HF2UCP: Converting a pytorch_model.bin or .safetensors checkpoint to UCP
#7212
+304
−21
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
Signed-off-by: Schwidola0607 <[email protected]>
Signed-off-by: Schwidola0607 <[email protected]>
Signed-off-by: Schwidola0607 <[email protected]>
Signed-off-by: Schwidola0607 <[email protected]>
Signed-off-by: Schwidola0607 <[email protected]>
Schwidola0607
requested review from
tjruwase,
tohtana and
loadams
as code owners
Signed-off-by: Schwidola0607 <[email protected]>
Signed-off-by: Schwidola0607 <[email protected]>
Signed-off-by: Schwidola0607 <[email protected]>
Signed-off-by: Schwidola0607 <[email protected]>
Signed-off-by: Schwidola0607 <[email protected]>
xylian86
reviewed
Jun 23, 2025
| @@ -0,0 +1,36 @@ | |||
| --- | |||
There was a problem hiding this comment.
How about moving this document to the Megatron-DeepSpeed repository?
The purpose of converting the Hugging Face checkpoint to UCP is to enable loading with 4D parallelism, so the Megatron-DeepSpeed repo might be a more suitable place for it.
| optim_sd = torch.load(optim_state_path, weights_only=False) | ||
|
|
||
| self._load_global_state(optim_sd) | ||
| if os.path.isfile(optim_state_path): |
There was a problem hiding this comment.
Suggestion:
Update the logic as follows:
if universal_checkpoint:
# Allow missing optimizer state path, but issue a warning
if optim_state_path is None:
warnings
else:
assert os.path.isfile()
| optim_sd = torch.load(optim_state_path, weights_only=False) | ||
| self._load_global_state_stage3(optim_sd) | ||
| if os.path.isfile(optim_state_path): | ||
| ignore_missing_optim_state = False |
There was a problem hiding this comment.
Suggestion:
Update the logic as follows:
if universal_checkpoint:
# Allow missing optimizer state path, but issue a warning
if optim_state_path is None:
warnings
else:
assert os.path.isfile()
| @@ -2947,7 +2947,7 @@ def load_checkpoint(self, | |||
|
|
|||
| Returns: | |||
| A tuple of ``load_path`` and ``client_state``. | |||
| *``load_path``: Path of the loaded checkpoint. ``None`` if loading the checkpoint failed. | |||
| *``load_path``: Path of the loaded checkpoint. ``None`` if loading the checkpoint failed or loading a HF based UCP | |||
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 for HF2UCP feature
Converting a pytorch_model.bin or .safetensors checkpoint to UCP will