executorch 1.0.0-x86_64-linux

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
checksums.yaml ADDED
@@ -0,0 +1,7 @@
1
+ ---
2
+ SHA256:
3
+ metadata.gz: 95f5d47d1b68bda775a78d0826a995a725733a6739055c5df748ede3b0fe0097
4
+ data.tar.gz: e13448a4d75a7dfea424dd281aa476d3120e534976400e2a1bbb6d41c814e45b
5
+ SHA512:
6
+ metadata.gz: 651f80ff56c5cbec156b9918efa6aa8df13cee2bb206aec1ef3ef30d096603c7ba112c734c477685c321f1837b695708e2b05ef7048e8f64a62c90a35fe35453
7
+ data.tar.gz: f47050d6136642d3c173496c25c6fb40ec23988d91c8ba44f38b00d902933957a9e5c4d92b67c2436413f09fcd0251ee9108ba53c86fdbe830fe6afd035ee5b3
@@ -0,0 +1 @@
1
+ 1.6.0
data/CHANGELOG.md ADDED
@@ -0,0 +1,33 @@
1
+ # Changelog
2
+
3
+ All notable changes to this project will be documented in this file.
4
+
5
+ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
6
+ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
7
+
8
+ ## [1.0.0] - 2026-11-19
9
+
10
+ Complete from-scratch rewrite targeting ExecuTorch 1.6.0 and Ruby >= 3.2.
11
+
12
+ ### Breaking Changes
13
+
14
+ - Replaced `Executorch::Model` with `Executorch::Module` (`Module.open`, `Module.from_bytes`, `#call`, `#forward`, `#execute`, `#meta`, `#load_method`, `#unload_method`, `#close`).
15
+ - Canonical tensor dtype symbols are now explicit-width (`:float32`, `:float64`, `:float16`, `:bfloat16`, `:uint8`, `:int8`, `:int16`, `:uint16`, `:int32`, `:uint32`, `:int64`, `:uint64`, `:bool`). Short aliases (`:float`, `:double`, `:int`, `:long`, etc.) remain accepted on input.
16
+ - Raised minimum Ruby version to `>= 3.2.0`.
17
+ - Removed `rice` dependency and leaked internal methods (`_original_to_a`, `Tensor.create`, `Tensor.from_array`, `NATIVE_VERSION`, `NativeError`).
18
+
19
+ ### Added
20
+
21
+ - Two-layer `noexcept` C++17 core beneath a direct Ruby C-API seam — eliminates `rb_raise` `longjmp`-past-destructor UB by construction.
22
+ - Strict symbol visibility (`-fvisibility=hidden` + ELF version script / macOS `-exported_symbol,_Init_executorch`) preventing ODR collisions with `torch.rb`, `onnxruntime-ruby`, or `ruby-vips`.
23
+ - GVL release (`rb_thread_call_without_gvl`) during module load, method load, and execution, paired with per-method thread-safety mutexes and `RB_GC_GUARD`.
24
+ - Native memory pressure reporting (`dsize` + `rb_gc_adjust_memory_usage`) and deterministic `#close` / block-form `Module.open`.
25
+ - Support for all 13 numeric/boolean dtypes, 0-D scalar tensors (`Tensor#item`), non-contiguous / `NHWC` `dim_order` normalization, value equality (`Tensor#==`), `Tensor.zeros`/`ones`/`full`, `Tensor.from_bytes`/`#to_bytes`, zero-copy `IO::Buffer` (`Tensor.from_buffer`/`#to_buffer`), pre-allocated `out:` output reuse, and optional `Numo::NArray`/`Torch::Tensor` adapters.
26
+ - `Module.open` defaults to `:mmap` (supporting `Pathname`, all `LoadMode`s, and external `.ptd` `data_files`), plus `Module.from_bytes` (`BufferDataLoader`) and `MethodMeta` / `TensorInfo` introspection.
27
+ - Runtime introspection (`Executorch.backends`, `.kernel_count`, `.num_threads`, `.num_threads=`, `.executorch_version`), PAL initialization (`runtime_init`), and thread-local PAL error log capture attached to typed `Executorch::Error` subclasses (`LoadError`, `MethodNotFoundError`, `OperatorMissingError`, `DelegateError`, `ShapeError`, `DTypeError`, `ExecutionError`, `MemoryError`, `ClosedError`).
28
+ - Precompiled platform gems (`x86_64-linux`, `aarch64-linux`, `arm64-darwin`) bundling the official ExecuTorch 1.6.0 shared SDK (XNNPACK, optimized & quantized kernels, CoreML/MLX on Apple Silicon) plus a pure-stdlib SHA-256-verified SDK fetcher for source installs — no Python, CMake, or Buck2 required at install or inference time.
29
+ - Deterministic `.pte`/`.ptd` test fixture export pipeline (`test/fixtures/export_fixtures.py`) and comprehensive correctness, ASan/UBSan, GC stress, concurrency, and adversarial safety test suites (0 skipped tests).
30
+
31
+ ## [0.1.0] - 2025-11-16
32
+
33
+ Initial prototype release.
data/LICENSE.txt ADDED
@@ -0,0 +1,176 @@
1
+ Apache License
2
+ Version 2.0, January 2004
3
+ http://www.apache.org/licenses/
4
+
5
+ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
6
+
7
+ 1. Definitions.
8
+
9
+ "License" shall mean the terms and conditions for use, reproduction,
10
+ and distribution as defined by Sections 1 through 9 of this document.
11
+
12
+ "Licensor" shall mean the copyright owner or entity authorized by
13
+ the copyright owner that is granting the License.
14
+
15
+ "Legal Entity" shall mean the union of the acting entity and all
16
+ other entities that control, are controlled by, or are under common
17
+ control with that entity. For the purposes of this definition,
18
+ "control" means (i) the power, direct or indirect, to cause the
19
+ direction or management of such entity, whether by contract or
20
+ otherwise, or (ii) ownership of fifty percent (50%) or more of the
21
+ outstanding shares, or (iii) beneficial ownership of such entity.
22
+
23
+ "You" (or "Your") shall mean an individual or Legal Entity
24
+ exercising permissions granted by this License.
25
+
26
+ "Source" form shall mean the preferred form for making modifications,
27
+ including but not limited to software source code, documentation
28
+ source, and configuration files.
29
+
30
+ "Object" form shall mean any form resulting from mechanical
31
+ transformation or translation of a Source form, including but
32
+ not limited to compiled object code, generated documentation,
33
+ and conversions to other media types.
34
+
35
+ "Work" shall mean the work of authorship, whether in Source or
36
+ Object form, made available under the License, as indicated by a
37
+ copyright notice that is included in or attached to the work
38
+ (an example is provided in the Appendix below).
39
+
40
+ "Derivative Works" shall mean any work, whether in Source or Object
41
+ form, that is based on (or derived from) the Work and for which the
42
+ editorial revisions, annotations, elaborations, or other modifications
43
+ represent, as a whole, an original work of authorship. For the purposes
44
+ of this License, Derivative Works shall not include works that remain
45
+ separable from, or merely link (or bind by name) to the interfaces of,
46
+ the Work and Derivative Works thereof.
47
+
48
+ "Contribution" shall mean any work of authorship, including
49
+ the original version of the Work and any modifications or additions
50
+ to that Work or Derivative Works thereof, that is intentionally
51
+ submitted to the Licensor for inclusion in the Work by the copyright owner
52
+ or by an individual or Legal Entity authorized to submit on behalf of
53
+ the copyright owner. For the purposes of this definition, "submitted"
54
+ means any form of electronic, verbal, or written communication sent
55
+ to the Licensor or its representatives, including but not limited to
56
+ communication on electronic mailing lists, source code control systems,
57
+ and issue tracking systems that are managed by, or on behalf of, the
58
+ Licensor for the purpose of discussing and improving the Work, but
59
+ excluding communication that is conspicuously marked or otherwise
60
+ designated in writing by the copyright owner as "Not a Contribution."
61
+
62
+ "Contributor" shall mean Licensor and any individual or Legal Entity
63
+ on behalf of whom a Contribution has been received by Licensor and
64
+ subsequently incorporated within the Work.
65
+
66
+ 2. Grant of Copyright License. Subject to the terms and conditions of
67
+ this License, each Contributor hereby grants to You a perpetual,
68
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
69
+ copyright license to reproduce, prepare Derivative Works of,
70
+ publicly display, publicly perform, sublicense, and distribute the
71
+ Work and such Derivative Works in Source or Object form.
72
+
73
+ 3. Grant of Patent License. Subject to the terms and conditions of
74
+ this License, each Contributor hereby grants to You a perpetual,
75
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
76
+ (except as stated in this section) patent license to make, have made,
77
+ use, offer to sell, sell, import, and otherwise transfer the Work,
78
+ where such license applies only to those patent claims licensable
79
+ by such Contributor that are necessarily infringed by their
80
+ Contribution(s) alone or by combination of their Contribution(s)
81
+ with the Work to which such Contribution(s) was submitted. If You
82
+ institute patent litigation against any entity (including a
83
+ cross-claim or counterclaim in a lawsuit) alleging that the Work
84
+ or a Contribution incorporated within the Work constitutes direct
85
+ or contributory patent infringement, then any patent licenses
86
+ granted to You under this License for that Work shall terminate
87
+ as of the date such litigation is filed.
88
+
89
+ 4. Redistribution. You may reproduce and distribute copies of the
90
+ Work or Derivative Works thereof in any medium, with or without
91
+ modifications, and in Source or Object form, provided that You
92
+ meet the following conditions:
93
+
94
+ (a) You must give any other recipients of the Work or
95
+ Derivative Works a copy of this License; and
96
+
97
+ (b) You must cause any modified files to carry prominent notices
98
+ stating that You changed the files; and
99
+
100
+ (c) You must retain, in the Source form of any Derivative Works
101
+ that You distribute, all copyright, patent, trademark, and
102
+ attribution notices from the Source form of the Work,
103
+ excluding those notices that do not pertain to any part of
104
+ the Derivative Works; and
105
+
106
+ (d) If the Work includes a "NOTICE" text file as part of its
107
+ distribution, then any Derivative Works that You distribute must
108
+ include a readable copy of the attribution notices contained
109
+ within such NOTICE file, excluding those notices that do not
110
+ pertain to any part of the Derivative Works, in at least one
111
+ of the following places: within a NOTICE text file distributed
112
+ as part of the Derivative Works; within the Source form or
113
+ documentation, if provided along with the Derivative Works; or,
114
+ within a display generated by the Derivative Works, if and
115
+ wherever such third-party notices normally appear. The contents
116
+ of the NOTICE file are for informational purposes only and
117
+ do not modify the License. You may add Your own attribution
118
+ notices within Derivative Works that You distribute, alongside
119
+ or as an addendum to the NOTICE text from the Work, provided
120
+ that such additional attribution notices cannot be construed
121
+ as modifying the License.
122
+
123
+ You may add Your own copyright statement to Your modifications and
124
+ may provide additional or different license terms and conditions
125
+ for use, reproduction, or distribution of Your modifications, or
126
+ for any such Derivative Works as a whole, provided Your use,
127
+ reproduction, and distribution of the Work otherwise complies with
128
+ the conditions stated in this License.
129
+
130
+ 5. Submission of Contributions. Unless You explicitly state otherwise,
131
+ any Contribution intentionally submitted for inclusion in the Work
132
+ by You to the Licensor shall be under the terms and conditions of
133
+ this License, without any additional terms or conditions.
134
+ Notwithstanding the above, nothing herein shall supersede or modify
135
+ the terms of any separate license agreement you may have executed
136
+ with Licensor regarding such Contributions.
137
+
138
+ 6. Trademarks. This License does not grant permission to use the trade
139
+ names, trademarks, service marks, or product names of the Licensor,
140
+ except as required for reasonable and customary use in describing the
141
+ origin of the Work and reproducing the content of the NOTICE file.
142
+
143
+ 7. Disclaimer of Warranty. Unless required by applicable law or
144
+ agreed to in writing, Licensor provides the Work (and each
145
+ Contributor provides its Contributions) on an "AS IS" BASIS,
146
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
147
+ implied, including, without limitation, any warranties or conditions
148
+ of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
149
+ PARTICULAR PURPOSE. You are solely responsible for determining the
150
+ appropriateness of using or redistributing the Work and assume any
151
+ risks associated with Your exercise of permissions under this License.
152
+
153
+ 8. Limitation of Liability. In no event and under no legal theory,
154
+ whether in tort (including negligence), contract, or otherwise,
155
+ unless required by applicable law (such as deliberate and grossly
156
+ negligent acts) or agreed to in writing, shall any Contributor be
157
+ liable to You for damages, including any direct, indirect, special,
158
+ incidental, or consequential damages of any character arising as a
159
+ result of this License or out of the use or inability to use the
160
+ Work (including but not limited to damages for loss of goodwill,
161
+ work stoppage, computer failure or malfunction, or any and all
162
+ other commercial damages or losses), even if such Contributor
163
+ has been advised of the possibility of such damages.
164
+
165
+ 9. Accepting Warranty or Additional Liability. While redistributing
166
+ the Work or Derivative Works thereof, You may choose to offer,
167
+ and charge a fee for, acceptance of support, warranty, indemnity,
168
+ or other liability obligations and/or rights consistent with this
169
+ License. However, in accepting such obligations, You may act only
170
+ on Your own behalf and on Your sole responsibility, not on behalf
171
+ of any other Contributor, and only if You agree to indemnify,
172
+ defend, and hold each Contributor harmless for any liability
173
+ incurred by, or claims asserted against, such Contributor by reason
174
+ of your accepting any such warranty or additional liability.
175
+
176
+ END OF TERMS AND CONDITIONS
data/README.md ADDED
@@ -0,0 +1,212 @@
1
+ # executorch-ruby
2
+
3
+ [![CI](https://github.com/benngarcia/executorch-ruby/actions/workflows/ci.yml/badge.svg)](https://github.com/benngarcia/executorch-ruby/actions/workflows/ci.yml)
4
+ [![Gem Version](https://badge.fury.io/rb/executorch.svg)](https://rubygems.org/gems/executorch)
5
+ [![License](https://img.shields.io/badge/license-Apache%202.0-blue.svg)](LICENSE.txt)
6
+
7
+ Native Ruby bindings for Meta's [ExecuTorch](https://pytorch.org/executorch/) runtime (`1.6.0`), built directly on the Ruby C API.
8
+
9
+ Run exported PyTorch models (`.pte` / `.ptd`) in Ruby with precompiled platform gems, GVL-free execution, zero-copy `IO::Buffer` tensors, and hardware delegates (XNNPACK, CoreML, MLX).
10
+
11
+ ## Installation
12
+
13
+ Add to your `Gemfile`:
14
+
15
+ ```ruby
16
+ gem "executorch", "~> 1.0"
17
+ ```
18
+
19
+ Or install directly:
20
+
21
+ ```bash
22
+ gem install executorch
23
+ ```
24
+
25
+ Requires **Ruby `>= 3.2`**.
26
+
27
+ ### Supported platforms
28
+
29
+ | Platform | Precompiled Gem | Bundled Delegates & Kernels |
30
+ |---|---|---|
31
+ | `x86_64-linux` | Yes | XNNPACK, Optimized CPU, Quantized |
32
+ | `aarch64-linux` | Yes | XNNPACK (KleidiAI), Optimized CPU, Quantized, TorchAO |
33
+ | `arm64-darwin` | Yes | XNNPACK, CoreML, MLX, Optimized CPU, Quantized, TorchAO |
34
+ | Source (`ruby`) | Yes | Fetches the official ExecuTorch 1.6.0 C++ SDK (or uses `--with-executorch-dir`) |
35
+
36
+ ## Quick Start
37
+
38
+ Export a PyTorch model to `.pte` using ExecuTorch:
39
+
40
+ ```python
41
+ import torch
42
+ from executorch.backends.xnnpack.partition.xnnpack_partitioner import XnnpackPartitioner
43
+ from executorch.exir import to_edge_transform_and_lower
44
+
45
+ class AddMul(torch.nn.Module):
46
+ def forward(self, x):
47
+ return x * 2.0 + 1.0
48
+
49
+ example_input = (torch.tensor([[1.0, 2.0, 3.0]], dtype=torch.float32),)
50
+ exported = torch.export.export(AddMul().eval(), example_input)
51
+ et_program = to_edge_transform_and_lower(
52
+ exported, partitioner=[XnnpackPartitioner()]
53
+ ).to_executorch()
54
+
55
+ with open("model.pte", "wb") as f:
56
+ et_program.write_to_file(f)
57
+ ```
58
+
59
+ Load and run it in Ruby:
60
+
61
+ ```ruby
62
+ require "executorch"
63
+
64
+ Executorch::Module.open("model.pte") do |mod|
65
+ meta = mod.meta(:forward)
66
+ p meta.backends # => ["XnnpackBackend"]
67
+ p meta.inputs # => [#<Executorch::TensorInfo dtype=:float32 shape=[1, 3] nbytes=12>]
68
+
69
+ input = Executorch::Tensor.new([[1.0, 2.0, 3.0]])
70
+ output = mod.call(input)
71
+ p output.to_a # => [[3.0, 5.0, 7.0]]
72
+ end
73
+ ```
74
+
75
+ ## Usage
76
+
77
+ ### Modules
78
+
79
+ `Executorch::Module.open` accepts a `String` or `Pathname`, defaults to `:mmap` loading (sharing read-only weight pages across forked Puma/Unicorn workers), and closes deterministically when given a block.
80
+
81
+ ```ruby
82
+ # Load modes: :mmap (default), :file, :mmap_mlock, :mmap_mlock_ignore_errors, :mmap_madvise
83
+ mod = Executorch::Module.open(Pathname.new("model.pte"), load_mode: :mmap)
84
+ mod.load_method(:forward) # optional pre-warm before serving requests
85
+ out = mod.call(input)
86
+ mod.close
87
+
88
+ # Multi-method programs
89
+ Executorch::Module.open("autoencoder.pte") do |m|
90
+ z = m.execute(:encode, x).first
91
+ reconstructed = m.execute(:decode, z).first
92
+ end
93
+
94
+ # External .ptd weight files
95
+ Executorch::Module.open("model.pte", data_files: ["weights.ptd"]) do |m|
96
+ logits = m.call(tokens)
97
+ end
98
+
99
+ # Load directly from an in-memory String (e.g. S3 or ActiveStorage)
100
+ mod = Executorch::Module.from_bytes(pte_bytes)
101
+ mod.close
102
+ ```
103
+
104
+ - `#call(*inputs, out: nil)` runs `"forward"` and returns the single output directly when there is one output, or an `Array` otherwise.
105
+ - `#forward(*inputs, out: nil)` and `#execute(name, *inputs, out: nil)` always return an `Array`.
106
+ - `#meta(name = :forward)` returns an `Executorch::MethodMeta` describing input/output shapes, dtypes, planned memory size, and lowered backends without loading the full method.
107
+
108
+ ### Tensors
109
+
110
+ Supported dtypes: `:float32`, `:float64`, `:float16`, `:bfloat16`, `:uint8`, `:int8`, `:int16`, `:uint16`, `:int32`, `:uint32`, `:int64`, `:uint64`, and `:bool`. Short aliases (`:float`, `:double`, `:half`, `:byte`, `:char`, `:short`, `:int`, `:long`) are accepted at construction.
111
+
112
+ ```ruby
113
+ t = Executorch::Tensor.new([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]])
114
+ t.shape # => [2, 3]
115
+ t.dtype # => :float32
116
+ t.nbytes # => 24
117
+ t.to_a # => [[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]]
118
+
119
+ # 0-D scalar tensors
120
+ scalar = Executorch::Tensor.new(42.5, dtype: :float32)
121
+ scalar.shape # => []
122
+ scalar.item # => 42.5
123
+
124
+ # Packed binary I/O & factories
125
+ raw = t.to_bytes
126
+ t2 = Executorch::Tensor.from_bytes(raw, shape: [2, 3], dtype: :float32)
127
+ t == t2 # => true
128
+
129
+ zeros = Executorch::Tensor.zeros([2, 3], dtype: :float32)
130
+ ones = Executorch::Tensor.ones([2, 3], dtype: :int64)
131
+ full = Executorch::Tensor.full([2, 3], 7, dtype: :uint8)
132
+ ```
133
+
134
+ Optional adapters are provided for `Numo::NArray` (`Tensor.from_numo` / `Tensor#to_numo`) and `Torch::Tensor` (`Tensor.from_torch`).
135
+
136
+ ### Zero-copy `IO::Buffer` and output reuse
137
+
138
+ Pass Ruby `IO::Buffer` instances directly into ExecuTorch without copying or allocating intermediate arrays. Pairing `Tensor.from_buffer` with `out:` writes inference outputs straight into a caller-owned buffer:
139
+
140
+ ```ruby
141
+ in_buf = IO::Buffer.for([1.0, 2.0, 3.0].pack("f*"))
142
+ out_buf = IO::Buffer.new(12)
143
+
144
+ input = Executorch::Tensor.from_buffer(in_buf, shape: [1, 3], dtype: :float32)
145
+ output = Executorch::Tensor.from_buffer(out_buf, shape: [1, 3], dtype: :float32)
146
+
147
+ Executorch::Module.open("model.pte") do |mod|
148
+ mod.call(input, out: output)
149
+ p out_buf.get_string.unpack("f*") # => [3.0, 5.0, 7.0]
150
+ end
151
+ ```
152
+
153
+ ### Concurrency, runtime introspection & errors
154
+
155
+ ```ruby
156
+ Executorch.backends # => ["XnnpackBackend"]
157
+ Executorch.kernel_count # => 263
158
+ Executorch.num_threads # => 24
159
+ Executorch.num_threads = 2 # size the ExecuTorch threadpool for multi-worker servers
160
+ Executorch.executorch_version # => "1.6.0"
161
+ ```
162
+
163
+ - **GVL release & thread safety:** Model loading and execution release the Ruby GVL (`rb_thread_call_without_gvl`). Calls on separate `Module` instances run in parallel across cores; concurrent calls to the same method on a shared `Module` serialize safely via a per-method mutex.
164
+ - **Typed errors:** Runtime failures inherit from `Executorch::Error < StandardError` (`LoadError`, `MethodNotFoundError`, `OperatorMissingError`, `DelegateError`, `ShapeError`, `DTypeError`, `ExecutionError`, `MemoryError`, `ClosedError`) and expose `#code` and `#pal_log`.
165
+
166
+ ## Architecture
167
+
168
+ The native extension is written directly against the Ruby C API (no binding generator or runtime gem dependencies) in two layers:
169
+
170
+ - **`ext/executorch/core/`** — A `noexcept` C++17 core with zero Ruby headers. Handles ExecuTorch lifecycle, per-method locking, thread-local PAL log capture, overflow-checked shape math, and strided/`NHWC` layout normalization.
171
+ - **`ext/executorch/rb_*.cpp`** — A thin Ruby C-API layer managing `TypedData`, `dsize` + `rb_gc_adjust_memory_usage`, `GC.compact`-safe global roots, `rb_io_buffer_lock` around GVL-free execution, and single-pass C++ nested-array flattening. All C++ stack objects are destroyed before any Ruby exception is raised.
172
+ - **Symbol isolation:** Compiled with `-fvisibility=hidden` and a linker export map so only `Init_executorch` is exported, avoiding symbol collisions with `torch.rb`, `onnxruntime`, or `ruby-vips`.
173
+
174
+ ## Custom ExecuTorch builds
175
+
176
+ To compile against a local ExecuTorch installation instead of the default SDK:
177
+
178
+ ```bash
179
+ bundle config set --local build.executorch --with-executorch-dir=/absolute/path/to/executorch
180
+ ```
181
+
182
+ Or pass `EXECUTORCH_DIR=/absolute/path/to/executorch`. Set `EXECUTORCH_NO_DOWNLOAD=1` to disable automatic SDK fetching.
183
+
184
+ ## Changelog
185
+
186
+ See [CHANGELOG.md](CHANGELOG.md) for release notes and migration details.
187
+
188
+ ## Development
189
+
190
+ ```bash
191
+ bundle install
192
+ bundle exec rake sdk:fetch # Fetches verified ExecuTorch 1.6.0 C++ SDK into ~/.cache/executorch-ruby/
193
+ bundle exec rake compile # Compiles native C++17 extension
194
+ bundle exec rake fixtures:export # Exports deterministic .pte/.ptd test models via uv
195
+ bundle exec rake test # Runs full test suite
196
+ ruby bench/run_bench.rb # Runs local latency benchmarks
197
+ ```
198
+
199
+ To run the suite under AddressSanitizer + UndefinedBehaviorSanitizer on Linux:
200
+
201
+ ```bash
202
+ bundle exec rake clean
203
+ EXECUTORCH_SANITIZE=1 bundle exec rake compile
204
+ LD_PRELOAD="$(g++ -print-file-name=libasan.so):$(g++ -print-file-name=libubsan.so)" \
205
+ ASAN_OPTIONS=detect_leaks=0:use_sigaltstack=0:halt_on_error=1 \
206
+ UBSAN_OPTIONS=print_stacktrace=1:halt_on_error=1 \
207
+ bundle exec rake test
208
+ ```
209
+
210
+ ## License
211
+
212
+ Apache License 2.0. See [LICENSE.txt](LICENSE.txt).
Binary file
Binary file
Binary file
@@ -0,0 +1,70 @@
1
+ # frozen_string_literal: true
2
+
3
+ module Executorch
4
+ # Immutable metadata describing a single input or output slot of a method.
5
+ class TensorInfo
6
+ attr_reader :tag, :dtype, :shape, :dim_order, :nbytes
7
+
8
+ def initialize(tag, dtype, shape, dim_order, nbytes, memory_planned)
9
+ @tag = tag
10
+ @dtype = dtype
11
+ @shape = shape&.freeze
12
+ @dim_order = dim_order&.freeze
13
+ @nbytes = nbytes
14
+ @memory_planned = !!memory_planned
15
+ freeze
16
+ end
17
+
18
+ def tensor?
19
+ @tag == :tensor
20
+ end
21
+
22
+ def memory_planned?
23
+ @memory_planned
24
+ end
25
+
26
+ def ==(other)
27
+ other.is_a?(TensorInfo) &&
28
+ @tag == other.tag &&
29
+ @dtype == other.dtype &&
30
+ @shape == other.shape &&
31
+ @dim_order == other.dim_order &&
32
+ @nbytes == other.nbytes &&
33
+ @memory_planned == other.memory_planned?
34
+ end
35
+ alias eql? ==
36
+
37
+ def inspect
38
+ if tensor?
39
+ "#<Executorch::TensorInfo dtype=#{@dtype.inspect} shape=#{@shape.inspect} nbytes=#{@nbytes}>"
40
+ else
41
+ "#<Executorch::TensorInfo tag=#{@tag.inspect}>"
42
+ end
43
+ end
44
+ alias to_s inspect
45
+ end
46
+
47
+ # Immutable metadata snapshot for an ExecuTorch method, readable without
48
+ # paying the memory-planning cost of loading the full method.
49
+ class MethodMeta
50
+ attr_reader :name, :inputs, :outputs, :backends, :planned_memory_bytes
51
+
52
+ def initialize(name, inputs, outputs, backends, planned_memory_bytes)
53
+ @name = name.dup.freeze
54
+ @inputs = inputs.freeze
55
+ @outputs = outputs.freeze
56
+ @backends = backends.freeze
57
+ @planned_memory_bytes = planned_memory_bytes
58
+ freeze
59
+ end
60
+
61
+ def uses_backend?(backend_name)
62
+ @backends.include?(backend_name.to_s)
63
+ end
64
+
65
+ def inspect
66
+ "#<Executorch::MethodMeta name=#{@name.inspect} inputs=#{@inputs.size} outputs=#{@outputs.size} backends=#{@backends.inspect} planned_memory_bytes=#{@planned_memory_bytes}>"
67
+ end
68
+ alias to_s inspect
69
+ end
70
+ end
@@ -0,0 +1,56 @@
1
+ # frozen_string_literal: true
2
+
3
+ module Executorch
4
+ class Module
5
+ class << self
6
+ # Open an ExecuTorch .pte program from disk.
7
+ # Defaults to :mmap for copy-on-write page sharing across forked workers.
8
+ # Supports an optional block for deterministic resource cleanup.
9
+ #
10
+ # @param path [String, Pathname] Path to .pte file
11
+ # @param load_mode [Symbol] :mmap (default), :file, :mmap_mlock,
12
+ # :mmap_mlock_ignore_errors, or :mmap_madvise
13
+ # @param data_files [Array<String, Pathname>, nil] Optional .ptd external weight files
14
+ def open(path, load_mode: :mmap, data_files: nil)
15
+ mod = _open_file(path, load_mode: load_mode, data_files: data_files)
16
+ return mod unless block_given?
17
+
18
+ begin
19
+ yield mod
20
+ ensure
21
+ mod.close
22
+ end
23
+ end
24
+ alias load open
25
+ end
26
+
27
+ # Execute a named method and always return an Array of outputs.
28
+ #
29
+ # @param method_name [String, Symbol] Method name (e.g. :forward, :encode)
30
+ # @param inputs [Array<Tensor, Integer, Float, Boolean, String, nil>]
31
+ # @param out [Tensor, Array<Tensor>, nil] Optional pre-allocated output tensor(s)
32
+ # @return [Array]
33
+ def execute(method_name, *inputs, out: nil)
34
+ _execute(method_name, inputs, out)
35
+ end
36
+
37
+ # Execute the "forward" method and always return an Array of outputs.
38
+ def forward(*inputs, out: nil)
39
+ _execute("forward", inputs, out)
40
+ end
41
+
42
+ # Ergonomic invocation of "forward": returns the single output directly when
43
+ # the method has 1 output, or an Array when it has 0 or >1 outputs.
44
+ def call(*inputs, out: nil)
45
+ outputs = _execute("forward", inputs, out)
46
+ outputs.size == 1 ? outputs.first : outputs
47
+ end
48
+
49
+ def inspect
50
+ return "#<Executorch::Module (closed)>" if closed?
51
+
52
+ "#<Executorch::Module source=#{source.inspect} methods=#{method_names.inspect}>"
53
+ end
54
+ alias to_s inspect
55
+ end
56
+ end
@@ -0,0 +1,251 @@
1
+ # frozen_string_literal: true
2
+
3
+ require "digest"
4
+ require "fileutils"
5
+ require "net/http"
6
+ require "rbconfig"
7
+ require "tmpdir"
8
+ require "uri"
9
+ require "zlib"
10
+
11
+ module Executorch
12
+ # Fetches and extracts the official prebuilt ExecuTorch C++ SDK (headers,
13
+ # shared runtime/backend/kernel libraries, pkg-config, and CMake configs)
14
+ # from official PyPI wheels with strict SHA-256 verification and zero
15
+ # non-stdlib dependencies.
16
+ module SDK
17
+ VERSION = File.read(File.expand_path("../../.executorch-version", __dir__)).strip.freeze
18
+
19
+ # Official ExecuTorch 1.6.0 PyPI wheel manifest (only include/, lib/, share/cmake/ are extracted).
20
+ MANIFEST = {
21
+ "x86_64-linux" => {
22
+ url: "https://files.pythonhosted.org/packages/ef/24/f8fde978e1e5fd629d68c74a3f4acee30acb05f4f90c433cc1a8e29d8eed/executorch-1.6.0-cp312-cp312-manylinux_2_28_x86_64.whl",
23
+ sha256: "d05efe9afe00495112ecf304ca548fcd3239d39e96ac25381b528cdbad948b54",
24
+ },
25
+ "aarch64-linux" => {
26
+ url: "https://files.pythonhosted.org/packages/d5/7f/9028996ec1dead9c57ed2c6384037e3567c2f514e4ed0c54480d2d3eeb48/executorch-1.6.0-cp312-cp312-manylinux_2_28_aarch64.whl",
27
+ sha256: "ff595d236f4052a196a03a867ca0f53d2e3b003d00ab871433660360e194e78e",
28
+ },
29
+ "arm64-darwin" => {
30
+ url: "https://files.pythonhosted.org/packages/c6/47/1e2120ca1ee178e81471d57255ea0f87b6e04272e32ba20e5ed60ebdb46d/executorch-1.6.0-cp312-cp312-macosx_14_0_arm64.whl",
31
+ sha256: "02c220037b41a0a6074a94499a747400ef1a5075a3f2b4d83911871c7291a341",
32
+ },
33
+ }.freeze
34
+
35
+ # Core runtime + CPU/hardware delegate shared libs bundled in fat gems.
36
+ VENDOR_LIBS = {
37
+ "x86_64-linux" => %w[
38
+ libexecutorch.so
39
+ libexecutorch_threadpool.so
40
+ libexecutorch_kernels_optimized.so
41
+ libexecutorch_kernels_quantized.so
42
+ libexecutorch_backend_xnnpack.so
43
+ libexecutorch_etdump.so
44
+ ],
45
+ "aarch64-linux" => %w[
46
+ libexecutorch.so
47
+ libexecutorch_threadpool.so
48
+ libexecutorch_kernels_optimized.so
49
+ libexecutorch_kernels_quantized.so
50
+ libexecutorch_kernels_torchao.so
51
+ libexecutorch_backend_xnnpack.so
52
+ libexecutorch_etdump.so
53
+ ],
54
+ "arm64-darwin" => %w[
55
+ libexecutorch.dylib
56
+ libexecutorch_threadpool.dylib
57
+ libexecutorch_kernels_optimized.dylib
58
+ libexecutorch_kernels_quantized.dylib
59
+ libexecutorch_kernels_torchao.dylib
60
+ libexecutorch_backend_xnnpack.dylib
61
+ libexecutorch_backend_coreml.dylib
62
+ libexecutorch_backend_mlx.dylib
63
+ libexecutorch_etdump.dylib
64
+ mlx.metallib
65
+ ],
66
+ }.freeze
67
+
68
+ module_function
69
+
70
+ def platform_key(os: RbConfig::CONFIG["target_os"], cpu: RbConfig::CONFIG["target_cpu"])
71
+ norm_cpu = case cpu
72
+ when /x86_64|amd64/i then "x86_64"
73
+ when /aarch64|arm64/i then os =~ /darwin/i ? "arm64" : "aarch64"
74
+ else cpu
75
+ end
76
+ norm_os = case os
77
+ when /darwin/i then "darwin"
78
+ when /linux/i then "linux"
79
+ else os
80
+ end
81
+ "#{norm_cpu}-#{norm_os}"
82
+ end
83
+
84
+ def default_cache_dir(key = platform_key)
85
+ base = ENV["XDG_CACHE_HOME"]
86
+ base = File.join(Dir.home, ".cache") if base.nil? || base.empty?
87
+ File.join(base, "executorch-ruby", "sdk-#{VERSION}-#{key}")
88
+ end
89
+
90
+ def valid_sdk?(dir)
91
+ return false unless dir && File.directory?(dir)
92
+
93
+ File.exist?(File.join(dir, "include", "executorch", "extension", "module", "module.h")) &&
94
+ File.directory?(File.join(dir, "lib"))
95
+ end
96
+
97
+ def ensure_sdk(dest_dir: nil, key: platform_key, io: $stdout)
98
+ dest = dest_dir || default_cache_dir(key)
99
+ return dest if valid_sdk?(dest)
100
+
101
+ entry = MANIFEST[key]
102
+ raise "No prebuilt ExecuTorch #{VERSION} SDK manifest for platform '#{key}'. Set --with-executorch-dir or EXECUTORCH_DIR." unless entry
103
+
104
+ io&.puts "Fetching official ExecuTorch #{VERSION} C++ SDK for #{key}..."
105
+ Dir.mktmpdir("executorch-sdk") do |tmp|
106
+ whl_path = File.join(tmp, "executorch.whl")
107
+ download_with_redirects(entry[:url], whl_path)
108
+ actual_sha = Digest::SHA256.file(whl_path).hexdigest
109
+ unless actual_sha == entry[:sha256]
110
+ raise "SHA-256 mismatch for ExecuTorch SDK (#{key}): expected #{entry[:sha256]}, got #{actual_sha}"
111
+ end
112
+
113
+ stage = File.join(tmp, "stage")
114
+ extract_sdk_from_wheel(whl_path, stage)
115
+ FileUtils.mkdir_p(File.dirname(dest))
116
+ FileUtils.rm_rf(dest)
117
+ FileUtils.mv(stage, dest)
118
+ end
119
+
120
+ raise "Extracted SDK at #{dest} is incomplete" unless valid_sdk?(dest)
121
+
122
+ io&.puts "Installed ExecuTorch #{VERSION} C++ SDK to #{dest}"
123
+ dest
124
+ end
125
+
126
+ def download_with_redirects(url, dest, limit = 5)
127
+ raise "Too many HTTP redirects downloading #{url}" if limit <= 0
128
+
129
+ uri = URI.parse(url)
130
+ Net::HTTP.start(uri.host, uri.port, use_ssl: uri.scheme == "https", open_timeout: 30, read_timeout: 120) do |http|
131
+ req = Net::HTTP::Get.new(uri)
132
+ http.request(req) do |res|
133
+ case res
134
+ when Net::HTTPSuccess
135
+ File.open(dest, "wb") { |f| res.read_body { |chunk| f.write(chunk) } }
136
+ when Net::HTTPRedirection
137
+ download_with_redirects(res["location"], dest, limit - 1)
138
+ else
139
+ raise "Failed downloading #{url}: HTTP #{res.code} #{res.message}"
140
+ end
141
+ end
142
+ end
143
+ end
144
+
145
+ # Minimal pure-Ruby ZIP reader (central directory) extracting ONLY
146
+ # executorch/{include,lib,share/cmake}/ from the wheel.
147
+ def extract_sdk_from_wheel(whl_path, dest_dir)
148
+ File.open(whl_path, "rb") do |f|
149
+ size = f.size
150
+ search_len = [size, 65_557].min
151
+ f.seek(size - search_len)
152
+ tail = f.read(search_len)
153
+ eocd_idx = tail.rindex("PK\x05\x06")
154
+ raise "Invalid wheel archive (missing EOCD)" unless eocd_idx
155
+
156
+ _sig, _disk, _cd_disk, _entries_disk, total_entries, cd_size, cd_offset =
157
+ tail.byteslice(eocd_idx, 22).unpack("a4vvvvVV")
158
+
159
+ # Handle ZIP64 if needed
160
+ if total_entries == 0xFFFF || cd_offset == 0xFFFFFFFF
161
+ loc_idx = tail.rindex("PK\x06\x07", eocd_idx)
162
+ raise "ZIP64 locator missing" unless loc_idx
163
+
164
+ zip64_offset = tail.byteslice(loc_idx + 8, 8).unpack1("Q<")
165
+ f.seek(zip64_offset)
166
+ z64 = f.read(56)
167
+ total_entries = z64.byteslice(32, 8).unpack1("Q<")
168
+ cd_size = z64.byteslice(40, 8).unpack1("Q<")
169
+ cd_offset = z64.byteslice(48, 8).unpack1("Q<")
170
+ end
171
+
172
+ f.seek(cd_offset)
173
+ cd = f.read(cd_size)
174
+ pos = 0
175
+
176
+ total_entries.times do
177
+ break if pos + 46 > cd.bytesize
178
+
179
+ sig = cd.byteslice(pos, 4)
180
+ raise "Invalid central directory header" unless sig == "PK\x01\x02"
181
+
182
+ method, _time, _date, _crc, comp_size, uncomp_size, name_len, extra_len, comment_len,
183
+ _disk_start, _int_attr, _ext_attr, local_offset =
184
+ cd.byteslice(pos + 10, 36).unpack("vvvVVVvvvvvVV")
185
+
186
+ name = cd.byteslice(pos + 46, name_len)
187
+ extra = cd.byteslice(pos + 46 + name_len, extra_len)
188
+ pos += 46 + name_len + extra_len + comment_len
189
+
190
+ next unless name.start_with?("executorch/include/", "executorch/lib/", "executorch/share/cmake/")
191
+ next if name.end_with?("/")
192
+
193
+ # Resolve ZIP64 extra fields if 32-bit fields overflowed
194
+ if comp_size == 0xFFFFFFFF || uncomp_size == 0xFFFFFFFF || local_offset == 0xFFFFFFFF
195
+ ep = 0
196
+ while ep + 4 <= extra.bytesize
197
+ eid, esz = extra.byteslice(ep, 4).unpack("vv")
198
+ if eid == 0x0001
199
+ ef = extra.byteslice(ep + 4, esz)
200
+ off = 0
201
+ if uncomp_size == 0xFFFFFFFF
202
+ uncomp_size = ef.byteslice(off, 8).unpack1("Q<")
203
+ off += 8
204
+ end
205
+ if comp_size == 0xFFFFFFFF
206
+ comp_size = ef.byteslice(off, 8).unpack1("Q<")
207
+ off += 8
208
+ end
209
+ if local_offset == 0xFFFFFFFF
210
+ local_offset = ef.byteslice(off, 8).unpack1("Q<")
211
+ end
212
+ break
213
+ end
214
+ ep += 4 + esz
215
+ end
216
+ end
217
+
218
+ rel = name.delete_prefix("executorch/")
219
+ # Prevent path traversal
220
+ raise "Unsafe archive path: #{rel}" if rel.include?("..") || rel.start_with?("/")
221
+
222
+ out_path = File.join(dest_dir, rel)
223
+ FileUtils.mkdir_p(File.dirname(out_path))
224
+
225
+ f.seek(local_offset)
226
+ lh = f.read(30)
227
+ raise "Invalid local file header for #{name}" unless lh&.start_with?("PK\x03\x04")
228
+
229
+ l_name_len, l_extra_len = lh.byteslice(26, 4).unpack("vv")
230
+ f.seek(local_offset + 30 + l_name_len + l_extra_len)
231
+ raw = f.read(comp_size)
232
+
233
+ data = case method
234
+ when 0 then raw
235
+ when 8
236
+ z = Zlib::Inflate.new(-Zlib::MAX_WBITS)
237
+ begin
238
+ z.inflate(raw)
239
+ ensure
240
+ z.close
241
+ end
242
+ else
243
+ raise "Unsupported ZIP compression method #{method} for #{name}"
244
+ end
245
+
246
+ File.binwrite(out_path, data)
247
+ end
248
+ end
249
+ end
250
+ end
251
+ end
@@ -0,0 +1,85 @@
1
+ # frozen_string_literal: true
2
+
3
+ module Executorch
4
+ class Tensor
5
+ class << self
6
+ # Create a tensor filled with ones.
7
+ def ones(shape, dtype: :float32)
8
+ full(shape, 1, dtype: dtype)
9
+ end
10
+
11
+ # Create a tensor filled with a constant scalar value.
12
+ def full(shape, value, dtype: :float32)
13
+ raise TypeError, "shape must be an Array" unless shape.is_a?(Array)
14
+
15
+ numel = shape.empty? ? 1 : shape.reduce(1, :*)
16
+ raise ShapeError, "shape dimensions must be non-negative" if shape.any? { |d| !d.is_a?(Integer) || d < 0 }
17
+
18
+ new(Array.new(numel, value), shape: shape, dtype: dtype)
19
+ end
20
+
21
+ # Build a Tensor from a Numo::NArray (zero-copy byte path).
22
+ def from_numo(narray)
23
+ dtype = NUMO_TO_DTYPE[narray.class.name]
24
+ raise DTypeError, "Unsupported Numo::NArray type: #{narray.class.name}" unless dtype
25
+
26
+ from_bytes(narray.to_binary, shape: narray.shape, dtype: dtype)
27
+ end
28
+
29
+ # Build a Tensor from a Torch::Tensor (torch.rb interop).
30
+ def from_torch(torch_tensor)
31
+ dtype = TORCH_TO_DTYPE[torch_tensor.dtype]
32
+ raise DTypeError, "Unsupported Torch::Tensor dtype: #{torch_tensor.dtype.inspect}" unless dtype
33
+
34
+ new(torch_tensor.contiguous.to_a, shape: torch_tensor.shape, dtype: dtype)
35
+ end
36
+ end
37
+
38
+ NUMO_TO_DTYPE = {
39
+ "Numo::SFloat" => :float32,
40
+ "Numo::DFloat" => :float64,
41
+ "Numo::Int8" => :int8,
42
+ "Numo::Int16" => :int16,
43
+ "Numo::Int32" => :int32,
44
+ "Numo::Int64" => :int64,
45
+ "Numo::UInt8" => :uint8,
46
+ "Numo::UInt16" => :uint16,
47
+ "Numo::UInt32" => :uint32,
48
+ "Numo::UInt64" => :uint64,
49
+ }.freeze
50
+
51
+ TORCH_TO_DTYPE = {
52
+ float32: :float32,
53
+ float64: :float64,
54
+ float16: :float16,
55
+ bfloat16: :bfloat16,
56
+ int8: :int8,
57
+ int16: :int16,
58
+ int32: :int32,
59
+ int64: :int64,
60
+ uint8: :uint8,
61
+ bool: :bool,
62
+ }.freeze
63
+
64
+ def empty?
65
+ numel == 0
66
+ end
67
+
68
+ def to_numo
69
+ raise LoadError, "Numo::NArray is not loaded (require 'numo/narray')" unless defined?(::Numo::NArray)
70
+
71
+ klass_name = NUMO_TO_DTYPE.key(dtype)
72
+ raise DTypeError, "No Numo::NArray equivalent for dtype #{dtype.inspect}" unless klass_name
73
+
74
+ klass = Object.const_get(klass_name)
75
+ klass.from_binary(to_bytes, shape)
76
+ end
77
+
78
+ def inspect
79
+ return "#<Executorch::Tensor (closed)>" if closed?
80
+
81
+ "#<Executorch::Tensor shape=#{shape.inspect} dtype=#{dtype.inspect} nbytes=#{nbytes}>"
82
+ end
83
+ alias to_s inspect
84
+ end
85
+ end
@@ -0,0 +1,5 @@
1
+ # frozen_string_literal: true
2
+
3
+ module Executorch
4
+ VERSION = "1.0.0"
5
+ end
data/lib/executorch.rb ADDED
@@ -0,0 +1,16 @@
1
+ # frozen_string_literal: true
2
+
3
+ require_relative "executorch/version"
4
+ require_relative "executorch/method_meta"
5
+
6
+ # Load precompiled Ruby-version-specific native extension first (fat gems),
7
+ # falling back to standard single-version native extension path.
8
+ begin
9
+ ruby_abi = RUBY_VERSION[/\A\d+\.\d+/]
10
+ require_relative "executorch/#{ruby_abi}/executorch"
11
+ rescue LoadError
12
+ require_relative "executorch/executorch"
13
+ end
14
+
15
+ require_relative "executorch/tensor"
16
+ require_relative "executorch/module"
metadata ADDED
@@ -0,0 +1,65 @@
1
+ --- !ruby/object:Gem::Specification
2
+ name: executorch
3
+ version: !ruby/object:Gem::Version
4
+ version: 1.0.0
5
+ platform: x86_64-linux
6
+ authors:
7
+ - Benjamin Garcia
8
+ bindir: bin
9
+ cert_chain: []
10
+ date: 1980-01-02 00:00:00.000000000 Z
11
+ dependencies: []
12
+ description: Run PyTorch models exported with ExecuTorch (.pte / .ptd) natively in
13
+ Ruby with XNNPACK/CoreML acceleration, GVL-free execution, and zero-copy IO::Buffer
14
+ tensors.
15
+ email:
16
+ - hey@bengarcia.dev
17
+ executables: []
18
+ extensions: []
19
+ extra_rdoc_files: []
20
+ files:
21
+ - ".executorch-version"
22
+ - CHANGELOG.md
23
+ - LICENSE.txt
24
+ - README.md
25
+ - lib/executorch.rb
26
+ - lib/executorch/3.2/executorch.so
27
+ - lib/executorch/3.3/executorch.so
28
+ - lib/executorch/3.4/executorch.so
29
+ - lib/executorch/method_meta.rb
30
+ - lib/executorch/module.rb
31
+ - lib/executorch/sdk.rb
32
+ - lib/executorch/tensor.rb
33
+ - lib/executorch/vendor/libexecutorch.so
34
+ - lib/executorch/vendor/libexecutorch_backend_xnnpack.so
35
+ - lib/executorch/vendor/libexecutorch_etdump.so
36
+ - lib/executorch/vendor/libexecutorch_kernels_optimized.so
37
+ - lib/executorch/vendor/libexecutorch_kernels_quantized.so
38
+ - lib/executorch/vendor/libexecutorch_threadpool.so
39
+ - lib/executorch/version.rb
40
+ homepage: https://github.com/benngarcia/executorch-ruby
41
+ licenses:
42
+ - Apache-2.0
43
+ metadata:
44
+ source_code_uri: https://github.com/benngarcia/executorch-ruby
45
+ changelog_uri: https://github.com/benngarcia/executorch-ruby/blob/master/CHANGELOG.md
46
+ bug_tracker_uri: https://github.com/benngarcia/executorch-ruby/issues
47
+ rubygems_mfa_required: 'true'
48
+ rdoc_options: []
49
+ require_paths:
50
+ - lib
51
+ required_ruby_version: !ruby/object:Gem::Requirement
52
+ requirements:
53
+ - - ">="
54
+ - !ruby/object:Gem::Version
55
+ version: 3.2.0
56
+ required_rubygems_version: !ruby/object:Gem::Requirement
57
+ requirements:
58
+ - - ">="
59
+ - !ruby/object:Gem::Version
60
+ version: '0'
61
+ requirements: []
62
+ rubygems_version: 3.6.9
63
+ specification_version: 4
64
+ summary: Native Ruby bindings for Meta's ExecuTorch runtime
65
+ test_files: []