fix torch_dtype, contiguous, and load_state_dict regression #36512
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
Contributor
|
Hi 👋, thank you for opening this pull request! The pull request is converted to draft by default. When it is ready for review, please click the |
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
…ssion' into fix-dtype-and-contiguous-regression
SunMarc
changed the title
SunMarc
commented
Mar 3, 2025
| folder = None | ||
|
|
||
| model.expected_keys = expected_keys | ||
| model_to_load.expected_keys = expected_keys |
Member
Author
There was a problem hiding this comment.
since we are calling _fix_state_dict_keys_on_load on model_to_load
garrett361
pushed a commit
to garrett361/transformers
that referenced
this pull request
Mar 4, 2025
…ace#36512) * fix regression * fix param * fix load_state_dict * style * better fix for module * fix tests * quick fix for now * rm print
garrett361
pushed a commit
to garrett361/transformers
that referenced
this pull request
Mar 4, 2025
…ace#36512) * fix regression * fix param * fix load_state_dict * style * better fix for module * fix tests * quick fix for now * rm print
fxmarty-amd
reviewed
Mar 10, 2025
| distribute_module(module_to_tp, device_mesh, None, input_fn, output_fn) | ||
| else: | ||
| module_to_tp.load_state_dict({param_type: param[:]}, strict=False, assign=True) | ||
| param = param[:] |
Contributor
There was a problem hiding this comment.
@SunMarc why is param = param[:] needed?
edit - ok, this is for safetensors. Unfortuantely safetensors get_slice does not play well with 0-dim tensors :( huggingface/safetensors#380
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.
What does this PR do ?
This PR fixes the regression we have from #36335:
load_state_dictmap_location default to "cpu" + couples of small fixes -> fixes CI !get_submodule_get_module_from_nametoget_module_nameis it also handles the case where the param name reference to a direct param from the model.Should fixes tests in peft and in the quantization CI
Perf with TP 2 GPUS: