populora 0.0.1__tar.gz

populora-0.0.1/.gitignore

@@ -0,0 +1,220 @@
+uv_publish.sh
+
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[codz]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#   Usually these files are written by a python script from a template
+#   before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py.cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+# Pipfile.lock
+
+# UV
+#   Similar to Pipfile.lock, it is generally recommended to include uv.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+# uv.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+# poetry.lock
+# poetry.toml
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#   pdm recommends including project-wide configuration in pdm.toml, but excluding .pdm-python.
+#   https://pdm-project.org/en/latest/usage/project/#working-with-version-control
+# pdm.lock
+# pdm.toml
+.pdm-python
+.pdm-build/
+
+# pixi
+#   Similar to Pipfile.lock, it is generally recommended to include pixi.lock in version control.
+# pixi.lock
+#   Pixi creates a virtual environment in the .pixi directory, just like venv module creates one
+#   in the .venv directory. It is recommended not to include this directory in version control.
+.pixi
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# Redis
+*.rdb
+*.aof
+*.pid
+
+# RabbitMQ
+mnesia/
+rabbitmq/
+rabbitmq-data/
+
+# ActiveMQ
+activemq-data/
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.envrc
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#   JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#   be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#   and can be added to the global gitignore or merged into this file.  For a more nuclear
+#   option (not recommended) you can uncomment the following to ignore the entire idea folder.
+# .idea/
+
+# Abstra
+#   Abstra is an AI-powered process automation framework.
+#   Ignore directories containing user credentials, local state, and settings.
+#   Learn more at https://abstra.io/docs
+.abstra/
+
+# Visual Studio Code
+#   Visual Studio Code specific template is maintained in a separate VisualStudioCode.gitignore 
+#   that can be found at https://github.com/github/gitignore/blob/main/Global/VisualStudioCode.gitignore
+#   and can be added to the global gitignore or merged into this file. However, if you prefer, 
+#   you could uncomment the following to ignore the entire vscode folder
+# .vscode/
+# Temporary file for partial code execution
+tempCodeRunnerFile.py
+
+# Ruff stuff:
+.ruff_cache/
+
+# PyPI configuration file
+.pypirc
+
+# Marimo
+marimo/_static/
+marimo/_lsp/
+__marimo__/
+
+# Streamlit
+.streamlit/secrets.toml

populora-0.0.1/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Phil Wang
+
+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.

populora-0.0.1/PKG-INFO

@@ -0,0 +1,87 @@
+Metadata-Version: 2.4
+Name: populora
+Version: 0.0.1
+Summary: Implementation of PopuLoRA
+Project-URL: Homepage, https://pypi.org/project/populora/
+Project-URL: Repository, https://codeberg.org/lucidrains/populora
+Author-email: Phil Wang <lucidrains@gmail.com>
+License: MIT License
+        
+        Copyright (c) 2026 Phil Wang
+        
+        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.
+License-File: LICENSE
+Keywords: artificial intelligence,deep learning,evolution,reinforcement learning
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.10
+Requires-Dist: einops>=0.8.1
+Requires-Dist: einx>=0.3.0
+Requires-Dist: torch-einops-utils>=0.0.29
+Requires-Dist: torch>=2.5
+Provides-Extra: examples
+Provides-Extra: test
+Requires-Dist: pytest; extra == 'test'
+Requires-Dist: x-transformers; extra == 'test'
+Description-Content-Type: text/markdown
+
+## PopuLoRA (wip)
+
+Implementation and explorations into [PopuLoRA](https://arxiv.org/abs/2605.16727v1), [Co-Evolving LLM Populations for Reasoning Self-Play](https://vmax.ai/team/populora-co-evolving-llm-populations-for-reasoning-self-play), from Roger Castanyer et al at [vmax.ai](https://vmax.ai/)
+
+## Citations
+
+```bibtex
+@misc{castanyer2026populoracoevolvingllmpopulations,
+    title   = {PopuLoRA: Co-Evolving LLM Populations for Reasoning Self-Play},
+    author  = {Roger Creus Castanyer and Geoffrey Bradway and Lorenz Wolf and Maxwill Lin and Augustine N. Mavor-Parker and Matthew James Sargent},
+    year    = {2026},
+    eprint  = {2605.16727},
+    archivePrefix = {arXiv},
+    primaryClass = {cs.AI},
+    url     = {https://arxiv.org/abs/2605.16727},
+}
+```
+
+```bibtex
+@misc{schmidhuber2012powerplaytrainingincreasinglygeneral,
+    title    = {POWERPLAY: Training an Increasingly General Problem Solver by Continually Searching for the Simplest Still Unsolvable Problem},
+    author   = {Jürgen Schmidhuber},
+    year     = {2012},
+    eprint   = {1112.5309},
+    archivePrefix = {arXiv},
+    primaryClass = {cs.AI},
+    url      = {https://arxiv.org/abs/1112.5309},
+}
+```
+
+```bibtex
+@misc{bahlousboldi2026vectorpolicyoptimizationtraining,
+    title   = {Vector Policy Optimization: Training for Diversity Improves Test-Time Search},
+    author  = {Ryan Bahlous-Boldi and Isha Puri and Idan Shenfeld and Akarsh Kumar and Mehul Damani and Sebastian Risi and Omar Khattab and Zhang-Wei Hong and Pulkit Agrawal},
+    year    = {2026},
+    eprint  = {2605.22817},
+    archivePrefix = {arXiv},
+    primaryClass = {cs.LG},
+    url     = {https://arxiv.org/abs/2605.22817},
+}
+```

populora-0.0.1/README.md

@@ -0,0 +1,41 @@
+## PopuLoRA (wip)
+
+Implementation and explorations into [PopuLoRA](https://arxiv.org/abs/2605.16727v1), [Co-Evolving LLM Populations for Reasoning Self-Play](https://vmax.ai/team/populora-co-evolving-llm-populations-for-reasoning-self-play), from Roger Castanyer et al at [vmax.ai](https://vmax.ai/)
+
+## Citations
+
+```bibtex
+@misc{castanyer2026populoracoevolvingllmpopulations,
+    title   = {PopuLoRA: Co-Evolving LLM Populations for Reasoning Self-Play},
+    author  = {Roger Creus Castanyer and Geoffrey Bradway and Lorenz Wolf and Maxwill Lin and Augustine N. Mavor-Parker and Matthew James Sargent},
+    year    = {2026},
+    eprint  = {2605.16727},
+    archivePrefix = {arXiv},
+    primaryClass = {cs.AI},
+    url     = {https://arxiv.org/abs/2605.16727},
+}
+```
+
+```bibtex
+@misc{schmidhuber2012powerplaytrainingincreasinglygeneral,
+    title    = {POWERPLAY: Training an Increasingly General Problem Solver by Continually Searching for the Simplest Still Unsolvable Problem},
+    author   = {Jürgen Schmidhuber},
+    year     = {2012},
+    eprint   = {1112.5309},
+    archivePrefix = {arXiv},
+    primaryClass = {cs.AI},
+    url      = {https://arxiv.org/abs/1112.5309},
+}
+```
+
+```bibtex
+@misc{bahlousboldi2026vectorpolicyoptimizationtraining,
+    title   = {Vector Policy Optimization: Training for Diversity Improves Test-Time Search},
+    author  = {Ryan Bahlous-Boldi and Isha Puri and Idan Shenfeld and Akarsh Kumar and Mehul Damani and Sebastian Risi and Omar Khattab and Zhang-Wei Hong and Pulkit Agrawal},
+    year    = {2026},
+    eprint  = {2605.22817},
+    archivePrefix = {arXiv},
+    primaryClass = {cs.LG},
+    url     = {https://arxiv.org/abs/2605.22817},
+}
+```

populora-0.0.1/populora/__init__.py

@@ -0,0 +1 @@
+from populora.populora import PopuLoRA

populora-0.0.1/populora/populora.py

@@ -0,0 +1,132 @@
+from __future__ import annotations
+
+import torch
+from torch.nn import Linear, Module, Parameter, ParameterDict, init
+
+from einops import einsum, repeat, rearrange
+from torch_einops_utils import tree_map_tensor
+
+from contextlib import contextmanager
+
+# helpers
+
+def exists(v):
+    return v is not None
+
+# main class
+
+class PopuLoRA(Module):
+    def __init__(
+        self,
+        model: Module,
+        *,
+        pop_size: int,
+        low_rank: int,
+        lora_targets: tuple[str, ...] | list[str],
+        requires_grad: bool = False
+    ):
+        super().__init__()
+        self.model = model
+        self.pop_size = pop_size
+
+        self.w_down = ParameterDict()
+        self.w_up = ParameterDict()
+        self._hooks = []
+
+        for path in lora_targets:
+            linear = model.get_submodule(path)
+            assert isinstance(linear, Linear), f'{path} must point to a Linear module'
+
+            key = path.replace('.', '_')
+            dim, dim_inner = linear.in_features, linear.out_features
+
+            self.w_down[key] = Parameter(torch.empty(pop_size, dim, low_rank), requires_grad = requires_grad)
+            self.w_up[key] = Parameter(torch.empty(pop_size, dim_inner, low_rank), requires_grad = requires_grad)
+
+            init.normal_(self.w_down[key], std = dim ** -0.5)
+            init.normal_(self.w_up[key], std = low_rank ** -0.5)
+
+            self._hooks.append(linear.register_forward_hook(self._create_hook(key)))
+
+        self._individual = None
+
+    @contextmanager
+    def _route(self, individual, individuals, all_individuals):
+        assert sum((exists(individual), exists(individuals), all_individuals)) <= 1
+
+        if all_individuals:
+            individual = ...
+        elif exists(individuals):
+            individual = list(individuals)
+
+        self._individual = individual
+
+        try:
+            yield
+        finally:
+            self._individual = None
+
+    @contextmanager
+    def _eval_and_no_grad(self, enabled):
+        if not enabled:
+            yield
+            return
+
+        is_training = self.training
+        self.eval()
+
+        with torch.no_grad():
+            try:
+                yield
+            finally:
+                self.train(is_training)
+
+    def _create_hook(self, lora_key: str):
+        def hook(_, args, output):
+            if not exists(self._individual) and self._individual is not ...:
+                return output
+
+            w_down, w_up = self.w_down[lora_key], self.w_up[lora_key]
+            x, = args
+
+            if isinstance(self._individual, list) or self._individual is ...:
+                w_down_i, w_up_i = w_down[self._individual], w_up[self._individual]
+                p = w_down_i.shape[0]
+
+                x = rearrange(x, '(p b) ... -> p b ...', p = p)
+                lora_out = einsum(x, w_down_i, w_up_i, 'p b ... d, p d r, p e r -> p b ... e')
+                lora_out = rearrange(lora_out, 'p b ... -> (p b) ...')
+            else:
+                lora_out = einsum(x, w_down[self._individual], w_up[self._individual], '... d, d r, e r -> ... e')
+
+            return output + lora_out
+
+        return hook
+
+    def forward(
+        self,
+        *args,
+        individual: int | None = None,
+        individuals: tuple[int, ...] | list[int] | None = None,
+        all_individuals: bool = False,
+        ignore_args_kwargs: tuple[int | str, ...] = tuple(),
+        eval_and_no_grad: bool = False,
+        **kwargs
+    ):
+        if all_individuals or exists(individuals):
+            ignore = set(ignore_args_kwargs)
+            p = self.pop_size if all_individuals else len(individuals)
+            repeat_batch = lambda t: repeat(t, 'b ... -> (p b) ...', p = p)
+
+            args = tuple(
+                tree_map_tensor(repeat_batch, a) if i not in ignore else a
+                for i, a in enumerate(args)
+            )
+
+            kwargs = {
+                k: tree_map_tensor(repeat_batch, v) if k not in ignore else v
+                for k, v in kwargs.items()
+            }
+
+        with self._route(individual, individuals, all_individuals), self._eval_and_no_grad(eval_and_no_grad):
+            return self.model(*args, **kwargs)

populora-0.0.1/pyproject.toml

@@ -0,0 +1,69 @@
+[project]
+name = "populora"
+version = "0.0.1"
+description = "Implementation of PopuLoRA"
+authors = [
+    { name = "Phil Wang", email = "lucidrains@gmail.com" }
+]
+readme = "README.md"
+requires-python = ">= 3.10"
+license = { file = "LICENSE" }
+keywords = [
+    'artificial intelligence',
+    'deep learning',
+    'reinforcement learning',
+    'evolution',
+]
+
+classifiers=[
+    'Development Status :: 4 - Beta',
+    'Intended Audience :: Developers',
+    'Topic :: Scientific/Engineering :: Artificial Intelligence',
+    'License :: OSI Approved :: MIT License',
+    'Programming Language :: Python :: 3.10',
+]
+
+dependencies = [
+    "einx>=0.3.0",
+    "einops>=0.8.1",
+    "torch>=2.5",
+    "torch-einops-utils>=0.0.29",
+]
+
+[project.urls]
+Homepage = "https://pypi.org/project/populora/"
+Repository = "https://codeberg.org/lucidrains/populora"
+
+[project.optional-dependencies]
+examples = []
+test = [
+    "pytest",
+    "x-transformers"
+]
+
+[tool.pytest.ini_options]
+pythonpath = [
+    "."
+]
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[tool.rye]
+managed = true
+dev-dependencies = []
+
+[tool.hatch.metadata]
+allow-direct-references = true
+
+[tool.hatch.build]
+include = [
+    "populora",
+    "pyproject.toml",
+    "README.md",
+    "LICENSE"
+]
+
+[tool.hatch.build.targets.wheel]
+packages = ["populora"]