NOTICE

fmha_sm100
Copyright (c) 2026 MiniMax

This product includes software developed by third parties. See the
"Third-party licenses" section of README.md for the full list of bundled
and runtime-dependency components and license summaries. The
authoritative license text for each third-party component is shipped
with that component itself (vendored source headers, the LICENSE file
distributed with the pip-installed package, or the upstream project
repository).

================================================================================
Bundled / derived source (shipped in this repository)
================================================================================

This project bundles or derives from the following third-party components.
Each retains its original license and copyright. In compliance with the
BSD-3-Clause §2 and Apache-2.0 §4 redistribution clauses, the original
copyright notices and license terms are reproduced below (and are also
preserved at the top of every vendored source file via the
SPDX-License-Identifier header).

--------------------------------------------------------------------------------
NVIDIA CUTLASS — BSD-3-Clause
--------------------------------------------------------------------------------

Copyright (c) 2017-2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved.

Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:

  1. Redistributions of source code must retain the above copyright notice,
     this list of conditions and the following disclaimer.
  2. Redistributions in binary form must reproduce the above copyright notice,
     this list of conditions and the following disclaimer in the documentation
     and/or other materials provided with the distribution.
  3. Neither the name of the copyright holder nor the names of its
     contributors may be used to endorse or promote products derived from
     this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
POSSIBILITY OF SUCH DAMAGE.

Project:    https://github.com/NVIDIA/cutlass
Where:      python/fmha_sm100/cutlass/ (git submodule: include/ and
            tools/util/include/)
            plus BSD-3-tagged headers under python/fmha_sm100/csrc/include/.
            The SM100 MMA descriptor encodings in
            python/fmha_sm100/cute/src/common/mma_sm100_desc.py
            mirror NVIDIA CUTLASS hardware descriptors.
Full text:  python/fmha_sm100/cutlass/LICENSE.txt (submodule), or
            https://github.com/NVIDIA/cutlass/blob/main/LICENSE

--------------------------------------------------------------------------------
FlashInfer — Apache License 2.0
--------------------------------------------------------------------------------

Copyright (c) 2023-2025 by FlashInfer team.

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    https://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.

Project:    https://github.com/flashinfer-ai/flashinfer
Where:      Apache-2.0-tagged sources under python/fmha_sm100/csrc/
            and python/fmha_sm100/csrc/include/ (e.g. allocator.h,
            exception.h, utils.cuh, cutlass_utils.cuh,
            fmha_cutlass_sm100.cuh, sparse_topk_select.cuh, plan.cuh,
            sm100_fmha_reduction.hpp, tvm_ffi_utils.h).
Full text:  https://www.apache.org/licenses/LICENSE-2.0

--------------------------------------------------------------------------------
NVIDIA TensorRT-LLM — Apache License 2.0
--------------------------------------------------------------------------------

Copyright (c) 2019-2026 NVIDIA CORPORATION. All rights reserved.

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    https://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.

Project:    https://github.com/NVIDIA/TensorRT-LLM
Where:      python/fmha_sm100/csrc/include/sparse_topk_select.cuh
            portions of the indexerTopK histogram-step + insertion-sort
            algorithm, derived from
            tensorrt_llm/cpp/tensorrt_llm/kernels/indexerTopK.cu.
Full text:  https://www.apache.org/licenses/LICENSE-2.0

--------------------------------------------------------------------------------
NAVER Corp. (CLOVA) — Apache License 2.0
--------------------------------------------------------------------------------

Copyright (c) 2021 NAVER Corp.
Licensed under the Apache License, Version 2.0; you may not use this file
except in compliance with the License. You may obtain a copy of the
License at https://www.apache.org/licenses/LICENSE-2.0

Project:    https://github.com/NAVER-CLOVA
Where:      Co-authored portions of
            python/fmha_sm100/csrc/include/sparse_topk_select.cuh
            (indexerTopK algorithm, alongside the NVIDIA copyright above).
Full text:  https://www.apache.org/licenses/LICENSE-2.0

--------------------------------------------------------------------------------
Changes to bundled / derived source
--------------------------------------------------------------------------------

This project does not modify the text of vendored third-party source files
beyond what is required to integrate them (path adjustments in
`#include` directives and minor macro renames to avoid collision with
project-local symbols; no semantic changes). All such integration
modifications are confined to `csrc/` and to the import shim under
`python/fmha_sm100/sparse.py`. The SM100 MMA descriptor encodings in
`python/fmha_sm100/cute/src/common/mma_sm100_desc.py` are
re-typed from the CUTLASS hardware descriptors — they are numerical
constants, not CUTLASS source code, but the original copyright is
attributed here for completeness.

================================================================================
Runtime dependencies (installed via pip, not vendored)
================================================================================

For runtime dependencies, the authoritative license text is distributed
with each installed package; consult `pip show <pkg>` for the license
metadata. The full list and license summaries are in the
"Third-party licenses" section of README.md.

================================================================================
Trademarks
================================================================================

NVIDIA, the NVIDIA logo, Blackwell, SM100, B200, B300, GB200, GB300,
TensorRT, TensorRT-LLM, cuDNN, cuBLAS, and CUDA are trademarks and/or
registered trademarks of NVIDIA Corporation. Apache, Apache TVM, and
Apache TVM FFI are trademarks of The Apache Software Foundation. All
other trademarks referenced in this project are the property of their
respective owners. The use of any trademark in this NOTICE or in
README.md is for identification purposes only and does not imply
endorsement.