation needs to be moved to line 39 right before the function def line.
|
@tjruwase this PR is an approach to abstract the generic part of 1bit-adam and implment accelerator dependent part with DeepSpeed custom op builder. So 1bit-adam does not need to depend on accelerator specific libraries. @inkcherry I remember you investigated in 1bit adam portability before, FYI this PR implement a portable version of 1bit adam support. |
25 tasks
|
Hi @tjruwase , could you please help to review this PR? Thanks! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
In the process of adding onebit optimizers support for XPU devices, we have noticed that for different accelerator, the main difference of implementation of
compressed_allreducelies onpackbitsandunpackbits. CUDA uses cupy and NPU uses torch_npu. Instead of replace these to xpu only functions, we provided a CompressedBackend to do thecompressed_allreducework where users can add their own packbits/unpackbits kernels, which is a general path for all kinds of accelerators.In this PR, we: