Sherlock Huang
Sherlock Huang
Hi @naveenthangudu TorchScript is a legacy path that is currently maintained by pytorch community. We would recommend integrating new backends via the PT2 path. See https://dev-discuss.pytorch.org/t/registering-new-compiler-backend-in-pytorch2-0/1092/4 https://colab.research.google.com/drive/1Zh-Uo3TcTH8yYJF-LLo5rjlHVMtqvMdf#scrollTo=KA_FS0D831T5
@suo I added SymInt as permissible return type. See `sample_symint_export.txt` as an example.
@pytorchbot merge
This change has a pretty large blast radius in the internal code base, and require fixes over 100+ of typecheck failures in various projects. I am afraid that I don't...
test failures looks relevant.
This is cool, and would be pretty useful. I also want this to be a standalone graph passes ``` def enrich_debuginfo_pass(gm, sample_input): with config.pattch("codegen_record_function", True): gm.recompile() return gm ``` or...