Metadata-Version: 2.4
Name: drift-models
Version: 0.1.0
Summary: Community PyTorch reproduction of Generative Modeling via Drifting
Project-URL: Homepage, https://github.com/kmccleary3301/drift_models
Project-URL: Repository, https://github.com/kmccleary3301/drift_models
Project-URL: Issues, https://github.com/kmccleary3301/drift_models/issues
Author: Drifting Models Repro Maintainers
License: MIT License
        
        Copyright (c) 2026 Drifting Models Repro Maintainers
        
        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: diffusion,drifting-models,generative-models,pytorch,research-reproduction
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Science/Research
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3 :: Only
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
Requires-Dist: numpy>=1.26
Requires-Dist: pillow>=10
Requires-Dist: torch>=2.3
Requires-Dist: tqdm>=4.66
Provides-Extra: all
Requires-Dist: accelerate>=0.31.0; extra == 'all'
Requires-Dist: build>=1.2.2; extra == 'all'
Requires-Dist: diffusers>=0.31.0; extra == 'all'
Requires-Dist: huggingface-hub>=0.24.0; extra == 'all'
Requires-Dist: mkdocs-material>=9.5.39; extra == 'all'
Requires-Dist: mkdocs>=1.6.1; extra == 'all'
Requires-Dist: pyright>=1.1.380; extra == 'all'
Requires-Dist: pytest>=8.0.0; extra == 'all'
Requires-Dist: ruff>=0.6.0; extra == 'all'
Requires-Dist: safetensors>=0.4.0; extra == 'all'
Requires-Dist: torchvision>=0.18; extra == 'all'
Requires-Dist: transformers>=4.40.0; extra == 'all'
Requires-Dist: twine>=5.1.1; extra == 'all'
Provides-Extra: dev
Requires-Dist: build>=1.2.2; extra == 'dev'
Requires-Dist: pyright>=1.1.380; extra == 'dev'
Requires-Dist: pytest>=8.0.0; extra == 'dev'
Requires-Dist: ruff>=0.6.0; extra == 'dev'
Requires-Dist: twine>=5.1.1; extra == 'dev'
Provides-Extra: docs
Requires-Dist: mkdocs-material>=9.5.39; extra == 'docs'
Requires-Dist: mkdocs>=1.6.1; extra == 'docs'
Provides-Extra: eval
Requires-Dist: torchvision>=0.18; extra == 'eval'
Provides-Extra: imagenet
Requires-Dist: torchvision>=0.18; extra == 'imagenet'
Provides-Extra: sdvae
Requires-Dist: accelerate>=0.31.0; extra == 'sdvae'
Requires-Dist: diffusers>=0.31.0; extra == 'sdvae'
Requires-Dist: huggingface-hub>=0.24.0; extra == 'sdvae'
Requires-Dist: safetensors>=0.4.0; extra == 'sdvae'
Requires-Dist: transformers>=4.40.0; extra == 'sdvae'
Description-Content-Type: text/markdown

# Drifting Models Reproduction (PyTorch)

[![CI](https://img.shields.io/github/actions/workflow/status/kmccleary3301/drift_models/ci.yml?branch=main&label=CI)](https://github.com/kmccleary3301/drift_models/actions/workflows/ci.yml)
[![Nightly](https://img.shields.io/github/actions/workflow/status/kmccleary3301/drift_models/nightly.yml?label=Nightly)](https://github.com/kmccleary3301/drift_models/actions/workflows/nightly.yml)
[![Runtime Health](https://img.shields.io/badge/runtime%20health-preflight%20enabled-2ea44f)](docs/runtime_health.md)
[![PyPI](https://img.shields.io/pypi/v/drift-models)](https://pypi.org/project/drift-models/)
[![Python](https://img.shields.io/pypi/pyversions/drift-models)](https://pypi.org/project/drift-models/)
[![License](https://img.shields.io/github/license/kmccleary3301/drift_models)](./LICENSE)

Community reproduction of *Generative Modeling via Drifting* in PyTorch.

## Project status and claim boundaries

- This repository is **not an official release** from the paper authors.
- We are actively hardening paper-faithful semantics and evidence artifacts.
- We do **not** currently claim full paper-level metric reproduction.
- Pixel pipeline remains **experimental** and should not be treated as parity-closed.

See:
- `docs/faithfulness_status.md`
- `docs/reproduction_report.md`
- `docs/experiment_log.md`
- `docs/eval_contract.md`

## Quickstart (60 seconds)

### Option A: `uv` (recommended)

```bash
uv sync --extra dev --extra eval --extra sdvae
uv run python scripts/runtime_preflight.py --device auto --check-torchvision --strict
uv run python scripts/train_toy.py --config configs/toy/quick.yaml --output-dir outputs/toy_quick --device cpu
```

### Option B: `pip`

```bash
python -m venv .venv
source .venv/bin/activate
pip install -U pip
pip install -e ".[dev,eval,sdvae]"
python scripts/runtime_preflight.py --device auto --check-torchvision --strict
python scripts/train_toy.py --config configs/toy/quick.yaml --output-dir outputs/toy_quick --device cpu
```

## Installation guides

- Linux + NVIDIA CUDA: `docs/install_linux_cuda.md`
- CPU-only: `docs/install_cpu_only.md`
- macOS (Apple Silicon / MPS): `docs/install_macos.md`
- Windows + WSL2: `docs/install_windows_wsl2.md`

## Common workflows

- Toy trajectory training: `docs/getting_started.md`
- Latent smoke training: `docs/getting_started.md`
- Sampling/eval smoke: `docs/getting_started.md`
- Full command catalog: `docs/commands.md`

## Compatibility tiers

Compatibility and support policy is documented in:
- `docs/compatibility_matrix.md`

## Runtime health

- Runtime preflight is enforced in CI on Linux/macOS/Windows and nightly on Linux.
- Preflight JSON reports are uploaded as workflow artifacts for each run.
- CI also generates an aggregated runtime summary + failure triage and posts it as a sticky PR comment.
- Runtime diagnostics guide: `docs/runtime_health.md`
- Local preflight entrypoint: `scripts/runtime_preflight.py`

## Reproducibility and evidence

- Run metadata contracts: `docs/provenance_contract.md`
- Claim/evidence mapping: `docs/claim_to_evidence_matrix.md`
- Release parity gate: `docs/release_gate_checklist.md`
- Public release gate: `docs/RELEASE_CHECKLIST.md`
- Branch protection policy: `docs/branch_protection.md`
- PyPI/TestPyPI publish setup: `docs/pypi_trusted_publishing.md`

## Contributing and governance

- Contribution guide: `CONTRIBUTING.md`
- Code of conduct: `CODE_OF_CONDUCT.md`
- Security policy: `SECURITY.md`
- Changelog: `CHANGELOG.md`

## Citation

If you use this repository, cite the original paper and this implementation repo.

Paper: *Generative Modeling via Drifting* (arXiv:2602.04770).
