ℹ️ Unify autocast behavior to torch.autocast and make it cover XPU
#3541
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
device-agnostic to cover xpu Signed-off-by: YAO Matrix <matrix.yao@intel.com>
yao-matrix
commented
Jun 5, 2025
|
|
||
| ```python | ||
| training_args = DPOConfig(..., optimize_cuda_cache=True) | ||
| training_args = DPOConfig(..., optimize_device_cache=True) |
There was a problem hiding this comment.
there is no optimize_cuda_cache anymore, so update the doc here
| @@ -82,7 +82,7 @@ class ScriptArguments: | |||
| batch_size=script_args.batch_size, | |||
| mini_batch_size=script_args.mini_batch_size, | |||
| gradient_accumulation_steps=script_args.gradient_accumulation_steps, | |||
| optimize_cuda_cache=True, | |||
| optimize_device_cache=True, | |||
| torch.cuda.empty_cache() | ||
| elif torch_device == "xpu": | ||
| torch.xpu.empty_cache() | ||
| backend_empty_cache(torch_device) |
There was a problem hiding this comment.
use the device-agnostic utility from transformers.testing_utils rather than if-else
| @unittest.skipIf( | ||
| get_device_properties()[0] == "cuda" and get_device_properties()[1] < 8, | ||
| "Skipping because bf16 not supported on CUDA GPU with capability < 8.0", | ||
| ) |
There was a problem hiding this comment.
add skipIf per the comments and remove condition-less skip
| if is_torch_xpu_available(): | ||
| return f"xpu:{state.local_process_index}" | ||
| if torch.cuda.is_available() or is_torch_xpu_available(): | ||
| return state.local_process_index | ||
| elif is_torch_npu_available(): |
There was a problem hiding this comment.
we don't need this WA anymore, xpu now support integer device index
kashif
approved these changes
Jun 9, 2025
qgallouedec
reviewed
Jun 9, 2025
qgallouedec
reviewed
Jun 9, 2025
qgallouedec
approved these changes
Jun 9, 2025
Co-authored-by: Quentin Gallouédec <45557362+qgallouedec@users.noreply.github.com>
Co-authored-by: Quentin Gallouédec <45557362+qgallouedec@users.noreply.github.com>
|
@qgallouedec is the test failing due to the CI issue? |
|
Yes, fixing it in #3551 |
|
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. |
torch.autocast and make it cover XPUtorch.autocast and make it cover XPU
|
Let's wait for #3553 to be merged |
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.
@kashif , pls help review and comment, thx very much.