[BUG]: RuntimeError(f"Could not compute metadata for {kind} target {target}: {e}")
Nobody has claimed this yet.
Assessment
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Newbie friendliness
- 35/100
- Issue type
- Bug
- Clarity
- Needs clarification
- Activity status
- Quiet
- Tech stack
- python, pytorch
- Domain
- machine-learning
Research direction
Start with colossalai/fx/tracer/tracer.py, especially _meta_data_computing and create_proxy, and compare the failing get_attr path with the installed PyTorch FX behavior shown in the traceback. A minimal reproduction and environment details are still needed; done means metadata computation succeeds for the reported transformer target or the incompatibility is documented with a confirmed fix path.
Written by the indexing model from the issue text.
Description
Is there an existing issue for this bug?
- I have searched the existing issues
The bug has not been fixed in the latest main branch
- I have checked the latest main branch
Do you feel comfortable sharing a concise (minimal) script that reproduces the error? :)
Yes, I will share a minimal reproducible script.
🐛 Describe the bug
Is it because the PyTorch version is not right?
Traceback (most recent call last):
File "/usr/local/lib/python3.12/dist-packages/colossalai/fx/tracer/tracer.py", line 300, in _meta_data_computing
attr_itr = getattr(attr_itr, atom)
^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/torch/fx/_symbolic_trace.py", line 852, in module_getattr_wrapper
return self.getattr(attr, attr_val, parameter_proxy_cache)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/torch/fx/_symbolic_trace.py", line 632, in getattr
maybe_parameter_proxy = maybe_get_proxy_for_attr(
^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/torch/fx/_symbolic_trace.py", line 626, in maybe_get_proxy_for_attr
val_proxy = self.create_proxy("get_attr", n, (), {}, **kwargs) # type: ignore[arg-type]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/colossalai/fx/tracer/tracer.py", line 142, in create_proxy
meta_out = self._meta_data_computing(
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/colossalai/fx/tracer/tracer.py", line 313, in _meta_data_computing
raise RuntimeError(f"Could not compute metadata for {kind} target {target}: {e}")
RuntimeError: Could not compute metadata for get_attr target transformer.h.0.attn.c_attn.bias: Could not compute metadata for get_attr target transformer.h.0.attn.c_attn.bias: Could not compute metadata for get_attr target transformer.h.0.attn.c_attn.bias: ......
Environment
No response
- Dominant language
- Python
- Stars
- 41.4k
- Forks
- 4.5k
- PR merge metrics
- No merged PRs in 30d
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from hpcaitech/ColossalAI
-
Difficulty 2/5 1-3 hours Newbie friendliness 75/100
hpcaitech/ColossalAI#6437 ·
-
bug
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
hpcaitech/ColossalAI#6428 ·
-
documentation
Difficulty 2/5 1-3 hours Newbie friendliness 76/100
hpcaitech/ColossalAI#6197 · 3 comments ·
-
documentation
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
hpcaitech/ColossalAI#5293 · 3 comments ·
-
Difficulty 4/5 3-5 days Newbie friendliness 25/100
hpcaitech/ColossalAI#6403 ·
All issues in hpcaitech/ColossalAI
Similar issues
-
Difficulty 2/5 1-3 hours Newbie friendliness 88/100
-
Difficulty 2/5 1-3 hours Newbie friendliness 82/100
-
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
-
enhancement
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
-
Difficulty 2/5 1-3 hours Newbie friendliness 74/100