PoPE-pytorch 0.0.1__tar.gz

pope_pytorch-0.0.1/.github/workflows/python-publish.yml

@@ -0,0 +1,36 @@
+# This workflow will upload a Python Package using Twine when a release is created
+# For more information see: https://help.github.com/en/actions/language-and-framework-guides/using-python-with-github-actions#publishing-to-package-registries
+
+# This workflow uses actions that are not certified by GitHub.
+# They are provided by a third-party and are governed by
+# separate terms of service, privacy policy, and support
+# documentation.
+
+name: Upload Python Package
+
+on:
+  release:
+    types: [published]
+
+jobs:
+  deploy:
+
+    runs-on: ubuntu-latest
+
+    steps:
+    - uses: actions/checkout@v2
+    - name: Set up Python
+      uses: actions/setup-python@v2
+      with:
+        python-version: '3.x'
+    - name: Install dependencies
+      run: |
+        python -m pip install --upgrade pip
+        pip install build
+    - name: Build package
+      run: python -m build
+    - name: Publish package
+      uses: pypa/gh-action-pypi-publish@27b31702a0e7fc50959f5ad993c78deac1bdfc29
+      with:
+        user: __token__
+        password: ${{ secrets.PYPI_API_TOKEN }}

pope_pytorch-0.0.1/.github/workflows/test.yml

@@ -0,0 +1,21 @@
+name: Pytest
+on: [push, pull_request]
+
+jobs:
+  build:
+
+    runs-on: ubuntu-latest
+
+    steps:
+    - uses: actions/checkout@v4
+    - name: Set up Python 3.10
+      uses: actions/setup-python@v5
+      with:
+        python-version: "3.10"
+    - name: Install dependencies
+      run: |
+        python -m pip install --upgrade pip
+        python -m pip install -e .[test]
+    - name: Test with pytest
+      run: |
+        python -m pytest tests/

pope_pytorch-0.0.1/.gitignore

@@ -0,0 +1,207 @@
+# 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
+
+# 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/
+
+# Ruff stuff:
+.ruff_cache/
+
+# PyPI configuration file
+.pypirc
+
+# Cursor
+#  Cursor is an AI-powered code editor. `.cursorignore` specifies files/directories to
+#  exclude from AI features like autocomplete and code analysis. Recommended for sensitive data
+#  refer to https://docs.cursor.com/context/ignore-files
+.cursorignore
+.cursorindexingignore
+
+# Marimo
+marimo/_static/
+marimo/_lsp/
+__marimo__/

pope_pytorch-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.

pope_pytorch-0.0.1/PKG-INFO

@@ -0,0 +1,97 @@
+Metadata-Version: 2.4
+Name: PoPE-pytorch
+Version: 0.0.1
+Summary: PoPE
+Project-URL: Homepage, https://pypi.org/project/PoPE-pytorch/
+Project-URL: Repository, https://github.com/lucidrains/PoPE-pytorch
+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,positional embedding
+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: torch-einops-utils>=0.0.14
+Requires-Dist: torch>=2.5
+Provides-Extra: examples
+Provides-Extra: test
+Requires-Dist: pytest; extra == 'test'
+Description-Content-Type: text/markdown
+
+<img src="./pope.png" width="400px"></img>
+
+## PoPE-pytorch
+
+Efficient implementation (and explorations) into [polar coordinate positional embedding (PoPE)](https://arxiv.org/abs/2509.10534) - from [Gopalakrishnan](https://agopal42.github.io/) et al. under Schmidhuber
+
+## Install
+
+```shell
+$ pip install PoPE-pytorch
+```
+
+## Usage
+
+```python
+import torch
+from PoPE_pytorch import PoPE
+
+# define pope
+
+pope = PoPE(64, heads = 8)
+
+# pass in sequence length
+
+pos_embed = pope(1024)
+
+# queries and keys in attention
+
+q = torch.randn(1, 8, 1024, 64)
+k = torch.randn(1, 8, 1024, 64)
+
+# training
+
+rotated_q, rotated_k = pope.apply_pope_to_qk(pos_embed, q, k)
+
+# inference
+
+rotated_q, rotated_k = pope.apply_pope_to_qk(pos_embed, q[..., -1:, :], k)
+```
+
+## Citations
+
+```bibtex
+@misc{gopalakrishnan2025decouplingwhatwherepolar,
+    title   = {Decoupling the "What" and "Where" With Polar Coordinate Positional Embeddings}, 
+    author  = {Anand Gopalakrishnan and Robert Csordás and Jürgen Schmidhuber and Michael C. Mozer},
+    year    = {2025},
+    eprint  = {2509.10534},
+    archivePrefix = {arXiv},
+    primaryClass = {cs.LG},
+    url     = {https://arxiv.org/abs/2509.10534}, 
+}
+```

pope_pytorch-0.0.1/PoPE_pytorch/PoPE.py

@@ -0,0 +1,121 @@
+from __future__ import annotations
+from collections import namedtuple
+from math import pi
+
+import torch
+from torch import arange, stack, is_tensor, Tensor
+from torch.nn import Module, Parameter
+from torch.amp import autocast
+
+import torch.nn.functional as F
+
+from einops import einsum, rearrange
+
+from torch_einops_utils import slice_right_at_dim
+
+# constants
+
+PolarEmbedReturn = namedtuple('PolarEmbedReturn', ('freqs', 'bias'))
+
+# helper functions
+
+def exists(v):
+    return v is not None
+
+def default(v, d):
+    return v if exists(v) else d
+
+# applying pope to qk
+
+@autocast('cuda', enabled = False)
+def apply_pope_to_qk(
+    pope: PolarEmbedReturn,
+    q, k,
+    to_magnitude = F.softplus
+):
+    q_len, k_len = q.shape[-2], k.shape[-2]
+    assert q_len <= k_len
+
+    freqs, bias = pope
+
+    if freqs.ndim == 3:
+        freqs = rearrange(freqs, 'b n d -> b 1 n d')
+
+    freq_with_bias = freqs + rearrange(bias, 'h d -> h 1 d')
+
+    # convert q and k to polar magnitudes with activation
+
+    q, k = to_magnitude(q), to_magnitude(k)
+
+    # apply rotations
+
+    freqs = slice_right_at_dim(freqs, q_len, dim = -2)
+
+    qcos, qsin = freqs.cos(), freqs.sin()
+
+    q = rearrange([q * qcos, q * qsin], 'two ... d -> ... (d two)')
+
+    # handle inference
+
+    kcos, ksin = freq_with_bias.cos(), freq_with_bias.sin()
+
+    k = rearrange([k * kcos, k * ksin], 'two ... d -> ... (d two)')
+
+    return q, k
+
+# main class
+
+class PoPE(Module):
+    def __init__(
+        self,
+        dim,
+        *,
+        heads,
+        theta = 10000,
+        bias_uniform_init = False
+    ):
+        super().__init__()
+
+        # freqs
+
+        inv_freqs = theta ** -(arange(dim).float() / dim)
+        self.register_buffer('inv_freqs', inv_freqs)
+
+        # the learned bias on the keys
+
+        self.bias = Parameter(torch.zeros(heads, dim))
+
+        if bias_uniform_init:
+            self.bias.uniform_(-2. * pi, 0.)
+
+        # convenience
+
+        self.apply_pope_to_qk = staticmethod(apply_pope_to_qk)
+
+    @property
+    def device(self):
+        return self.inv_freqs.device
+
+    @autocast('cuda', enabled = False)
+    def forward(
+        self,
+        pos_or_seq_len: Tensor | int
+    ):
+
+        # get positions depending on input
+
+        if is_tensor(pos_or_seq_len):
+            pos = pos_or_seq_len
+        else:
+            seq_len = pos_or_seq_len
+            pos = arange(seq_len, device = self.device, dtype = self.inv_freqs.dtype)
+
+        # freqs
+
+        freqs = einsum(pos, self.inv_freqs, '... i, j -> ... i j')
+
+        # the bias, with clamping
+
+        bias = self.bias.clamp(-2. * pi, 0.)
+
+        return PolarEmbedReturn(freqs, bias)

pope_pytorch-0.0.1/PoPE_pytorch/__init__.py

@@ -0,0 +1 @@
+from PoPE_pytorch.PoPE import PoPE

pope_pytorch-0.0.1/README.md

@@ -0,0 +1,53 @@
+<img src="./pope.png" width="400px"></img>
+
+## PoPE-pytorch
+
+Efficient implementation (and explorations) into [polar coordinate positional embedding (PoPE)](https://arxiv.org/abs/2509.10534) - from [Gopalakrishnan](https://agopal42.github.io/) et al. under Schmidhuber
+
+## Install
+
+```shell
+$ pip install PoPE-pytorch
+```
+
+## Usage
+
+```python
+import torch
+from PoPE_pytorch import PoPE
+
+# define pope
+
+pope = PoPE(64, heads = 8)
+
+# pass in sequence length
+
+pos_embed = pope(1024)
+
+# queries and keys in attention
+
+q = torch.randn(1, 8, 1024, 64)
+k = torch.randn(1, 8, 1024, 64)
+
+# training
+
+rotated_q, rotated_k = pope.apply_pope_to_qk(pos_embed, q, k)
+
+# inference
+
+rotated_q, rotated_k = pope.apply_pope_to_qk(pos_embed, q[..., -1:, :], k)
+```
+
+## Citations
+
+```bibtex
+@misc{gopalakrishnan2025decouplingwhatwherepolar,
+    title   = {Decoupling the "What" and "Where" With Polar Coordinate Positional Embeddings}, 
+    author  = {Anand Gopalakrishnan and Robert Csordás and Jürgen Schmidhuber and Michael C. Mozer},
+    year    = {2025},
+    eprint  = {2509.10534},
+    archivePrefix = {arXiv},
+    primaryClass = {cs.LG},
+    url     = {https://arxiv.org/abs/2509.10534}, 
+}
+```

pope_pytorch-0.0.1/pyproject.toml

@@ -0,0 +1,58 @@
+[project]
+name = "PoPE-pytorch"
+version = "0.0.1"
+description = "PoPE"
+authors = [
+    { name = "Phil Wang", email = "lucidrains@gmail.com" }
+]
+readme = "README.md"
+requires-python = ">= 3.10"
+license = { file = "LICENSE" }
+keywords = [
+    'artificial intelligence',
+    'deep learning',
+    'positional embedding',
+]
+
+classifiers=[
+    'Development Status :: 4 - Beta',
+    'Intended Audience :: Developers',
+    'Topic :: Scientific/Engineering :: Artificial Intelligence',
+    'License :: OSI Approved :: MIT License',
+    'Programming Language :: Python :: 3.10',
+]
+
+dependencies = [
+    "einops>=0.8.1",
+    "torch>=2.5",
+    "torch-einops-utils>=0.0.14"
+]
+
+[project.urls]
+Homepage = "https://pypi.org/project/PoPE-pytorch/"
+Repository = "https://github.com/lucidrains/PoPE-pytorch"
+
+[project.optional-dependencies]
+examples = []
+test = [
+    "pytest"
+]
+
+[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.targets.wheel]
+packages = ["PoPE_pytorch"]

pope_pytorch-0.0.1/tests/test_pope.py

@@ -0,0 +1,22 @@
+import pytest
+import torch
+
+def test_pope():
+    from PoPE_pytorch.PoPE import PoPE
+
+    pope = PoPE(64, heads = 8)
+
+    pos_embed = pope(1024)
+
+    q = torch.randn(1, 8, 1024, 64)
+    k = torch.randn(1, 8, 1024, 64)
+
+    rotated_q, rotated_k = pope.apply_pope_to_qk(pos_embed, q, k)
+
+    assert rotated_q.shape == (1, 8, 1024, 128)
+    assert rotated_k.shape == (1, 8, 1024, 128)
+
+    rotated_q, rotated_k = pope.apply_pope_to_qk(pos_embed, q[..., -1:, :], k)
+
+    assert rotated_q.shape == (1, 8, 1, 128)
+    assert rotated_k.shape == (1, 8, 1024, 128)