parser() — pytorch Function Reference

Architecture documentation for the parser() function in generate_kernels.py from the pytorch codebase.

Function python CoreTensor Registration calls 1

Entity Profile

CoreTensor→ Registration→ parser() — pytorch Function Reference

Dependency Diagram

graph TD
  2a7b91f6_0c36_b743_0ba0_d6b276e5dc23["parser()"]
  aad87944_33f6_d3c8_e945_0c0de37e4fe0["main()"]
  2a7b91f6_0c36_b743_0ba0_d6b276e5dc23 -->|calls| aad87944_33f6_d3c8_e945_0c0de37e4fe0
  style 2a7b91f6_0c36_b743_0ba0_d6b276e5dc23 fill:#6366f1,stroke:#818cf8,color:#fff

Relationship Graph

Source Code

aten/src/ATen/native/transformers/cuda/mem_eff_attention/kernels/generate_kernels.py lines 401–414

    parser = argparse.ArgumentParser(
        prog="generate_kernels",
        description="Generate the mem-eff kernels template instantiations",
    )
    # Set an optional output directory
    parser.add_argument(
        "-o",
        "--output_dir",
        required=False,
        help="Where to generate the kernels "
        " will default to <ATen/native/transformers/cuda/mem_eff_attention/kernels/> ",
    )
    args = parser.parse_args()
    main(args.output_dir)