/usr/lib/python3.12/site-packages/torch/cuda/__init__.py:68: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
  import pynvml  # type: ignore[import]
Usage: flashinfer [OPTIONS] [COMMAND] [ARGS]...

  FlashInfer CLI

Options:
  --download-cubin  Download raw cubin artifacts to the local cache.
  --help            Show this message and exit.

Commands:
  clear-cache                 Clear cache
  clear-cubin                 Clear cubin
  collect-env                 Dump a full environment report for bug...
  download-cubin              Download cubin artifacts into...
  download-kernels            Install matching flashinfer-cubin and...
  export-compile-commands     Export compile commands to...
  generate-tactics-blocklist  Probe the local GPU and write an offline...
  install-cubin-wheel         Install the matching flashinfer-cubin wheel.
  install-jit-cache-wheel     Install the matching flashinfer-jit-cache...
  list-cubins                 List downloaded cubins
  list-modules                List or inspect compilation modules
  module-status               Show module compilation status
  replay                      Replay API calls from dump directory
  show-config                 Show configuration