laya-apple 1.0.1__tar.gz

laya_apple-1.0.1/LICENSE

@@ -0,0 +1,176 @@
+                                 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

laya_apple-1.0.1/NOTICE

@@ -0,0 +1,33 @@
+laya-apple
+Copyright 2026 laya-apple contributors
+
+Independent project. Not an official Convai Innovations, Apple, or MLX release.
+
+This product includes software adapted from the following Apache-2.0 projects:
+
+Laya — https://github.com/NandhaKishorM/laya
+  Copyright Convai Innovations and Laya contributors.
+  Revision: 573e5b62696ba441230cd6be71d593331b5d23af
+  Adapted: prompt construction, option rendering, calibration temperatures and
+  clamping, result schema (laya_apple/prompt.py, laya_apple/result.py).
+
+laya-mlx — https://github.com/mizorewww/laya-mlx
+  Copyright 2026 laya-mlx contributors.
+  Revision: 0a859518634112655cb97c745dbf04f5191aaf13
+  Adapted: native MLX ModernBERT and Laya decision heads, weight-name mapping
+  (laya_apple/models/modernbert_mlx.py); Rust-tokenizer wrapper.
+  Modifications: packaging only; module docstring and imports.
+
+laya-coreml — https://github.com/mizorewww/laya-coreml
+  Copyright 2026 laya-coreml contributors.
+  Revision: 4619e0483f07adf39068532e85b42ec2347edb83
+  Adapted: export-only PyTorch DecisionModel (laya_apple/conversion/torch_reference.py),
+  channel-first BC1S Neural Engine graph (laya_apple/conversion/bc1s.py, from
+  experiments/ane_engineering/model.py), host embedding / action-head runtime
+  (laya_apple/backends/coreml_ane.py, from laya_coreml/ane.py), Core ML compute-plan
+  inspection (laya_apple/artifacts.py).
+  Modifications: generalised to ModernBERT-large checkpoints, strict artifact
+  verification, fixed-shape bucket management.
+
+Model weights are not included. They are downloaded from the upstream Hugging Face
+repositories at pinned revisions and remain under their respective licenses.

laya_apple-1.0.1/PKG-INFO

@@ -0,0 +1,267 @@
+Metadata-Version: 2.4
+Name: laya-apple
+Version: 1.0.1
+Summary: Correctness-validated heterogeneous Laya runtime for Apple Silicon (MLX GPU + Apple Neural Engine)
+Keywords: laya,apple-silicon,mlx,coreml,apple-neural-engine,ane,inference,heterogeneous-computing,machine-learning
+Author: oliver
+Author-email: oliver <tc3oliver@gmail.com>
+License-Expression: Apache-2.0
+License-File: LICENSE
+License-File: NOTICE
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Science/Research
+Classifier: Operating System :: MacOS
+Classifier: Operating System :: MacOS :: MacOS X
+Classifier: Environment :: GPU
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Typing :: Typed
+Requires-Dist: mlx>=0.32.2,<0.33 ; platform_machine == 'arm64' and sys_platform == 'darwin'
+Requires-Dist: numpy>=1.26
+Requires-Dist: tokenizers>=0.21,<1
+Requires-Dist: huggingface-hub>=0.34,<2
+Requires-Dist: safetensors>=0.6
+Requires-Dist: coremltools==9.0 ; extra == 'ane'
+Requires-Dist: numpy>=1.26,<2.2 ; extra == 'ane'
+Requires-Dist: coremltools==9.0 ; extra == 'convert'
+Requires-Dist: numpy>=1.26,<2.2 ; extra == 'convert'
+Requires-Dist: torch==2.7.0 ; extra == 'convert'
+Requires-Dist: pytest>=8 ; extra == 'dev'
+Requires-Dist: ruff>=0.12 ; extra == 'dev'
+Requires-Dist: psutil>=6 ; extra == 'dev'
+Requires-Dist: torch==2.7.0 ; extra == 'reference'
+Requires-Dist: transformers==5.17.0 ; extra == 'reference'
+Requires-Dist: laya==0.3.5 ; extra == 'reference'
+Requires-Python: >=3.11, <3.14
+Project-URL: Homepage, https://github.com/tc3oliver/laya-apple
+Project-URL: Documentation, https://github.com/tc3oliver/laya-apple/blob/main/docs/guide.md
+Project-URL: Changelog, https://github.com/tc3oliver/laya-apple/blob/main/CHANGELOG.md
+Project-URL: Issues, https://github.com/tc3oliver/laya-apple/issues
+Project-URL: Repository, https://github.com/tc3oliver/laya-apple
+Project-URL: Upstream, https://github.com/NandhaKishorM/laya
+Provides-Extra: ane
+Provides-Extra: convert
+Provides-Extra: dev
+Provides-Extra: reference
+Description-Content-Type: text/markdown
+
+# laya-apple
+
+[![CI](https://github.com/tc3oliver/laya-apple/actions/workflows/ci.yml/badge.svg)](https://github.com/tc3oliver/laya-apple/actions/workflows/ci.yml)
+![Python 3.11–3.13](https://img.shields.io/badge/python-3.11%E2%80%933.13-blue)
+[![License: Apache-2.0](https://img.shields.io/badge/license-Apache--2.0-green)](LICENSE)
+
+**Correctness-validated heterogeneous [Laya](https://github.com/NandhaKishorM/laya) runtime
+for Apple silicon.** It runs the MLX GPU and the Apple Neural Engine at the same time, and
+uses the Neural Engine only where it has been proven to give the same decisions as
+upstream Laya.
+
+**Mixed-workload throughput against GPU-only serving:**
+
+| Model | GPU-only | GPU + ANE | Speed-up | Answer mismatches |
+|---|---:|---:|---:|---:|
+| laya | 41.9 req/s | 122.5 req/s | **2.92×** | 0 |
+| laya-multilingual | 55.7 req/s | 241.8 req/s | **4.34×** | 0 |
+| laya-typed-decisions | 24.0 req/s | 109.6 req/s | **4.57×** | 0 |
+
+This is the v1.0 benchmark on one Apple M4 Max with macOS 26.6.2: one short and one long
+request stream through one `Laya` instance. Other Macs are untested, and you can
+[add yours](docs/community-benchmarks.md). The method and raw data are in
+[`benchmarks/v1.0.md`](benchmarks/v1.0.md).
+
+## Why this exists
+
+Laya answers typed questions about a context (`choice`, `score`, `noul`) in one forward
+pass. On a Mac there are two engines that can run it, with different strengths.
+
+1. **Correct ANE execution.** A fast Core ML export is not necessarily correct. On the
+   tested Mac, the ordinary Core ML export ran on the Neural Engine without any error and
+   changed up to 85 decisions against upstream PyTorch. laya-apple ships an ANE artifact
+   only after it passes a parity gate on the machine that uses it
+   ([`docs/correctness.md`](docs/correctness.md)).
+2. **Automatic routing.** Short, validated single-question requests go to the ANE. Long or
+   multi-question requests go to the MLX GPU. The router decides before a request runs and
+   records why.
+3. **Concurrent GPU + ANE serving.** Both engines serve independent requests at the same
+   time, so short requests stop queueing behind long ones.
+
+## Install
+
+From a clone (Apple silicon, Python 3.11–3.13; not on PyPI yet):
+
+```bash
+git clone https://github.com/tc3oliver/laya-apple && cd laya-apple
+uv sync --extra ane --extra convert
+uv run laya-apple artifacts build laya-typed-decisions   # optional: build + parity-validate ANE artifacts here (~5 min)
+```
+
+Without `artifacts build`, or with a plain `uv sync`, everything runs on the MLX GPU.
+
+## Quickstart (30 seconds)
+
+```python
+from laya_apple import Laya
+
+model = Laya.from_pretrained(
+    "convaiinnovations/laya-typed-decisions",
+    device="auto",
+)
+
+result = model.predict(
+    context="The customer was charged twice for the same invoice and is frustrated.",
+    questions={
+        "urgency": {
+            "type": "choice",
+            "instructions": "How urgent is this?",
+            "criteria": ["low", "medium", "high"],
+        }
+    },
+)
+print(result.answers["urgency"]["choice"], result.answers["urgency"]["probabilities"])
+
+rt = result.runtime
+print(rt.backend, rt.device, rt.routing_reason, f"{rt.latency_ms:.1f} ms")
+```
+
+On the tested machine:
+
+```text
+high {'low': 0.1713, 'medium': 0.3358, 'high': 0.4929}
+coreml ane validated_short_single_question_path 11.2 ms
+```
+
+- The first call downloads the pinned checkpoint. After that it works offline
+  (`local_files_only=True`).
+- Without ANE artifacts, the same request runs on MLX and `routing_reason` says why.
+- More: [`examples/`](examples/) (`basic.py`, `auto_routing.py`,
+  `heterogeneous_serving.py`) and the [user guide](docs/guide.md).
+
+## How auto routing works
+
+| Request | Goes to | Why (measured on the tested Mac) |
+|---|---|---|
+| One question, ≤ 128 tokens, validated artifact present | **ANE** | Faster: laya-typed-decisions L128 takes 9.9 ms on the ANE against 12.2 ms on MLX (forward P50) |
+| Longer context | **MLX GPU** | MLX is faster there: 19.2 ms at L256 and 71.0 ms at L1024 |
+| Several questions | **MLX GPU** | MLX batches the questions; the ANE runs them one at a time |
+| Unvalidated Mac, missing artifact, or no Core ML | **MLX GPU** | Recorded as `platform_not_validated`, `ane_artifact_unavailable` or `ane_runtime_unavailable` |
+
+**The production threshold is more conservative than the measured crossover.**
+laya-multilingual is slightly faster on the ANE at exactly 256 tokens (8.3 ms against
+8.6 ms), but that bucket stays explicit-only, because it does not beat MLX at the previous
+bucket, 128 tokens. Every result carries `routing_reason`.
+
+How the thresholds are derived: [`docs/support-matrix.md`](docs/support-matrix.md). How
+the pieces fit together: [`docs/architecture.md`](docs/architecture.md).
+
+## GPU + ANE heterogeneous serving
+
+```python
+with Laya.from_pretrained("convaiinnovations/laya-typed-decisions", execution="workers") as model:
+    futures = [model.submit(context=c, questions=q) for c, q in requests]   # thread-safe
+```
+
+- The GPU runs in a worker process, and the ANE on its own dispatcher.
+- Each request runs on one device, chosen by the router.
+- Under load, the router also compares queue backlogs.
+
+**Short-request P99 under open-loop bursty arrivals**, measured from arrival with queueing
+included (v1.0, same arrival sequence for both):
+
+| Model | GPU-only | GPU + ANE |
+|---|---:|---:|
+| laya (46.2 req/s offered) | 1538.0 ms | 108.5 ms |
+| laya-multilingual (83.8 req/s) | 2052.3 ms | 29.6 ms |
+| laya-typed-decisions (35.8 req/s) | 1592.9 ms | 79.5 ms |
+
+A single short request is not dramatically faster on the ANE (for example 9.9 against
+12.2 ms). The gain comes from using both engines at once.
+
+## Correctness
+
+Parity against upstream Laya on PyTorch CPU FP32, over the shipped golden rows (v1.0).
+Each cell gives hard mismatches, then the max probability error.
+
+| Implementation on the tested Mac | laya | laya-multilingual | laya-typed-decisions |
+|---|---|---|---|
+| Ordinary Core ML export · `CPU_AND_NE` | ❌ 12, 0.56 | ❌ 85, 1.0 | ❌ 19, 0.42 |
+| Ordinary Core ML export · `CPU_AND_GPU` | ✅ 0, 0.0066 | ✅ 0, 0.0059 | ✅ 0, 0.0028 |
+| **laya-apple MLX FP16** | ✅ 0, 0.0037 | ✅ 0, 0.0045 | ✅ 0, 0.0017 |
+| **laya-apple ANE FP16** | ✅ 0 (1 near-tie), 0.012 | ✅ 0, 0.013 | ✅ 0, 0.0077 |
+
+- **The FP16 gate:** probability error ≤ 0.02 and 0 hard mismatches.
+- **Near-tie flips** (upstream's top-two margin < 0.04) are listed, not hidden.
+- **Explicit ANE requests never fall back.** They run the validated artifact or raise, and
+  every loaded artifact is also timed against `CPU_ONLY` to catch a silent CPU placement.
+- Definitions, every configuration tested, and the fallback audit are in
+  [`docs/correctness.md`](docs/correctness.md) and
+  [`docs/no-silent-fallback.md`](docs/no-silent-fallback.md).
+
+## Supported models and platforms
+
+| Model | max_len | MLX GPU | ANE buckets (explicit) | ANE buckets used by `auto` |
+|---|---:|---|---|---|
+| [`convaiinnovations/laya`](https://huggingface.co/convaiinnovations/laya) | 512 | FP16 / FP32, any length | 64, 96, 128 | 64, 96, 128 |
+| [`convaiinnovations/laya-multilingual`](https://huggingface.co/convaiinnovations/laya-multilingual) | 1024 | FP16 / FP32, any length | 64, 96, 128, 256 | 64, 96, 128 |
+| [`convaiinnovations/laya-typed-decisions`](https://huggingface.co/convaiinnovations/laya-typed-decisions) | 1024 | FP16 / FP32, any length | 64, 96, 128 | 64, 96, 128 |
+
+**Tested:**
+- Apple M4 Max, macOS 26.6.2, MLX 0.32.2, coremltools 9.0;
+- Python 3.11–3.13.
+
+**Other Apple silicon:**
+- MLX is expected to work.
+- `auto` stays on MLX until artifacts are built and calibrated on that machine
+  (`laya-apple calibrate`).
+
+See [`docs/compatibility.md`](docs/compatibility.md) and the community matrix in
+[`docs/community-benchmarks.md`](docs/community-benchmarks.md).
+
+## Reproduction
+
+Each headline number above traces to a report, raw data, a command and an environment in
+[`docs/reproducibility.md`](docs/reproducibility.md). The full v1.0 suite, which compares
+PyTorch CPU/MPS, the ordinary Core ML export, MLX and laya-apple, is in
+[`benchmarks/v1.0.md`](benchmarks/v1.0.md). The quick check for your own Mac:
+
+```bash
+uv run python scripts/hardware_report.py --quick
+```
+
+## Contributing
+
+The most useful first contribution is a benchmark from a Mac other than an M4 Max: run
+the command above and open a PR with `hardware-results/`
+([how](docs/community-benchmarks.md)).
+
+- [`CONTRIBUTING.md`](CONTRIBUTING.md) covers setup, test tiers (which tests a change
+  actually needs), parity checks and backend changes.
+- Open work is labelled `good first issue`, `help wanted` and `research`.
+
+## Limitations
+
+- **One test machine.** Every benchmark is from one Apple M4 Max on macOS 26.6.2. Routing
+  thresholds are not assumed to hold on other Apple SoCs.
+- **Long contexts stay on MLX,** which is faster there. The ANE path is batch 1 only.
+- **Isolation is partial.** Under concurrency, each stream's P99 is above its solo value.
+- **Cold start** on a fresh artifact location costs 3–5 minutes of Core ML compile per
+  model. `ane_startup="background"` serves on MLX in the meantime.
+- **`choice` decisions can depend on option order.** This comes from upstream Laya, and
+  laya-apple reproduces it exactly ([`research/option-order/`](research/option-order/)).
+- **Not measured yet:** energy use, quantized artifacts and cross-SoC validation.
+
+## More
+
+- User guide: [`docs/guide.md`](docs/guide.md).
+- Stable API: [`docs/api.md`](docs/api.md).
+- Architecture: [`docs/architecture.md`](docs/architecture.md).
+- Changes: [`CHANGELOG.md`](CHANGELOG.md).
+- Security: [`SECURITY.md`](SECURITY.md).
+
+Apache-2.0; see [`LICENSE`](LICENSE) and [`NOTICE`](NOTICE). Model weights are downloaded
+from their pinned Hugging Face revisions and are not redistributed. This is an independent
+project, not an official release of Convai Innovations, Apple or MLX.