support save load optimizer master_weights #60027
Merged
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
… flatten_and_dedup_for_save_load
|
你的PR提交成功，感谢你对开源项目的贡献! |
… flatten_and_dedup_for_save_load
pangengzheng
changed the title
… flatten_and_dedup_for_save_load
zhiqiu
reviewed
Dec 26, 2023
| @@ -61,5 +61,47 @@ def compute_local_shape_and_global_offset( | |||
|
|
|||
|
|
|||
| def flatten_state_dict(state_dict): | |||
There was a problem hiding this comment.
Does it support multiple level, i.e., {'model': {'m': {'w': xxx}}} ?
| ) | ||
|
|
||
| tensor.set(load_para_np, framework._current_expected_place()) | ||
| var.set_value(state_dict[var_tmp.name]) |
There was a problem hiding this comment.
代码复用，set_value api的行为包含这部分删掉的代码逻辑，且支持设置distributed tensor赋值
| @@ -74,6 +73,16 @@ def dedup_storage_metadata(global_storage_metadata): | |||
| return out | |||
|
|
|||
|
|
|||
| def dedup_tensor(state_dict, local_storage_metadata, dedup_storage_metadata): | |||
There was a problem hiding this comment.
ok，不过这个方法可以算是此文件的私有方法，不对外。comment已加
Wanglongzhi2001
pushed a commit
to Wanglongzhi2001/Paddle
that referenced
this pull request
Jan 7, 2024
* exclude xpu * dedup tensor in state_dict * polish * support flatten and unflatten state_dict * test flatten * rename test * fix dedup tensor test * fix test * fix load state dict * rename * fix test * support save load optimizer master weights * add comment
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Outdated suggestions cannot be applied.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.
PR types
Others
PR changes
Others
Description
card-78318
support flatten state_dict, save load optimizer master_weights and deduplicate tensor when save state_dict