script: accelerate.py function: clip_grad_norm_ if parameters == [p for p in model.parameters()]: RuntimeError: Boolean value of Tensor with more than one value is ambiguous

[Xingyuxingyu]

System Info

accelerate 1.10.1
os: Linux #25-Ubuntu SMP Wed Mar 30 15:54:22 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux
python: 3.10
numpy: 1.26.0
torch version: 2.5.1+cu118

Information

• The official example scripts
• My own modified scripts

Tasks

• One of the scripts in the examples/ folder of Accelerate or an officially supported no_trainer script in the examples folder of the transformers repo (such as run_no_trainer_glue.py)
• My own task or dataset (give details below)

Reproduction

accelerate fsdp model
if accelerator.sync_gradients and train_config.max_grad_norm > 0:
all_norm = accelerator.clip_grad_norm_(model.parameters(), train_config.max_grad_norm)

Expected behavior

No RuntimeError

[S1ro1]

Hi, can you check the PR #3791 if that fixes it for you?

[SunMarc]

Hey @Xingyuxingyu, please share a bettter reproducer !

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.

[Xingyuxingyu]

这是来自QQ邮箱的假期自动回复邮件。您好，已收到您的邮件。我会尽快给您回复。

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.

[Xingyuxingyu]

这是来自QQ邮箱的假期自动回复邮件。您好，已收到您的邮件。我会尽快给您回复。