[Question] qwen3.5 27B train error

Your Question

File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1786, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/root/Megatron-LM/megatron/core/transformer/transformer_layer.py", line 475, in forward
hidden_states, context = self._forward_attention(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/root/Megatron-LM/megatron/core/transformer/transformer_layer.py", line 549, in _forward_attention
attention_output_with_bias = self.self_attention(
^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1786, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/root/Megatron-LM/megatron/core/ssm/gated_delta_net.py", line 302, in forward
raise NotImplementedError("GDN does not support packed sequence for now.")
NotImplementedError: GDN does not support packed sequence for now.

What I've Tried

File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1786, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/root/Megatron-LM/megatron/core/transformer/transformer_layer.py", line 475, in forward
hidden_states, context = self._forward_attention(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/root/Megatron-LM/megatron/core/transformer/transformer_layer.py", line 549, in _forward_attention
attention_output_with_bias = self.self_attention(
^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1786, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/root/Megatron-LM/megatron/core/ssm/gated_delta_net.py", line 302, in forward
raise NotImplementedError("GDN does not support packed sequence for now.")
NotImplementedError: GDN does not support packed sequence for now.

Environment (if relevant)

• slime version:
• Python version:
• PyTorch version:
• CUDA/ROCm version:
• GPU type and count:
• OS:

Additional Context

No response

Pre-submission Checklist

• I have read the CONTRIBUTING.md and understand the collaboration scope.
• I have read the documentation and FAQ and my question is not answered there.
• I have searched for existing issues and my question has not been asked before.

Start with megatron/core/ssm/gated_delta_net.py at the reported line 302, then trace the call from megatron/core/transformer/transformer_layer.py lines 475 and 549. Reproduce the training error after obtaining the missing environment and launch details; done means explaining the packed-sequence incompatibility and documenting a verified resolution.