[user empathy day 2][based] torch.compile issues #128071
Labels
dynamo-triage-june2024
dynamo-user-empathy-day
empathy-day
Label for issues from user empathy days
module: dynamo
module: startup-tracing-compile
Compilation mechanism or time spent in (re)compilation, tracing, startup
oncall: pt2
triaged
This issue has been looked at a team member, and triaged and prioritized into an appropriate module
🐛 Describe the bug
Pytorch-eager
Time = 2.8532286300323904 seconds for 10 iterations
Torch.compile backend=eager
Observations
torch/_dynamo/symbolic_convert.py:774] [49/974]
Time = 35.62640808802098 seconds for 10 iterations
Verdict - Does not work
Other fixes
torch.compile warning message for pybind'ed c++ functions is very spammy #128060
/home/anijain/local/miniconda3/envs/user_empathy_day2/lib/python3.10/site-packages/torch/_dynamo/variables/functions.py:661: UserWarning: Graph break due to unsupported builtin flash_attn_2_cuda.PyCapsule.fwd_kvcache. This function is either a Python builtin (e.g. _warnings.warn) or a third-party C/C++ Python extension (perhaps created with pybind). If it is a Python builtin, please file an issue on GitHub so the PyTorch team can add support for it and see the next case for a workaround. If it is a third-party C/C++ Python extension, please either wrap it into a PyTorch-understood custom operator (see https://pytorch.org/docs/main/notes/custom_operators.html for more details) or, if it is traceable, use torch.compiler.allow_in_graph.
warnings.warn(msg)
torch.compile with backend=aot_eager
Error logs
No response
Minified repro
No response
Versions
NA
cc @ezyang @chauhang @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @chenyang78 @kadeng @oulgen @jamesjwu @aorenste @laithsakka @bdhirsh @msaroufim
The text was updated successfully, but these errors were encountered: