RuntimeError: Expected to mark a variable ready only once. This error is caused by one of the following reasons: 1) Use of a module parameter outside the `forward` function. Please make sure model paramet
ers are not shared across multiple concurrent forward-backward passes. or try to use _set_static_graph() as a workaround if this module graph does not change during training loop.2) Reused parameters in m
ultiple reentrant backward passes. For example, if you use multiple `checkpoint` functions to wrap the same part of your model, it would result in the same set of parameters been used by different reentra
nt backward passes multiple times, and hence marking a variable ready multiple times. DDP does not support such use cases in default. You can try to use _set_static_graph() as a workaround if your module
graph does not change over iterations.
Parameter at index 166 with name model.layers.27.mlp_norm.weight has been marked as ready twice. This means that multiple autograd engine  hooks have fired for this particular parameter during this iterat
ion.
[rank0]: Traceback (most recent call last):
[rank0]:   File "/workspace/scoring_train_modernbert_custom.py", line 355, in <module>
[rank0]:     train_result = trainer.train()
[rank0]:                    ^^^^^^^^^^^^^^^
[rank0]:   File "/workspace/pi_scorer_trainer.py", line 2507, in train
[rank0]:     return inner_training_loop(
[rank0]:            ^^^^^^^^^^^^^^^^^^^^
[rank0]:   File "/workspace/pi_scorer_trainer.py", line 2955, in _inner_training_loop
[rank0]:     tr_loss_step = self.training_step(
[rank0]:                    ^^^^^^^^^^^^^^^^^^^
[rank0]:   File "/workspace/pi_scorer_trainer.py", line 4392, in training_step
[rank0]:     self.accelerator.backward(loss, **kwargs)
[rank0]:   File "/usr/local/lib/python3.11/dist-packages/accelerate/accelerator.py", line 2246, in backward
[rank0]:     loss.backward(**kwargs)
[rank0]:   File "/usr/local/lib/python3.11/dist-packages/torch/_tensor.py", line 521, in backward
[rank0]:     torch.autograd.backward(
[rank0]:   File "/usr/local/lib/python3.11/dist-packages/torch/autograd/__init__.py", line 289, in backward
[rank0]:     _engine_run_backward(
[rank0]:   File "/usr/local/lib/python3.11/dist-packages/torch/autograd/graph.py", line 769, in _engine_run_backward
[rank0]:     return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[rank0]:            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[rank0]:   File "/usr/local/lib/python3.11/dist-packages/torch/autograd/function.py", line 306, in apply
[rank0]:     return user_fn(self, *args)
[rank0]:            ^^^^^^^^^^^^^^^^^^^^
[rank0]:   File "/usr/local/lib/python3.11/dist-packages/torch/utils/checkpoint.py", line 313, in backward
[rank0]:     torch.autograd.backward(outputs_with_grad, args_with_grad)
[rank0]:   File "/usr/local/lib/python3.11/dist-packages/torch/autograd/__init__.py", line 289, in backward
[rank0]:     _engine_run_backward(
[rank0]:   File "/usr/local/lib/python3.11/dist-packages/torch/autograd/graph.py", line 769, in _engine_run_backward
[rank0]:     return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
[rank0]:            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[rank0]: RuntimeError: Expected to mark a variable ready only once. This error is caused by one of the following reasons: 1) Use of a module parameter outside the `forward` function. Please make sure mode
l parameters are not shared across multiple concurrent forward-backward passes. or try to use _set_static_graph() as a workaround if this module graph does not change during training loop.2) Reused parame
ters in multiple reentrant backward passes. For example, if you use multiple `checkpoint` functions to wrap the same part of your model, it would result in the same set of parameters been used by differen
t reentrant backward passes multiple times, and hence marking a variable ready multiple times. DDP does not support such use cases in default. You can try to use _set_static_graph() as a workaround if you
r module graph does not change over iterations.