Set version to 0.8.1

This change adds `LockedLayerRepository` as an alternative to
`LayerRepository`. `LockedLayerRepository` allows for locking all kernel
layers that are used at the project level. Example usage:

```
with use_kernel_mapping(
    {
        "SomeLayer": {
            "cuda": LockedLayerRepository(
                repo_id="some-org/some-layer",
                layer_name="SomeLayer",
            )
        },
    }
):
    layer = kernelize(layer, device="cuda", mode=Mode.INFERENCE)
```

This requires that the project has a `pyproject.toml` with kernel
version specifications and `kernel.lock` with the locked kernels.

.github/workflows/lint.yml

@@ -0,0 +1,10 @@
+name: Lints
+on: [push, pull_request]
+jobs:
+  lint:
+    name: Run lints
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - name: Run ruff
+        uses: astral-sh/ruff-action@v3

.github/workflows/publish.yml

@@ -0,0 +1,120 @@
+name: Publish Python 🐍 distribution 📦 to PyPI and TestPyPI
+
+on: push
+
+jobs:
+  build:
+    name: Build distribution 📦
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          persist-credentials: false
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.9"
+      - name: Install pypa/build
+        run: >-
+          python3 -m
+          pip install
+          build
+          --user
+      - name: Build a binary wheel and a source tarball
+        run: python3 -m build
+      - name: Store the distribution packages
+        uses: actions/upload-artifact@v4
+        with:
+          name: python-package-distributions
+          path: dist/
+
+  publish-to-pypi:
+    name: >-
+      Publish Python 🐍 distribution 📦 to PyPI
+    if: startsWith(github.ref, 'refs/tags/') # only publish to PyPI on tag pushes
+    needs:
+      - build
+    runs-on: ubuntu-latest
+    environment:
+      name: pypi
+      url: https://pypi.org/p/kernels
+    permissions:
+      id-token: write # IMPORTANT: mandatory for trusted publishing
+
+    steps:
+      - name: Download all the dists
+        uses: actions/download-artifact@v4
+        with:
+          name: python-package-distributions
+          path: dist/
+      - name: Publish distribution 📦 to PyPI
+        uses: pypa/gh-action-pypi-publish@release/v1
+
+  github-release:
+    name: >-
+      Sign the Python 🐍 distribution 📦 with Sigstore
+      and upload them to GitHub Release
+    needs:
+      - publish-to-pypi
+    runs-on: ubuntu-latest
+
+    permissions:
+      contents: write # IMPORTANT: mandatory for making GitHub Releases
+      id-token: write # IMPORTANT: mandatory for sigstore
+
+    steps:
+      - name: Download all the dists
+        uses: actions/download-artifact@v4
+        with:
+          name: python-package-distributions
+          path: dist/
+      - name: Sign the dists with Sigstore
+        uses: sigstore/gh-action-sigstore-python@v3.0.0
+        with:
+          inputs: >-
+            ./dist/*.tar.gz
+            ./dist/*.whl
+      - name: Create GitHub Release
+        env:
+          GITHUB_TOKEN: ${{ github.token }}
+        run: >-
+          gh release create
+          "$GITHUB_REF_NAME"
+          --repo "$GITHUB_REPOSITORY"
+          --notes ""
+      - name: Upload artifact signatures to GitHub Release
+        env:
+          GITHUB_TOKEN: ${{ github.token }}
+        # Upload to GitHub Release using the `gh` CLI.
+        # `dist/` contains the built packages, and the
+        # sigstore-produced signatures and certificates.
+        run: >-
+          gh release upload
+          "$GITHUB_REF_NAME" dist/**
+          --repo "$GITHUB_REPOSITORY"
+
+  publish-to-testpypi:
+    name: Publish Python 🐍 distribution 📦 to TestPyPI
+    needs:
+      - build
+    runs-on: ubuntu-latest
+
+    environment:
+      name: testpypi
+      url: https://test.pypi.org/p/kernels
+
+    permissions:
+      id-token: write # IMPORTANT: mandatory for trusted publishing
+
+    steps:
+      - name: Download all the dists
+        uses: actions/download-artifact@v4
+        with:
+          name: python-package-distributions
+          path: dist/
+      - name: Publish distribution 📦 to TestPyPI
+        uses: pypa/gh-action-pypi-publish@release/v1
+        with:
+          repository-url: https://test.pypi.org/legacy/
+          skip-existing: true # Only upload when the version is unique.

.github/workflows/test.yml

@@ -1,4 +1,4 @@
-name: Test hf-kernels
+name: Test kernels
 
 on:
   push:
@@ -24,7 +24,10 @@ jobs:
       max-parallel: 4
       matrix:
         python-version: ["3.10", "3.12"]
-        torch-version: ["2.5.1", "2.6.0"]
+        torch-version: ["2.6.0", "2.7.0"]
+
+    env:
+      UV_PYTHON_PREFERENCE: only-managed
 
     steps:
       - name: Checkout code
@@ -41,5 +44,28 @@ jobs:
       - name: Install the project
         run: uv sync --all-extras --dev
 
+      - name: Install setuptools for Triton-based test
+        run: uv pip install setuptools
+
+      - name: Check typing
+        run: uv run mypy src/kernels
+
       - name: Run tests
         run: uv run pytest tests
+
+      - name: Check kernel conversion
+        run: |
+          uv pip install wheel
+          uv run kernels to-wheel kernels-community/triton-layer-norm 0.0.1
+          uv pip install triton_layer_norm-0.0.1*.whl
+          uv run python -c "import triton_layer_norm"
+
+      - name: Check README generation
+        # For now, just checks that generation doesn't fail.
+        run: |
+          uv run kernels generate-readme kernels-community/triton-layer-norm
+
+      - name: Import check without torch
+        run: |
+          uv pip uninstall torch
+          python -c "import kernels"

LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   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
+
+       http://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.

README.md

@@ -1,11 +1,42 @@
-# hf-kernels
+# kernels
 
-Make sure you have `torch==2.5.1+cu124` installed.
+<div align="center">
+<img src="https://github.com/user-attachments/assets/64a652f3-0cd3-4829-b3c1-df13f7933569" width="450" height="450" alt="kernel-builder logo">
+<p align="center">
+    <a href="https://pypi.org/project/kernels"><img alt="PyPI - Version" src="https://img.shields.io/pypi/v/kernels"></a>
+    <a href="https://github.com/huggingface/kernels/tags"><img alt="GitHub tag" src="https://img.shields.io/github/v/tag/huggingface/kernels"></a>
+    <a href="https://github.com/huggingface/kernels/actions/workflows/docker-build-push.yaml"><img alt="Test kernels" src="https://img.shields.io/github/actions/workflow/status/huggingface/kernels/test.yml?label=test"></a>
+  
+</p>
+</div>
+<hr/>
+
+The Kernel Hub allows Python libraries and applications to load compute
+kernels directly from the [Hub](https://hf.co/). To support this kind
+of dynamic loading, Hub kernels differ from traditional Python kernel
+packages in that they are made to be:
+
+- Portable: a kernel can be loaded from paths outside `PYTHONPATH`.
+- Unique: multiple versions of the same kernel can be loaded in the
+  same Python process.
+- Compatible: kernels must support all recent versions of Python and
+  the different PyTorch build configurations (various CUDA versions
+  and C++ ABIs). Furthermore, older C library versions must be supported.
+
+## 🚀 Quick Start
+
+Install the `kernels` package with `pip` (requires `torch>=2.5` and CUDA):
+
+```bash
+pip install kernels
+```
+
+Here is how you would use the [activation](https://huggingface.co/kernels-community/activation) kernels from the Hugging Face Hub:
 
 ```python
 import torch
 
-from hf_kernels import get_kernel
+from kernels import get_kernel
 
 # Download optimized kernels from the Hugging Face hub
 activation = get_kernel("kernels-community/activation")
@@ -20,57 +51,15 @@ activation.gelu_fast(y, x)
 print(y)
 ```
 
-## Docker Reference
+You can [search for kernels](https://huggingface.co/models?other=kernel) on
+the Hub.
 
-build and run the reference [examples/basic.py](examples/basic.py) in a Docker container with the following commands:
+## 📚 Documentation
 
-```bash
+- [Using layers](docs/layers.md)
-docker build --platform linux/amd64 -t kernels-reference -f docker/Dockerfile.reference .
+- [Locking kernel/layer versions](docs/locking.md)
-docker run --gpus all -it --rm -e HF_TOKEN=$HF_TOKEN kernels-reference
+- [Environment variables](docs/env.md)
-```
+- [Using kernels in a Docker container](docs/docker.md)
-
+- [Kernel requirements](docs/kernel-requirements.md)
-## Locking kernel versions
+- [Frequently Asked Questions](docs/faq.md)
-
+- [Writing kernels](https://github.com/huggingface/kernel-builder/blob/main/docs/writing-kernels.md) using [kernel-builder](https://github.com/huggingface/kernel-builder/)
-Projects that use `setuptools` can lock the kernel versions that should be
-used. First specify the accepted versions in `pyproject.toml` and make
-sure that `hf-kernels` is a build dependency:
-
-```toml
-[build-system]
-requires = ["hf-kernels", "setuptools"]
-build-backend = "setuptools.build_meta"
-
-[tool.kernels.dependencies]
-"kernels-community/activation" = ">=0.0.1"
-```
-
-Then run `hf-kernel lock .` in the project directory. This generates a `kernels.lock` file with
-the locked revisions. The locked revision will be used when loading a kernel with
-`get_locked_kernel`:
-
-```python
-from hf_kernels import get_locked_kernel
-
-activation = get_locked_kernel("kernels-community/activation")
-```
-
-**Note:** the lock file is included in the package metadata, so it will only be visible
-to `hf-kernels` after doing an (editable or regular) installation of your project.
-
-## Pre-downloading locked kernels
-
-Locked kernels can be pre-downloaded by running `hf-kernel download .` in your
-project directory. This will download the kernels to your local Hugging Face
-Hub cache.
-
-The pre-downloaded kernels are used by the `get_locked_kernel` function.
-`get_locked_kernel` will download a kernel when it is not pre-downloaded. If you
-want kernel loading to error when a kernel is not pre-downloaded, you can use
-the `load_kernel` function instead:
-
-````python
-```python
-from hf_kernels import load_kernel
-
-activation = load_kernel("kernels-community/activation")
-````

docker/Dockerfile.reference

@@ -31,13 +31,13 @@ WORKDIR /app/kernel-test
 # install python depdencies
 RUN uv add torch==2.5.0 numpy
 
-# copy hf-kernels lib
+# copy kernels lib
-COPY src ./hf-kernels/src
+COPY src ./kernels/src
-COPY pyproject.toml ./hf-kernels/pyproject.toml
+COPY pyproject.toml ./kernels/pyproject.toml
-COPY README.md ./hf-kernels/README.md
+COPY README.md ./kernels/README.md
 
 # install library
-RUN uv pip install -e hf-kernels
+RUN uv pip install -e kernels
 
 # copy examples
 COPY examples ./examples
@@ -48,4 +48,4 @@ ENV NVIDIA_DRIVER_CAPABILITIES=compute,utility
 
 # command to run the script
 CMD ["uv", "run", "examples/basic.py"]
-# CMD ["ls", "hf-kernels"]
+# CMD ["ls", "kernels"]

docs/docker.md

@@ -0,0 +1,8 @@
+# Using kernels in a Docker container
+
+build and run the reference [examples/basic.py](examples/basic.py) in a Docker container with the following commands:
+
+```bash
+docker build --platform linux/amd64 -t kernels-reference -f docker/Dockerfile.reference .
+docker run --gpus all -it --rm -e HF_TOKEN=$HF_TOKEN kernels-reference
+```

docs/env.md

@@ -0,0 +1,10 @@
+# Environment variables
+
+## `KERNELS_CACHE`
+
+The directory to use as the local kernel cache. If not set, the cache
+of the `huggingface_hub` package is used.
+
+## `DISABLE_KERNEL_MAPPING`
+
+Disables kernel mappings for [`layers`](layers.md).

docs/faq.md

@@ -0,0 +1,13 @@
+# FAQ
+
+## Why is the kernelization step needed?
+
+In earlier versions of `kernels`, a layer's `forward` was replaced by
+`use_kernel_forward_from_hub` and `replace_kernel_forward_from_hub`. The
+new `forward` would dispatch to a kernel based on the device type,
+whether a model was training, etc. However, this approach was
+fundamentally incompatible with `torch.compile` since it relied
+on data-dependent branching.
+
+To avoid branching, we have to make dispatch decisions ahead of time,
+which is what the `kernelize` function does.

docs/kernel-requirements.md

@@ -0,0 +1,210 @@
+# Kernel requirements
+
+Kernels on the Hub must fulfill the requirements outlined on this page. By
+ensuring kernels are compliant, they can be used on a wide range of Linux
+systems and Torch builds.
+
+You can use [kernel-builder](https://github.com/huggingface/kernel-builder/)
+to build compliant kernels.
+
+## Directory layout
+
+A kernel repository on the Hub must contain a `build` directory. This
+directory contains build variants of a kernel in the form of directories
+following the template
+`<framework><version>-cxx<abiver>-<cu><cudaver>-<arch>-<os>`.
+For example `build/torch26-cxx98-cu118-x86_64-linux`.
+
+Each variant directory must contain a single directory with the same name
+as the repository (replacing `-` by `_`). For instance, kernels in the
+`kernels-community/activation` repository have a directories like
+`build/<variant>/activation`. This directory
+must be a Python package with an `__init__.py` file.
+
+## Build variants
+
+A kernel can be compliant for a specific compute framework (e.g. CUDA) or
+architecture (e.g. x86_64). For compliance with a compute framework and
+architecture combination, all the variants from the [build variant list](https://github.com/huggingface/kernel-builder/blob/main/docs/build-variants.md)
+must be available for that combination.
+
+## Versioning
+
+Kernels are versioned on the Hub using Git tags. Version tags must be of
+the form `v<major>.<minor>.<patch>`. Versions are used by [locking](./locking.md)
+to resolve the version constraints.
+
+## Native Python module
+
+Kernels will typically contain a native Python module with precompiled
+compute kernels and bindings. This module must fulfill the requirements
+outlined in this section. For all operating systems, a kernel must not
+have dynamic library dependencies outside:
+
+- Torch;
+- CUDA/ROCm libraries installed as dependencies of Torch.
+
+### Linux
+
+- Use [ABI3/Limited API](https://docs.python.org/3/c-api/stable.html#stable-application-binary-interface)
+  for compatibility with Python 3.9 and later.
+- Compatible with [`manylinux_2_28`](https://github.com/pypa/manylinux?tab=readme-ov-file#manylinux_2_28-almalinux-8-based).
+  This means that the extension **must not** use symbols versions higher than:
+
+  - GLIBC 2.28
+  - GLIBCXX 3.4.24
+  - CXXABI 1.3.11
+  - GCC 7.0.0
+
+These requirement can be checked with the ABI checker (see below).
+
+### macOS
+
+- Use [ABI3/Limited API](https://docs.python.org/3/c-api/stable.html#stable-application-binary-interface)
+  for compatibility with Python 3.9 and later.
+- macOS deployment target 15.0.
+- Metal 3.0 (`-std=metal3.0`).
+
+The ABI3 requirement can be checked with the ABI checker (see below).
+
+### ABI checker
+
+The manylinux_2_28 and Python ABI 3.9 version requirements can be checked with
+[`kernel-abi-check`](https://crates.io/crates/kernel-abi-check):
+
+```bash
+
+$ cargo install kernel-abi-check
+$ kernel-abi-check result/relu/_relu_e87e0ca_dirty.abi3.so
+🐍 Checking for compatibility with manylinux_2_28 and Python ABI version 3.9
+✅ No compatibility issues found
+```
+
+## Torch extension
+
+Torch native extension functions must be [registered](https://pytorch.org/tutorials/advanced/cpp_custom_ops.html#cpp-custom-ops-tutorial)
+in `torch.ops.<namespace>`. Since we allow loading of multiple versions of
+a module in the same Python process, `namespace` must be unique for each
+version of a kernel. Failing to do so will create clashes when different
+versions of the same kernel are loaded. Two suggested ways of doing this
+are:
+
+- Appending a truncated SHA-1 hash of the git commit that the kernel was
+  built from to the name of the extension.
+- Appending random material to the name of the extension.
+
+**Note:** we recommend against appending a version number or git tag.
+Version numbers are typically not bumped on each commit, so users
+might use two different commits that happen to have the same version
+number. Git tags are not stable, so they do not provide a good way
+of guaranteeing uniqueness of the namespace.
+
+## Layers
+
+A kernel can provide layers in addition to kernel functions. A layer from
+the Hub can replace the `forward` method of an existing layer for a certain
+device type. This makes it possible to provide more performant kernels for
+existing layers. See the [layers documentation](layers.md) for more information
+on how to use layers.
+
+### Writing layers
+
+To make the extension of layers safe, the layers must fulfill the following
+requirements:
+
+- The layers are subclasses of `torch.nn.Module`.
+- The layers are pure, meaning that they do not have their own state. This
+  means that:
+  - The layer must not define its own constructor.
+  - The layer must not use class variables.
+- No other methods must be defined than `forward`.
+- The `forward` method has a signature that is compatible with the
+  `forward` method that it is extending.
+
+There are two exceptions to the _no class variables rule_:
+
+1. The `has_backward` variable can be used to indicate whether the layer has
+   a backward pass implemented (`True` when absent).
+2. The `can_torch_compile` variable can be used to indicate whether the layer
+   supports `torch.compile` (`False` when absent).
+
+This is an example of a pure layer:
+
+```python
+class SiluAndMul(nn.Module):
+    # This layer does not implement backward.
+    has_backward: bool = False
+
+    def forward(self, x: torch.Tensor):
+        d = x.shape[-1] // 2
+        output_shape = x.shape[:-1] + (d,)
+        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
+        ops.silu_and_mul(out, x)
+        return out
+```
+
+For some layers, the `forward` method has to use state from the adopting class.
+In these cases, we recommend to use type annotations to indicate what member
+variables are expected. For instance:
+
+```python
+class LlamaRMSNorm(nn.Module):
+    weight: torch.Tensor
+    variance_epsilon: float
+
+    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
+        return rms_norm_fn(
+            hidden_states,
+            self.weight,
+            bias=None,
+            residual=None,
+            eps=self.variance_epsilon,
+            dropout_p=0.0,
+            prenorm=False,
+            residual_in_fp32=False,
+        )
+```
+
+This layer expects the adopting layer to have `weight` and `variance_epsilon`
+member variables and uses them in the `forward` method.
+
+### Exporting layers
+
+To accommodate portable loading, `layers` must be defined in the main
+`__init__.py` file. For example:
+
+```python
+from . import layers
+
+__all__ = [
+  # ...
+  "layers"
+  # ...
+]
+```
+
+## Python requirements
+
+- Python code must be compatible with Python 3.9 and later.
+- All Python code imports from the kernel itself must be relative. So,
+  for instance if in the example kernel `example`,
+  `module_b` needs a function from `module_a`, import as:
+
+  ```python
+  from .module_a import foo
+  ```
+
+  **Never use:**
+
+  ```python
+  # DO NOT DO THIS!
+
+  from example.module_a import foo
+  ```
+
+  The latter would import from the module `example` that is in Python's
+  global module dict. However, since we allow loading multiple versions
+  of a module, we uniquely name the module.
+
+- Only modules from the Python standard library, Torch, or the kernel itself
+  can be imported.

docs/layers.md

@@ -0,0 +1,272 @@
+# Layers
+
+A kernel can provide layers in addition to kernel functions. A layer from
+the Hub can replace the `forward` method of an existing layer for a certain
+device type. This makes it possible to provide more performant kernels for
+existing layers.
+
+See [Kernel requirements](kernel-requirements.md) for more information the
+requirements of Hub layers.
+
+## Making a layer extensible with kernels from the hub
+
+### Using a decorator
+
+A layer can be made extensible with the `use_kernel_forward_from_hub`
+decorator. For example:
+
+```python
+@use_kernel_forward_from_hub("SiluAndMul")
+class SiluAndMul(nn.Module):
+    def forward(self, input: torch.Tensor) -> torch.Tensor:
+        d = input.shape[-1] // 2
+        return F.silu(input[..., :d]) * input[..., d:]
+```
+
+The decorator does not change the behavior of the class -- it annotates
+the class with the given name (here `SiluAndMul`). The `kernelize` function
+described below uses this name to look up kernels for the layer.
+
+### External layers
+
+An existing layer that does not (yet) have the `use_kernel_forward_from_hub`
+decorator can be made extensible using the `replace_kernel_forward_from_hub`
+function:
+
+```python
+from somelibrary import SiluAndMul
+
+replace_kernel_forward_from_hub(SiluAndMul, "SiluAndMul")
+```
+
+**Warning:** we strongly recommend using layers with a decorator, since
+it signifies that the maintainer intends to keep the `forward` signature
+compatible with layers from the hub.
+
+## Kernelizing a model
+
+A model will not use Hub kernels by default, even if it contains extensible
+layers. To enable the use of Hub kernels in the model, it needs to be
+'kernelized' using the `kernelize` function. This function traverses the
+model graph and replaces the `forward` methods of extensible layers for which
+Hub kernels are registered. `kernelize` can be used as follows:
+
+```python
+model = MyModel(...)
+model = kernelize(model, mode=Mode.INFERENCE)
+```
+
+The `kernelize` function modifies the model in-place, the model itself is
+returned as a convenience. The `mode` specifies that the model will be used
+in inference. Similarly, you can ask `kernelize` to prepare the model for
+training:
+
+```python
+model = MyModel(...)
+model = kernelize(model, mode=Mode.TRAINING)
+```
+
+A model that is kernelized for training can also be used for inference, but
+not the other way around. If you want to change the mode of the kernelized
+model, you can just run `kernelize` on the model again with the new mode.
+
+If you want to compile a model with `torch.compile`, this should be indicated
+in the mode as well. You can do this by combining `Mode.INFERENCE` or
+`Mode.TRAINING` with `Mode.TORCH_COMPILE` using the set union (`|`) operator:
+
+```python
+model = MyModel(...)
+
+# Inference
+model = kernelize(model, mode=Mode.INFERENCE | Mode.TORCH_COMPILE)
+
+# Training
+model = kernelize(model, mode=Mode.TRAINING | Mode.TORCH_COMPILE)
+```
+
+When the `mode` argument is not specified,
+`Mode.TRAINING | Mode.TORCH_COMPILE` is used as the default. This mode
+aligns most closely with pure PyTorch layers which also support training
+and `torch.compile`. However, to select the most performant kernels, it
+is often good to make the mode specific as possible.
+
+### Kernel device
+
+Kernels can be registered per device type. For instance, separate `cuda` and
+`metal` kernels could be registered for the name `SiluAndMul`. By default,
+`kernelize` will try to infer the device type from the model's parameters.
+You can pass the device type to `kernelize` if the device type cannot be
+inferred (e.g. because the model has no parameters):
+
+```python
+model = MyModel(...)
+model = kernelize(model, device="cuda", mode=Mode.INFERENCE)
+```
+
+### Fallback `forward`
+
+If the `TRAINING` and/or `TORCH_COMPILE` modes are used, but a registered
+kernel does not support backward passes or `torch.compile` respectively,
+`kernenize` will fall back to the original, non-kernelized, layer. You
+can let `kernelize` raise an exception instead by using `use_fallback=False`:
+
+```python
+model = MyModel(...)
+model = kernelize(model, mode=Mode.INFERENCE | Mode.TORCH_COMPILE, use_fallback=False)
+```
+
+This can be useful if you want to guarantee that Hub kernels are used.
+
+### Inspecting kernels which kernels are used
+
+The kernels that are used are logged at the `INFO` level by `kernelize`.
+See the [Python logging](https://docs.python.org/3/library/logging.html)
+documentation for information on how to configure logging.
+
+## Registering a hub kernel for a layer
+
+`kernelize` relies on kernel mappings to find Hub kernels for layers.
+Kernel mappings map a kernel name such as `SiluAndMul` to a kernel on
+the Hub. For example:
+
+```python
+kernel_layer_mapping = {
+    "SiluAndMul": {
+        "cuda": LayerRepository(
+            repo_id="kernels-community/activation",
+            layer_name="SiluAndMul",
+        )
+    }
+}
+```
+
+You can register such a mapping using `register_kernel_mapping`:
+
+```python
+register_kernel_mapping(kernel_layer_mapping)
+```
+
+This will register the kernel mapping in the current context, which is
+normally global. It is recommended to scope the mapping to where it is
+used with the `use_kernel_mapping` context manager:
+
+```python
+with use_kernel_mapping(kernel_layer_mapping):
+    # Use the layer for which the mapping is applied.
+    model = kernelize(model)
+```
+
+This ensures that the mapping is not active anymore outside the
+`with`-scope.
+
+### Registering kernels for specific modes
+
+You might want to register two different kernels for a particular layer,
+where one kernel is optimized for a specific mode. You can do so by
+registering layer repositories for specific modes. For example:
+
+```python
+kernel_layer_mapping = {
+    "SiluAndMul": {
+        "cuda": {
+          Mode.INFERENCE: LayerRepository(
+              repo_id="kernels-community/activation-inference-optimized",
+              layer_name="SiluAndMul",
+          ),
+          Mode.TRAINING | Mode.TORCH_COMPILE: LayerRepository(
+              repo_id="kernels-community/activation-training-optimized",
+              layer_name="SiluAndMul",
+          ),
+      }
+    }
+}
+```
+
+The `kernelize` function will attempt to use the following registered
+kernels for a given mode:
+
+- `INFERENCE`: `INFERENCE` → `INFERENCE | TORCH_COMPILE` → `TRAINING` →
+  `TRAINING | TORCH_COMPILE` → `FALLBACK`
+- `INFERENCE | TORCH_COMPILE`: `INFERENCE | TORCH_COMPILE` →
+  `TRAINING | TORCH_COMPILE` → `FALLBACK`
+- `TRAINING`: `TRAINING` → `TRAINING | TORCH_COMPILE` → `FALLBACK`
+- `TRAINING | TORCH_COMPILE`: `TRAINING | TORCH_COMPILE` → `FALLBACK`
+
+`Mode.FALLBACK` is a special mode that is used when no other mode matches. It
+is also used when a kernel is registered without a mode, as described in the
+previous section.
+
+```python
+kernel_layer_mapping = {
+    "SiluAndMul": {
+        "cuda": {
+            Mode.FALLBACK: LayerRepository(
+                repo_id="kernels-community/activation",
+                layer_name="SiluAndMul",
+            ),
+            Mode.INFERENCE: LayerRepository(
+                repo_id="kernels-community/activation-inference-optimized",
+                layer_name="SiluAndMul",
+            ),
+            Mode.TRAINING: LayerRepository(
+                repo_id="kernels-community/activation-training-optimized",
+                layer_name="SiluAndMul",
+            ),
+        }
+    }
+}
+```
+
+In this case, both `Mode.INFERENCE | Mode.TORCH_COMPILE` and
+`Mode.TRAINING | Mode.TORCH_COMPILE` will use the `Mode.FALLBACK` kernel,
+since the other kernels do not support `torch.compile`.
+
+### Registering kernels for specific CUDA capabilities
+
+Some kernels only work with newer CUDA architectures. For instance, some
+kernels require capability 9.0 for the TMA unit on Hopper GPUs. `kernels`
+supports registering layers for a range of CUDA capabilities. To do so,
+you need to register the layer for a `Device` with type `cuda` and
+set the supported range of CUDA capabilities with using `CUDAProperties`:
+
+```python
+kernel_layer_mapping = {
+    "SiluAndMul": {
+        Device(
+            type="cuda",
+            properties=CUDAProperties(
+                min_capability=75, max_capability=89
+            ),
+        ): LayerRepository(
+            repo_id="kernels-community/activation",
+            layer_name="SiluAndMul",
+        ),
+        Device(
+            type="cuda",
+            properties=CUDAProperties(
+                min_capability=90, max_capability=sys.maxsize
+            ),
+        ): LayerRepository(
+            repo_id="kernels-community/activation-hopper",
+            layer_name="SiluAndMul",
+        ),
+    }
+}
+```
+
+Capabilities behave as follows:
+
+- The minimum and maximum capabilities are inclusive.
+- When a new kernel is registered with the same min/max capabilities as
+  an existing kernel, the new kernel will replace the old kernel.
+- When there are multiple kernels that support a capability, the kernel
+  with the smaller capability interval will be used. E.g. given:
+
+  - `KernelA` with `min_capability=80` and `max_capability=89`;
+  - `KernelB` with `min_capability=75` and `max_capability=89`;
+  - `kernelize` runs on a system with capability 8.6.
+
+  Then `KernelA` will be used because the interval 80..89 is smaller
+  than 75..89. The motivation is that kernels with smaller ranges
+  tend to be more optimized for a specific set of GPUs. **This behavior
+  might still change in the future.**

docs/locking.md

@@ -0,0 +1,62 @@
+# Locking kernel/layer versions
+
+Projects that use `setuptools` can lock the kernel versions that should be
+used. First specify the accepted versions in `pyproject.toml` and make
+sure that `kernels` is a build dependency:
+
+```toml
+[build-system]
+requires = ["kernels", "setuptools"]
+build-backend = "setuptools.build_meta"
+
+[tool.kernels.dependencies]
+"kernels-community/activation" = ">=0.0.1"
+```
+
+Then run `kernels lock .` in the project directory. This generates a `kernels.lock` file with
+the locked revisions. The locked revision will be used when loading a kernel with
+`get_locked_kernel`:
+
+```python
+from kernels import get_locked_kernel
+
+activation = get_locked_kernel("kernels-community/activation")
+```
+
+**Note:** the lock file is included in the package metadata, so it will only be visible
+to `kernels` after doing an (editable or regular) installation of your project.
+
+## Locked kernel layers
+
+Locking is also supported for kernel layers. To use locked layers, register them
+with the `LockedLayerRepository` class:
+
+```python
+kernel_layer_mapping = {
+    "SiluAndMul": {
+        "cuda": LockedLayerRepository(
+            repo_id="kernels-community/activation",
+            layer_name="SiluAndMul",
+        )
+    }
+}
+
+register_kernel_mapping(kernel_layer_mapping)
+```
+
+## Pre-downloading locked kernels
+
+Locked kernels can be pre-downloaded by running `kernels download .` in your
+project directory. This will download the kernels to your local Hugging Face
+Hub cache.
+
+The pre-downloaded kernels are used by the `get_locked_kernel` function.
+`get_locked_kernel` will download a kernel when it is not pre-downloaded. If you
+want kernel loading to error when a kernel is not pre-downloaded, you can use
+the `load_kernel` function instead:
+
+```python
+from kernels import load_kernel
+
+activation = load_kernel("kernels-community/activation")
+```

examples/basic.py

@@ -1,6 +1,6 @@
 import torch
 
-from hf_kernels import get_kernel
+from kernels import get_kernel
 
 print("Starting examples/basic.py demo")
 

flake.lock

@@ -0,0 +1,133 @@
+{
+  "nodes": {
+    "flake-compat": {
+      "locked": {
+        "lastModified": 1733328505,
+        "narHash": "sha256-NeCCThCEP3eCl2l/+27kNNK7QrwZB1IJCrXfrbv5oqU=",
+        "owner": "edolstra",
+        "repo": "flake-compat",
+        "rev": "ff81ac966bb2cae68946d5ed5fc4994f96d0ffec",
+        "type": "github"
+      },
+      "original": {
+        "owner": "edolstra",
+        "repo": "flake-compat",
+        "type": "github"
+      }
+    },
+    "flake-utils": {
+      "inputs": {
+        "systems": "systems"
+      },
+      "locked": {
+        "lastModified": 1731533236,
+        "narHash": "sha256-l0KFg5HjrsfsO/JpG+r7fRrqm12kzFHyUHqHCVpMMbI=",
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "rev": "11707dc2f618dd54ca8739b309ec4fc024de578b",
+        "type": "github"
+      },
+      "original": {
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "type": "github"
+      }
+    },
+    "flake-utils_2": {
+      "inputs": {
+        "systems": "systems_2"
+      },
+      "locked": {
+        "lastModified": 1731533236,
+        "narHash": "sha256-l0KFg5HjrsfsO/JpG+r7fRrqm12kzFHyUHqHCVpMMbI=",
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "rev": "11707dc2f618dd54ca8739b309ec4fc024de578b",
+        "type": "github"
+      },
+      "original": {
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "type": "github"
+      }
+    },
+    "hf-nix": {
+      "inputs": {
+        "flake-compat": "flake-compat",
+        "flake-utils": "flake-utils_2",
+        "nixpkgs": "nixpkgs"
+      },
+      "locked": {
+        "lastModified": 1750775451,
+        "narHash": "sha256-HiGqtwzIgUH7Xkh+wgpvHRZGooqrW0z663E6nauczA4=",
+        "owner": "huggingface",
+        "repo": "hf-nix",
+        "rev": "5943c3169e861618a6634bc8dbdb498e413ab9b7",
+        "type": "github"
+      },
+      "original": {
+        "owner": "huggingface",
+        "repo": "hf-nix",
+        "type": "github"
+      }
+    },
+    "nixpkgs": {
+      "locked": {
+        "lastModified": 1747820358,
+        "narHash": "sha256-fTqsZsUX6M3yeEvgyQvXcbGmT2CaRVyVwsi8eK29Oj4=",
+        "owner": "danieldk",
+        "repo": "nixpkgs",
+        "rev": "d3c1681180717528068082103bf323147de6ab0b",
+        "type": "github"
+      },
+      "original": {
+        "owner": "danieldk",
+        "ref": "cudatoolkit-12.9-kernel-builder",
+        "repo": "nixpkgs",
+        "type": "github"
+      }
+    },
+    "root": {
+      "inputs": {
+        "flake-utils": "flake-utils",
+        "hf-nix": "hf-nix",
+        "nixpkgs": [
+          "hf-nix",
+          "nixpkgs"
+        ]
+      }
+    },
+    "systems": {
+      "locked": {
+        "lastModified": 1681028828,
+        "narHash": "sha256-Vy1rq5AaRuLzOxct8nz4T6wlgyUR7zLU309k9mBC768=",
+        "owner": "nix-systems",
+        "repo": "default",
+        "rev": "da67096a3b9bf56a91d16901293e51ba5b49a27e",
+        "type": "github"
+      },
+      "original": {
+        "owner": "nix-systems",
+        "repo": "default",
+        "type": "github"
+      }
+    },
+    "systems_2": {
+      "locked": {
+        "lastModified": 1681028828,
+        "narHash": "sha256-Vy1rq5AaRuLzOxct8nz4T6wlgyUR7zLU309k9mBC768=",
+        "owner": "nix-systems",
+        "repo": "default",
+        "rev": "da67096a3b9bf56a91d16901293e51ba5b49a27e",
+        "type": "github"
+      },
+      "original": {
+        "owner": "nix-systems",
+        "repo": "default",
+        "type": "github"
+      }
+    }
+  },
+  "root": "root",
+  "version": 7
+}

flake.nix

@@ -0,0 +1,57 @@
+{
+  inputs = {
+    hf-nix.url = "github:huggingface/hf-nix";
+    nixpkgs.follows = "hf-nix/nixpkgs";
+    flake-utils.url = "github:numtide/flake-utils";
+  };
+  outputs =
+    {
+      self,
+      nixpkgs,
+      flake-utils,
+      hf-nix,
+    }:
+    flake-utils.lib.eachDefaultSystem (
+      system:
+      let
+        pkgs = import nixpkgs {
+          inherit system;
+          config = hf-nix.lib.config system;
+          overlays = [
+            hf-nix.overlays.default
+          ];
+        };
+      in
+      {
+        formatter = pkgs.nixfmt-tree;
+        devShells = with pkgs; rec {
+          default = mkShell {
+            buildInputs =
+              [
+                black
+                mypy
+                pyright
+                ruff
+              ]
+              ++ (with python3.pkgs; [
+                docutils
+                huggingface-hub
+                pytest
+                pytest-benchmark
+                pyyaml
+                torch
+                types-pyyaml
+                venvShellHook
+              ]);
+
+            venvDir = "./.venv";
+
+            postVenvCreation = ''
+              unset SOURCE_DATE_EPOCH
+              ( python -m pip install --no-build-isolation --no-dependencies -e . )
+            '';
+          };
+        };
+      }
+    );
+}

pyproject.toml

@@ -1,20 +1,21 @@
 [project]
-name = "hf-kernels"
+name = "kernels"
-version = "0.1.6"
+version = "0.8.1"
-description = "Download cuda kernels"
+description = "Download compute kernels"
 authors = [
   { name = "OlivierDehaene", email = "olivier@huggingface.co" },
   { name = "Daniel de Kok", email = "daniel@huggingface.co" },
   { name = "David Holtz", email = "david@huggingface.co" },
   { name = "Nicolas Patry", email = "nicolas@huggingface.co" },
 ]
+license = { text = "Apache-2.0" }
 readme = "README.md"
 requires-python = ">= 3.9"
 dependencies = [
-  "huggingface-hub>=0.26.3",
+  "huggingface_hub>=0.26.0,<1.0",
-  "packaging>=24.2",
+  "packaging>=20.0",
-  "tomli>=2.0.1; python_version<'3.11'",
+  "pyyaml>=6",
-  "torch>=2.4",
+  "tomli>=2.0; python_version<'3.11'",
 ]
 
 [build-system]
@@ -23,18 +24,47 @@ build-backend = "setuptools.build_meta"
 
 [dependency-groups]
 dev = [
+  "mypy >= 1.15.0",
   "pytest >=8",
   # Whatever version is compatible with pytest.
   "pytest-benchmark",
+  "torch >=2.5",
+  "types-pyyaml"
 ]
 
+[project.optional-dependencies]
+torch = ["torch"]
+
 [project.scripts]
-hf-kernels = "hf_kernels.cli:main"
+kernels = "kernels.cli:main"
 
 [project.entry-points."egg_info.writers"]
-"hf-kernels.lock" = "hf_kernels.lockfile:write_egg_lockfile"
+"kernels.lock" = "kernels.lockfile:write_egg_lockfile"
 
-#[build-system]
+
-#requires = ["torch", "huggingface_hub", "numpy", "tomli;python_version<='3.10'"]
+[tool.ruff]
-#build-backend = "hf_kernels.build"
+exclude = [
-#backend-path = ["src"]
+  ".eggs",
+  ".git",
+  ".git-rewrite",
+  ".hg",
+  ".mypy_cache",
+  ".nox",
+  ".pants.d",
+  ".pytype",
+  ".ruff_cache",
+  ".svn",
+  ".tox",
+  ".venv",
+  ".venv*",
+  "__pypackages__",
+  "_build",
+  "build",
+  "dist",
+  "venv",
+]
+line-length = 119
+# Ignored rules:
+# "E501" -> line length violation
+lint.ignore = ["E501"]
+lint.select = ["E", "F", "I", "W"]

pytest.ini

@@ -0,0 +1,4 @@
+[pytest]
+markers =
+    darwin_only: marks tests that should only run on macOS
+    linux_only: marks tests that should only run on Linux

src/hf_kernels/__init__.py

@@ -1,3 +0,0 @@
-from hf_kernels.utils import get_kernel, install_kernel, load_kernel, get_locked_kernel
-
-__all__ = ["get_kernel", "get_locked_kernel", "load_kernel", "install_kernel"]

src/hf_kernels/build.py

@@ -1,144 +0,0 @@
-"""
-Python shims for the PEP 517 and PEP 660 build backend.
-
-Major imports in this module are required to be lazy:
-```
-$ hyperfine \
-     "/usr/bin/python3 -c \"print('hi')\"" \
-     "/usr/bin/python3 -c \"from subprocess import check_call; print('hi')\""
-Base: Time (mean ± σ):      11.0 ms ±   1.7 ms    [User: 8.5 ms, System: 2.5 ms]
-With import: Time (mean ± σ):      15.2 ms ±   2.0 ms    [User: 12.3 ms, System: 2.9 ms]
-Base 1.38 ± 0.28 times faster than with import
-```
-
-The same thing goes for the typing module, so we use Python 3.10 type annotations that
-don't require importing typing but then quote them so earlier Python version ignore
-them while IDEs and type checker can see through the quotes.
-"""
-
-from hf_kernels.compat import tomllib
-
-TYPE_CHECKING = False
-if TYPE_CHECKING:
-    from collections.abc import Mapping, Sequence  # noqa:I001
-    from typing import Any  # noqa:I001
-
-
-def warn_config_settings(config_settings: "Mapping[Any, Any] | None" = None) -> None:
-    import sys
-
-    if config_settings:
-        print("Warning: Config settings are not supported", file=sys.stderr)
-
-
-def call(
-    args: "Sequence[str]", config_settings: "Mapping[Any, Any] | None" = None
-) -> str:
-    """Invoke a uv subprocess and return the filename from stdout."""
-    import shutil
-    import subprocess
-    import sys
-
-    warn_config_settings(config_settings)
-    # Unlike `find_uv_bin`, this mechanism must work according to PEP 517
-    import os
-
-    cwd = os.getcwd()
-    filename = os.path.join(cwd, "pyproject.toml")
-    with open(filename, "rb") as f:
-        data = tomllib.load(f)
-
-    for kernel, _ in (
-        data.get("tool", {}).get("hf-kernels", {}).get("dependencies", {}).items()
-    ):
-        from hf_kernels.utils import install_kernel
-
-        install_kernel(kernel, revision="main")
-    uv_bin = shutil.which("uv")
-    if uv_bin is None:
-        raise RuntimeError("uv was not properly installed")
-    # Forward stderr, capture stdout for the filename
-    result = subprocess.run([uv_bin, *args], stdout=subprocess.PIPE)
-    if result.returncode != 0:
-        sys.exit(result.returncode)
-    # If there was extra stdout, forward it (there should not be extra stdout)
-    stdout = result.stdout.decode("utf-8").strip().splitlines(keepends=True)
-    sys.stdout.writelines(stdout[:-1])
-    # Fail explicitly instead of an irrelevant stacktrace
-    if not stdout:
-        print("uv subprocess did not return a filename on stdout", file=sys.stderr)
-        sys.exit(1)
-    return stdout[-1].strip()
-
-
-def build_sdist(
-    sdist_directory: str, config_settings: "Mapping[Any, Any] | None" = None
-) -> str:
-    """PEP 517 hook `build_sdist`."""
-    args = ["build-backend", "build-sdist", sdist_directory]
-    return call(args, config_settings)
-
-
-def build_wheel(
-    wheel_directory: str,
-    config_settings: "Mapping[Any, Any] | None" = None,
-    metadata_directory: "str | None" = None,
-) -> str:
-    """PEP 517 hook `build_wheel`."""
-    args = ["build-backend", "build-wheel", wheel_directory]
-    if metadata_directory:
-        args.extend(["--metadata-directory", metadata_directory])
-    return call(args, config_settings)
-
-
-def get_requires_for_build_sdist(
-    config_settings: "Mapping[Any, Any] | None" = None,
-) -> "Sequence[str]":
-    """PEP 517 hook `get_requires_for_build_sdist`."""
-    warn_config_settings(config_settings)
-    return []
-
-
-def get_requires_for_build_wheel(
-    config_settings: "Mapping[Any, Any] | None" = None,
-) -> "Sequence[str]":
-    """PEP 517 hook `get_requires_for_build_wheel`."""
-    warn_config_settings(config_settings)
-    return []
-
-
-def prepare_metadata_for_build_wheel(
-    metadata_directory: str, config_settings: "Mapping[Any, Any] | None" = None
-) -> str:
-    """PEP 517 hook `prepare_metadata_for_build_wheel`."""
-    args = ["build-backend", "prepare-metadata-for-build-wheel", metadata_directory]
-    return call(args, config_settings)
-
-
-def build_editable(
-    wheel_directory: str,
-    config_settings: "Mapping[Any, Any] | None" = None,
-    metadata_directory: "str | None" = None,
-) -> str:
-    """PEP 660 hook `build_editable`."""
-    args = ["build-backend", "build-editable", wheel_directory]
-
-    if metadata_directory:
-        args.extend(["--metadata-directory", metadata_directory])
-    return call(args, config_settings)
-
-
-def get_requires_for_build_editable(
-    config_settings: "Mapping[Any, Any] | None" = None,
-) -> "Sequence[str]":
-    """PEP 660 hook `get_requires_for_build_editable`."""
-    warn_config_settings(config_settings)
-    return []
-
-
-def prepare_metadata_for_build_editable(
-    metadata_directory: str, config_settings: "Mapping[Any, Any] | None" = None
-) -> str:
-    """PEP 660 hook `prepare_metadata_for_build_editable`."""
-    args = ["build-backend", "prepare-metadata-for-build-editable", metadata_directory]
-    return call(args, config_settings)

src/hf_kernels/cli.py

@@ -1,92 +0,0 @@
-import argparse
-import dataclasses
-import json
-import sys
-from pathlib import Path
-
-from hf_kernels.compat import tomllib
-from hf_kernels.lockfile import KernelLock, get_kernel_locks
-from hf_kernels.utils import install_kernel, install_kernel_all_variants
-
-
-def main():
-    parser = argparse.ArgumentParser(
-        prog="hf-kernel", description="Manage compute kernels"
-    )
-    subparsers = parser.add_subparsers(required=True)
-
-    download_parser = subparsers.add_parser("download", help="Download locked kernels")
-    download_parser.add_argument(
-        "project_dir",
-        type=Path,
-        help="The project directory",
-    )
-    download_parser.add_argument(
-        "--all-variants",
-        action="store_true",
-        help="Download all build variants of the kernel",
-    )
-    download_parser.set_defaults(func=download_kernels)
-
-    lock_parser = subparsers.add_parser("lock", help="Lock kernel revisions")
-    lock_parser.add_argument(
-        "project_dir",
-        type=Path,
-        help="The project directory",
-    )
-    lock_parser.set_defaults(func=lock_kernels)
-
-    args = parser.parse_args()
-    args.func(args)
-
-
-def download_kernels(args):
-    lock_path = args.project_dir / "hf-kernels.lock"
-
-    if not lock_path.exists():
-        print(f"No hf-kernels.lock file found in: {args.project_dir}", file=sys.stderr)
-        sys.exit(1)
-
-    with open(args.project_dir / "hf-kernels.lock", "r") as f:
-        lock_json = json.load(f)
-
-    all_successful = True
-
-    for kernel_lock_json in lock_json:
-        kernel_lock = KernelLock.from_json(kernel_lock_json)
-        print(
-            f"Downloading `{kernel_lock.repo_id}` at with SHA: {kernel_lock.sha}",
-            file=sys.stderr,
-        )
-        if args.all_variants:
-            install_kernel_all_variants(kernel_lock.repo_id, kernel_lock.sha)
-        else:
-            try:
-                install_kernel(kernel_lock.repo_id, kernel_lock.sha)
-            except FileNotFoundError as e:
-                print(e, file=sys.stderr)
-                all_successful = False
-
-    if not all_successful:
-        sys.exit(1)
-
-
-def lock_kernels(args):
-    with open(args.project_dir / "pyproject.toml", "rb") as f:
-        data = tomllib.load(f)
-
-    kernel_versions = data.get("tool", {}).get("kernels", {}).get("dependencies", None)
-
-    all_locks = []
-    for kernel, version in kernel_versions.items():
-        all_locks.append(get_kernel_locks(kernel, version))
-
-    with open(args.project_dir / "hf-kernels.lock", "w") as f:
-        json.dump(all_locks, f, cls=_JSONEncoder, indent=2)
-
-
-class _JSONEncoder(json.JSONEncoder):
-    def default(self, o):
-        if dataclasses.is_dataclass(o):
-            return dataclasses.asdict(o)
-        return super().default(o)

src/hf_kernels/utils.py

@@ -1,178 +0,0 @@
-import ctypes
-import importlib
-import importlib.metadata
-import inspect
-import json
-import os
-from pathlib import Path
-import platform
-import sys
-from importlib.metadata import Distribution
-from types import ModuleType
-from typing import List, Optional, Tuple
-
-from huggingface_hub import hf_hub_download, snapshot_download
-from packaging.version import parse
-
-from hf_kernels.compat import tomllib
-from hf_kernels.lockfile import KernelLock
-
-CACHE_DIR: Optional[str] = os.environ.get("HF_KERNELS_CACHE", None)
-
-
-def build_variant():
-    import torch
-
-    if torch.version.cuda is None:
-        raise AssertionError("This kernel requires CUDA to be installed. Torch was not compiled with CUDA enabled.")
-
-    torch_version = parse(torch.__version__)
-    cuda_version = parse(torch.version.cuda)
-    cxxabi = "cxx11" if torch.compiled_with_cxx11_abi() else "cxx98"
-    cpu = platform.machine()
-    os = platform.system().lower()
-
-    return f"torch{torch_version.major}{torch_version.minor}-{cxxabi}-cu{cuda_version.major}{cuda_version.minor}-{cpu}-{os}"
-
-
-def import_from_path(module_name: str, file_path):
-    # We cannot use the module name as-is, after adding it to `sys.modules`,
-    # it would also be used for other imports. So, we make a module name that
-    # depends on the path for it to be unique using the hex-encoded hash of
-    # the path.
-    path_hash = "{:x}".format(ctypes.c_size_t(hash(file_path)).value)
-    module_name = f"{module_name}_{path_hash}"
-    spec = importlib.util.spec_from_file_location(module_name, file_path)
-    module = importlib.util.module_from_spec(spec)
-    sys.modules[module_name] = module
-    spec.loader.exec_module(module)
-    return module
-
-
-def install_kernel(
-    repo_id: str, revision: str, local_files_only: bool = False
-) -> Tuple[str, str]:
-    """Download a kernel for the current environment to the cache."""
-    package_name = repo_id.split('/')[-1]
-    package_name = package_name.replace('-', '_')
-    repo_path = snapshot_download(
-        repo_id,
-        allow_patterns=f"build/{build_variant()}/*",
-        cache_dir=CACHE_DIR,
-        revision=revision,
-        local_files_only=local_files_only,
-    )
-
-    variant_path = f"{repo_path}/build/{build_variant()}"
-    module_init_path = f"{variant_path}/{package_name}/__init__.py"
-
-    if not os.path.exists(module_init_path):
-        raise FileNotFoundError(
-            f"Kernel `{repo_id}` at revision {revision} does not have build: {build_variant()}"
-        )
-
-    return package_name, variant_path
-
-
-def install_kernel_all_variants(
-    repo_id: str, revision: str, local_files_only: bool = False
-):
-    snapshot_download(
-        repo_id,
-        allow_patterns="build/*",
-        cache_dir=CACHE_DIR,
-        revision=revision,
-        local_files_only=local_files_only,
-    )
-
-
-def get_metadata(repo_id: str, revision: str, local_files_only: bool = False):
-    with open(
-        hf_hub_download(
-            repo_id,
-            "build.toml",
-            cache_dir=CACHE_DIR,
-            revision=revision,
-            local_files_only=local_files_only,
-        ),
-        "rb",
-    ) as f:
-        return tomllib.load(f)
-
-
-def get_kernel(repo_id: str, revision: str = "main"):
-    package_name, package_path = install_kernel(repo_id, revision=revision)
-    return import_from_path(package_name, f"{package_path}/{package_name}/__init__.py")
-
-
-def load_kernel(repo_id: str):
-    """Get a pre-downloaded, locked kernel."""
-    locked_sha = _get_caller_locked_kernel(repo_id)
-
-    if locked_sha is None:
-        raise ValueError(f"Kernel `{repo_id}` is not locked")
-
-    filename = hf_hub_download(
-        repo_id,
-        "build.toml",
-        cache_dir=CACHE_DIR,
-        local_files_only=True,
-        revision=locked_sha,
-    )
-    with open(filename, "rb") as f:
-        metadata = tomllib.load(f)
-    package_name = metadata["torch"]["name"]
-
-    repo_path = os.path.dirname(filename)
-    package_path = f"{repo_path}/build/{build_variant()}"
-    return import_from_path(package_name, f"{package_path}/{package_name}/__init__.py")
-
-
-def get_locked_kernel(repo_id: str, local_files_only: bool = False):
-    """Get a kernel using a lock file."""
-    locked_sha = _get_caller_locked_kernel(repo_id)
-
-    if locked_sha is None:
-        raise ValueError(f"Kernel `{repo_id}` is not locked")
-
-    package_name, package_path = install_kernel(
-        repo_id, locked_sha, local_files_only=local_files_only
-    )
-
-    return import_from_path(package_name, f"{package_path}/{package_name}/__init__.py")
-
-
-def _get_caller_locked_kernel(repo_id: str) -> Optional[str]:
-    for dist in _get_caller_distributions():
-        lock_json = dist.read_text("hf-kernels.lock")
-        if lock_json is not None:
-            for kernel_lock_json in json.loads(lock_json):
-                kernel_lock = KernelLock.from_json(kernel_lock_json)
-                if kernel_lock.repo_id == repo_id:
-                    return kernel_lock.sha
-    return None
-
-
-def _get_caller_distributions() -> List[Distribution]:
-    module = _get_caller_module()
-    if module is None:
-        return []
-
-    # Look up all possible distributions that this module could be from.
-    package = module.__name__.split(".")[0]
-    dist_names = importlib.metadata.packages_distributions().get(package)
-    if dist_names is None:
-        return []
-
-    return [importlib.metadata.distribution(dist_name) for dist_name in dist_names]
-
-
-def _get_caller_module() -> Optional[ModuleType]:
-    stack = inspect.stack()
-    # Get first module in the stack that is not the current module.
-    first_module = inspect.getmodule(stack[0][0])
-    for frame in stack[1:]:
-        module = inspect.getmodule(frame[0])
-        if module is not None and module != first_module:
-            return module
-    return first_module

src/kernels/__init__.py

@@ -0,0 +1,37 @@
+from kernels.layer import (
+    CUDAProperties,
+    Device,
+    LayerRepository,
+    Mode,
+    kernelize,
+    register_kernel_mapping,
+    replace_kernel_forward_from_hub,
+    use_kernel_forward_from_hub,
+    use_kernel_mapping,
+)
+from kernels.utils import (
+    get_kernel,
+    get_local_kernel,
+    get_locked_kernel,
+    has_kernel,
+    install_kernel,
+    load_kernel,
+)
+
+__all__ = [
+    "CUDAProperties",
+    "Device",
+    "LayerRepository",
+    "Mode",
+    "get_kernel",
+    "get_local_kernel",
+    "get_locked_kernel",
+    "has_kernel",
+    "install_kernel",
+    "kernelize",
+    "load_kernel",
+    "register_kernel_mapping",
+    "replace_kernel_forward_from_hub",
+    "use_kernel_forward_from_hub",
+    "use_kernel_mapping",
+]

src/kernels/_interval_tree.py

@@ -0,0 +1,200 @@
+# AVL-balanced interval trees. We could use the intervaltree
+# packages, but it seems unmaintained and does not have type
+# annotations.
+
+from typing import Generic, List, Optional, Tuple, TypeVar
+
+T = TypeVar("T")
+
+
+class _Node(Generic[T]):
+    """A node in the interval tree."""
+
+    def __init__(self, start: int, end: int, data: T):
+        self.start: int = start
+        self.end: int = end
+        self.data: T = data
+        self.max_end: int = end
+        self.left: Optional["_Node[T]"] = None
+        self.right: Optional["_Node[T]"] = None
+        self.height: int = 1
+
+    def __repr__(self) -> str:
+        return f"Node({self.start}, {self.end})"
+
+
+class IntervalTree(Generic[T]):
+    """A data structure to hold and query (unique) intervals."""
+
+    root: Optional[_Node[T]]
+
+    def __init__(self):
+        self.root = None
+
+    def insert(self, start: int, end: int, data: T) -> None:
+        """
+        Inserts a new interval into the tree.
+
+        Args:
+            start: The starting point of the interval.
+            end: The ending point of the interval.
+            data: The data associated with this interval.
+        """
+        self.root = self._insert(self.root, start, end, data)
+
+    def _get_height(self, node: Optional[_Node[T]]) -> int:
+        if not node:
+            return 0
+        return node.height
+
+    def _get_balance(self, node: Optional[_Node[T]]) -> int:
+        if not node:
+            return 0
+        return self._get_height(node.left) - self._get_height(node.right)
+
+    def _update_node_attributes(self, node: _Node[T]) -> None:
+        node.height = 1 + max(self._get_height(node.left), self._get_height(node.right))
+        node.max_end = node.end
+        if node.left:
+            node.max_end = max(node.max_end, node.left.max_end)
+        if node.right:
+            node.max_end = max(node.max_end, node.right.max_end)
+
+    def _right_rotate(self, y: _Node[T]) -> _Node[T]:
+        """Performs a right rotation."""
+        x = y.left
+        assert x is not None
+        T2 = x.right
+
+        x.right = y
+        y.left = T2
+
+        self._update_node_attributes(y)
+        self._update_node_attributes(x)
+
+        return x
+
+    def _left_rotate(self, x: _Node[T]) -> _Node[T]:
+        """Performs a left rotation."""
+        y = x.right
+        assert y is not None
+        T2 = y.left
+
+        y.left = x
+        x.right = T2
+
+        self._update_node_attributes(x)
+        self._update_node_attributes(y)
+
+        return y
+
+    def _insert(
+        self, node: Optional[_Node[T]], start: int, end: int, data: T
+    ) -> _Node[T]:
+        """Recursive helper to insert a new node and balance the tree."""
+        if not node:
+            return _Node(start, end, data)
+
+        # Replace the data if the interval already exists.
+        if start == node.start and end == node.end:
+            node.data = data
+            return node
+
+        if start < node.start:
+            node.left = self._insert(node.left, start, end, data)
+        else:
+            node.right = self._insert(node.right, start, end, data)
+
+        self._update_node_attributes(node)
+
+        balance = self._get_balance(node)
+
+        # Left Left Case
+        if balance > 1 and node.left and start < node.left.start:
+            return self._right_rotate(node)
+
+        # Right Right Case
+        if balance < -1 and node.right and start >= node.right.start:
+            return self._left_rotate(node)
+
+        # Left Right Case
+        if balance > 1 and node.left and start >= node.left.start:
+            node.left = self._left_rotate(node.left)
+            return self._right_rotate(node)
+
+        # Right Left Case
+        if balance < -1 and node.right and start < node.right.start:
+            node.right = self._right_rotate(node.right)
+            return self._left_rotate(node)
+
+        return node
+
+    def search(self, point: int) -> List[T]:
+        """
+        Searches for all intervals that contain the given point.
+
+        Args:
+            point: The point to search for.
+
+        Returns:
+            A list of data items from all matching intervals.
+        """
+        results: List[T] = []
+        self._search(self.root, point, results)
+        return results
+
+    def _search(self, node: Optional[_Node[T]], point: int, results: List[T]) -> None:
+        """Recursive helper to find all overlapping intervals."""
+        if node is None or point > node.max_end:
+            return
+
+        if node.left:
+            self._search(node.left, point, results)
+
+        if node.start <= point <= node.end:
+            results.append(node.data)
+
+        if point >= node.start and node.right:
+            self._search(node.right, point, results)
+
+    def find_smallest_interval(self, point: int) -> Optional[T]:
+        """
+        Finds the item with the most specific (smallest) range for a given point.
+
+        Args:
+            point: The capability to look up.
+
+        Returns:
+            The data of the best-matching item, or None if no match is found.
+        """
+        matches: List[Tuple[int, int, T]] = []
+        self._find_with_intervals(self.root, point, matches)
+
+        if not matches:
+            return None
+
+        # Return the smallest interval, sort by memory location when
+        # there are multiple matches with the same interval size. This
+        # is just to ensure that we can compare against a trivial
+        # implementation in tests.
+        best_match = min(matches, key=lambda x: (x[1] - x[0], id(x[2])))
+        return best_match[2]
+
+    def _find_with_intervals(
+        self,
+        node: Optional[_Node[T]],
+        point: int,
+        results: List[Tuple[int, int, T]],
+    ) -> None:
+        """A modified search that collects interval ranges along with data."""
+        if node is None or point > node.max_end:
+            return
+
+        if node.left:
+            self._find_with_intervals(node.left, point, results)
+
+        if node.start <= point <= node.end:
+            results.append((node.start, node.end, node.data))
+
+        if point >= node.start and node.right:
+            self._find_with_intervals(node.right, point, results)

src/kernels/_vendored/convert_rst_to_mdx.py

@@ -0,0 +1,751 @@
+# coding=utf-8
+# Copyright 2021 The HuggingFace Team. 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
+#
+#     http://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.
+
+# Vendored from https://github.com/huggingface/doc-builder/blob/main/src/doc_builder/convert_rst_to_mdx.py
+
+import re
+
+# Re pattern to catch things inside ` ` in :obj:`thing`.
+_re_obj = re.compile(r":obj:`([^`]+)`")
+# Re pattern to catch things inside ` ` in :math:`thing`.
+_re_math = re.compile(r":math:`([^`]+)`")
+# Re pattern to catch things between single backquotes.
+_re_single_backquotes = re.compile(r"(^|[^`])`([^`]+)`([^`]|$)")
+# Re pattern to catch things between double backquotes.
+_re_double_backquotes = re.compile(r"(^|[^`])``([^`]+)``([^`]|$)")
+# Re pattern to catch things inside ` ` in :func/class/meth:`thing`.
+_re_func_class = re.compile(r":(?:func|class|meth):`([^`]+)`")
+
+
+def convert_rst_formatting(text):
+    """
+    Convert rst syntax for formatting to markdown in a given text.
+    """
+    # Remove :class:, :func: and :meth: markers. To code-links and put double backquotes
+    # (to not be caught by the italic conversion).
+    text = _re_func_class.sub(r"[``\1``]", text)
+    # Remove :obj: markers. What's after is in a single backquotes so we put in double backquotes
+    # (to not be caught by the italic conversion).
+    text = _re_obj.sub(r"``\1``", text)
+    # Remove :math: markers.
+    text = _re_math.sub(r"\\\\(\1\\\\)", text)
+    # Convert content in single backquotes to italic.
+    text = _re_single_backquotes.sub(r"\1*\2*\3", text)
+    # Convert content in double backquotes to single backquotes.
+    text = _re_double_backquotes.sub(r"\1`\2`\3", text)
+    # Remove remaining ::
+    text = re.sub(r"::\n", "", text)
+
+    # Remove new lines inside blocks in backsticks as they will be kept.
+    lines = text.split("\n")
+    in_code = False
+    text = None
+    for line in lines:
+        if in_code:
+            splits = line.split("`")
+            in_code = len(splits) > 1 and len(splits) % 2 == 1
+            if len(splits) == 1:
+                # Some forgotten lone backstick
+                text += "\n" + line
+            else:
+                text += " " + line.lstrip()
+        else:
+            if text is not None:
+                text += "\n" + line
+            else:
+                text = line
+            splits = line.split("`")
+            in_code = len(splits) % 2 == 0
+    return text
+
+
+# Re pattern to catch description and url in links of the form `description <url>`_.
+_re_links = re.compile(r"`([^`]+\S)\s+</*([^/][^>`]*)>`_+")
+# Re pattern to catch description and url in links of the form :prefix_link:`description <url>`_.
+_re_prefix_links = re.compile(r":prefix_link:`([^`]+\S)\s+</*([^/][^>`]*)>`")
+# Re pattern to catch reference in links of the form :doc:`reference`.
+_re_simple_doc = re.compile(r":doc:`([^`<]*)`")
+# Re pattern to catch description and reference in links of the form :doc:`description <reference>`.
+_re_doc_with_description = re.compile(r":doc:`([^`<]+\S)\s+</*([^/][^>`]*)>`")
+# Re pattern to catch reference in links of the form :ref:`reference`.
+_re_simple_ref = re.compile(r":ref:`([^`<]*)`")
+# Re pattern to catch description and reference in links of the form :ref:`description <reference>`.
+_re_ref_with_description = re.compile(r":ref:`([^`<]+\S)\s+<([^>]*)>`")
+
+
+def convert_rst_links(text, page_info):
+    """
+    Convert the rst links in text to markdown.
+    """
+    if "package_name" not in page_info:
+        raise ValueError("`page_info` must contain at least the package_name.")
+    package_name = page_info["package_name"]
+    version = page_info.get("version", "main")
+    language = page_info.get("language", "en")
+    no_prefix = page_info.get("no_prefix", False)
+
+    prefix = "" if no_prefix else f"/docs/{package_name}/{version}/{language}/"
+    # Links of the form :doc:`page`
+    text = _re_simple_doc.sub(rf"[\1]({prefix}\1)", text)
+    # Links of the form :doc:`text <page>`
+    text = _re_doc_with_description.sub(rf"[\1]({prefix}\2)", text)
+
+    if "page" in page_info and not no_prefix:
+        page = str(page_info["page"])
+        if page.endswith(".html"):
+            page = page[:-5]
+        prefix = f"{prefix}{page}"
+    else:
+        prefix = ""
+    # Refs of the form :ref:`page`
+    text = _re_simple_ref.sub(rf"[\1]({prefix}#\1)", text)
+    # Refs of the form :ref:`text <page>`
+    text = _re_ref_with_description.sub(rf"[\1]({prefix}#\2)", text)
+
+    # Links with a prefix
+    # TODO: when it exists, use the API to deal with prefix links properly.
+    prefix = f"https://github.com/huggingface/{package_name}/tree/main/"
+    text = _re_prefix_links.sub(rf"[\1]({prefix}\2)", text)
+    # Other links
+    text = _re_links.sub(r"[\1](\2)", text)
+    # Relative links or Transformers links need to remove the .html
+    if (
+        "(https://https://huggingface.co/" in text
+        or re.search(r"\(\.+/", text) is not None
+    ):
+        text = text.replace(".html", "")
+    return text
+
+
+# Re pattern that catches examples blocks of the form `Example::`.
+_re_example = re.compile(r"^\s*(\S.*)::\s*$")
+# Re pattern that catches rst blocks of the form `.. block_name::`.
+_re_block = re.compile(r"^\s*\.\.\s+(\S+)::")
+# Re pattern that catches what's after the :: in rst blocks of the form `.. block_name:: something`.
+_re_block_info = re.compile(r"^\s*\.\.\s+\S+::\s*(\S.*)$")
+
+
+def is_empty_line(line):
+    return len(line) == 0 or line.isspace()
+
+
+def find_indent(line):
+    """
+    Returns the number of spaces that start a line indent.
+    """
+    search = re.search(r"^(\s*)(?:\S|$)", line)
+    if search is None:
+        return 0
+    return len(search.groups()[0])
+
+
+_re_rst_option = re.compile(r"^\s*:(\S+):(.*)$")
+
+
+def convert_special_chars(text):
+    """
+    Converts { and < that have special meanings in MDX.
+    """
+    text = text.replace("{", "&amp;lcub;")
+    # We don't want to replace those by the HTML code, so we temporarily set them at LTHTML
+    text = re.sub(
+        r"<(img|br|hr|Youtube)", r"LTHTML\1", text
+    )  # html void elements with no closing counterpart
+    _re_lt_html = re.compile(r"<(\S+)([^>]*>)(((?!</\1>).)*)<(/\1>)", re.DOTALL)
+    while _re_lt_html.search(text):
+        text = _re_lt_html.sub(r"LTHTML\1\2\3LTHTML\5", text)
+    text = re.sub(r"(^|[^<])<([^<]|$)", r"\1&amp;lt;\2", text)
+    text = text.replace("LTHTML", "<")
+    return text
+
+
+def parse_options(block_content):
+    """
+    Parses the option in some rst block content.
+    """
+    block_lines = block_content.split("\n")
+    block_indent = find_indent(block_lines[0])
+    current_option = None
+    result = {}
+    for line in block_lines:
+        if _re_rst_option.search(line) is not None:
+            current_option, value = _re_rst_option.search(line).groups()
+            result[current_option] = value.lstrip()
+        elif find_indent(line) > block_indent:
+            result[current_option] += " " + line.lstrip()
+
+    return result
+
+
+def apply_min_indent(text, min_indent):
+    """
+    Make sure all lines in a text are have a minimum indentation.
+
+    Args:
+        text (`str`): The text to treat.
+        min_indent (`int`): The minimal indentation.
+
+    Returns:
+        `str`: The processed text.
+    """
+    lines = text.split("\n")
+    idx = 0
+    while idx < len(lines):
+        if is_empty_line(lines[idx]):
+            idx += 1
+            continue
+        indent = find_indent(lines[idx])
+        if indent < min_indent:
+            while idx < len(lines) and (
+                find_indent(lines[idx]) >= indent or is_empty_line(lines[idx])
+            ):
+                if not is_empty_line(lines[idx]):
+                    lines[idx] = " " * (min_indent - indent) + lines[idx]
+                idx += 1
+        else:
+            idx += 1
+
+    return "\n".join(lines)
+
+
+def convert_rst_blocks(text, page_info):
+    """
+    Converts rst special blocks (examples, notes) into MDX.
+    """
+    if "package_name" not in page_info:
+        raise ValueError("`page_info` must contain at least the package_name.")
+    package_name = page_info["package_name"]
+    version = page_info.get("version", "main")
+    language = page_info.get("language", "en")
+
+    lines = text.split("\n")
+    idx = 0
+    new_lines = []
+    while idx < len(lines):
+        block_type = None
+        block_info = None
+        if _re_block.search(lines[idx]) is not None:
+            block_type = _re_block.search(lines[idx]).groups()[0]
+            if _re_block_info.search(lines[idx]) is not None:
+                block_info = _re_block_info.search(lines[idx]).groups()[0]
+        elif _re_example.search(lines[idx]) is not None:
+            block_type = "code-block-example"
+            block_info = "python"
+            example_name = _re_example.search(lines[idx]).groups()[0]
+            new_lines.append(f"<exampletitle>{example_name}:</exampletitle>\n")
+        elif lines[idx].strip() == "..":
+            block_type = "comment"
+        elif lines[idx].strip() == "::":
+            block_type = "code-block"
+
+        if block_type is not None:
+            block_indent = find_indent(lines[idx])
+            # Find the next nonempty line
+            idx += 1
+            while idx < len(lines) and is_empty_line(lines[idx]):
+                idx += 1
+            # Grab the indent of the return line, this block will stop when we unindent under it (or has already)
+            example_indent = (
+                find_indent(lines[idx]) if idx < len(lines) else block_indent
+            )
+
+            if example_indent == block_indent:
+                block_content = ""
+            else:
+                block_lines = []
+                while idx < len(lines) and (
+                    is_empty_line(lines[idx])
+                    or find_indent(lines[idx]) >= example_indent
+                ):
+                    block_lines.append(lines[idx][example_indent:])
+                    idx += 1
+                block_content = "\n".join(block_lines)
+
+            if block_type in ["code", "code-block"]:
+                prefix = "```" if block_info is None else f"```{block_info}"
+                new_lines.append(f"{prefix}\n{block_content.strip()}\n```\n")
+            elif block_type == "code-block-example":
+                prefix = f"<example>```{block_info}"
+                new_lines.append(f"{prefix}\n{block_content.strip()}\n```\n</example>")
+            elif block_type == "note":
+                new_lines.append(
+                    apply_min_indent(
+                        f"<Tip>\n\n{block_content.strip()}\n\n</Tip>\n", block_indent
+                    )
+                )
+            elif block_type == "warning":
+                new_lines.append(
+                    apply_min_indent(
+                        "<Tip warning={true}>\n\n"
+                        + f"{block_content.strip()}\n\n</Tip>\n",
+                        block_indent,
+                    )
+                )
+            elif block_type == "raw":
+                new_lines.append(block_content.strip() + "\n")
+            elif block_type == "math":
+                new_lines.append(f"$${block_content.strip()}$$\n")
+            elif block_type == "comment":
+                new_lines.append(f"<!--{block_content.strip()}\n-->\n")
+            elif block_type == "autofunction":
+                if block_info is not None:
+                    new_lines.append(f"[[autodoc]] {block_info}\n")
+            elif block_type == "autoclass":
+                if block_info is not None:
+                    block = f"[[autodoc]] {block_info}\n"
+                    options = parse_options(block_content)
+                    if "special-members" in options:
+                        special_members = options["special-members"].split(", ")
+                        for special_member in special_members:
+                            block += f"    - {special_member}\n"
+                    if "members" in options:
+                        members = options["members"]
+                        if len(members) == 0:
+                            block += "    - all\n"
+                        else:
+                            for member in members.split(", "):
+                                block += f"    - {member}\n"
+                    new_lines.append(block)
+            elif block_type == "image":
+                options = parse_options(block_content)
+                target = options.pop("target", None)
+                if block_info is not None:
+                    options["src"] = block_info
+                else:
+                    if target is None:
+                        raise ValueError("Image source not defined.")
+                    options["src"] = target
+                # Adapt path
+                options["src"] = options["src"].replace(
+                    "/imgs/", f"/docs/{package_name}/{version}/{language}/imgs/"
+                )
+                html_code = " ".join(
+                    [f'{key}="{value}"' for key, value in options.items()]
+                )
+                new_lines.append(f"<img {html_code}/>\n")
+
+            else:
+                new_lines.append(
+                    f"{block_type},{block_info}\n{block_content.rstrip()}\n"
+                )
+
+        else:
+            new_lines.append(lines[idx])
+            idx += 1
+
+    return "\n".join(new_lines)
+
+
+# Re pattern that catches rst args blocks of the form `Parameters:`.
+_re_args = re.compile(r"^\s*(Args?|Arguments?|Attributes?|Params?|Parameters?):\s*$")
+# Re pattern that catches return blocks of the form `Return:`.
+_re_returns = re.compile(r"^\s*(Return|Yield|Raise)s?:\s*$")
+
+
+def split_return_line(line):
+    """
+    Split the return line with format `type: some doc`. Type may contain colons in the form of :obj: or :class:.
+    """
+    splits_on_colon = line.split(":")
+    idx = 1
+    while idx < len(splits_on_colon) and splits_on_colon[idx] in ["obj", "class"]:
+        idx += 2
+    if idx >= len(splits_on_colon):
+        if len(splits_on_colon) % 2 == 1 and re.search(r"`\w+`$", line.rstrip()):
+            return line, ""
+        return None, line
+    return ":".join(splits_on_colon[:idx]), ":".join(splits_on_colon[idx:])
+
+
+def split_raise_line(line):
+    """
+    Split the raise line with format `SomeError some doc`.
+    """
+    splits_on_colon = line.strip().split(" ")
+    error_type, doc = splits_on_colon[0], " ".join(splits_on_colon[1:])
+    if error_type and error_type[-1] == ":":
+        error_type = error_type[:-1]
+    return error_type, doc
+
+
+def split_arg_line(line):
+    """
+    Split the return line with format `type: some doc`. Type may contain colons in the form of :obj: or :class:.
+    """
+    splits_on_colon = line.split(":")
+    idx = 1
+    while idx < len(splits_on_colon) and splits_on_colon[idx] in ["obj", "class"]:
+        idx += 2
+    if idx >= len(splits_on_colon):
+        return line, ""
+    return ":".join(splits_on_colon[:idx]), ":".join(splits_on_colon[idx:])
+
+
+class InvalidRstDocstringError(ValueError):
+    pass
+
+
+_re_parameters = re.compile(
+    r"<parameters>(((?!<parameters>).)*)</parameters>", re.DOTALL
+)
+_re_md_link = re.compile(r"\[(.+)\]\(.+\)", re.DOTALL)
+
+
+def parse_rst_docstring(docstring):
+    """
+    Parses a docstring written in rst, in particular the list of arguments and the return type.
+    """
+    lines = docstring.split("\n")
+    idx = 0
+    while idx < len(lines):
+        # Parameters section
+        if _re_args.search(lines[idx]) is not None:
+            # Title of the section.
+            lines[idx] = "<parameters>\n"
+            # Find the next nonempty line
+            idx += 1
+            while is_empty_line(lines[idx]):
+                idx += 1
+            # Grab the indent of the list of parameters, this block will stop when we unindent under it or we see the
+            # Returns or Raises block.
+            param_indent = find_indent(lines[idx])
+            while (
+                idx < len(lines)
+                and find_indent(lines[idx]) == param_indent
+                and _re_returns.search(lines[idx]) is None
+            ):
+                intro, doc = split_arg_line(lines[idx])
+                # Line starting with a > after indent indicate a "section title" in the parameters.
+                if intro.lstrip().startswith(">"):
+                    lines[idx] = intro.lstrip()
+                else:
+                    lines[idx] = (
+                        re.sub(r"^\s*(\S+)(\s)?", r"- **\1**\2", intro) + " --" + doc
+                    )
+                idx += 1
+                while idx < len(lines) and (
+                    is_empty_line(lines[idx]) or find_indent(lines[idx]) > param_indent
+                ):
+                    idx += 1
+            lines.insert(idx, "</parameters>\n")
+            idx += 1
+
+        # Returns section
+        elif _re_returns.search(lines[idx]) is not None:
+            # tag is either `return` or `yield`
+            tag = _re_returns.match(lines[idx]).group(1).lower()
+            # Title of the section.
+            lines[idx] = f"<{tag}s>\n"
+            # Find the next nonempty line
+            idx += 1
+            while is_empty_line(lines[idx]):
+                idx += 1
+
+            # Grab the indent of the return line, this block will stop when we unindent under it.
+            return_indent = find_indent(lines[idx])
+            raised_errors = []
+            # The line may contain the return type.
+            if tag in ["return", "yield"]:
+                return_type, return_description = split_return_line(lines[idx])
+                lines[idx] = return_description
+                idx += 1
+                while idx < len(lines) and (
+                    is_empty_line(lines[idx])
+                    or find_indent(lines[idx]) >= return_indent
+                ):
+                    idx += 1
+            else:
+                while idx < len(lines) and find_indent(lines[idx]) == return_indent:
+                    return_type, return_description = split_raise_line(lines[idx])
+                    raised_error = re.sub(r"^\s*`?([\w\.]*)`?$", r"``\1``", return_type)
+                    lines[idx] = "- " + raised_error + " -- " + return_description
+                    md_link = _re_md_link.match(raised_error)
+                    if md_link:
+                        raised_error = md_link[1]
+                        raised_error = re.sub(
+                            r"^\s*`?([\w\.]*)`?$", r"``\1``", raised_error
+                        )
+                    if raised_error not in raised_errors:
+                        raised_errors.append(raised_error)
+                    idx += 1
+                    while idx < len(lines) and (
+                        is_empty_line(lines[idx])
+                        or find_indent(lines[idx]) > return_indent
+                    ):
+                        idx += 1
+
+            lines.insert(idx, f"</{tag}s>\n")
+            idx += 1
+
+            # Return block finished, we insert the return type if one was specified
+            if tag in ["return", "yield"] and return_type is not None:
+                lines[idx - 1] += f"\n<{tag}type>{return_type}</{tag}type>\n"
+            elif len(raised_errors) > 0:
+                # raised errors
+                lines[
+                    idx - 1
+                ] += f"\n<raisederrors>{' or '.join(raised_errors)}</raisederrors>\n"
+
+        else:
+            idx += 1
+
+    result = "\n".join(lines)
+
+    # combine multiple <parameters> blocks into one block
+    if result.count("<parameters>") > 1:
+        parameters_blocks = _re_parameters.findall(result)
+        parameters_blocks = [pb[0].strip() for pb in parameters_blocks]
+        parameters_str = "\n".join(parameters_blocks)
+        result = _re_parameters.sub("", result)
+        result += f"\n<parameters>{parameters_str}</parameters>\n"
+
+    return result
+
+
+_re_list = re.compile(r"^\s*(-|\*|\d+\.)\s")
+_re_autodoc = re.compile(r"^\s*\[\[autodoc\]\]\s+(\S+)\s*$")
+
+
+def remove_indent(text):
+    """
+    Remove indents in text, except the one linked to lists (or sublists).
+    """
+    lines = text.split("\n")
+    # List of indents to remember for nested lists
+    current_indents = []
+    # List of new indents to remember for nested lists
+    new_indents = []
+    is_inside_code = False
+    code_indent = 0
+    for idx, line in enumerate(lines):
+        # Line is an item in a list.
+        if _re_list.search(line) is not None:
+            indent = find_indent(line)
+            # Is it a new list / new level of nestedness?
+            if len(current_indents) == 0 or indent > current_indents[-1]:
+                current_indents.append(indent)
+                new_indent = 0 if len(new_indents) == 0 else new_indents[-1]
+                lines[idx] = " " * new_indent + line[indent:]
+                new_indent += len(_re_list.search(line).groups()[0]) + 1
+                new_indents.append(new_indent)
+            # Otherwise it's an existing level of list (current one, or previous one)
+            else:
+                # Let's find the proper level of indentation
+                level = len(current_indents) - 1
+                while level >= 0 and current_indents[level] != indent:
+                    level -= 1
+                current_indents = current_indents[: level + 1]
+                new_indents = new_indents[:level]
+                new_indent = 0 if len(new_indents) == 0 else new_indents[-1]
+                lines[idx] = " " * new_indent + line[indent:]
+                new_indent += len(_re_list.search(line).groups()[0]) + 1
+                new_indents.append(new_indent)
+
+        # Line is an autodoc, we keep the indent for the list just after if there is one.
+        elif _re_autodoc.search(line) is not None:
+            indent = find_indent(line)
+            current_indents = [indent]
+            new_indents = [4]
+            lines[idx] = line.strip()
+
+        # Deal with empty lines separately
+        elif is_empty_line(line):
+            lines[idx] = ""
+
+        # Code blocks
+        elif line.lstrip().startswith("```"):
+            is_inside_code = not is_inside_code
+            if is_inside_code:
+                code_indent = find_indent(line)
+            lines[idx] = line[code_indent:]
+        elif is_inside_code:
+            lines[idx] = line[code_indent:]
+
+        else:
+            indent = find_indent(line)
+            if len(current_indents) > 0 and indent > current_indents[-1]:
+                lines[idx] = " " * new_indents[-1] + line[indent:]
+            elif len(current_indents) > 0:
+                # Let's find the proper level of indentation
+                level = len(current_indents) - 1
+                while level >= 0 and current_indents[level] > indent:
+                    level -= 1
+                current_indents = current_indents[: level + 1]
+                if level >= 0:
+                    if current_indents[level] < indent:
+                        new_indents = new_indents[: level + 1]
+                    else:
+                        new_indents = new_indents[:level]
+                    new_indent = 0 if len(new_indents) == 0 else new_indents[-1]
+                    lines[idx] = " " * new_indent + line[indent:]
+                    new_indents.append(new_indent)
+                else:
+                    new_indents = []
+                    lines[idx] = line[indent:]
+            else:
+                lines[idx] = line[indent:]
+
+    return "\n".join(lines)
+
+
+def base_rst_to_mdx(text, page_info, unindent=True):
+    """
+    Convert a text from rst to mdx, with the base operations necessary for both docstrings and rst docs.
+    """
+    text = convert_rst_links(text, page_info)
+    text = convert_special_chars(text)
+    text = convert_rst_blocks(text, page_info)
+    # Convert * in lists to - to avoid the formatting conversion treat them as bold.
+    text = re.sub(r"^(\s*)\*(\s)", r"\1-\2", text, flags=re.MULTILINE)
+    text = convert_rst_formatting(text)
+    return remove_indent(text) if unindent else text
+
+
+def convert_rst_docstring_to_mdx(docstring, page_info):
+    """
+    Convert a docstring written in rst to mdx.
+    """
+    text = parse_rst_docstring(docstring)
+    return base_rst_to_mdx(text, page_info)
+
+
+def process_titles(lines):
+    """Converts rst titles to markdown titles."""
+    title_chars = """= - ` : ' " ~ ^ _ * + # < >""".split(" ")
+    title_levels = {}
+    new_lines = []
+    for line in lines:
+        if (
+            len(new_lines) > 0
+            and len(line) >= len(new_lines[-1])
+            and len(set(line)) == 1
+            and line[0] in title_chars
+            and line != "::"
+        ):
+            char = line[0]
+            level = title_levels.get(char, len(title_levels) + 1)
+            if level not in title_levels:
+                title_levels[char] = level
+            new_lines[-1] = f"{'#' * level} {new_lines[-1]}"
+        else:
+            new_lines.append(line)
+    return new_lines
+
+
+# Matches lines with a pattern of a table new line in rst.
+_re_ignore_line_table = re.compile(r"^(\+[\-\s]+)+\+\s*$")
+# Matches lines with a pattern of a table new line in rst, with a first column empty.
+_re_ignore_line_table1 = re.compile(r"^\|\s+(\+[\-\s]+)+\+\s*$")
+# Matches lines with a pattern of a first table line in rst.
+_re_sep_line_table = re.compile(r"^(\+[=\s]+)+\+\s*$")
+# Re pattern that catches anchors of the type .. reference:
+_re_anchor_section = re.compile(r"^\.\.\s+_(\S+):")
+
+
+def split_pt_tf_code_blocks(text):
+    """
+    Split PyTorch and TensorFlow specific block codes.
+    """
+    lines = text.split("\n")
+    new_lines = []
+    idx = 0
+    while idx < len(lines):
+        if lines[idx].startswith("```"):
+            code_lines = {"common": [lines[idx]], "pytorch": [], "tensorflow": []}
+            is_pytorch = False
+            is_tensorflow = False
+            idx += 1
+            while idx < len(lines) and lines[idx].strip() != "```":
+                if "## PYTORCH CODE" in lines[idx]:
+                    is_pytorch = True
+                    is_tensorflow = False
+                elif "## TENSORFLOW CODE" in lines[idx]:
+                    is_tensorflow = True
+                    is_pytorch = False
+                elif is_pytorch:
+                    code_lines["pytorch"].append(lines[idx])
+                elif is_tensorflow:
+                    code_lines["tensorflow"].append(lines[idx])
+                else:
+                    code_lines["common"].append(lines[idx])
+                idx += 1
+            if len(code_lines["pytorch"]) > 0 or len(code_lines["tensorflow"]) > 0:
+                block_lines = ["<frameworkcontent>", "<pt>"]
+                block_lines.extend(code_lines["common"].copy() + code_lines["pytorch"])
+                block_lines.extend(["```", "</pt>", "<tf>"])
+                block_lines.extend(
+                    code_lines["common"].copy() + code_lines["tensorflow"]
+                )
+                block_lines.extend(["```", "</tf>", "</frameworkcontent>"])
+                new_lines.extend(block_lines)
+            else:
+                block_lines = code_lines["common"] + ["```"]
+                new_lines.extend(block_lines)
+            idx += 1
+        else:
+            new_lines.append(lines[idx])
+            idx += 1
+    return "\n".join(new_lines)
+
+
+def convert_rst_to_mdx(rst_text, page_info, add_imports=True):
+    """
+    Convert a document written in rst to mdx.
+    """
+    lines = rst_text.split("\n")
+    lines = process_titles(lines)
+    if add_imports:
+        new_lines = [
+            '<script lang="ts">',
+            '	import Tip from "$lib/Tip.svelte";',
+            '	import Youtube from "$lib/Youtube.svelte";',
+            '	import Docstring from "$lib/Docstring.svelte";',
+            '	import CodeBlock from "$lib/CodeBlock.svelte";',
+            '	import CodeBlockFw from "$lib/CodeBlockFw.svelte";',
+            '	import DocNotebookDropdown from "$lib/DocNotebookDropdown.svelte";',
+            '	import CourseFloatingBanner from "$lib/CourseFloatingBanner.svelte";',
+            '	import IconCopyLink from "$lib/IconCopyLink.svelte";',
+            '	import FrameworkContent from "$lib/FrameworkContent.svelte";',
+            '	import Markdown from "$lib/Markdown.svelte";',
+            '	import ExampleCodeBlock from "$lib/ExampleCodeBlock.svelte";',
+            '	import Added from "$lib/Added.svelte";',
+            '	import Changed from "$lib/Changed.svelte";',
+            '	import Deprecated from "$lib/Deprecated.svelte";',
+            '	import PipelineIcon from "$lib/PipelineIcon.svelte";',
+            '	import PipelineTag from "$lib/PipelineTag.svelte";',
+            "	",
+            '	export let fw: "pt" | "tf"',
+            "</script>",
+            "<svelte:head>",
+            '<meta name="hf:doc:metadata" content={JSON.stringify(metadata)} >',
+            "</svelte:head>",
+            "",
+        ]
+    else:
+        new_lines = []
+    for line in lines:
+        if _re_ignore_line_table.search(line) is not None:
+            continue
+        elif _re_ignore_line_table1.search(line) is not None:
+            continue
+        elif _re_sep_line_table.search(line) is not None:
+            line = line.replace("=", "-").replace("+", "|")
+        elif _re_anchor_section.search(line) is not None:
+            anchor_name = _re_anchor_section.search(line).groups()[0]
+            line = f"<a id='{anchor_name}'></a>"
+        new_lines.append(line)
+    text = "\n".join(new_lines)
+
+    return split_pt_tf_code_blocks(base_rst_to_mdx(text, page_info))

src/kernels/_versions.py

@@ -0,0 +1,52 @@
+from typing import Dict, Optional
+
+from huggingface_hub import HfApi
+from huggingface_hub.hf_api import GitRefInfo
+from packaging.specifiers import SpecifierSet
+from packaging.version import InvalidVersion, Version
+
+
+def _get_available_versions(repo_id: str) -> Dict[Version, GitRefInfo]:
+    """Get kernel versions that are available in the repository."""
+    versions = {}
+    for tag in HfApi().list_repo_refs(repo_id).tags:
+        if not tag.name.startswith("v"):
+            continue
+        try:
+            versions[Version(tag.name[1:])] = tag
+        except InvalidVersion:
+            continue
+
+    return versions
+
+
+def resolve_version_spec_as_ref(repo_id: str, version_spec: str) -> GitRefInfo:
+    """
+    Get the locks for a kernel with the given version spec.
+
+    The version specifier can be any valid Python version specifier:
+    https://packaging.python.org/en/latest/specifications/version-specifiers/#version-specifiers
+    """
+    versions = _get_available_versions(repo_id)
+    requirement = SpecifierSet(version_spec)
+    accepted_versions = sorted(requirement.filter(versions.keys()))
+
+    if len(accepted_versions) == 0:
+        raise ValueError(
+            f"No version of `{repo_id}` satisfies requirement: {version_spec}"
+        )
+
+    return versions[accepted_versions[-1]]
+
+
+def select_revision_or_version(
+    repo_id: str, revision: Optional[str], version: Optional[str]
+) -> str:
+    if revision is not None and version is not None:
+        raise ValueError("Either a revision or a version must be specified, not both.")
+    elif revision is None and version is None:
+        revision = "main"
+    elif version is not None:
+        revision = resolve_version_spec_as_ref(repo_id, version).target_commit
+    assert revision is not None
+    return revision

src/kernels/cli.py

@@ -0,0 +1,160 @@
+import argparse
+import dataclasses
+import json
+import sys
+from pathlib import Path
+
+from kernels.compat import tomllib
+from kernels.lockfile import KernelLock, get_kernel_locks
+from kernels.utils import install_kernel, install_kernel_all_variants
+
+from .doc import generate_readme_for_kernel
+from .wheel import build_variant_to_wheel
+
+
+def main():
+    parser = argparse.ArgumentParser(
+        prog="kernel", description="Manage compute kernels"
+    )
+    subparsers = parser.add_subparsers(required=True)
+
+    download_parser = subparsers.add_parser("download", help="Download locked kernels")
+    download_parser.add_argument(
+        "project_dir",
+        type=Path,
+        help="The project directory",
+    )
+    download_parser.add_argument(
+        "--all-variants",
+        action="store_true",
+        help="Download all build variants of the kernel",
+    )
+    download_parser.set_defaults(func=download_kernels)
+
+    lock_parser = subparsers.add_parser("lock", help="Lock kernel revisions")
+    lock_parser.add_argument(
+        "project_dir",
+        type=Path,
+        help="The project directory",
+    )
+    lock_parser.set_defaults(func=lock_kernels)
+
+    to_wheel_parser = subparsers.add_parser(
+        "to-wheel", help="Convert a kernel to a wheel file"
+    )
+    to_wheel_parser.add_argument("repo_id", type=str, help="The kernel repo ID")
+    to_wheel_parser.add_argument("version", type=str, help="The kernel version")
+    to_wheel_parser.add_argument(
+        "--python-version",
+        type=str,
+        default="3.9",
+        help="The minimum Python version. Must match the Python version that the kernel was compiled for.",
+    )
+    to_wheel_parser.add_argument(
+        "--manylinux-version",
+        type=str,
+        default="2.28",
+        help="The manylinux version. Must match the manylinux version that the kernel was compiled for.",
+    )
+    to_wheel_parser.set_defaults(func=kernels_to_wheel)
+
+    # Add generate-readme subcommand parser
+    generate_readme_parser = subparsers.add_parser(
+        "generate-readme",
+        help="Generate README snippets for a kernel's public functions",
+    )
+    generate_readme_parser.add_argument(
+        "repo_id",
+        type=str,
+        help="The kernel repo ID (e.g., kernels-community/activation)",
+    )
+    generate_readme_parser.add_argument(
+        "--revision",
+        type=str,
+        default="main",
+        help="The kernel revision (branch, tag, or commit SHA, defaults to 'main')",
+    )
+    generate_readme_parser.set_defaults(
+        func=lambda args: generate_readme_for_kernel(
+            repo_id=args.repo_id, revision=args.revision
+        )
+    )
+
+    args = parser.parse_args()
+    args.func(args)
+
+
+def download_kernels(args):
+    lock_path = args.project_dir / "kernels.lock"
+
+    if not lock_path.exists():
+        print(f"No kernels.lock file found in: {args.project_dir}", file=sys.stderr)
+        sys.exit(1)
+
+    with open(args.project_dir / "kernels.lock", "r") as f:
+        lock_json = json.load(f)
+
+    all_successful = True
+
+    for kernel_lock_json in lock_json:
+        kernel_lock = KernelLock.from_json(kernel_lock_json)
+        print(
+            f"Downloading `{kernel_lock.repo_id}` at with SHA: {kernel_lock.sha}",
+            file=sys.stderr,
+        )
+        if args.all_variants:
+            install_kernel_all_variants(
+                kernel_lock.repo_id, kernel_lock.sha, variant_locks=kernel_lock.variants
+            )
+        else:
+            try:
+                install_kernel(
+                    kernel_lock.repo_id,
+                    kernel_lock.sha,
+                    variant_locks=kernel_lock.variants,
+                )
+            except FileNotFoundError as e:
+                print(e, file=sys.stderr)
+                all_successful = False
+
+    if not all_successful:
+        sys.exit(1)
+
+
+def kernels_to_wheel(args):
+    variants_path = install_kernel_all_variants(
+        repo_id=args.repo_id, revision=f"v{args.version}"
+    )
+    for variant_path in variants_path.iterdir():
+        if not variant_path.is_dir():
+            continue
+        wheel_path = build_variant_to_wheel(
+            manylinux_version=args.manylinux_version,
+            python_version=args.python_version,
+            repo_id=args.repo_id,
+            version=args.version,
+            variant_path=variant_path,
+            wheel_dir=Path("."),
+        )
+        print(f"☸️ {wheel_path.name}", file=sys.stderr)
+
+
+def lock_kernels(args):
+    with open(args.project_dir / "pyproject.toml", "rb") as f:
+        data = tomllib.load(f)
+
+    kernel_versions = data.get("tool", {}).get("kernels", {}).get("dependencies", None)
+
+    all_locks = []
+    for kernel, version in kernel_versions.items():
+        all_locks.append(get_kernel_locks(kernel, version))
+
+    with open(args.project_dir / "kernels.lock", "w") as f:
+        json.dump(all_locks, f, cls=_JSONEncoder, indent=2)
+
+
+class _JSONEncoder(json.JSONEncoder):
+    def default(self, o):
+        if dataclasses.is_dataclass(o):
+            return dataclasses.asdict(o)
+        return super().default(o)

src/kernels/doc.py

@@ -0,0 +1,242 @@
+import inspect
+import re
+import sys
+from types import ModuleType
+
+import yaml
+
+from ._vendored.convert_rst_to_mdx import convert_rst_docstring_to_mdx
+from .utils import get_kernel
+
+_RE_PARAMETERS = re.compile(
+    r"<parameters>(((?!<parameters>).)*)</parameters>", re.DOTALL
+)
+_RE_RETURNS = re.compile(r"<returns>(((?!<returns>).)*)</returns>", re.DOTALL)
+_RE_RETURNTYPE = re.compile(
+    r"<returntype>(((?!<returntype>).)*)</returntype>", re.DOTALL
+)
+
+
+def _extract_description_before_tags(docstring_mdx: str) -> str:
+    """Extract the description part of a docstring before any tags."""
+    params_pos = docstring_mdx.find("<parameters>")
+    returns_pos = docstring_mdx.find("<returns>")
+    returntype_pos = docstring_mdx.find("<returntype>")
+    positions = [pos for pos in [params_pos, returns_pos, returntype_pos] if pos != -1]
+
+    if positions:
+        first_tag_pos = min(positions)
+        return docstring_mdx[:first_tag_pos].strip()
+    else:
+        return docstring_mdx.strip()
+
+
+def _print_parameters_section(docstring_mdx: str, *, header_level: int) -> None:
+    """Print the parameters section from a docstring."""
+    matches = _RE_PARAMETERS.findall(docstring_mdx)
+    if matches:
+        header = "#" * header_level
+        print(f"\n{header} Parameters")
+        for match in matches:
+            print(f"\n{match[0].strip()}")
+
+
+def _print_returns_section(
+    docstring_mdx: str, *, context_name: str, header_level: int
+) -> None:
+    """Print the returns section from a docstring."""
+    return_matches = _RE_RETURNS.findall(docstring_mdx)
+    returntype_matches = _RE_RETURNTYPE.findall(docstring_mdx)
+
+    if return_matches or returntype_matches:
+        header = "#" * header_level
+        print(f"\n{header} Returns")
+
+        if returntype_matches:
+            if len(returntype_matches) > 1:
+                raise ValueError(
+                    f"More than one <returntype> tag found in docstring for {context_name}"
+                )
+            print(f"\n**Type**: {returntype_matches[0][0].strip()}")
+
+        if return_matches:
+            for match in return_matches:
+                print(f"\n{match[0].strip()}")
+
+
+def _get_docstring(obj, use_dict_check: bool = False) -> str:
+    """Get docstring from an object, with fallback to default message."""
+    # Check whether the class/method itself has docs and not just
+    # the superclass.
+    if use_dict_check:
+        has_doc = obj.__dict__.get("__doc__", None) is not None
+    else:
+        has_doc = getattr(obj, "__doc__", None) is not None
+
+    # We use inspect.getdoc because it does normalization.
+    doc = inspect.getdoc(obj)
+
+    return doc if has_doc and doc is not None else "No documentation available."
+
+
+def _process_and_print_docstring(
+    docstring: str, *, kernel_name: str, context_name: str, header_level: int
+) -> None:
+    """Convert docstring to MDX and print description, parameters, and returns sections."""
+    docstring_mdx = convert_rst_docstring_to_mdx(
+        docstring, page_info={"package_name": kernel_name}
+    )
+
+    # Print the description
+    description = _extract_description_before_tags(docstring_mdx)
+    print(f"\n{description}")
+
+    # Print parameters and returns sections
+    _print_parameters_section(docstring_mdx, header_level=header_level)
+    _print_returns_section(
+        docstring_mdx, context_name=context_name, header_level=header_level
+    )
+
+
+def generate_readme_for_kernel(repo_id: str, *, revision: str = "main") -> None:
+    kernel_module = get_kernel(repo_id=repo_id, revision=revision)
+    kernel_name = repo_id.split("/")[-1].replace("-", "_")
+
+    generate_metadata(kernel_module)
+    generate_kernel_doc(kernel_module, kernel_name)
+    generate_function_doc(kernel_module, kernel_name)
+    generate_layers_doc(kernel_module, kernel_name)
+
+
+def generate_metadata(module: ModuleType) -> None:
+    metadata = getattr(module, "__kernel_metadata__", {})
+    if "tags" not in metadata:
+        metadata["tags"] = ["kernel"]
+    else:
+        if "kernel" not in metadata["tags"]:
+            metadata["tags"].append("kernel")
+
+    print("---")
+    print(yaml.dump(metadata), end="")
+    print("---")
+
+
+def generate_kernel_doc(module: ModuleType, kernel_name: str) -> None:
+    docstring = module.__doc__.strip() if module.__doc__ is not None else None
+    if docstring:
+        title, rest = docstring.split("\n", 1)
+        print(f"# {title.strip()}")
+        print(
+            f"\n{convert_rst_docstring_to_mdx(rest.strip(), page_info={'package_name': kernel_name})}"
+        )
+
+
+def generate_function_doc(kernel_module: ModuleType, kernel_name: str) -> None:
+    print("\n## Functions")
+
+    # Track if we found any functions
+    found_functions = False
+
+    for name, func in inspect.getmembers(kernel_module, inspect.isfunction):
+        # Do not include imported functions.
+        if func.__module__ != kernel_module.__name__:
+            continue
+
+        # Exclude private functions.
+        if name.startswith("_"):
+            continue
+
+        found_functions = True
+
+        try:
+            sig = inspect.signature(func)
+            docstring = _get_docstring(func)
+        except ValueError:
+            print(
+                f"Warning: Could not retrieve signature for {name} in {kernel_module.__name__}",
+                file=sys.stderr,
+            )
+            continue
+
+        print(f"\n### Function `{name}`")
+        print(f"\n`{sig}`")
+
+        _process_and_print_docstring(
+            docstring, kernel_name=kernel_name, context_name=name, header_level=3
+        )
+
+    if not found_functions:
+        print("\nNo public top-level functions.")
+
+
+def generate_layers_doc(kernel_module: ModuleType, kernel_name: str) -> None:
+    # Check if layers module is available
+    layers_module = getattr(kernel_module, "layers", None)
+    if layers_module is None:
+        return
+
+    print("\n## Layers")
+
+    # Track if we found any classes
+    found_classes = False
+
+    for class_name, cls in inspect.getmembers(layers_module, inspect.isclass):
+        # Exclude classes that were imported.
+        if cls.__module__ != layers_module.__name__:
+            continue
+
+        found_classes = True
+
+        try:
+            # Get docstring, but not from superclasses.
+            class_docstring = _get_docstring(cls, use_dict_check=True)
+        except Exception:
+            print(
+                f"Warning: Could not retrieve documentation for class {class_name} in {layers_module.__name__}",
+                file=sys.stderr,
+            )
+            continue
+
+        print(f"\n### Class `{class_name}`")
+
+        # Always print class description (helper handles conversion and formatting)
+        class_docstring_mdx = convert_rst_docstring_to_mdx(
+            class_docstring, page_info={"package_name": kernel_name}
+        )
+        description = _extract_description_before_tags(class_docstring_mdx)
+        print(f"\n{description}")
+
+        # Document methods
+        print("\n#### Methods")
+
+        for method_name, method in inspect.getmembers(cls, inspect.isfunction):
+            # Note: also skip __init__, since extension layers cannot have a constructor.
+            if method_name.startswith("_"):
+                continue
+
+            # Skip methods from superclasses.
+            if method_name not in cls.__dict__:
+                continue
+
+            try:
+                sig = inspect.signature(method)
+                method_docstring = _get_docstring(method)
+            except ValueError:
+                print(
+                    f"Warning: Could not retrieve signature for {method_name} in {class_name}",
+                    file=sys.stderr,
+                )
+                continue
+
+            print(f"\n##### Method `{method_name}`")
+            print(f"\n`{sig}`")
+
+            _process_and_print_docstring(
+                method_docstring,
+                kernel_name=kernel_name,
+                context_name=method_name,
+                header_level=6,
+            )
+
+    if not found_classes:
+        print("\nNo layers defined.")

src/kernels/layer.py

@@ -0,0 +1,758 @@
+from __future__ import annotations
+
+import functools
+import inspect
+import logging
+import os
+import sys
+import warnings
+from abc import ABC, abstractmethod
+from contextvars import ContextVar
+from copy import deepcopy
+from dataclasses import dataclass
+from enum import Flag, auto
+from functools import lru_cache
+from pathlib import Path
+from types import MethodType
+from typing import (
+    TYPE_CHECKING,
+    Dict,
+    Optional,
+    Protocol,
+    Tuple,
+    Type,
+    Union,
+)
+
+from ._interval_tree import IntervalTree
+from ._versions import select_revision_or_version
+from .utils import _get_caller_locked_kernel, _get_locked_kernel, get_kernel
+
+if TYPE_CHECKING:
+    import torch
+    from torch import nn
+
+
+_DISABLE_KERNEL_MAPPING: bool = bool(int(os.environ.get("DISABLE_KERNEL_MAPPING", "0")))
+
+
+class Mode(Flag):
+    """
+    Kernelize mode
+
+    The `Mode` flag is used by `kernelize` to select kernels for the given
+    mode. Mappings can be registered for specific modes.
+
+    * `INFERENCE`: The kernel is used for inference.
+    * `TRAINING`: The kernel is used for training.
+    * `TORCH_COMPILE`: The kernel is used with `torch.compile`.
+    * `FALLBACK`: In a kernel mapping, this kernel is used when no other mode
+       matches.
+
+    Different modes can be combined. For instance, `INFERENCE | TORCH_COMPILE`
+    should be used for layers that are used for inference *with* `torch.compile`.
+    """
+
+    _NONE = 0
+    FALLBACK = auto()
+    TRAINING = auto()
+    INFERENCE = auto()
+    TORCH_COMPILE = auto()
+
+    def __or__(self, other: Mode) -> Mode:
+        union = super().__or__(other)
+
+        if Mode.INFERENCE in union and Mode.TRAINING in union:
+            raise ValueError("Mode.INFERENCE and Mode.TRAINING are mutually exclusive.")
+
+        if Mode.FALLBACK in union and union != Mode.FALLBACK:
+            raise ValueError("Mode.FALLBACK cannot be combined with other modes.")
+
+        return union
+
+
+@dataclass(frozen=True)
+class Device:
+    type: str
+    properties: Optional[CUDAProperties] = None
+
+    def __post_init__(self):
+        if self.properties is not None and isinstance(self.properties, CUDAProperties):
+            if self.type != "cuda":
+                raise ValueError("CUDAProperties is only supported for 'cuda' devices.")
+
+    def create_repo(self) -> _DeviceRepos:
+        """Create an appropriate repository set for this device type."""
+        if self.type == "cuda":
+            return _CUDARepos()
+        elif self.type == "mps":
+            return _MPSRepos()
+        else:
+            raise ValueError(f"Unknown device type: {self.type}")
+
+    def __eq__(self, other):
+        if not isinstance(other, Device):
+            return NotImplemented
+        return self.type == other.type and self.properties == other.properties
+
+    def __hash__(self):
+        return hash((self.type, self.properties))
+
+
+@dataclass(frozen=True)
+class CUDAProperties:
+    min_capability: int
+    max_capability: int
+
+    def __eq__(self, other):
+        if not isinstance(other, CUDAProperties):
+            return NotImplemented
+        return (
+            self.min_capability == other.min_capability
+            and self.max_capability == other.max_capability
+        )
+
+    def __hash__(self):
+        return hash((self.min_capability, self.max_capability))
+
+
+class LayerRepositoryProtocol(Protocol):
+    @property
+    def layer_name(self) -> str: ...
+
+    @property
+    def repo_id(self) -> str: ...
+
+    @property
+    def revision(self) -> str: ...
+
+
+class LayerRepository:
+    """
+    Repository and name of a layer.
+    """
+
+    def __init__(
+        self,
+        repo_id: str,
+        *,
+        layer_name: str,
+        revision: Optional[str] = None,
+        version: Optional[str] = None,
+    ):
+        """
+        Construct a layer repository.
+
+        Args:
+            repo_id (`str`): The Hub repository containing the layer.
+            revision (`str`, *optional*, defaults to `"main"`): The specific
+                revision (branch, tag, or commit) to download.
+                Cannot be used together with `version`.
+            version (`str`, *optional*): The kernel version to download. This
+                can be a Python version specifier, such as `">=1.0.0,<2.0.0"`.
+                Cannot be used together with `revision`.
+        """
+
+        if revision is not None and version is not None:
+            raise ValueError(
+                "Either a revision or a version must be specified, not both."
+            )
+
+        self.repo_id = repo_id
+        self.layer_name = layer_name
+
+        # We are going to resolve these lazily, since we do not want
+        # to do a network request for every registered LayerRepository.
+        self._revision = revision
+        self._version = version
+
+    @property
+    @functools.lru_cache()
+    def revision(self) -> str:
+        return select_revision_or_version(
+            repo_id=self.repo_id, revision=self._revision, version=self._version
+        )
+
+    def __eq__(self, other):
+        return (
+            isinstance(other, LayerRepository)
+            and self.layer_name == other.layer_name
+            and self.repo_id == other.repo_id
+            and self._revision == other._revision
+            and self._version == other._version
+        )
+
+    def __hash__(self):
+        return hash((self.layer_name, self.repo_id, self._revision, self._version))
+
+
+class LockedLayerRepository:
+    """
+    Repository and name of a layer.
+
+    In contrast to `LayerRepository`, this class uses repositories that
+    are locked inside a project.
+    """
+
+    def __init__(
+        self,
+        repo_id: str,
+        *,
+        lockfile: Optional[Path] = None,
+        layer_name: str,
+    ):
+        """
+        Construct a layer repository.
+
+        Args:
+            repo_id (`str`): The Hub repository containing the layer.
+        """
+        self.repo_id = repo_id
+        self.lockfile = lockfile
+        self.layer_name = layer_name
+
+    @property
+    @functools.lru_cache()
+    def revision(self) -> str:
+        if self.lockfile is None:
+            locked_sha = _get_caller_locked_kernel(self.repo_id)
+        else:
+            with open(self.lockfile, "r") as f:
+                locked_sha = _get_locked_kernel(self.repo_id, f.read())
+
+        if locked_sha is None:
+            raise ValueError(f"Kernel `{self.repo_id}` is not locked")
+
+        return locked_sha
+
+    def __eq__(self, other):
+        return (
+            isinstance(other, LockedLayerRepository)
+            and self.layer_name == other.layer_name
+            and self.repo_id == other.repo_id
+        )
+
+    def __hash__(self):
+        return hash((self.layer_name, self.repo_id))
+
+
+_CACHED_LAYER: Dict[LayerRepositoryProtocol, Type["nn.Module"]] = {}
+
+
+class _DeviceRepos(ABC):
+    """
+    Device-specific kernel layer repositories.
+    """
+
+    @property
+    @abstractmethod
+    def repos(
+        self,
+    ) -> Optional[Dict[Mode, LayerRepositoryProtocol]]: ...
+
+    @abstractmethod
+    def insert(self, device: Device, repos: Dict[Mode, LayerRepositoryProtocol]):
+        """
+        Insert a repository for a specific device and mode.
+        """
+        ...
+
+
+class _MPSRepos(_DeviceRepos):
+    _repos: Dict[Mode, LayerRepositoryProtocol]
+
+    def __init__(self):
+        super().__init__()
+        self._repos = {}
+
+    @property
+    def repos(
+        self,
+    ) -> Optional[Dict[Mode, LayerRepositoryProtocol]]:
+        return self._repos
+
+    def insert(self, device: Device, repos: Dict[Mode, LayerRepositoryProtocol]):
+        if device.type != "mps":
+            raise ValueError(f"Device type must be 'mps', got {device.type}")
+
+        self._repos = repos
+
+
+class _CUDARepos(_DeviceRepos):
+    _repos: IntervalTree[Dict[Mode, LayerRepositoryProtocol]]
+
+    def __init__(self):
+        super().__init__()
+        self.repos_by_capability = IntervalTree()
+
+    @property
+    def repos(
+        self,
+    ) -> Optional[Dict[Mode, LayerRepositoryProtocol]]:
+        capability = _find_capability()
+        return self.repos_by_capability.find_smallest_interval(capability)
+
+    def insert(self, device: Device, repos: Dict[Mode, LayerRepositoryProtocol]):
+        assert device.properties is None or isinstance(
+            device.properties, CUDAProperties
+        )
+
+        min_capability = (
+            0 if device.properties is None else device.properties.min_capability
+        )
+        max_capability = (
+            sys.maxsize
+            if device.properties is None
+            else device.properties.max_capability
+        )
+
+        self.repos_by_capability.insert(min_capability, max_capability, repos)
+
+
+_KERNEL_MAPPING: ContextVar[Dict[str, Dict[str, _DeviceRepos]]] = ContextVar(
+    "_KERNEL_MAPPING", default={}
+)
+
+
+def use_kernel_mapping(
+    mapping: Dict[
+        str,
+        Dict[
+            Union[Device, str],
+            Union[LayerRepositoryProtocol, Dict[Mode, LayerRepositoryProtocol]],
+        ],
+    ],
+    *,
+    inherit_mapping: bool = True,
+):
+    """
+    Context manager that sets a mapping for a duration of the context.
+
+    When `inherit_mapping` is set to `True` the current mapping will be
+    extended by `mapping` inside the context. If it is `False`, only
+    `mapping` is used inside the context.
+    """
+
+    class ContextManager:
+        def __enter__(self):
+            # Mappings always stack on previous mappings.
+            if inherit_mapping:
+                self.token = _KERNEL_MAPPING.set(deepcopy(_KERNEL_MAPPING.get()))
+            else:
+                self.token = _KERNEL_MAPPING.set({})
+            register_kernel_mapping(mapping)
+
+        def __exit__(self, exc_type, exc_value, traceback):
+            _KERNEL_MAPPING.reset(self.token)
+
+    return ContextManager()
+
+
+def register_kernel_mapping(
+    mapping: Dict[
+        str,
+        Dict[
+            Union[Device, str],
+            Union[LayerRepositoryProtocol, Dict[Mode, LayerRepositoryProtocol]],
+        ],
+    ],
+):
+    """
+    Allows one to register a mapping between a layer name and the corresponding
+    kernel(s) to use, depending on the device. This should be used in conjunction
+    with `kernelize`.
+
+    Example usage:
+
+    ```python
+    from kernels import LayerRepository, register_kernel_mapping
+
+    kernel_layer_mapping = {
+      "LlamaRMSNorm": {
+          "cuda": LayerRepository(
+              repo_id="kernels-community/activation",
+              layer_name="RmsNorm",
+              revision="layers",
+          ),
+      },
+    }
+    register_kernel_mapping(kernel_layer_mapping)
+    ```
+    """
+    # Merge with existing mappings.
+    for new_kernel, new_device_repos in mapping.items():
+        device_repo = _KERNEL_MAPPING.get().setdefault(new_kernel, {})
+        for new_device, new_repo in new_device_repos.items():
+            device = (
+                Device(type=new_device) if isinstance(new_device, str) else new_device
+            )
+
+            if isinstance(new_repo, dict):
+                kernel_options = new_repo
+            else:
+                kernel_options = {Mode.FALLBACK: new_repo}
+
+            feature_repos = device_repo.setdefault(device.type, device.create_repo())
+            feature_repos.insert(device, kernel_options)
+
+
+def replace_kernel_forward_from_hub(
+    cls,
+    layer_name: str,
+):
+    """
+    Decorator that prepares a layer class to use a kernel from the Hugging Face Hub.
+
+    This decorator stores the layer name and original forward method, which will be used
+    by the kernelize function to replace the forward implementation with the appropriate
+    kernel from the hub.
+
+    Args:
+        cls: The layer class to decorate
+        layer_name: The name of the layer to use for kernel lookup
+    """
+    cls.kernel_layer_name = layer_name
+
+
+_MODE_FALLBACK_PRIORITY = {
+    Mode.INFERENCE: [
+        Mode.INFERENCE,
+        Mode.INFERENCE | Mode.TORCH_COMPILE,
+        Mode.TRAINING,
+        Mode.TRAINING | Mode.TORCH_COMPILE,
+        Mode.FALLBACK,
+    ],
+    Mode.TRAINING: [
+        Mode.TRAINING,
+        Mode.TRAINING | Mode.TORCH_COMPILE,
+        Mode.FALLBACK,
+    ],
+    Mode.INFERENCE
+    | Mode.TORCH_COMPILE: [
+        Mode.INFERENCE | Mode.TORCH_COMPILE,
+        Mode.TRAINING | Mode.TORCH_COMPILE,
+        Mode.FALLBACK,
+    ],
+    Mode.TRAINING
+    | Mode.TORCH_COMPILE: [
+        Mode.TRAINING | Mode.TORCH_COMPILE,
+        Mode.FALLBACK,
+    ],
+}
+
+
+def _select_repository(
+    repositories: Dict[Mode, LayerRepositoryProtocol],
+    *,
+    mode: Mode,
+) -> Optional[Tuple[LayerRepositoryProtocol, Mode]]:
+    # Get the fallback priority list for the requested mode
+    if mode not in _MODE_FALLBACK_PRIORITY:
+        raise ValueError(f"Unsupported mode: {mode}")
+
+    fallback_modes = _MODE_FALLBACK_PRIORITY[mode]
+
+    # Try each mode in priority order
+    for fallback_mode in fallback_modes:
+        if fallback_mode in repositories:
+            return (repositories[fallback_mode], fallback_mode)
+
+    return None
+
+
+def kernelize(
+    model: "nn.Module",
+    *,
+    mode: Mode = Mode.TRAINING | Mode.TORCH_COMPILE,
+    device: Optional[Union[str, "torch.device"]] = None,
+    use_fallback: bool = True,
+):
+    """
+    Iterate over all modules in the model and replace the `forward` method of
+    extensible layers for which kernels are registered using `register_kernel_mapping`
+    or `use_kernel_mapping`.
+
+    Args:
+        model: The PyTorch model to kernelize
+        mode: the mode that the kernel is going to be used in (e.g.
+            `Mode.TRAINING | Mode.TORCH_COMPILE` kernelizes the model for training
+            and `torch.compile`).
+        device: The device type to load kernels for. The device type will be inferred
+            from the parameters of the model when not provided.
+        use_fallback: Whether to use the original forward method of modules when no
+            compatible kernel could be found. If set to `False`, an exception will
+            be raised in such cases.
+
+    Returns:
+        The kernelized model
+    """
+    import torch
+
+    if mode == Mode.FALLBACK:
+        raise ValueError("Mode.FALLBACK can only be used to register kernel mappings.")
+
+    # Type check ignored because this causes a false negative on Python < 3.11.
+    # Looks similar to: https://github.com/python/mypy/issues/9642
+    # Remove once we start doing typing checks on >= 3.11.
+    if Mode.INFERENCE not in mode and Mode.TRAINING not in mode:  # type: ignore[operator]
+        raise ValueError("kernelize mode must contain Mode.INFERENCE or Mode.TRAINING.")
+
+    if device is None:
+        device_type = _find_device(model)
+    elif isinstance(device, str):
+        device_type = Device(type=torch.device(device).type)
+    else:
+        device_type = Device(device.type)
+
+    assert isinstance(device_type, Device)
+
+    for _, module in model.named_modules():
+        module_class = type(module)
+        if not hasattr(module_class, "kernel_layer_name"):
+            continue
+        layer_name = module_class.kernel_layer_name
+
+        if _DISABLE_KERNEL_MAPPING:
+            _replace_forward(module, module_class)
+            continue
+
+        kernel = _KERNEL_MAPPING.get().get(str(layer_name))
+
+        if kernel is None:
+            warnings.warn(
+                "\n"
+                f"No kernel mapping found for layer `{layer_name}`. "
+                f"Check if the layer name matches one of the kernels in the mapping or add the kernel "
+                f"you want to use to the mapping. Defaulting to original forward implementation."
+            )
+            if not use_fallback:
+                raise ValueError(f"No layer mapping for `{layer_name}`")
+            _replace_forward(module, module_class)
+            continue
+
+        # Get kernel options for the device
+        property_repos = kernel.get(device_type.type)
+
+        if property_repos is None:
+            if not use_fallback:
+                raise ValueError(
+                    f"No layer mapping for `{layer_name}` with device type `{device_type}`"
+                )
+            _replace_forward(module, module_class)
+            continue
+
+        repos = property_repos.repos
+
+        if repos is None:
+            if not use_fallback:
+                raise ValueError(
+                    f"No layer mapping for `{layer_name}` device `{device_type}` with the right properties"
+                )
+            _replace_forward(module, module_class)
+            continue
+
+        repo_with_mode = _select_repository(
+            repos,
+            mode=mode,
+        )
+
+        if repo_with_mode is None:
+            if not use_fallback:
+                raise ValueError(
+                    f"No repository for `{layer_name}` for configuration mode={mode}"
+                )
+            _replace_forward(module, module_class)
+            continue
+
+        repo, repo_mode = repo_with_mode
+
+        logging.info(
+            f"Using layer `{repo.layer_name}` from repo `{repo.repo_id}` (revision: {repo.revision}) for layer `{layer_name}`"
+        )
+        logging.debug(f"kernelize mode: {mode}, repo mode: {repo_mode}")
+
+        layer = _get_layer_memoize(repo, module_class)
+
+        # Ideally we would do validation on the mapping where we check that
+        # e.g. if a repo class is registered for TRAINING | TORCH_COMPILE,
+        # the actual layer is compatible with that. Unfortunately, this would
+        # mean that we have to pre-download everything.
+        _validate_layer_has_mode(
+            layer_name=layer_name, module=layer, repo=repo, repo_mode=repo_mode
+        )
+
+        _conditionally_replace_forward(
+            module=module,
+            layer=layer,
+            mode=mode,
+            use_fallback=use_fallback,
+        )
+
+    return model
+
+
+def use_kernel_forward_from_hub(layer_name: str):
+    """
+    Make a layer extensible using the name `layer_name`.
+    """
+
+    def decorator(cls):
+        replace_kernel_forward_from_hub(cls, layer_name)
+        return cls
+
+    return decorator
+
+
+def _get_kernel_layer(
+    *, repo_id: str, layer_name: str, revision: str
+) -> Type["nn.Module"]:
+    """Get a layer from a kernel."""
+
+    kernel = get_kernel(repo_id, revision=revision)
+
+    if getattr(kernel, "layers", None) is None:
+        raise ValueError(
+            f"Kernel `{repo_id}` at revision `{revision}` does not define any layers."
+        )
+
+    layer = getattr(kernel.layers, layer_name, None)
+    if layer is None:
+        raise ValueError(f"Layer `{layer_name}` not found in kernel `{repo_id}`.")
+    return layer
+
+
+def _validate_layer(*, check_cls, cls):
+    import torch.nn as nn
+
+    # The layer must have at least have the following properties: (1) it
+    # must be stateless; (2) the forward signature should correspond to
+    # the signature it is replacing; (3) forward should not call other
+    # methods.
+
+    if not issubclass(cls, nn.Module):
+        raise TypeError(f"Layer `{cls}` is not a Torch layer.")
+
+    # We verify statelessness by checking that the does not have its own
+    # constructor (since the constructor could add member variables)...
+    if cls.__init__ is not nn.Module.__init__:
+        raise TypeError("Layer must not override nn.Module constructor.")
+
+    # ... or predefined member variables.
+    torch_module_members = {name for name, _ in inspect.getmembers(nn.Module)}
+    cls_members = {name for name, _ in inspect.getmembers(cls)}
+    difference = cls_members - torch_module_members
+    # verify if : difference ⊄ {"can_torch_compile", "has_backward"}
+    if not difference <= {"can_torch_compile", "has_backward"}:
+        raise TypeError("Layer must not contain additional members.")
+
+    # Check whether the forward signatures are similar.
+    params = inspect.signature(cls.forward).parameters
+    ref_params = inspect.signature(check_cls.forward).parameters
+
+    if len(params) != len(ref_params):
+        raise TypeError(
+            "Forward signature does not match: different number of arguments."
+        )
+
+    for param, ref_param in zip(params.values(), ref_params.values()):
+        if param.kind != ref_param.kind:
+            raise TypeError(
+                f"Forward signature does not match: different kind of arguments ({param} ({param.kind}) and {ref_param} ({ref_param.kind})"
+            )
+
+
+def _find_device(model: "nn.Module") -> Device:
+    try:
+        param = next(model.parameters())
+    except StopIteration:
+        raise ValueError(
+            "Cannot determine model device, provide as `device` argument to `kernelize`."
+        )
+
+    return Device(type=param.device.type)
+
+
+@lru_cache
+def _find_capability() -> int:
+    import torch
+
+    major, minor = torch.cuda.get_device_capability(device=None)
+    return major * 10 + minor
+
+
+def _conditionally_replace_forward(
+    *,
+    module: "nn.Module",
+    layer: Type["nn.Module"],
+    mode: Mode,
+    use_fallback: bool,
+):
+    module_class = type(module)
+
+    # Switch to fallback if the mode is not supported by the layer.
+    # Note that this is useful even after _validate_layer_has_mode because
+    # layers registered with the FALLBACK mode never get rejected by
+    # _validate_layer_has_mode. For such layers, we want to fall back in
+    # case the layer does not support the given mode.
+    needs_fallback = Mode.TORCH_COMPILE in mode and not getattr(
+        layer, "can_torch_compile", False
+    )
+    needs_fallback |= Mode.TRAINING in mode and not getattr(layer, "has_backward", True)
+
+    if needs_fallback:
+        if use_fallback:
+            _replace_forward(module, module_class)
+        else:
+            raise ValueError(f"Available kernel does not support mode: {mode}")
+    else:
+        _replace_forward(module, layer)
+
+
+def _replace_forward(module: "nn.Module", layer: Type["nn.Module"]):
+    module.forward = MethodType(layer.forward, module)  # type: ignore[method-assign]
+
+
+def _validate_layer_has_mode(
+    *,
+    layer_name: str,
+    module: Type["nn.Module"],
+    repo: LayerRepositoryProtocol,
+    repo_mode: Mode,
+):
+    """
+    Check that a repository supports the mode that it was registered for.
+    """
+
+    if Mode.TRAINING in repo_mode and not getattr(module, "has_backward", True):
+        raise ValueError(
+            f"Layer `{repo.layer_name}` ({repo.repo_id}, revision: {repo.revision}) does not support backward.\n"
+            f"Was registered for `{layer_name}` with mode `{repo_mode}`"
+        )
+
+    if Mode.TORCH_COMPILE in repo_mode and not getattr(
+        module, "can_torch_compile", False
+    ):
+        raise ValueError(
+            f"Layer `{repo.layer_name}` ({repo.repo_id}, revision: {repo.revision}) does not support torch.compile.\n"
+            f"Was registered for `{layer_name}` with mode `{repo_mode}`"
+        )
+
+    return True
+
+
+def _get_layer_memoize(
+    repo: LayerRepositoryProtocol, module_class: Type["nn.Module"]
+) -> Type["nn.Module"]:
+    layer = _CACHED_LAYER.get(repo, None)
+    if layer is not None:
+        return layer
+
+    layer = _get_kernel_layer(
+        repo_id=repo.repo_id,
+        layer_name=repo.layer_name,
+        revision=repo.revision,
+    )
+    _validate_layer(check_cls=module_class, cls=layer)
+    _CACHED_LAYER[repo] = layer
+
+    return layer

src/kernels/lockfile.py

@@ -1,63 +1,42 @@
+import hashlib
 from dataclasses import dataclass
 from pathlib import Path
-from typing import Dict, List
+from typing import Dict, List, Tuple
 
 from huggingface_hub import HfApi
-from packaging.specifiers import SpecifierSet
-from packaging.version import InvalidVersion, Version
 
-from hf_kernels.compat import tomllib
+from kernels._versions import resolve_version_spec_as_ref
+from kernels.compat import tomllib
 
 
 @dataclass
-class FileLock:
+class VariantLock:
-    filename: str
+    hash: str
-    blob_id: str
+    hash_type: str = "git_lfs_concat"
 
 
 @dataclass
 class KernelLock:
     repo_id: str
     sha: str
-    files: List[FileLock]
+    variants: Dict[str, VariantLock]
 
     @classmethod
     def from_json(cls, o: Dict):
-        files = [FileLock(**f) for f in o["files"]]
+        variants = {
-        return cls(repo_id=o["repo_id"], sha=o["sha"], files=files)
+            variant: VariantLock(**lock) for variant, lock in o["variants"].items()
+        }
+        return cls(repo_id=o["repo_id"], sha=o["sha"], variants=variants)
 
 
-def _get_available_versions(repo_id: str):
+def get_kernel_locks(repo_id: str, version_spec: str) -> KernelLock:
-    """Get kernel versions that are available in the repository."""
-    versions = {}
-    for tag in HfApi().list_repo_refs(repo_id).tags:
-        if not tag.name.startswith("v"):
-            continue
-        try:
-            versions[Version(tag.name[1:])] = tag
-        except InvalidVersion:
-            continue
-
-    return versions
-
-
-def get_kernel_locks(repo_id: str, version_spec: str):
     """
     Get the locks for a kernel with the given version spec.
 
     The version specifier can be any valid Python version specifier:
     https://packaging.python.org/en/latest/specifications/version-specifiers/#version-specifiers
     """
-    versions = _get_available_versions(repo_id)
+    tag_for_newest = resolve_version_spec_as_ref(repo_id, version_spec)
-    requirement = SpecifierSet(version_spec)
-    accepted_versions = sorted(requirement.filter(versions.keys()))
-
-    if len(accepted_versions) == 0:
-        raise ValueError(
-            f"No version of `{repo_id}` satisfies requirement: {version_spec}"
-        )
-
-    tag_for_newest = versions[accepted_versions[-1]]
 
     r = HfApi().repo_info(
         repo_id=repo_id, revision=tag_for_newest.target_commit, files_metadata=True
@@ -72,17 +51,36 @@ def get_kernel_locks(repo_id: str, version_spec: str):
             f"Cannot get sibling information for {repo_id} for tag {tag_for_newest.name}"
         )
 
-    file_locks = []
+    variant_files: Dict[str, List[Tuple[bytes, str]]] = {}
     for sibling in r.siblings:
         if sibling.rfilename.startswith("build/torch"):
             if sibling.blob_id is None:
                 raise ValueError(f"Cannot get blob ID for {sibling.rfilename}")
 
-            file_locks.append(
+            path = Path(sibling.rfilename)
-                FileLock(filename=sibling.rfilename, blob_id=sibling.blob_id)
+            variant = path.parts[1]
-            )
+            filename = Path(*path.parts[2:])
 
-    return KernelLock(repo_id=repo_id, sha=r.sha, files=file_locks)
+            hash = sibling.lfs.sha256 if sibling.lfs is not None else sibling.blob_id
+
+            files = variant_files.setdefault(variant, [])
+
+            # Encode as posix for consistent slash handling, then encode
+            # as utf-8 for byte-wise sorting later.
+            files.append((filename.as_posix().encode("utf-8"), hash))
+
+    variant_locks = {}
+    for variant, files in variant_files.items():
+        m = hashlib.sha256()
+        for filename_bytes, hash in sorted(files):
+            # Filename as bytes.
+            m.update(filename_bytes)
+            # Git blob or LFS file hash as bytes.
+            m.update(bytes.fromhex(hash))
+
+        variant_locks[variant] = VariantLock(hash=f"sha256-{m.hexdigest()}")
+
+    return KernelLock(repo_id=repo_id, sha=r.sha, variants=variant_locks)
 
 
 def write_egg_lockfile(cmd, basename, filename):
@@ -101,7 +99,7 @@ def write_egg_lockfile(cmd, basename, filename):
     if kernel_versions is None:
         return
 
-    lock_path = cwd / "hf-kernels.lock"
+    lock_path = cwd / "kernels.lock"
     if not lock_path.exists():
         logging.warning(f"Lock file {lock_path} does not exist")
         # Ensure that the file gets deleted in editable installs.

src/kernels/utils.py

@@ -0,0 +1,424 @@
+import ctypes
+import hashlib
+import importlib
+import importlib.metadata
+import inspect
+import json
+import logging
+import os
+import platform
+import sys
+from importlib.metadata import Distribution
+from pathlib import Path
+from types import ModuleType
+from typing import Dict, List, Optional, Tuple
+
+from huggingface_hub import file_exists, snapshot_download
+from packaging.version import parse
+
+from kernels._versions import select_revision_or_version
+from kernels.lockfile import KernelLock, VariantLock
+
+
+def _get_cache_dir() -> Optional[str]:
+    """Returns the kernels cache directory."""
+    cache_dir = os.environ.get("HF_KERNELS_CACHE", None)
+    if cache_dir is not None:
+        logging.warning(
+            "HF_KERNELS_CACHE will be removed in the future, use KERNELS_CACHE instead"
+        )
+        return cache_dir
+
+    return os.environ.get("KERNELS_CACHE", None)
+
+
+CACHE_DIR: Optional[str] = _get_cache_dir()
+
+
+def build_variant() -> str:
+    import torch
+
+    if torch.version.cuda is not None:
+        cuda_version = parse(torch.version.cuda)
+        compute_framework = f"cu{cuda_version.major}{cuda_version.minor}"
+    elif torch.version.hip is not None:
+        rocm_version = parse(torch.version.hip.split("-")[0])
+        compute_framework = f"rocm{rocm_version.major}{rocm_version.minor}"
+    elif torch.backends.mps.is_available():
+        compute_framework = "metal"
+    elif hasattr(torch, "xpu") and torch.xpu.is_available():
+        compute_framework = "xpu"
+    else:
+        raise AssertionError(
+            "Torch was not compiled with CUDA, Metal, XPU, or ROCm enabled."
+        )
+
+    torch_version = parse(torch.__version__)
+    cpu = platform.machine()
+    os = platform.system().lower()
+
+    if os == "darwin":
+        cpu = "aarch64" if cpu == "arm64" else cpu
+        return f"torch{torch_version.major}{torch_version.minor}-{compute_framework}-{cpu}-{os}"
+
+    cxxabi = "cxx11" if torch.compiled_with_cxx11_abi() else "cxx98"
+
+    return f"torch{torch_version.major}{torch_version.minor}-{cxxabi}-{compute_framework}-{cpu}-{os}"
+
+
+def universal_build_variant() -> str:
+    # Once we support other frameworks, detection goes here.
+    return "torch-universal"
+
+
+def import_from_path(module_name: str, file_path: Path) -> ModuleType:
+    # We cannot use the module name as-is, after adding it to `sys.modules`,
+    # it would also be used for other imports. So, we make a module name that
+    # depends on the path for it to be unique using the hex-encoded hash of
+    # the path.
+    path_hash = "{:x}".format(ctypes.c_size_t(hash(file_path)).value)
+    module_name = f"{module_name}_{path_hash}"
+    spec = importlib.util.spec_from_file_location(module_name, file_path)
+    if spec is None:
+        raise ImportError(f"Cannot load spec for {module_name} from {file_path}")
+    module = importlib.util.module_from_spec(spec)
+    if module is None:
+        raise ImportError(f"Cannot load module {module_name} from spec")
+    sys.modules[module_name] = module
+    spec.loader.exec_module(module)  # type: ignore
+    return module
+
+
+def install_kernel(
+    repo_id: str,
+    revision: str,
+    local_files_only: bool = False,
+    variant_locks: Optional[Dict[str, VariantLock]] = None,
+) -> Tuple[str, Path]:
+    """
+    Download a kernel for the current environment to the cache.
+
+    The output path is validated againt `hash` when set.
+    """
+    package_name = package_name_from_repo_id(repo_id)
+    variant = build_variant()
+    universal_variant = universal_build_variant()
+    repo_path = Path(
+        snapshot_download(
+            repo_id,
+            allow_patterns=[f"build/{variant}/*", f"build/{universal_variant}/*"],
+            cache_dir=CACHE_DIR,
+            revision=revision,
+            local_files_only=local_files_only,
+        )
+    )
+
+    try:
+        return _load_kernel_from_path(repo_path, package_name, variant_locks)
+    except FileNotFoundError:
+        # Redo with more specific error message.
+        raise FileNotFoundError(
+            f"Kernel `{repo_id}` at revision {revision} does not have build: {variant}"
+        )
+
+
+def _load_kernel_from_path(
+    repo_path: Path,
+    package_name: str,
+    variant_locks: Optional[Dict[str, VariantLock]] = None,
+) -> Tuple[str, Path]:
+    variant = build_variant()
+    universal_variant = universal_build_variant()
+
+    variant_path = repo_path / "build" / variant
+    universal_variant_path = repo_path / "build" / universal_variant
+
+    if not variant_path.exists() and universal_variant_path.exists():
+        # Fall back to universal variant.
+        variant = universal_variant
+        variant_path = universal_variant_path
+
+    if variant_locks is not None:
+        variant_lock = variant_locks.get(variant)
+        if variant_lock is None:
+            raise ValueError(f"No lock found for build variant: {variant}")
+        validate_kernel(repo_path=repo_path, variant=variant, hash=variant_lock.hash)
+
+    module_init_path = variant_path / package_name / "__init__.py"
+
+    if not os.path.exists(module_init_path):
+        raise FileNotFoundError(
+            f"Kernel at path `{repo_path}` does not have build: {variant}"
+        )
+
+    return package_name, variant_path
+
+
+def install_kernel_all_variants(
+    repo_id: str,
+    revision: str,
+    local_files_only: bool = False,
+    variant_locks: Optional[Dict[str, VariantLock]] = None,
+) -> Path:
+    repo_path = Path(
+        snapshot_download(
+            repo_id,
+            allow_patterns="build/*",
+            cache_dir=CACHE_DIR,
+            revision=revision,
+            local_files_only=local_files_only,
+        )
+    )
+
+    if variant_locks is not None:
+        for entry in (repo_path / "build").iterdir():
+            variant = entry.parts[-1]
+
+            variant_lock = variant_locks.get(variant)
+            if variant_lock is None:
+                raise ValueError(f"No lock found for build variant: {variant}")
+
+            validate_kernel(
+                repo_path=repo_path, variant=variant, hash=variant_lock.hash
+            )
+
+    return repo_path / "build"
+
+
+def get_kernel(
+    repo_id: str, revision: Optional[str] = None, version: Optional[str] = None
+) -> ModuleType:
+    """
+    Load a kernel from the kernel hub.
+    This function downloads a kernel to the local Hugging Face Hub cache
+    directory (if it was not downloaded before) and then loads the kernel.
+    Args:
+        repo_id (`str`): The Hub repository containing the kernel.
+        revision (`str`, *optional*, defaults to `"main"`): The specific
+            revision (branch, tag, or commit) to download.
+            Cannot be used together with `version`.
+        version (`str`, *optional*): The kernel version to download. This
+            can be a Python version specifier, such as `">=1.0.0,<2.0.0"`.
+            Cannot be used together with `revision`.
+    Returns:
+        `ModuleType`: The imported kernel module.
+    Example:
+        ```python
+        from kernels import get_kernel
+        kernel = get_kernel("username/my-kernel")
+        result = kernel.kernel_function(input_data)
+        ```
+    """
+    revision = select_revision_or_version(repo_id, revision, version)
+    package_name, package_path = install_kernel(repo_id, revision=revision)
+    return import_from_path(package_name, package_path / package_name / "__init__.py")
+
+
+def get_local_kernel(repo_path: Path, package_name: str) -> ModuleType:
+    """
+    Import a kernel from a local kernel repository path.
+    """
+    package_name, package_path = _load_kernel_from_path(repo_path, package_name)
+    return import_from_path(package_name, package_path / package_name / "__init__.py")
+
+
+def has_kernel(
+    repo_id: str, revision: Optional[str] = None, version: Optional[str] = None
+) -> bool:
+    """
+    Check whether a kernel build exists for the current environment
+    (Torch version and compute framework).
+
+    Args:
+        repo_id (`str`): The Hub repository containing the kernel.
+        revision (`str`, *optional*, defaults to `"main"`): The specific
+            revision (branch, tag, or commit) to download.
+            Cannot be used together with `version`.
+        version (`str`, *optional*): The kernel version to download. This
+            can be a Python version specifier, such as `">=1.0.0,<2.0.0"`.
+            Cannot be used together with `revision`.
+    Returns:
+        `bool`: `true` if a kernel is avaialble for the current environment.
+    """
+    revision = select_revision_or_version(repo_id, revision, version)
+
+    package_name = package_name_from_repo_id(repo_id)
+    variant = build_variant()
+    universal_variant = universal_build_variant()
+
+    if file_exists(
+        repo_id,
+        revision=revision,
+        filename=f"build/{universal_variant}/{package_name}/__init__.py",
+    ):
+        return True
+
+    return file_exists(
+        repo_id,
+        revision=revision,
+        filename=f"build/{variant}/{package_name}/__init__.py",
+    )
+
+
+def load_kernel(repo_id: str, *, lockfile: Optional[Path] = None) -> ModuleType:
+    """
+    Get a pre-downloaded, locked kernel.
+
+    If `lockfile` is not specified, the lockfile will be loaded from the
+    caller's package metadata.
+    """
+    if lockfile is None:
+        locked_sha = _get_caller_locked_kernel(repo_id)
+    else:
+        with open(lockfile, "r") as f:
+            locked_sha = _get_locked_kernel(repo_id, f.read())
+
+    if locked_sha is None:
+        raise ValueError(
+            f"Kernel `{repo_id}` is not locked. Please lock it with `kernels lock <project>` and then reinstall the project."
+        )
+
+    package_name = package_name_from_repo_id(repo_id)
+
+    variant = build_variant()
+    universal_variant = universal_build_variant()
+
+    repo_path = Path(
+        snapshot_download(
+            repo_id,
+            allow_patterns=[f"build/{variant}/*", f"build/{universal_variant}/*"],
+            cache_dir=CACHE_DIR,
+            revision=locked_sha,
+            local_files_only=True,
+        )
+    )
+
+    variant_path = repo_path / "build" / variant
+    universal_variant_path = repo_path / "build" / universal_variant
+    if not variant_path.exists() and universal_variant_path.exists():
+        # Fall back to universal variant.
+        variant = universal_variant
+        variant_path = universal_variant_path
+
+    module_init_path = variant_path / package_name / "__init__.py"
+    if not os.path.exists(module_init_path):
+        raise FileNotFoundError(
+            f"Locked kernel `{repo_id}` does not have build `{variant}` or was not downloaded with `kernels download <project>`"
+        )
+
+    return import_from_path(package_name, variant_path / package_name / "__init__.py")
+
+
+def get_locked_kernel(repo_id: str, local_files_only: bool = False) -> ModuleType:
+    """Get a kernel using a lock file."""
+    locked_sha = _get_caller_locked_kernel(repo_id)
+
+    if locked_sha is None:
+        raise ValueError(f"Kernel `{repo_id}` is not locked")
+
+    package_name, package_path = install_kernel(
+        repo_id, locked_sha, local_files_only=local_files_only
+    )
+
+    return import_from_path(package_name, package_path / package_name / "__init__.py")
+
+
+def _get_caller_locked_kernel(repo_id: str) -> Optional[str]:
+    for dist in _get_caller_distributions():
+        lock_json = dist.read_text("kernels.lock")
+        if lock_json is None:
+            continue
+        locked_sha = _get_locked_kernel(repo_id, lock_json)
+        if locked_sha is not None:
+            return locked_sha
+    return None
+
+
+def _get_locked_kernel(repo_id: str, lock_json: str) -> Optional[str]:
+    for kernel_lock_json in json.loads(lock_json):
+        kernel_lock = KernelLock.from_json(kernel_lock_json)
+        if kernel_lock.repo_id == repo_id:
+            return kernel_lock.sha
+    return None
+
+
+def _get_caller_distributions() -> List[Distribution]:
+    module = _get_caller_module()
+    if module is None:
+        return []
+
+    # Look up all possible distributions that this module could be from.
+    package = module.__name__.split(".")[0]
+    dist_names = importlib.metadata.packages_distributions().get(package)
+    if dist_names is None:
+        return []
+
+    return [importlib.metadata.distribution(dist_name) for dist_name in dist_names]
+
+
+def _get_caller_module() -> Optional[ModuleType]:
+    stack = inspect.stack()
+    # Get first module in the stack that is not the current module.
+    first_module = inspect.getmodule(stack[0][0])
+    for frame in stack[1:]:
+        module = inspect.getmodule(frame[0])
+        if module is not None and module != first_module:
+            return module
+    return first_module
+
+
+def validate_kernel(*, repo_path: Path, variant: str, hash: str):
+    """Validate the given build variant of a kernel against a hasht."""
+    variant_path = repo_path / "build" / variant
+
+    # Get the file paths. The first element is a byte-encoded relative path
+    # used for sorting. The second element is the absolute path.
+    files: List[Tuple[bytes, Path]] = []
+    # Ideally we'd use Path.walk, but it's only available in Python 3.12.
+    for dirpath, _, filenames in os.walk(variant_path):
+        for filename in filenames:
+            file_abs = Path(dirpath) / filename
+
+            # Python likes to create files when importing modules from the
+            # cache, only hash files that are symlinked blobs.
+            if file_abs.is_symlink():
+                files.append(
+                    (
+                        file_abs.relative_to(variant_path).as_posix().encode("utf-8"),
+                        file_abs,
+                    )
+                )
+
+    m = hashlib.sha256()
+
+    for filename_bytes, full_path in sorted(files):
+        m.update(filename_bytes)
+
+        blob_filename = full_path.resolve().name
+        if len(blob_filename) == 40:
+            # SHA-1 hashed, so a Git blob.
+            m.update(git_hash_object(full_path.read_bytes()))
+        elif len(blob_filename) == 64:
+            # SHA-256 hashed, so a Git LFS blob.
+            m.update(hashlib.sha256(full_path.read_bytes()).digest())
+        else:
+            raise ValueError(f"Unexpected blob filename length: {len(blob_filename)}")
+
+    computedHash = f"sha256-{m.hexdigest()}"
+    if computedHash != hash:
+        raise ValueError(
+            f"Lock file specifies kernel with hash {hash}, but downloaded kernel has hash: {computedHash}"
+        )
+
+
+def git_hash_object(data: bytes, object_type: str = "blob"):
+    """Calculate git SHA1 of data."""
+    header = f"{object_type} {len(data)}\0".encode()
+    m = hashlib.sha1()
+    m.update(header)
+    m.update(data)
+    return m.digest()
+
+
+def package_name_from_repo_id(repo_id: str) -> str:
+    return repo_id.split("/")[-1].replace("-", "_")

src/kernels/wheel.py

@@ -0,0 +1,186 @@
+import email.policy
+import os
+from dataclasses import dataclass
+from email.message import Message
+from importlib.metadata import PackageNotFoundError, version
+from pathlib import Path
+from typing import Optional
+
+try:
+    KERNELS_VERSION = version("kernels")
+except PackageNotFoundError:
+    KERNELS_VERSION = "unknown"
+
+
+@dataclass
+class Metadata:
+    name: str
+    version: str
+    cuda_version: Optional[str]
+    cxx_abi_version: Optional[str]
+    torch_version: Optional[str]
+    os: Optional[str]
+    platform: Optional[str]
+
+    @property
+    def is_universal(self) -> bool:
+        return self.platform is None
+
+
+def build_variant_to_wheel(
+    repo_id: str,
+    *,
+    version: str,
+    variant_path: Path,
+    wheel_dir: Path,
+    manylinux_version: str = "2.28",
+    python_version: str = "3.9",
+) -> Path:
+    """
+    Create a wheel file from the variant path.
+    """
+    name = repo_id.split("/")[-1].replace("_", "-")
+    metadata = extract_metadata(name, version, variant_path)
+    return build_wheel(
+        metadata,
+        variant_path=variant_path,
+        wheel_dir=wheel_dir,
+        manylinux_version=manylinux_version,
+        python_version=python_version,
+    )
+
+
+def extract_metadata(name: str, version: str, variant_path: Path) -> Metadata:
+    """
+    Extract metadata from the variant path.
+    """
+    if variant_path.name == "torch-universal":
+        return Metadata(
+            name=name,
+            version=version,
+            cuda_version=None,
+            cxx_abi_version=None,
+            torch_version=None,
+            os=None,
+            platform=None,
+        )
+
+    if not variant_path.name.startswith("torch"):
+        raise ValueError("Currently only conversion of Torch kernels is supported.")
+
+    variant_parts = variant_path.name.removeprefix("torch").split("-")
+    if len(variant_parts) != 5:
+        raise ValueError(f"Invalid variant name: {variant_path.name}")
+
+    torch_version = f"{variant_parts[0][:-1]}.{variant_parts[0][-1:]}"
+    cpp_abi_version = variant_parts[1].removeprefix("cxx")
+    cuda_version = variant_parts[2].removeprefix("cu")
+    platform = variant_parts[3].replace("-", "_")
+    os = variant_parts[4]
+
+    return Metadata(
+        name=name,
+        version=version,
+        cuda_version=cuda_version,
+        cxx_abi_version=cpp_abi_version,
+        torch_version=torch_version,
+        os=os,
+        platform=platform,
+    )
+
+
+def build_wheel(
+    metadata: Metadata,
+    *,
+    variant_path: Path,
+    wheel_dir: Path,
+    manylinux_version: str = "2.28",
+    python_version: str = "3.9",
+) -> Path:
+    """
+    Build the wheel file.
+    """
+    try:
+        from wheel.wheelfile import WheelFile  # type: ignore
+    except ImportError:
+        raise ImportError(
+            "The 'wheel' package is required to build wheels. Please install it with: `pip install wheel`"
+        )
+
+    name = metadata.name.replace("-", "_")
+    python_version_flat = python_version.replace(".", "")
+
+    if metadata.is_universal:
+        python_tag = f"py{python_version_flat}"
+        abi_tag = "none"
+        platform_tag = "any"
+        wheel_filename = (
+            f"{name}-{metadata.version}-{python_tag}-{abi_tag}-{platform_tag}.whl"
+        )
+        dist_info_dir_name = f"{name}-{metadata.version}.dist-info"
+        root_is_purelib = "true"
+        requires_dist_torch = "torch"
+    else:
+        python_tag = f"cp{python_version_flat}"
+        abi_tag = "abi3"
+
+        if (
+            metadata.torch_version is None
+            or metadata.cuda_version is None
+            or metadata.cxx_abi_version is None
+            or metadata.os is None
+            or metadata.platform is None
+        ):
+            raise ValueError(
+                "Torch version, CUDA version, C++ ABI version, OS, and platform must be specified for non-universal wheels."
+            )
+
+        local_version = f"torch{metadata.torch_version.replace('.', '')}cu{metadata.cuda_version}cxx{metadata.cxx_abi_version}"
+
+        if metadata.os == "linux":
+            platform_tag = (
+                f"manylinux_{manylinux_version.replace('.', '_')}_{metadata.platform}"
+            )
+        else:
+            platform_tag = f"{metadata.os}_{metadata.platform.replace('-', '_')}"
+
+        wheel_filename = f"{name}-{metadata.version}+{local_version}-{python_tag}-{abi_tag}-{platform_tag}.whl"
+        dist_info_dir_name = f"{name}-{metadata.version}+{local_version}.dist-info"
+        root_is_purelib = "false"
+        requires_dist_torch = f"torch=={metadata.torch_version}.*"
+
+    wheel_path = wheel_dir / wheel_filename
+
+    wheel_msg = Message(email.policy.compat32)
+    wheel_msg.add_header("Wheel-Version", "1.0")
+    wheel_msg.add_header("Generator", f"kernels ({KERNELS_VERSION})")
+    wheel_msg.add_header("Root-Is-Purelib", root_is_purelib)
+    wheel_msg.add_header("Tag", f"{python_tag}-{abi_tag}-{platform_tag}")
+
+    metadata_msg = Message(email.policy.compat32)
+    metadata_msg.add_header("Metadata-Version", "2.1")
+    metadata_msg.add_header("Name", name)
+    metadata_msg.add_header("Version", metadata.version)
+    metadata_msg.add_header("Summary", f"{name} kernel")
+    metadata_msg.add_header("Requires-Python", ">=3.9")
+    metadata_msg.add_header("Requires-Dist", requires_dist_torch)
+
+    source_pkg_dir = variant_path / name
+
+    with WheelFile(wheel_path, "w") as wheel_file:
+        for root, dirnames, filenames in os.walk(source_pkg_dir):
+            for filename in filenames:
+                if filename.endswith(".pyc"):
+                    continue
+
+                abs_filepath = os.path.join(root, filename)
+                entry_name = os.path.relpath(abs_filepath, variant_path)
+                wheel_file.write(abs_filepath, entry_name)
+
+        wheel_metadata_path = os.path.join(dist_info_dir_name, "WHEEL")
+        wheel_file.writestr(wheel_metadata_path, str(wheel_msg).encode("utf-8"))
+
+        metadata_path = os.path.join(dist_info_dir_name, "METADATA")
+        wheel_file.writestr(metadata_path, str(metadata_msg).encode("utf-8"))
+
+    return wheel_path

tests/conftest.py

@@ -0,0 +1,10 @@
+import sys
+
+import pytest
+
+
+def pytest_runtest_setup(item):
+    if "linux_only" in item.keywords and not sys.platform.startswith("linux"):
+        pytest.skip("skipping Linux-only test on non-Linux platform")
+    if "darwin_only" in item.keywords and not sys.platform.startswith("darwin"):
+        pytest.skip("skipping macOS-only test on non-macOS platform")

tests/kernel_locking/kernels.lock

@@ -0,0 +1,94 @@
+[
+  {
+    "repo_id": "kernels-community/activation",
+    "sha": "fd6842e88f1f23f198551d78a4541b8eb07e0538",
+    "variants": {
+      "torch25-cxx11-cu118-x86_64-linux": {
+        "hash": "sha256-61e3e51b5b59b30d4a6ba943a5e6e4ef5a9c8260cc4bca40b9fb462c0777842b",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch25-cxx11-cu121-x86_64-linux": {
+        "hash": "sha256-baa6b872040730bd1d676c011381f6f626fb96189837b828f587c806af8994fa",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch25-cxx11-cu124-x86_64-linux": {
+        "hash": "sha256-c1ec7457847fa1f0e4ab43234dfc3cd0959977e03dc2ffe89b4f6b90970c7965",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch25-cxx98-cu118-x86_64-linux": {
+        "hash": "sha256-412f9c841f20741e42f2c6cdb8c7da0e33ab436b219975acffe18b62b97ecd7c",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch25-cxx98-cu121-x86_64-linux": {
+        "hash": "sha256-2fde7f97859506e000c1072b3916c0a75bc8cee750a9853ea8b68199e7b57bcd",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch25-cxx98-cu124-x86_64-linux": {
+        "hash": "sha256-93309986f39a64a5630378108154866f0545178fa8dfef9b8f8ccfef9a78608e",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch26-cxx11-cu118-x86_64-linux": {
+        "hash": "sha256-3284d3c64b76d92c1ee930bce8013aff307f16eefb16c2d5dea9f2ca70e71e1f",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch26-cxx11-cu124-x86_64-linux": {
+        "hash": "sha256-36a8c93773c08ddf8ef624a8a6b2866be26d1861450dfe1ecac0bed59f9ffa47",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch26-cxx11-cu126-aarch64-linux": {
+        "hash": "sha256-f5afb734520f587717665659798ff738a69e5ae1e34d4bd95624edd18fb165cd",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch26-cxx11-cu126-x86_64-linux": {
+        "hash": "sha256-940841a7cb44f76c9a896d8b39f5bc0e0420f1c4c05ae9423da96778de4d1f2c",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch26-cxx98-cu118-x86_64-linux": {
+        "hash": "sha256-8e0f907830c3acc8c6bebfc162c744012ff6973e8110d7bf8ecd74b492418204",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch26-cxx98-cu124-x86_64-linux": {
+        "hash": "sha256-0833414cbe658baec55b7ff63537cddccc973fe99e3c03008cced5e66e38b6c1",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch26-cxx98-cu126-aarch64-linux": {
+        "hash": "sha256-d94fa59a13a5b623b2071aadcd1e6c8477c4d557fd06ad144f15b46b1fc71aab",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch26-cxx98-cu126-x86_64-linux": {
+        "hash": "sha256-64784f5f2f9e232d0f2fd824fbc47eadde505e3c232f351bead5b04c429c65c2",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch27-cxx11-cu118-x86_64-linux": {
+        "hash": "sha256-bcba3765f061649bac0e5a9159bea8349ced4780e24a2330aa62ce0f8d3a9d78",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch27-cxx11-cu126-aarch64-linux": {
+        "hash": "sha256-e4625df5706af025c70bd824d952b928d9a2965eeaefda72fc47be0fae680c5e",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch27-cxx11-cu126-x86_64-linux": {
+        "hash": "sha256-7d7d3e655f34a7b03d5603d7c1ab723ef3efc823291762421a8b3a4aa51bd405",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch27-cxx11-cu128-aarch64-linux": {
+        "hash": "sha256-60e076194dcd55b32c5aca72f09816cba0fff52f340c8a063b17ff0577154d99",
+        "hash_type": "git_lfs_concat"
+      },
+      "torch27-cxx11-cu128-x86_64-linux": {
+        "hash": "sha256-f0a3802382efdcd78b40601187a9c416579a24ef2ed5a60d2296ef0951a89597",
+        "hash_type": "git_lfs_concat"
+      }
+    }
+  },
+  {
+    "repo_id": "kernels-community/triton-scaled-mm",
+    "sha": "af10d8c1affe8efce93d228c3e6e64ff673d493f",
+    "variants": {
+      "torch-universal": {
+        "hash": "sha256-b843c5f30b52b6c1c56fca28cb0cf453be71d6ce7d308f383dce71a8050f7b52",
+        "hash_type": "git_lfs_concat"
+      }
+    }
+  }
+]

tests/kernel_locking/pyproject.toml

@@ -0,0 +1,3 @@
+[tool.kernels.dependencies]
+"kernels-community/activation" = ">=0.0.2"
+"kernels-community/triton-scaled-mm" = ">=0.0.2"

tests/layer_locking/kernels.lock

@@ -0,0 +1,12 @@
+[
+  {
+    "repo_id": "kernels-test/versions",
+    "sha": "dc142fd6c9920c993d32be6358b78957c58681c3",
+    "variants": {
+      "torch-universal": {
+        "hash": "sha256-35ce0ccfe68e392cbc06feef72268f4c41a74b9920496a2c6ee8978db7f7c17c",
+        "hash_type": "git_lfs_concat"
+      }
+    }
+  }
+]

tests/layer_locking/pyproject.toml

@@ -0,0 +1,2 @@
+[tool.kernels.dependencies]
+"kernels-test/versions" = ">=0.1.0,<0.2.0"

tests/test_basic.py

@@ -1,6 +1,7 @@
 import pytest
 import torch
-from hf_kernels import get_kernel
+
+from kernels import get_kernel, get_local_kernel, has_kernel, install_kernel
 
 
 @pytest.fixture
@@ -8,6 +9,24 @@ def kernel():
     return get_kernel("kernels-community/activation")
 
 
+@pytest.fixture
+def local_kernel():
+    package_name, path = install_kernel("kernels-community/activation", "main")
+    # Path is the build variant path (build/torch-<...>), so the grandparent
+    # is the kernel repository path.
+    return get_local_kernel(path.parent.parent, package_name)
+
+
+@pytest.fixture
+def metal_kernel():
+    return get_kernel("kernels-test/relu-metal")
+
+
+@pytest.fixture
+def universal_kernel():
+    return get_kernel("kernels-community/triton-scaled-mm")
+
+
 @pytest.fixture
 def device():
     if not torch.cuda.is_available():
@@ -15,6 +34,7 @@ def device():
     return "cuda"
 
 
+@pytest.mark.linux_only
 def test_gelu_fast(kernel, device):
     x = torch.arange(1, 10, dtype=torch.float16, device=device).view(3, 3)
     y = torch.empty_like(x)
@@ -28,3 +48,78 @@ def test_gelu_fast(kernel, device):
     )
 
     assert torch.allclose(y, expected)
+
+
+@pytest.mark.linux_only
+def test_local_kernel(local_kernel, device):
+    x = torch.arange(1, 10, dtype=torch.float16, device=device).view(3, 3)
+    y = torch.empty_like(x)
+
+    local_kernel.gelu_fast(y, x)
+
+    expected = torch.tensor(
+        [[0.8408, 1.9551, 2.9961], [4.0000, 5.0000, 6.0000], [7.0000, 8.0000, 9.0000]],
+        device=device,
+        dtype=torch.float16,
+    )
+
+    assert torch.allclose(y, expected)
+
+
+@pytest.mark.darwin_only
+@pytest.mark.parametrize("dtype", [torch.float16, torch.float32])
+def test_relu_metal(metal_kernel, dtype):
+    x = torch.arange(-10, 10, dtype=dtype, device="mps")
+    y = metal_kernel.relu(x)
+    assert torch.allclose(y, torch.relu(x))
+
+
+@pytest.mark.linux_only
+@pytest.mark.parametrize(
+    "kernel_exists",
+    [
+        ("kernels-community/activation", "main", True),
+        ("kernels-community/triton-layer-norm", "main", True),
+        # Repo only contains Torch 2.4 kernels (and we don't
+        # support/test against this version).
+        ("kernels-test/only-torch-2.4", "main", False),
+        ("google-bert/bert-base-uncased", "87565a309", False),
+    ],
+)
+def test_has_kernel(kernel_exists):
+    repo_id, revision, kernel = kernel_exists
+    assert has_kernel(repo_id, revision=revision) == kernel
+
+
+def test_version():
+    kernel = get_kernel("kernels-test/versions")
+    assert kernel.version() == "0.2.0"
+    kernel = get_kernel("kernels-test/versions", version="<1.0.0")
+    assert kernel.version() == "0.2.0"
+    kernel = get_kernel("kernels-test/versions", version="<0.2.0")
+    assert kernel.version() == "0.1.1"
+    kernel = get_kernel("kernels-test/versions", version=">0.1.0,<0.2.0")
+    assert kernel.version() == "0.1.1"
+
+    with pytest.raises(ValueError, match=r"No version.*satisfies requirement"):
+        get_kernel("kernels-test/versions", version=">0.2.0")
+
+    with pytest.raises(ValueError, match=r"Either a revision or a version.*not both"):
+        kernel = get_kernel(
+            "kernels-test/versions", revision="v0.1.0", version="<1.0.0"
+        )
+
+
+@pytest.mark.linux_only
+def test_universal_kernel(universal_kernel):
+    torch.manual_seed(0)
+    A = torch.randint(-10, 10, (64, 128), dtype=torch.int8, device="cuda")
+    B = torch.randint(-10, 10, (128, 96), dtype=torch.int8, device="cuda")
+    scale_a = torch.tensor(0.4, dtype=torch.float16, device="cuda")
+    scale_b = torch.tensor(0.6, dtype=torch.float16, device="cuda")
+
+    out = universal_kernel.triton_scaled_mm(A, B, scale_a, scale_b, torch.float16)
+    out_check = (A * scale_a) @ (B * scale_b)
+    out_check = out_check.to(torch.float16)
+
+    torch.testing.assert_close(out, out_check, rtol=1e-1, atol=1e-1)

tests/test_benchmarks.py

@@ -1,6 +1,7 @@
 import pytest
 import torch
-from hf_kernels import get_kernel
+
+from kernels import get_kernel
 
 
 @pytest.fixture
@@ -15,18 +16,21 @@ def device():
     return "cuda"
 
 
+@pytest.mark.linux_only
 def test_gelu_small(kernel, device, benchmark):
     x = torch.randn(32, 32, dtype=torch.float16, device=device)
     y = torch.empty_like(x)
     benchmark(kernel.gelu_fast, y, x)
 
 
+@pytest.mark.linux_only
 def test_gelu_medium(kernel, device, benchmark):
     x = torch.randn(128, 128, dtype=torch.float16, device=device)
     y = torch.empty_like(x)
     benchmark(kernel.gelu_fast, y, x)
 
 
+@pytest.mark.linux_only
 def test_gelu_large(kernel, device, benchmark):
     x = torch.randn(512, 512, dtype=torch.float16, device=device)
     y = torch.empty_like(x)

tests/test_interval_tree.py

@@ -0,0 +1,230 @@
+import random
+from typing import Generic, List, Optional, Tuple, TypeVar
+
+import pytest
+
+from kernels._interval_tree import IntervalTree, _Node
+
+T = TypeVar("T")
+
+
+class SimpleIntervalStore(Generic[T]):
+    """A simple O(n) implementation that stores intervals in a list."""
+
+    def __init__(self):
+        self.intervals: List[Tuple[int, int, T]] = []
+
+    def insert(self, start: int, end: int, data: T) -> None:
+        """Insert an interval into the store."""
+        # Replace data if the interval already exists.
+        for i, (existing_start, existing_end, existing_data) in enumerate(
+            self.intervals
+        ):
+            if existing_start == start and existing_end == end:
+                self.intervals[i] = (start, end, data)
+                return
+
+        self.intervals.append((start, end, data))
+
+    def find_smallest_interval(self, point: int) -> Optional[T]:
+        """Find the best match using linear search."""
+        matches = []
+        for start, end, data in self.intervals:
+            if start <= point <= end:
+                matches.append((start, end, data))
+
+        if not matches:
+            return None
+
+        # Return the smallest interval, sort by memory location when
+        # there are multiple matches with the same interval size. This
+        # mirrors the ordering in the intervan tree.
+        best_match = min(matches, key=lambda x: (x[1] - x[0], id(x[2])))
+        return best_match[2]
+
+
+def is_balanced(tree: IntervalTree[T]) -> bool:
+    """Check if the AVL tree is properly balanced."""
+
+    def check_balance(node: Optional[_Node[T]]) -> Tuple[bool, int]:
+        if node is None:
+            return True, 0
+
+        # Left and right subtrees should be balanced.
+        left_balanced, left_height = check_balance(node.left)
+        if not left_balanced:
+            return False, -1
+
+        right_balanced, right_height = check_balance(node.right)
+        if not right_balanced:
+            return False, -1
+
+        # The difference in height should not exceed 1.
+        if abs(left_height - right_height) > 1:
+            return False, -1
+
+        # Check if the height is correct.
+        expected_height = 1 + max(left_height, right_height)
+        if node.height != expected_height:
+            return False, -1
+
+        return True, expected_height
+
+    balanced, _ = check_balance(tree.root)
+    return balanced
+
+
+@pytest.fixture
+def populated_tree() -> IntervalTree[str]:
+    """Provides a pre-populated IntervalTree for testing."""
+    tree = IntervalTree[str]()
+    kernels = [
+        (80, 89, "Kernel_A_General_80_89"),
+        (86, 89, "Kernel_B_Ampere_86_89"),
+        (80, 86, "Kernel_C_Older_Ampere_80_86"),
+        (70, 75, "Kernel_D_Volta_70_75"),
+        (86, 87, "Kernel_E_Specific_86_87"),
+    ]
+    for start, end, name in kernels:
+        tree.insert(start, end, name)
+    return tree
+
+
+def test_find_smallest_interval_match_with_multiple_overlaps(populated_tree):
+    # Check that the smallest inteval is selected when there are
+    # multiple matching intervals.
+    assert populated_tree.find_smallest_interval(86) == "Kernel_E_Specific_86_87"
+
+
+def test_find_single_match(populated_tree):
+    assert populated_tree.find_smallest_interval(72) == "Kernel_D_Volta_70_75"
+    assert populated_tree.find_smallest_interval(75) == "Kernel_D_Volta_70_75"
+
+
+def test_no_match_outside_all_ranges(populated_tree):
+    # Check that no interval is found when the value is out of range
+    # (too small/too large).
+    assert populated_tree.find_smallest_interval(65) is None
+    assert populated_tree.find_smallest_interval(95) is None
+
+
+def test_no_match_in_gap_between_ranges(populated_tree):
+    # Check that no interval is found when the value is between two
+    # intervals.
+    assert populated_tree.find_smallest_interval(78) is None
+
+
+def test_boundary_conditions_start_and_end(populated_tree):
+    # Test exact upper/lower bounds of intervals.
+    assert populated_tree.find_smallest_interval(80) == "Kernel_C_Older_Ampere_80_86"
+    assert populated_tree.find_smallest_interval(89) == "Kernel_B_Ampere_86_89"
+
+
+def test_empty_tree():
+    # Searching in an empty tree should return None.
+    empty_tree = IntervalTree[str]()
+    assert empty_tree.find_smallest_interval(100) is None
+
+
+def test_multiple_equally_specific_matches():
+    # Check that we pick the match in a stable way when there is are
+    # multiple matching intervals with the same size.
+    tree = IntervalTree[str]()
+    str1 = "First_Narrow_Kernel"
+    str2 = "Second_Narrow_Kernel"
+    tree.insert(10, 20, "Wide_Kernel")
+    tree.insert(12, 17, str1)
+    tree.insert(14, 19, str2)
+
+    if id(str1) < id(str2):
+        assert tree.find_smallest_interval(15) == str1
+    else:
+        assert tree.find_smallest_interval(15) == str2
+
+
+def test_property_based_interval_tree():
+    # Quick-check property-based testing:
+    #
+    # - Verify that the tree is balanced after each insertion.
+    # - Verify the query against a simple list-based implementation.
+
+    random.seed(42)  # For reproducible tests
+
+    test_points = list(range(0, 101))
+
+    for _ in range(5):
+        tree = IntervalTree[str]()
+        simple = SimpleIntervalStore[str]()
+
+        intervals = []
+        for i in range(100):
+            start = random.randint(0, 90)
+            end = random.randint(start, 100)
+            data = f"interval_{i}_s{start}_e{end}"
+            intervals.append((start, end, data))
+
+        for i, (start, end, data) in enumerate(intervals):
+            tree.insert(start, end, data)
+            simple.insert(start, end, data)
+
+            # Check that tree is still balanced
+            assert is_balanced(
+                tree
+            ), f"Tree became unbalanced after inserting interval {i}: ({start}, {end})"
+
+            for point in test_points:
+                tree_result = tree.find_smallest_interval(point)
+                simple_result = simple.find_smallest_interval(point)
+
+                assert tree_result == simple_result, (
+                    f"Mismatch for point {point} after inserting {i+1} intervals. "
+                    f"Tree: {tree_result}, Simple: {simple_result}. "
+                    f"Last inserted: ({start}, {end})"
+                )
+
+
+def test_property_based_edge_cases():
+    random.seed(123)
+
+    tree = IntervalTree[str]()
+    simple = SimpleIntervalStore[str]()
+
+    # Single-point intervals.
+    for i in range(10):
+        point = random.randint(0, 100)
+        data = f"single_point_{i}_{point}"
+        tree.insert(point, point, data)
+        simple.insert(point, point, data)
+
+        assert is_balanced(
+            tree
+        ), f"Tree unbalanced after inserting single point {point}"
+
+        # Test the exact point and neighbors
+        for test_point in [point - 1, point, point + 1]:
+            if 0 <= test_point <= 100:
+                tree_result = tree.find_smallest_interval(test_point)
+                simple_result = simple.find_smallest_interval(test_point)
+                assert tree_result == simple_result
+
+
+def test_unique_intervals_override():
+    """Test that inserting an interval with the same start/end overrides the previous value."""
+    tree = IntervalTree[str]()
+
+    tree.insert(10, 20, "original_value")
+    assert tree.find_smallest_interval(15) == "original_value"
+
+    tree.insert(10, 20, "new_value")
+    assert tree.find_smallest_interval(15) == "new_value"
+
+    tree.insert(10, 25, "different_interval")
+    results = tree.search(15)
+    assert "new_value" in results
+    assert "different_interval" in results
+    assert len(results) == 2
+
+    tree.insert(10, 20, "final_value")
+    assert tree.find_smallest_interval(15) == "final_value"
+
+    assert is_balanced(tree)

tests/test_kernel_locking.py

@@ -0,0 +1,60 @@
+from dataclasses import dataclass
+from pathlib import Path
+
+import pytest
+import torch.nn as nn
+
+from kernels import load_kernel
+from kernels.cli import download_kernels
+from kernels.layer import (
+    LockedLayerRepository,
+    Mode,
+    kernelize,
+    use_kernel_forward_from_hub,
+    use_kernel_mapping,
+)
+
+
+# Mock download arguments class.
+@dataclass
+class DownloadArgs:
+    all_variants: bool
+    project_dir: Path
+
+
+def test_download_all_hash_validation():
+    project_dir = Path(__file__).parent / "kernel_locking"
+    download_kernels(DownloadArgs(all_variants=True, project_dir=project_dir))
+
+
+@pytest.mark.linux_only
+def test_load_locked():
+    project_dir = Path(__file__).parent / "kernel_locking"
+    # Also validates that hashing works correctly.
+    download_kernels(DownloadArgs(all_variants=False, project_dir=project_dir))
+    load_kernel("kernels-community/activation", lockfile=project_dir / "kernels.lock")
+
+
+def test_layer_locked():
+    project_dir = Path(__file__).parent / "layer_locking"
+
+    @use_kernel_forward_from_hub("Version")
+    class Version(nn.Module):
+        def forward(self) -> str:
+            return "0.0.0"
+
+    version = Version()
+
+    with use_kernel_mapping(
+        {
+            "Version": {
+                "cuda": LockedLayerRepository(
+                    repo_id="kernels-test/versions",
+                    layer_name="Version",
+                    lockfile=project_dir / "kernels.lock",
+                )
+            },
+        }
+    ):
+        version = kernelize(version, device="cuda", mode=Mode.INFERENCE)
+        assert version() == "0.1.1"

tests/test_layer.py

@@ -0,0 +1,953 @@
+import sys
+from contextlib import nullcontext
+
+import pytest
+import torch
+import torch.nn as nn
+from torch.nn import functional as F
+
+from kernels import (
+    Device,
+    LayerRepository,
+    Mode,
+    kernelize,
+    register_kernel_mapping,
+    use_kernel_forward_from_hub,
+)
+from kernels.layer import (
+    _KERNEL_MAPPING,
+    CUDAProperties,
+    _validate_layer,
+    use_kernel_mapping,
+)
+
+kernel_layer_mapping = {
+    "SiluAndMul": {
+        Device(type="cuda"): LayerRepository(
+            repo_id="kernels-community/activation",
+            layer_name="SiluAndMul",
+        )
+    },
+    "SiluAndMulNoCompile": {
+        "cuda": LayerRepository(
+            repo_id="kernels-test/op-without-fake-test",
+            layer_name="SiluAndMul",
+        )
+    },
+    "SiluAndMulStringDevice": {
+        "cuda": LayerRepository(
+            repo_id="kernels-community/activation",
+            layer_name="SiluAndMul",
+        )
+    },
+}
+
+register_kernel_mapping(kernel_layer_mapping)
+
+
+class SiluAndMul(nn.Module):
+    def __init__(self):
+        super().__init__()
+        # Used to check that we called hub kernel.
+        self.n_calls = 0
+
+    def forward(self, input: torch.Tensor) -> torch.Tensor:
+        self.n_calls += 1
+        d = input.shape[-1] // 2
+        return F.silu(input[..., :d]) * input[..., d:]
+
+
+@use_kernel_forward_from_hub("SiluAndMulNoCompile")
+class SiluAndMulNoCompileKernel(SiluAndMul):
+    pass
+
+
+@use_kernel_forward_from_hub("SiluAndMul")
+class SiluAndMulWithKernel(SiluAndMul):
+    pass
+
+
+@use_kernel_forward_from_hub("SiluAndMulStringDevice")
+class SiluAndMulStringDevice(SiluAndMul):
+    pass
+
+
+@use_kernel_forward_from_hub("Linear")
+class TorchLinearWithCounter(nn.Linear):
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        # Used to check that we called hub kernel.
+        self.n_calls = 0
+
+    def forward(self, input: torch.Tensor) -> torch.Tensor:
+        self.n_calls += 1
+        return super().forward(input)
+
+
+def test_arg_kinds():
+    @use_kernel_forward_from_hub("ArgKind")
+    class ArgKind(nn.Module):
+        def forward(
+            self,
+            arg1,
+            arg2,
+            *,
+            kwarg1,
+            kwarg2=42,
+        ):
+            return (arg1, arg2, kwarg1, kwarg2)
+
+    arg_kind = ArgKind()
+    assert arg_kind("foo", "bar", kwarg1="baz") == ("foo", "bar", "baz", 42)
+    assert arg_kind("foo", "bar", kwarg1="baz", kwarg2=5) == ("foo", "bar", "baz", 5)
+
+
+@pytest.mark.linux_only
+@pytest.mark.parametrize("cls", [SiluAndMulWithKernel, SiluAndMulStringDevice])
+@pytest.mark.parametrize("device", ["cuda", "cpu"])
+def test_hub_forward(cls, device):
+    torch.random.manual_seed(0)
+
+    silu_and_mul = SiluAndMul()
+    X = torch.randn((32, 64), device=device)
+    Y = silu_and_mul(X)
+
+    silu_and_mul_with_kernel = kernelize(cls(), device=device, mode=Mode.INFERENCE)
+    Y_kernel = silu_and_mul_with_kernel(X)
+
+    torch.testing.assert_close(Y_kernel, Y)
+
+    assert silu_and_mul.n_calls == 1
+    if device == "cuda":
+        assert silu_and_mul_with_kernel.n_calls == 0
+    else:
+        assert silu_and_mul_with_kernel.n_calls == 1
+
+
+@pytest.mark.linux_only
+def test_capability():
+    linear = TorchLinearWithCounter(32, 32).to("cuda")
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                Device(
+                    type="cuda",
+                    properties=CUDAProperties(
+                        min_capability=75, max_capability=sys.maxsize
+                    ),
+                ): LayerRepository(
+                    repo_id="kernels-test/backward-marker-test",
+                    layer_name="LinearBackward",
+                )
+            }
+        }
+    ):
+        kernelize(linear, mode=Mode.INFERENCE)
+        X = torch.randn(10, 32, device="cuda")
+        linear(X)
+
+        # Check that we called out to the kernel.
+        assert linear.n_calls == 0
+
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                Device(
+                    type="cuda",
+                    properties=CUDAProperties(
+                        min_capability=sys.maxsize, max_capability=sys.maxsize
+                    ),
+                ): LayerRepository(
+                    repo_id="kernels-test/backward-marker-test",
+                    layer_name="LinearBackward",
+                )
+            }
+        }
+    ):
+        kernelize(linear, mode=Mode.INFERENCE)
+        X = torch.randn(10, 32, device="cuda")
+        linear(X)
+
+        # Check that we didn't call out to the kernel because there is
+        # is no kernel with a matching capability..
+        assert linear.n_calls == 1
+
+
+def test_layer_fallback_works():
+    @use_kernel_forward_from_hub("SiluAndMulNonExisting")
+    class SiluAndMulWithKernelFallback(SiluAndMul):
+        pass
+
+    # Check that we don't raise an exception for a non-existing kernel.
+    silu_and_mul = SiluAndMulWithKernelFallback()
+    kernelize(silu_and_mul, device="cuda", mode=Mode.INFERENCE)
+
+
+@pytest.mark.linux_only
+@pytest.mark.parametrize("cls", [SiluAndMulWithKernel, SiluAndMulNoCompileKernel])
+@pytest.mark.parametrize("device", ["cuda"])
+def test_torch_compile_layer_without_fallback(cls, device):
+    silu_and_mul = SiluAndMul()
+
+    X = torch.randn((32, 64), dtype=torch.float32, device=device)
+    Y = silu_and_mul(X)
+
+    silu_and_mul_with_kernel = cls()
+    silu_and_mul_with_kernel.eval()
+
+    ctx = (
+        pytest.raises(ValueError, match="does not support mode")
+        if cls is SiluAndMulNoCompileKernel
+        else nullcontext()
+    )
+    with ctx:
+        silu_and_mul_with_kernel = kernelize(
+            silu_and_mul_with_kernel,
+            device=device,
+            mode=Mode.INFERENCE | Mode.TORCH_COMPILE,
+            use_fallback=False,
+        )
+    silu_and_mul_compiled = torch.compile(silu_and_mul_with_kernel, fullgraph=True)
+
+    Y_compiled = silu_and_mul_compiled(X)
+
+    torch.testing.assert_close(Y_compiled, Y)
+
+
+@pytest.mark.linux_only
+@pytest.mark.parametrize("cls", [SiluAndMulWithKernel, SiluAndMulNoCompileKernel])
+@pytest.mark.parametrize("device", ["cuda"])
+def test_torch_compile_layer_with_fallback(cls, device):
+    silu_and_mul = SiluAndMul()
+
+    X = torch.randn((32, 64), dtype=torch.float32, device=device)
+    Y = silu_and_mul(X)
+
+    silu_and_mul_with_kernel = cls()
+    silu_and_mul_with_kernel.eval()
+    silu_and_mul_with_kernel = kernelize(
+        silu_and_mul_with_kernel,
+        device=device,
+        mode=Mode.INFERENCE | Mode.TORCH_COMPILE,
+    )
+    silu_and_mul_compiled = torch.compile(silu_and_mul_with_kernel, fullgraph=True)
+
+    Y_compiled = silu_and_mul_compiled(X)
+
+    torch.testing.assert_close(Y_compiled, Y)
+
+
+@pytest.mark.linux_only
+def test_mapping_contexts():
+    assert set(_KERNEL_MAPPING.get().keys()) == {
+        "SiluAndMul",
+        "SiluAndMulStringDevice",
+        "SiluAndMulNoCompile",
+    }
+
+    extra_mapping1 = {
+        "TestKernel": {
+            Device(type="cuda"): LayerRepository(
+                repo_id="kernels-community/activation",
+                layer_name="SiluAndMul",
+                revision="layers",
+            )
+        }
+    }
+
+    with use_kernel_mapping(extra_mapping1):
+        assert set(_KERNEL_MAPPING.get().keys()) == {
+            "SiluAndMul",
+            "SiluAndMulStringDevice",
+            "SiluAndMulNoCompile",
+            "TestKernel",
+        }
+
+        extra_mapping2 = {
+            "SiluAndMul": {
+                Device(type="cuda"): LayerRepository(
+                    repo_id="kernels-community/non-existing",
+                    layer_name="SiluAndMul",
+                    revision="layers",
+                )
+            }
+        }
+
+        with use_kernel_mapping(extra_mapping2):
+            assert set(_KERNEL_MAPPING.get().keys()) == {
+                "SiluAndMul",
+                "SiluAndMulStringDevice",
+                "SiluAndMulNoCompile",
+                "TestKernel",
+            }
+            assert (
+                _KERNEL_MAPPING.get()["SiluAndMul"]["cuda"].repos[Mode.FALLBACK].repo_id
+                == "kernels-community/non-existing"
+            )
+
+        assert set(_KERNEL_MAPPING.get().keys()) == {
+            "SiluAndMul",
+            "SiluAndMulStringDevice",
+            "SiluAndMulNoCompile",
+            "TestKernel",
+        }
+        assert (
+            _KERNEL_MAPPING.get()["SiluAndMul"]["cuda"].repos[Mode.FALLBACK].repo_id
+            == "kernels-community/activation"
+        )
+
+        with use_kernel_mapping(extra_mapping2, inherit_mapping=False):
+            assert set(_KERNEL_MAPPING.get().keys()) == {
+                "SiluAndMul",
+            }
+            assert (
+                _KERNEL_MAPPING.get()["SiluAndMul"]["cuda"].repos[Mode.FALLBACK].repo_id
+                == "kernels-community/non-existing"
+            )
+
+        assert set(_KERNEL_MAPPING.get().keys()) == {
+            "SiluAndMul",
+            "SiluAndMulStringDevice",
+            "SiluAndMulNoCompile",
+            "TestKernel",
+        }
+        assert (
+            _KERNEL_MAPPING.get()["SiluAndMul"]["cuda"].repos[Mode.FALLBACK].repo_id
+            == "kernels-community/activation"
+        )
+
+    assert set(_KERNEL_MAPPING.get().keys()) == {
+        "SiluAndMul",
+        "SiluAndMulStringDevice",
+        "SiluAndMulNoCompile",
+    }
+
+
+def test_validate_kernel_layer():
+    class BadLayer(nn.Module):
+        def __init__(self, *args, **kwargs):
+            super().__init__(*args, **kwargs)
+            self.foo = 42
+
+    with pytest.raises(TypeError, match="not override"):
+        _validate_layer(cls=BadLayer, check_cls=SiluAndMul)
+
+    class BadLayer2(nn.Module):
+        foo: int = 42
+
+    with pytest.raises(TypeError, match="not contain additional members"):
+        _validate_layer(cls=BadLayer2, check_cls=SiluAndMul)
+
+    class BadLayer3(nn.Module):
+        def forward(self, x: torch.Tensor, foo: int) -> torch.Tensor: ...
+
+    with pytest.raises(TypeError, match="different number of arguments"):
+        _validate_layer(cls=BadLayer3, check_cls=SiluAndMul)
+
+    class BadLayer4(nn.Module):
+        def forward(self, *, x: torch.Tensor) -> torch.Tensor: ...
+
+    with pytest.raises(TypeError, match="different kind of arguments"):
+        _validate_layer(cls=BadLayer4, check_cls=SiluAndMul)
+
+
+@pytest.mark.linux_only
+def test_invalid_mode_for_mapping_rejected():
+    linear = TorchLinearWithCounter(32, 32).to("cuda")
+
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                "cuda": {
+                    Mode.TRAINING: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearNoBackward",
+                    )
+                }
+            }
+        }
+    ):
+        with pytest.raises(ValueError, match="does not support backward"):
+            kernelize(linear, mode=Mode.TRAINING)
+
+
+@pytest.mark.linux_only
+def test_kernel_modes():
+    linear = TorchLinearWithCounter(32, 32).to("cuda")
+
+    # Case 1: layer without further specification, becomes the
+    #         base layer.
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                "cuda": LayerRepository(
+                    repo_id="kernels-test/backward-marker-test",
+                    layer_name="LinearBackward",
+                )
+            }
+        }
+    ):
+        kernelize(linear, mode=Mode.INFERENCE)
+        X = torch.randn(10, 32, device="cuda")
+        linear(X)
+        assert linear.n_calls == 0
+
+        kernelize(linear, mode=Mode.TRAINING)
+        linear(X)
+        assert linear.n_calls == 0
+
+        kernelize(linear, mode=Mode.TRAINING | Mode.TORCH_COMPILE)
+        linear(X)
+        assert linear.n_calls == 0
+
+        # Same as previous, since TRAINING | TORCH_COMPILE is the default.
+        kernelize(linear)
+        linear(X)
+        assert linear.n_calls == 0
+
+    # Case 2: register a kernel just for training. If no base kernel
+    #         layer is registered, we fall back to the original layer.
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                "cuda": {
+                    Mode.TRAINING: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    )
+                }
+            }
+        }
+    ):
+        kernelize(linear, mode=Mode.INFERENCE)
+        X = torch.randn(10, 32, device="cuda")
+        linear(X)
+        assert linear.n_calls == 0
+
+        kernelize(linear, mode=Mode.TRAINING)
+        linear(X)
+        # Training has a kernel, so fallback.
+        assert linear.n_calls == 0
+
+        kernelize(linear, mode=Mode.TRAINING | Mode.TORCH_COMPILE)
+        linear(X)
+        # TRAINING | TORCH_COMPILE cannot fall back to TRAINING kernel, so uses original.
+        assert linear.n_calls == 1
+
+        # Same as previous, since TRAINING | TORCH_COMPILE is the default.
+        kernelize(linear)
+        linear(X)
+        # TRAINING | TORCH_COMPILE cannot fall back to TRAINING kernel, so uses original.
+        assert linear.n_calls == 2
+
+    # Case 3: register a kernel just for training and one for fallback.
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                "cuda": {
+                    Mode.FALLBACK: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    ),
+                    Mode.TRAINING: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    ),
+                }
+            }
+        }
+    ):
+        kernelize(linear, mode=Mode.INFERENCE)
+        X = torch.randn(10, 32, device="cuda")
+        linear(X)
+        # Falls back to TRAINING.
+        assert linear.n_calls == 2
+
+        kernelize(linear, mode=Mode.TRAINING)
+        linear(X)
+        # Falls back to the TRAINING kernel.
+        assert linear.n_calls == 2
+
+        kernelize(linear, mode=Mode.TRAINING | Mode.TORCH_COMPILE)
+        linear(X)
+        # TRAINING | TORCH_COMPILE falls back to FALLBACK kernel.
+        assert linear.n_calls == 2
+
+        # Same as previous, since TRAINING | TORCH_COMPILE is the default.
+        kernelize(linear)
+        linear(X)
+        # TRAINING | TORCH_COMPILE falls back to FALLBACK kernel.
+        assert linear.n_calls == 2
+
+    # Case 4: register a kernel with two preferences.
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                "cuda": {
+                    Mode.TRAINING
+                    | Mode.TORCH_COMPILE: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    )
+                }
+            }
+        }
+    ):
+        kernelize(linear, mode=Mode.INFERENCE)
+        X = torch.randn(10, 32, device="cuda")
+        linear(X)
+        # Falls back to the TRAINING | TORCH_COMPILE kernel.
+        assert linear.n_calls == 2
+
+        kernelize(linear, mode=Mode.TRAINING)
+        linear(X)
+        # TRAINING can fall back to TRAINING | TORCH_COMPILE kernel.
+        assert linear.n_calls == 2
+
+        kernelize(linear, mode=Mode.TRAINING | Mode.TORCH_COMPILE)
+        linear(X)
+        # Uses TRAINING | TORCH_COMPILE kernel.
+        assert linear.n_calls == 2
+
+        kernelize(linear)
+        linear(X)
+        # Same as previous, since TRAINING | TORCH_COMPILE is the default.
+        assert linear.n_calls == 2
+
+
+@pytest.mark.linux_only
+def test_fallback_used_when_training():
+    linear = TorchLinearWithCounter(32, 32).to("cuda")
+
+    # Case 1: kernel with explicit backward support should always
+    #         use the kernel.
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                Device(type="cuda"): LayerRepository(
+                    repo_id="kernels-test/backward-marker-test",
+                    layer_name="LinearBackward",
+                )
+            }
+        }
+    ):
+        linear.train()
+        kernelize(linear, mode=Mode.INFERENCE)
+        X = torch.randn(10, 32, device="cuda")
+        linear(X)
+        assert linear.n_calls == 0
+
+        linear.eval()
+        linear(X)
+        assert linear.n_calls == 0
+
+    # Case 2: kernel with implicit backward support should always
+    #         use the kernel.
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                Device(type="cuda"): LayerRepository(
+                    repo_id="kernels-test/backward-marker-test",
+                    layer_name="LinearImplicitBackward",
+                )
+            }
+        }
+    ):
+        linear.train()
+        kernelize(linear, mode=Mode.INFERENCE)
+        X = torch.randn(10, 32, device="cuda")
+        linear(X)
+        assert linear.n_calls == 0
+
+        linear.eval()
+        linear(X)
+        assert linear.n_calls == 0
+
+
+def test_invalid_mode_rejected():
+    with pytest.raises(ValueError, match="mutually exclusive"):
+        _ = Mode.INFERENCE | Mode.TRAINING
+
+    with pytest.raises(ValueError, match="cannot be combined with other modes"):
+        _ = Mode.FALLBACK | Mode.TORCH_COMPILE
+
+    with pytest.raises(
+        ValueError, match="can only be used to register kernel mappings"
+    ):
+        kernelize(torch.nn.Linear(32, 32), mode=Mode.FALLBACK)
+
+    with pytest.raises(ValueError, match="mode must contain"):
+        kernelize(torch.nn.Linear(32, 32), mode=Mode.TORCH_COMPILE)
+
+
+@pytest.mark.linux_only
+def test_kernel_modes_inference():
+    """Test inference-specific fallback scenarios."""
+    linear = TorchLinearWithCounter(32, 32).to("cuda")
+
+    # Case 1: register a kernel just for inference
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                "cuda": {
+                    Mode.INFERENCE: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    )
+                }
+            }
+        }
+    ):
+        kernelize(linear, mode=Mode.INFERENCE)
+        X = torch.randn(10, 32, device="cuda")
+        linear(X)
+        assert linear.n_calls == 0
+
+        kernelize(linear, mode=Mode.INFERENCE | Mode.TORCH_COMPILE)
+        linear(X)
+        # INFERENCE | TORCH_COMPILE cannot fall back to INFERENCE kernel, so uses original
+        assert linear.n_calls == 1
+
+        kernelize(linear, mode=Mode.TRAINING)
+        linear(X)
+        # No training kernel, so fallback to original
+        assert linear.n_calls == 2
+
+    # Case 2: register a kernel just for inference + torch.compile
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                "cuda": {
+                    Mode.INFERENCE
+                    | Mode.TORCH_COMPILE: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    )
+                }
+            }
+        }
+    ):
+        kernelize(linear, mode=Mode.INFERENCE | Mode.TORCH_COMPILE)
+        X = torch.randn(10, 32, device="cuda")
+        linear(X)
+        assert linear.n_calls == 2
+
+        kernelize(linear, mode=Mode.INFERENCE)
+        linear(X)
+        # INFERENCE falls back to INFERENCE | TORCH_COMPILE kernel
+        assert linear.n_calls == 2
+
+        kernelize(linear, mode=Mode.TRAINING)
+        linear(X)
+        # No training kernel, so fallback to original
+        assert linear.n_calls == 3
+
+    # Case 3: register both inference kernels
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                "cuda": {
+                    Mode.INFERENCE: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    ),
+                    Mode.INFERENCE
+                    | Mode.TORCH_COMPILE: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    ),
+                }
+            }
+        }
+    ):
+        kernelize(linear, mode=Mode.INFERENCE)
+        X = torch.randn(10, 32, device="cuda")
+        linear(X)
+        # Uses exact INFERENCE kernel
+        assert linear.n_calls == 3
+
+        kernelize(linear, mode=Mode.INFERENCE | Mode.TORCH_COMPILE)
+        linear(X)
+        # Uses exact INFERENCE | TORCH_COMPILE kernel
+        assert linear.n_calls == 3
+
+        kernelize(linear, mode=Mode.TRAINING)
+        linear(X)
+        # No training kernel, so fallback to original
+        assert linear.n_calls == 4
+
+
+@pytest.mark.linux_only
+def test_kernel_modes_mixed():
+    """Test mixed training and inference kernel scenarios."""
+    linear = TorchLinearWithCounter(32, 32).to("cuda")
+
+    # Case 1: register both base inference and training kernels
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                "cuda": {
+                    Mode.INFERENCE: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    ),
+                    Mode.TRAINING: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    ),
+                }
+            }
+        }
+    ):
+        kernelize(linear, mode=Mode.INFERENCE)
+        X = torch.randn(10, 32, device="cuda")
+        linear(X)
+        assert linear.n_calls == 0
+
+        kernelize(linear, mode=Mode.TRAINING)
+        linear(X)
+        assert linear.n_calls == 0
+
+        kernelize(linear, mode=Mode.INFERENCE | Mode.TORCH_COMPILE)
+        linear(X)
+        # INFERENCE | TORCH_COMPILE cannot fall back to INFERENCE kernel, so uses original
+        assert linear.n_calls == 1
+
+        kernelize(linear, mode=Mode.TRAINING | Mode.TORCH_COMPILE)
+        linear(X)
+        # TRAINING | TORCH_COMPILE cannot fall back to TRAINING kernel, so uses original
+        assert linear.n_calls == 2
+
+    # Case 2: register all four kernel modes
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                "cuda": {
+                    Mode.INFERENCE: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    ),
+                    Mode.TRAINING: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    ),
+                    Mode.INFERENCE
+                    | Mode.TORCH_COMPILE: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    ),
+                    Mode.TRAINING
+                    | Mode.TORCH_COMPILE: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    ),
+                }
+            }
+        }
+    ):
+        kernelize(linear, mode=Mode.INFERENCE)
+        X = torch.randn(10, 32, device="cuda")
+        linear(X)
+        # Uses exact INFERENCE kernel
+        assert linear.n_calls == 2
+
+        kernelize(linear, mode=Mode.TRAINING)
+        linear(X)
+        # Uses exact TRAINING kernel
+        assert linear.n_calls == 2
+
+        kernelize(linear, mode=Mode.INFERENCE | Mode.TORCH_COMPILE)
+        linear(X)
+        # Uses exact INFERENCE | TORCH_COMPILE kernel
+        assert linear.n_calls == 2
+
+        kernelize(linear, mode=Mode.TRAINING | Mode.TORCH_COMPILE)
+        linear(X)
+        # Uses exact TRAINING | TORCH_COMPILE kernel
+        assert linear.n_calls == 2
+
+
+@pytest.mark.linux_only
+def test_kernel_modes_cross_fallback():
+    """Test cross-mode fallback scenarios from inference to training modes."""
+    linear = TorchLinearWithCounter(32, 32).to("cuda")
+
+    # Case 1: Only training kernel registered - inference should fall back to training
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                "cuda": {
+                    Mode.TRAINING: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    )
+                }
+            }
+        }
+    ):
+        kernelize(linear, mode=Mode.INFERENCE)
+        X = torch.randn(10, 32, device="cuda")
+        linear(X)
+        # INFERENCE falls back to TRAINING kernel
+        assert linear.n_calls == 0
+
+        kernelize(linear, mode=Mode.TRAINING)
+        linear(X)
+        # TRAINING uses the kernel directly
+        assert linear.n_calls == 0
+
+    # Case 2: Only training + torch.compile kernel registered
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                "cuda": {
+                    Mode.TRAINING
+                    | Mode.TORCH_COMPILE: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    )
+                }
+            }
+        }
+    ):
+        kernelize(linear, mode=Mode.INFERENCE)
+        X = torch.randn(10, 32, device="cuda")
+        linear(X)
+        # INFERENCE falls back to TRAINING | TORCH_COMPILE kernel
+        assert linear.n_calls == 0
+
+        kernelize(linear, mode=Mode.INFERENCE | Mode.TORCH_COMPILE)
+        linear(X)
+        # INFERENCE | TORCH_COMPILE falls back to TRAINING | TORCH_COMPILE kernel
+        assert linear.n_calls == 0
+
+        kernelize(linear, mode=Mode.TRAINING)
+        linear(X)
+        # TRAINING falls back to TRAINING | TORCH_COMPILE kernel
+        assert linear.n_calls == 0
+
+        kernelize(linear, mode=Mode.TRAINING | Mode.TORCH_COMPILE)
+        linear(X)
+        # TRAINING | TORCH_COMPILE uses the kernel directly
+        assert linear.n_calls == 0
+
+    # Case 3: Test that training modes don't fall back to inference modes
+    with use_kernel_mapping(
+        {
+            "Linear": {
+                "cuda": {
+                    Mode.INFERENCE: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    ),
+                    Mode.INFERENCE
+                    | Mode.TORCH_COMPILE: LayerRepository(
+                        repo_id="kernels-test/backward-marker-test",
+                        layer_name="LinearBackward",
+                    ),
+                }
+            }
+        }
+    ):
+        kernelize(linear, mode=Mode.TRAINING)
+        X = torch.randn(10, 32, device="cuda")
+        linear(X)
+        # TRAINING should NOT fall back to inference kernels, use original
+        assert linear.n_calls == 1
+
+        kernelize(linear, mode=Mode.TRAINING | Mode.TORCH_COMPILE)
+        linear(X)
+        # TRAINING | TORCH_COMPILE should NOT fall back to inference kernels, use original
+        assert linear.n_calls == 2
+
+
+def test_layer_versions():
+    @use_kernel_forward_from_hub("Version")
+    class Version(nn.Module):
+        def forward(self) -> str:
+            return "0.0.0"
+
+    version = Version()
+
+    with use_kernel_mapping(
+        {
+            "Version": {
+                Device(type="cuda"): LayerRepository(
+                    repo_id="kernels-test/versions",
+                    layer_name="Version",
+                )
+            }
+        }
+    ):
+        version = kernelize(version, device="cuda", mode=Mode.INFERENCE)
+        assert version() == "0.2.0"
+
+    with use_kernel_mapping(
+        {
+            "Version": {
+                Device(type="cuda"): LayerRepository(
+                    repo_id="kernels-test/versions",
+                    layer_name="Version",
+                    version="<1.0.0",
+                )
+            }
+        }
+    ):
+        version = kernelize(version, device="cuda", mode=Mode.INFERENCE)
+        assert version() == "0.2.0"
+
+    with use_kernel_mapping(
+        {
+            "Version": {
+                Device(type="cuda"): LayerRepository(
+                    repo_id="kernels-test/versions",
+                    layer_name="Version",
+                    version="<0.2.0",
+                )
+            }
+        }
+    ):
+        version = kernelize(version, device="cuda", mode=Mode.INFERENCE)
+        assert version() == "0.1.1"
+
+    with use_kernel_mapping(
+        {
+            "Version": {
+                Device(type="cuda"): LayerRepository(
+                    repo_id="kernels-test/versions",
+                    layer_name="Version",
+                    version=">0.1.0,<0.2.0",
+                )
+            }
+        }
+    ):
+        version = kernelize(version, device="cuda", mode=Mode.INFERENCE)
+        assert version() == "0.1.1"
+
+    with use_kernel_mapping(
+        {
+            "Version": {
+                Device(type="cuda"): LayerRepository(
+                    repo_id="kernels-test/versions",
+                    layer_name="Version",
+                    version=">0.2.0",
+                )
+            }
+        }
+    ):
+        with pytest.raises(ValueError, match=r"No version.*satisfies requirement"):
+            kernelize(version, device="cuda", mode=Mode.INFERENCE)
+
+    with pytest.raises(ValueError, match=r"Either a revision or a version.*not both"):
+        use_kernel_mapping(
+            {
+                "Version": {
+                    Device(type="cuda"): LayerRepository(
+                        repo_id="kernels-test/versions",
+                        layer_name="Version",
+                        revision="v0.1.0",
+                        version="<1.0.0",
+                    )
+                }
+            }
+        )