Metadata-Version: 2.2
Name: tiny-metal-nn
Version: 0.1.0a2
Summary: Apple Metal native runtime for hash-grid + MLP neural-field training (Python binding ships in v1.0).
Keywords: metal,neural-field,hash-grid,apple-silicon,tinycudann,instant-ngp
Author: tiny-metal-nn contributors
License:                                  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 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 Support. 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 support.
         
            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 2026 tiny-metal-nn contributors
         
            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.
         
Classifier: Development Status :: 3 - Alpha
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Operating System :: MacOS
Classifier: License :: OSI Approved :: Apache Software License
Requires-Python: >=3.10
Provides-Extra: dev
Requires-Dist: pytest>=8.0; extra == "dev"
Requires-Dist: numpy>=1.26; extra == "dev"
Requires-Dist: torch>=2.4; extra == "dev"
Requires-Dist: libcst>=1.4; extra == "dev"
Description-Content-Type: text/markdown

# tiny-metal-nn

`tiny-metal-nn` (`tmnn`) is a Metal-native, [tiny-cuda-nn](https://github.com/NVlabs/tiny-cuda-nn)-inspired C++ library for hash-grid + MLP neural-field training and inference on Apple GPUs.

`tiny-cuda-nn` is widely used in the CUDA ecosystem (instant-ngp, NeuS, Plenoxels, NerfAcc, and others) but does not run on Apple Silicon. `tmnn` is an attempt to provide a runtime in that style on Metal — fully fused MLPs, hash-grid encodings, and JSON config compatible with tcnn — without trying to be a general ML framework, a PyTorch / MLX replacement, or a generic GPU compute substrate.

See [`STATUS.md`](STATUS.md) for what is and is not in the box today, and what we explicitly do not yet claim.

## When tmnn might fit

- You are on Apple Silicon and want a small-network runtime with an API surface close to tcnn (JSON config, fused MLP families, an image-fitting sample), so that a tcnn-based codebase can be ported with relatively small diffs.
- You want explicit family selection (`FullyFusedMetal` / `TiledMetal` / `SafeDebugMetal`), inspectable planner reasons, manifest-backed kernel prewarm, numerics telemetry, and a frozen optimizer-checkpoint contract — operational surfaces we have intentionally exposed.

For a general Apple-native tensor framework, [MLX](https://github.com/ml-explore/mlx) is broader and more mature. On NVIDIA, [tiny-cuda-nn](https://github.com/NVlabs/tiny-cuda-nn) is the natural choice. A side-by-side comparison, including what we have not yet measured, is in [`docs/VS-MLX-AND-TCNN.md`](docs/VS-MLX-AND-TCNN.md).

## Status

- ~17K LOC C++ source; ~10.5K LOC tests
- `samples/mlp_learning_an_image.cpp` — flagship runnable sample, deliberately matches tiny-cuda-nn's filename and target shape so migration diffs stay small
- Pre-1.0. The v0.1.0 tag is pending external comparison benchmarks (vs MLX, vs tcnn) and at least two more flagship samples (NeRF synthetic, hash-grid SDF fitting). See [`STATUS.md`](STATUS.md) for the current honest scope and the "Roadmap" section below.

## Architecture

```
JSON config → factory_json::canonicalize_model_config
  → NetworkPlan (planner) → kernel selection + prewarm
  → MetalContext (Metal device + pipeline registry + batch pool)
  → fully-fused or tiled MLP training/inference loop
  → checkpoint contract (frozen, optimizer state portable across versions)
```

Targets:

| Target | Purpose |
|---|---|
| `tiny_metal_nn_core` | header-only public API |
| `tiny_metal_nn_runtime` | Metal runtime (MetalContext, pipeline registry, batch pool, training-step lifecycle) |
| `tiny_metal_nn_kernels` | MSL kernel generation (KernelSpec, KernelCompiler, MLPKernelEmitter) |
| `tiny_metal_nn_extensions` | built-in adapters (DNL, RMHE, 4D, standard SDF, multi-output MLP) |

See [`docs/ARCHITECTURE.md`](docs/ARCHITECTURE.md) for the design.

## Build (C++)

`tmnn` uses [vcpkg](https://github.com/microsoft/vcpkg) for `nlohmann_json` and
`gtest`. Either place a vcpkg checkout at `./.deps/vcpkg`, or point
`VCPKG_ROOT` at an existing one:

```bash
git clone https://github.com/microsoft/vcpkg.git ~/vcpkg
~/vcpkg/bootstrap-vcpkg.sh
export VCPKG_ROOT=~/vcpkg
```

Then build:

```bash
git clone <repo-url> tiny-metal-nn
cd tiny-metal-nn
cmake -S . -B build
cmake --build build -j
./build/samples/mlp_learning_an_image                           # flagship sample (built-in config)
./build/samples/mlp_learning_an_image my_config.json            # ...with a custom JSON config
ctest --test-dir build -V                                       # tests
```

## Python binding (optional)

A pybind11 binding ships alongside the C++ library. Install via the
project's build backend (scikit-build-core, which invokes the same CMake
project with `-DTMNN_BUILD_PYTHON_MODULE=ON`):

```bash
python3.13 -m venv .venv
.venv/bin/pip install scikit-build-core pybind11
VCPKG_ROOT=~/vcpkg .venv/bin/pip install -e ".[dev]" --no-build-isolation
.venv/bin/python -c "import tiny_metal_nn as tmnn; print(tmnn.__version__)"
```

The editable install reuses a persistent CMake binary directory, so subsequent
edits to `src/python/` (or any C++ source the binding pulls in) auto-rebuild
on the next `import tiny_metal_nn`. See [`docs/QUICKSTART.md`](docs/QUICKSTART.md)
for a Python "hello world" and [`docs/ARCHITECTURE.md`](docs/ARCHITECTURE.md)
§ 6 for the binding's design.

For users moving from `tinycudann`, a migration tool
(`tools/migrate_tcnn.py`) handles the mechanical conversion (imports + the
canonical 5-line training-loop body); the harder cases (custom losses,
non-canonical shapes) are flagged with actionable diagnostics. A worked
example pair lives at `examples/sphere_sdf/`. See
[`docs/TCNN-MIGRATION-GUIDE.md`](docs/TCNN-MIGRATION-GUIDE.md) § 10.

For an instrumented build with AddressSanitizer + UndefinedBehaviorSanitizer
(off by default; consumers must rebuild any downstream code with the same
sanitizer flags):

```bash
cmake -S . -B build-asan -DTMNN_ENABLE_SANITIZERS=ON
cmake --build build-asan -j
ctest --test-dir build-asan -j
```

macOS Apple Silicon is the primary target. Linux / non-Apple builds use a Metal stub (compiles, but GPU-runtime tests skip).

## Performance

A standalone GPU-measured benchmark binary lives at
`tests/benchmarks/tmnn_runtime_benchmarks.cpp`. Build and run it on your
hardware to get real numbers:

```bash
cmake --build build --target tmnn_runtime_benchmarks
./build/tests/tmnn_runtime_benchmarks --smoke    # quick check
./build/tests/tmnn_runtime_benchmarks            # full run
```

The binary emits planner / autotune-search / hot-step / morton-sort medians on
the local Metal device. We are intentionally not publishing pre-measured
"reference numbers" in this README until a same-hardware comparison story
against MLX (architecture-equivalent) and `tiny-cuda-nn` (cross-platform)
lands; see [`docs/VS-MLX-AND-TCNN.md`](docs/VS-MLX-AND-TCNN.md) for the
positioning we can honestly defend today, and [`STATUS.md`](STATUS.md) for
what is still on the roadmap.

## Dependencies

External (vcpkg manifest):

- `nlohmann_json` ≥ 3.11.3 for JSON config + checkpoint serialization
- `gtest` ≥ 1.15.0 (test-only)

Apple frameworks (system, macOS): Metal, Foundation.

## Documentation

- [`STATUS.md`](STATUS.md) — single source of truth for what works and what does not
- [`docs/QUICKSTART.md`](docs/QUICKSTART.md) — five-minute first program (C++ and Python)
- [`docs/ARCHITECTURE.md`](docs/ARCHITECTURE.md) — internal design (C++ runtime + Python binding + migration tooling)
- [`docs/VS-MLX-AND-TCNN.md`](docs/VS-MLX-AND-TCNN.md) — honest comparison with named alternatives
- [`docs/BENCHMARKS.md`](docs/BENCHMARKS.md) — benchmark methodology and how to run the benchmark binary
- [`docs/TCNN-MIGRATION-GUIDE.md`](docs/TCNN-MIGRATION-GUIDE.md) — porting tcnn projects to tmnn (C++ + Python)
- [`docs/CHECKPOINT-CONTRACT.md`](docs/CHECKPOINT-CONTRACT.md) — optimizer-state checkpoint format
- [`docs/EXTENSIBILITY-DESIGN.md`](docs/EXTENSIBILITY-DESIGN.md) — extension SDK
- [`docs/ERROR-HANDLING.md`](docs/ERROR-HANDLING.md) — `Result<T>` + `DiagnosticCode` contract

## Roadmap to v0.1.0

The following work is on the path to a stable v0.1.0 tag. Until those land,
expect breaking changes on `main` without deprecation notice:

1. Comparison benchmark vs MLX on the same hardware / same workload, with
   architecture-equivalent models on both sides
2. At least two additional flagship samples (NeRF synthetic, hash-grid SDF
   fitting)
3. CI matrix workflows shipped (`.github/workflows/test.yml`,
   `.github/workflows/release.yml`); pending first run + minutes-budget
   review on a public-repo runner
4. PyPI listing for the Python wheel (workflow gated off until a
   maintainer configures trusted publishing)
5. `MAINTAINERS.md` with response SLA

For the current honest scope ("what works today" vs "what does not work yet"),
see [`STATUS.md`](STATUS.md).

## Related projects

- [tiny-cuda-nn](https://github.com/NVlabs/tiny-cuda-nn) — the upstream model for tmnn's API surface and sample shape; the natural choice on NVIDIA hardware
- [MLX](https://github.com/ml-explore/mlx) — Apple's official ML framework; broader scope, the natural choice for general Apple-native tensor work
- [slangcsg](https://github.com/randallyanh/slangcsg) (currently private) — sibling project consuming `tmnn_core` + `tmnn_runtime` for differentiable CSG / SDF compute

## License

Apache-2.0. See [`LICENSE`](LICENSE).
