xref: /aosp_15_r20/external/pytorch/benchmarks/tensorexpr/HowToRun.md (revision da0073e96a02ea20f0ac840b70461e3646d07c45)
1From the root of pytorch repo, run:
2```
3python -m benchmarks.tensorexpr --help
4```
5to show documentation.
6
7An example of an actual command line that one might use as a starting point:
8```
9python -m benchmarks.tensorexpr --device gpu --mode fwd --jit-mode trace --cuda-fuser=te
10```
11