Kernels:
Trusted publisher
Download build/torch-cuda/modules/backends/__init__.py from kernels-community/fla: direct link, hf CLI and curl.
- Browser
- Download file 647 Bytes
-
https://huggingface.co/kernels/kernels-community/fla/resolve/v1/build/torch-cuda/modules/backends/__init__.py
- Command line
-
hf download hf://kernels-community/fla@v1/build/torch-cuda/modules/backends/__init__.py
-
curl -L -o __init__.py https://huggingface.co/kernels/kernels-community/fla/resolve/v1/build/torch-cuda/modules/backends/__init__.py
647 Bytes
| # Copyright (c) 2023-2026, Songlin Yang, Yu Zhang, Zhiyuan Li | |
| # | |
| # This source code is licensed under the MIT license found in the | |
| # LICENSE file in the root directory of this source tree. | |
| # For a list of all contributors, visit: | |
| # https://github.com/fla-org/flash-linear-attention/graphs/contributors | |
| """Module-level backends for FLA components such as rotary and cross-entropy.""" | |
| from ...modules.backends.triton_ascend import TritonAscendBackend | |
| from ...ops.backends import BackendRegistry, dispatch | |
| modules_registry = BackendRegistry("modules") | |
| modules_registry.register(TritonAscendBackend()) | |
| __all__ = ['dispatch', 'modules_registry'] | |