USN 0.1.0__tar.gz

usn-0.1.0/CHANGELOG.md

@@ -0,0 +1,24 @@
+# Changelog
+
+All notable changes to this project will be documented in this file.
+
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
+and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+## [0.1.0] - 2024-01-01
+
+### Added
+
+- Initial project structure and packaging
+- Core type definitions (UnifiedState, ModelState, BlockOutput, etc.)
+- USN modules: InputProjection, TemporalMixing, ExponentialGating, SelectiveWriting, StateUpdate, StateReadout, ChannelMixing
+- USNBlock layer with pre-norm residual architecture
+- Parallel scan and chunked parallel scan implementations
+- USNModel with embedding, blocks, and output head
+- Configuration system with presets (tiny through xxl)
+- Training system with mixed precision, curriculum, and distributed support
+- Inference system with greedy, top-k, top-p, and beam search decoding
+- `.usn` single-file serialization format
+- Triton fused kernels with PyTorch fallbacks
+- CLI interface (train, generate, benchmark, info, export, validate)
+- Comprehensive test suite with property-based tests

usn-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 BUEORM
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

usn-0.1.0/MANIFEST.in

@@ -0,0 +1,6 @@
+include LICENSE
+include README.md
+include CHANGELOG.md
+recursive-include docs *
+recursive-include examples *
+recursive-include usn py.typed *.pyi

usn-0.1.0/PKG-INFO

@@ -0,0 +1,96 @@
+Metadata-Version: 2.4
+Name: USN
+Version: 0.1.0
+Summary: Unified State Network (USN) Architecture Library - A production-grade autoregressive sequence modeling architecture with O(n) training and O(1) inference.
+Author: BUEORM
+License-Expression: MIT
+Project-URL: Homepage, https://github.com/BUEORM/USN
+Project-URL: Documentation, https://github.com/BUEORM/USN/docs
+Project-URL: Repository, https://github.com/BUEORM/USN
+Keywords: deep-learning,sequence-modeling,state-space,autoregressive,parallel-scan
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Science/Research
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: torch>=2.0.0
+Requires-Dist: numpy>=1.24.0
+Requires-Dist: tokenizers>=0.15.0
+Requires-Dist: pyyaml>=6.0
+Requires-Dist: tqdm>=4.60.0
+Provides-Extra: dev
+Requires-Dist: pytest>=7.0; extra == "dev"
+Requires-Dist: pytest-cov; extra == "dev"
+Requires-Dist: mypy; extra == "dev"
+Requires-Dist: ruff; extra == "dev"
+Requires-Dist: pre-commit; extra == "dev"
+Requires-Dist: hypothesis>=6.0; extra == "dev"
+Provides-Extra: docs
+Requires-Dist: sphinx; extra == "docs"
+Requires-Dist: sphinx-rtd-theme; extra == "docs"
+Provides-Extra: cuda
+Requires-Dist: triton>=2.1.0; extra == "cuda"
+Provides-Extra: all
+Requires-Dist: USN[cuda,dev,docs]; extra == "all"
+Dynamic: license-file
+
+# USN - Unified State Network Architecture Library
+
+A production-grade Python package implementing the Unified State Network (USN) architecture for autoregressive sequence modeling.
+
+USN replaces attention mechanisms with a unified persistent state partitioned into semantic (vector) and relational (matrix) subspaces, achieving **O(n) training complexity** via associative parallel scan and **O(1) inference memory** via constant-size state.
+
+## Features
+
+- Novel state-space architecture with no attention mechanism
+- O(n) training via parallel associative scan
+- O(1) inference memory via constant-size persistent state
+- 4-level acceleration hierarchy (Triton → torch.compile → custom autograd → eager)
+- Single-file `.usn` serialization format
+- CLI interface for training, generation, and benchmarking
+
+## Installation
+
+```bash
+pip install USN
+```
+
+For development:
+
+```bash
+pip install USN[dev]
+```
+
+For CUDA acceleration:
+
+```bash
+pip install USN[cuda]
+```
+
+For all optional dependencies:
+
+```bash
+pip install USN[all]
+```
+
+## Quick Start
+
+```python
+import usn
+
+# Create a model from a preset configuration
+config = usn.USNConfig.tiny()
+model = usn.create_model(config)
+
+# Print model summary
+print(model.summary())
+```
+
+## License
+
+MIT License - Copyright (c) 2024 BUEORM

usn-0.1.0/README.md

@@ -0,0 +1,55 @@
+# USN - Unified State Network Architecture Library
+
+A production-grade Python package implementing the Unified State Network (USN) architecture for autoregressive sequence modeling.
+
+USN replaces attention mechanisms with a unified persistent state partitioned into semantic (vector) and relational (matrix) subspaces, achieving **O(n) training complexity** via associative parallel scan and **O(1) inference memory** via constant-size state.
+
+## Features
+
+- Novel state-space architecture with no attention mechanism
+- O(n) training via parallel associative scan
+- O(1) inference memory via constant-size persistent state
+- 4-level acceleration hierarchy (Triton → torch.compile → custom autograd → eager)
+- Single-file `.usn` serialization format
+- CLI interface for training, generation, and benchmarking
+
+## Installation
+
+```bash
+pip install USN
+```
+
+For development:
+
+```bash
+pip install USN[dev]
+```
+
+For CUDA acceleration:
+
+```bash
+pip install USN[cuda]
+```
+
+For all optional dependencies:
+
+```bash
+pip install USN[all]
+```
+
+## Quick Start
+
+```python
+import usn
+
+# Create a model from a preset configuration
+config = usn.USNConfig.tiny()
+model = usn.create_model(config)
+
+# Print model summary
+print(model.summary())
+```
+
+## License
+
+MIT License - Copyright (c) 2024 BUEORM

usn-0.1.0/USN.egg-info/PKG-INFO

@@ -0,0 +1,96 @@
+Metadata-Version: 2.4
+Name: USN
+Version: 0.1.0
+Summary: Unified State Network (USN) Architecture Library - A production-grade autoregressive sequence modeling architecture with O(n) training and O(1) inference.
+Author: BUEORM
+License-Expression: MIT
+Project-URL: Homepage, https://github.com/BUEORM/USN
+Project-URL: Documentation, https://github.com/BUEORM/USN/docs
+Project-URL: Repository, https://github.com/BUEORM/USN
+Keywords: deep-learning,sequence-modeling,state-space,autoregressive,parallel-scan
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Science/Research
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: torch>=2.0.0
+Requires-Dist: numpy>=1.24.0
+Requires-Dist: tokenizers>=0.15.0
+Requires-Dist: pyyaml>=6.0
+Requires-Dist: tqdm>=4.60.0
+Provides-Extra: dev
+Requires-Dist: pytest>=7.0; extra == "dev"
+Requires-Dist: pytest-cov; extra == "dev"
+Requires-Dist: mypy; extra == "dev"
+Requires-Dist: ruff; extra == "dev"
+Requires-Dist: pre-commit; extra == "dev"
+Requires-Dist: hypothesis>=6.0; extra == "dev"
+Provides-Extra: docs
+Requires-Dist: sphinx; extra == "docs"
+Requires-Dist: sphinx-rtd-theme; extra == "docs"
+Provides-Extra: cuda
+Requires-Dist: triton>=2.1.0; extra == "cuda"
+Provides-Extra: all
+Requires-Dist: USN[cuda,dev,docs]; extra == "all"
+Dynamic: license-file
+
+# USN - Unified State Network Architecture Library
+
+A production-grade Python package implementing the Unified State Network (USN) architecture for autoregressive sequence modeling.
+
+USN replaces attention mechanisms with a unified persistent state partitioned into semantic (vector) and relational (matrix) subspaces, achieving **O(n) training complexity** via associative parallel scan and **O(1) inference memory** via constant-size state.
+
+## Features
+
+- Novel state-space architecture with no attention mechanism
+- O(n) training via parallel associative scan
+- O(1) inference memory via constant-size persistent state
+- 4-level acceleration hierarchy (Triton → torch.compile → custom autograd → eager)
+- Single-file `.usn` serialization format
+- CLI interface for training, generation, and benchmarking
+
+## Installation
+
+```bash
+pip install USN
+```
+
+For development:
+
+```bash
+pip install USN[dev]
+```
+
+For CUDA acceleration:
+
+```bash
+pip install USN[cuda]
+```
+
+For all optional dependencies:
+
+```bash
+pip install USN[all]
+```
+
+## Quick Start
+
+```python
+import usn
+
+# Create a model from a preset configuration
+config = usn.USNConfig.tiny()
+model = usn.create_model(config)
+
+# Print model summary
+print(model.summary())
+```
+
+## License
+
+MIT License - Copyright (c) 2024 BUEORM

usn-0.1.0/USN.egg-info/SOURCES.txt

@@ -0,0 +1,89 @@
+CHANGELOG.md
+LICENSE
+MANIFEST.in
+README.md
+pyproject.toml
+setup.py
+USN.egg-info/PKG-INFO
+USN.egg-info/SOURCES.txt
+USN.egg-info/dependency_links.txt
+USN.egg-info/entry_points.txt
+USN.egg-info/requires.txt
+USN.egg-info/top_level.txt
+docs/DOCUMENTATION.md
+examples/quick_start.py
+tests/test_backends.py
+tests/test_beam_search_repetition.py
+tests/test_generator_streaming.py
+tests/test_optim_factory.py
+tests/test_trainer_eval_checkpoint.py
+usn/__init__.py
+usn/__init__.pyi
+usn/exceptions.py
+usn/py.typed
+usn/backends/__init__.py
+usn/backends/acceleration.py
+usn/backends/detection.py
+usn/backends/fallbacks.py
+usn/backends/triton_kernels.py
+usn/cli/__init__.py
+usn/cli/main.py
+usn/config/__init__.py
+usn/config/generation_config.py
+usn/config/model_config.py
+usn/config/training_config.py
+usn/core/__init__.py
+usn/core/activations.py
+usn/core/base.py
+usn/core/interfaces.py
+usn/core/types.py
+usn/datasets/__init__.py
+usn/datasets/collate.py
+usn/datasets/math_dataset.py
+usn/datasets/usn_dataset.py
+usn/inference/__init__.py
+usn/inference/generator.py
+usn/layers/__init__.py
+usn/layers/block.py
+usn/layers/chunked_scan.py
+usn/layers/norm.py
+usn/layers/parallel_scan.py
+usn/losses/__init__.py
+usn/losses/cross_entropy.py
+usn/models/__init__.py
+usn/models/embedding.py
+usn/models/usn_model.py
+usn/modules/__init__.py
+usn/modules/channel_mixing.py
+usn/modules/exponential_gating.py
+usn/modules/input_projection.py
+usn/modules/selective_writing.py
+usn/modules/state_readout.py
+usn/modules/state_update.py
+usn/modules/temporal_mixing.py
+usn/optim/__init__.py
+usn/optim/factory.py
+usn/optim/schedulers.py
+usn/serialization/__init__.py
+usn/serialization/export.py
+usn/serialization/format_spec.py
+usn/serialization/migration.py
+usn/serialization/reader.py
+usn/serialization/validator.py
+usn/serialization/writer.py
+usn/tokenizers/__init__.py
+usn/tokenizers/bpe_tokenizer.py
+usn/tokenizers/char_tokenizer.py
+usn/tokenizers/word_tokenizer.py
+usn/training/__init__.py
+usn/training/curriculum.py
+usn/training/distributed.py
+usn/training/stability.py
+usn/training/trainer.py
+usn/utils/__init__.py
+usn/utils/counting.py
+usn/utils/diagnostics.py
+usn/utils/profiling.py
+usn/utils/seed.py
+usn/utils/timing.py
+usn/utils/visualization.py

usn-0.1.0/USN.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

usn-0.1.0/USN.egg-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+usn = usn.cli.main:cli

usn-0.1.0/USN.egg-info/requires.txt

@@ -0,0 +1,23 @@
+torch>=2.0.0
+numpy>=1.24.0
+tokenizers>=0.15.0
+pyyaml>=6.0
+tqdm>=4.60.0
+
+[all]
+USN[cuda,dev,docs]
+
+[cuda]
+triton>=2.1.0
+
+[dev]
+pytest>=7.0
+pytest-cov
+mypy
+ruff
+pre-commit
+hypothesis>=6.0
+
+[docs]
+sphinx
+sphinx-rtd-theme

usn-0.1.0/USN.egg-info/top_level.txt

@@ -0,0 +1 @@
+usn