Skip to main content

NEPAdapters

NEPAdapters 为 NEP 模型提供统一的推理运行时:同一个模型可以通过 Python、C/C++ 或 LAMMPS 调用,并显式选择 CPU 或 CUDA 后端。

如果选中的后端不可用,或该后端不支持当前模型,接口会明确报错,不会静默改用其他后端。普通 NEP、qNEP、spin NEP、dipole、polarizability 和 DFT-D3 使用各自明确的接口,避免输出含义或形状随模型类型悄悄变化。

项目来源与致谢

NEPAdapters 最初从 NepTrainKit 内置的 NEP 运行时中拆分而来。拆分的目的不是重新包装 NEP,也不是替代 GPUMD,而是让计算后端能够脱离桌面应用独立维护、测试和发布,使 NepTrainKit 可以按需更新运行时,也让同一套能力能够被 Python、C/C++ 和 LAMMPS 复用。

各部分的来源和关系如下:

  • CPU 后端以 Zheyong Fan、Junjie Wang、Eric Lindgren 等作者维护的 NEP_CPU 独立 C++ 实现为基础,在保留原版权和 GPL 许可证声明的前提下,继续进行接口适配、正确性修复、OpenMP 并行和性能优化。
  • CUDA 后端由 NEPAdapters 独立实现,没有直接引入 GPUMD 的 CUDA 源码树。它遵循 NEP 论文和 GPUMD 的模型格式、数学定义与结果语义,并以 NEP_CPU 参考结果和 CPU/CUDA 一致性测试约束正确性。这里的“独立实现”指代码和数据流设计独立,不代表重新发明了 NEP 方法。
  • 适配层与前端源自 NepTrainKit 的实际使用需求,负责稳定 API、跨平台打包、运行时能力检查,以及 Python、C/C++ 和 LAMMPS 集成。

NEP 方法及其科学贡献来自 Zheyong Fan 等作者的原创工作,并在 GPUMD 中持续发展。NEPAdapters 的工作重点是工程实现与集成;它不是新的势函数方法,也不是 NEP_CPU 或 GPUMD 的官方发行组成部分。

感谢 NEP 与 GPUMD 作者和社区长期公开算法、代码、文档与验证工作。没有这些基础,NEPAdapters 和 NepTrainKit 都无从建立。GPUMD 的正式文档见 gpumd.org。

先选你要的入口

使用场景 从这里开始
Python / NumPy / ASE Python 接口指南
CMake 构建、测试和安装 构建与安装
LAMMPS CPU 或 CUDA LAMMPS 前端指南
C/C++ API 公共头文件 与 架构说明
测试和正确性验证 测试说明

Python 快速开始

仅使用 NumPy 接口:

python -m pip install nep-adapters

同时使用 ASE:

python -m pip install 'nep-adapters[ase]'

使用 ASE 做一次普通 NEP 计算:

from ase import Atoms
from nep_adapters.ase import NepAseCalculator

atoms = Atoms(
    "Fe2",
    positions=[[0.0, 0.0, 0.0], [1.43, 1.43, 1.43]],
    cell=[2.86, 2.86, 2.86],
    pbc=True,
)
atoms.calc = NepAseCalculator("nep.txt", backend="cpu")

print(atoms.get_potential_energy())
print(atoms.get_forces())

批量预测多结构 XYZ,并把结果作为 ASE 单点计算器挂回每个结构:

from ase.io import read, write
from nep_adapters import NEPCalculator
from nep_adapters.ase import attach_single_point

structures = read("input.xyz", index=":")

calculator = NEPCalculator("nep.txt", backend="cpu")
attach_single_point(structures, calculator)
calculator.close()

write("predicted.xyz", structures, format="extxyz")

这里的 batch 是指一次把完整的 structures 列表传给计算后端。 calculator.predict_structures([atoms]) 仍然只计算一帧,不会获得跨结构的 batch 收益;attach_single_point(structures, calculator) 会对完整列表执行一次 predict_structures(structures),再为每个结构挂载 SinglePointCalculator。 写出的 extxyz 包含 energy、forces 和 ASE 约定的 stress;调用 calculator.close() 后这些结果仍可读取,不会再次调用 NEP 模型。 相比逐个结构挂载 NepAseCalculator 并分别触发计算,这种方式可以减少大量 结构的逐帧调用开销;小 batch 不保证更快,实际收益取决于结构数、每帧原子数 和后端。当前 batch 接口要求所有结构均为全周期。

CPU batch 什么时候更快

如果已经拿到一个结构列表,不要在 Python 中逐帧调用:

# 逐帧循环:每一帧都会重新进入 Python/native 边界
for atoms in structures:
    prediction = calculator.predict_structures([atoms])

# batch:一次提交完整列表
prediction = calculator.predict_structures(structures)

下面是 32 原子 BCC Fe、NEP89 模型的 CPU 实测结果。每个点取五组成对 A/B 的中位数;测试机器为 Intel Xeon Gold 6530,线程均绑定到同一个 CPU socket。 表中的倍数表示 batch 相对 Python 逐帧循环的加速比:

CPU 核数 512 帧 2048 帧
8 1.97× 1.92×
16 2.70× 2.60×
32 3.84× 3.97×

32 核预测 2048 帧时,逐帧循环约为 4109 frame/s,batch 约为 16267 frame/s。8–128 帧的小 batch 主要减少 Python 调用和数据组装开销, 本算例中通常只有 1.02–1.14×;结构数达到跨结构并行阈值后,收益才会明显。 这些数值用于说明趋势,不是所有模型和机器的固定加速比。测试中 energy 完全 一致,force 和 virial 的最大绝对差为 9.44e-16。

预编译 wheel 支持 CPython 3.10–3.14:

平台 wheel 包含的后端 运行要求
Linux x86_64 CPU + CUDA CPU 可直接使用;CUDA 需要 NVIDIA 驱动和受支持的 GPU
Windows x86_64 CPU + CUDA CPU 可直接使用;CUDA 需要 NVIDIA 驱动和受支持的 GPU
macOS x86_64 / arm64 CPU 不提供 CUDA 后端

Linux 和 Windows 的预编译 wheel 已包含 CUDA 运行时,不要求另外安装 CUDA Toolkit。普通 import nep_adapters 或选择 backend="cpu" 不会初始化 CUDA;只有显式选择 backend="cuda" 或检查 CUDA 状态时才会加载 GPU 扩展。CUDA 加载或计算失败时不会改用 CPU。

CUDA wheel 内置 sm_60 和 sm_89 SASS,并保留 compute_60 PTX。没有对应 SASS 的较新架构会在驱动支持时通过 PTX 即时编译运行,因此兼容范围更广,但可能增加首次加载时间或产生一定性能损失。

可以运行下面的命令检查扩展、驱动、设备,并实际完成一次 CUDA 内存分配、核函数启动、同步和结果回传:

python -c "from nep_adapters import backend_status; s=backend_status('cuda'); print(s); assert s.installed and s.available, s"

需要修改源码或只为本机 GPU 编译时,从仓库安装:

git clone https://github.com/MagTheoryLab/NEPAdapters.git
cd NEPAdapters
python -m pip install '.[ase]'

源码安装会依次查找 CUDACXX、CUDAToolkit_ROOT、CUDA_PATH、 CUDA_HOME 和 PATH 中的 NVCC。检测到 CUDA Toolkit 时自动构建 CPU+CUDA,否则只构建 CPU。不指定 CMAKE_CUDA_ARCHITECTURES 时使用 native,只为构建机器上的 GPU 编译。NEP_CUDA=1 和 NEP_CUDA=0 可分别强制启用或禁用 CUDA。

更多 NumPy batch、qNEP、spin、descriptor 和错误处理示例见 Python 接口指南。

CMake 快速开始

只构建 CPU core 和测试:

cmake -S . -B .build/release \
  -DCMAKE_BUILD_TYPE=Release \
  -DNEP_ADAPTERS_BUILD_TESTS=ON
cmake --build .build/release -j2
ctest --test-dir .build/release --output-on-failure

Python 源码安装会自动查找 NVCC;独立 CMake 和 LAMMPS 构建仍按目标显式 打开 CUDA,避免有 Toolkit 的机器意外改变 CPU-only 构建。完整组合、安装命令 和选项默认值见 构建与安装。

Intel oneAPI CPU 构建

使用 IntelLLVM icpx 并开启 NEP_ADAPTERS_CPU_ENABLE_NATIVE_ARCH=ON(即允许 -march=native)时, 建议在 Release 编译参数中加入 -fno-vectorize:

cmake -S . -B .build/intel-cpu \
  -DCMAKE_BUILD_TYPE=Release \
  -DCMAKE_CXX_COMPILER=icpx \
  -DCMAKE_CXX_FLAGS_RELEASE="-O3 -DNDEBUG -fno-vectorize" \
  -DNEP_ADAPTERS_CPU_ENABLE_NATIVE_ARCH=ON \
  -DNEP_ADAPTERS_BUILD_TESTS=ON
cmake --build .build/intel-cpu -j2
ctest --test-dir .build/intel-cpu --output-on-failure

在 Intel Xeon CPU Max 9470C 与 IntelLLVM 2024.2.1 的组合上,-march=native 会触发 loop vectorizer 对 CPU 内部 cell list 索引更新的错误优化,进而破坏 batch、descriptor、dipole 和 polarizability 结果。源码已对已知的 loop-carried dependency 单独禁止向量化;保留全局 -fno-vectorize 是该环境下 经过正确性与性能测试的推荐配置。官方预编译 wheel 使用可移植构建配置,不使用 icpx,也不启用 CPU native architecture,因此无需追加该参数。

LAMMPS 安装方式

LAMMPS 有两种支持的安装方式:

  • 运行时插件(推荐):单独构建共享库,通过 LAMMPS_PLUGIN_PATH 加载,不需要重新编译 LAMMPS;
  • 源码树内置:运行 tools/install_lammps_source.py,把受管的 pair style 源文件复制到 lammps/src/,再通过随附的 CMake 配置把 core 和 engine 静态编入 lmp。运行时不需要设置插件环境变量。

不要手工只复制 .cpp 和 .h 文件;安装脚本还会写入连接依赖所需的 CMake 配置、记录文件哈希,并提供安全卸载。两种方式的完整命令见 LAMMPS 前端指南。

当前支持范围

前端 / 后端 普通 NEP qNEP spin NEP 其他响应模型
Python / CPU 支持 charge/BEC 支持 支持 dipole、polarizability、DFT-D3 支持;DFT-D3 仅普通非 spin、非 charge 模型
Python / CUDA 支持 NEP4/NEP5 支持 direct;预编译 wheel 不包含 PPPM 支持 不支持 dipole、polarizability、DFT-D3
LAMMPS / CPU 插件 支持,已完成普通模型 MPI 冒烟测试 尚未列入已验证支持范围 已验证单节点 1/2/4/8 个 MPI 进程,覆盖能量、力、磁力、virial、spin 读回和短程 dynspin 不适用
LAMMPS / CUDA Kokkos 插件 支持 不支持,运行时会明确报错 已验证单节点 1/2/4/8 个 MPI 进程 不适用

补充边界:

  • 所有 Python batch 接口目前只接受全周期 pbc=(1, 1, 1)。
  • LAMMPS spin 不使用单独的 nep/spin/* 名称;仍使用 nep/cpu 或 nep/gpu/kk,并配合 atom_style spin 或 spin/kk。
  • 构建 LAMMPS CPU plugin 且检测到 MPI launcher 时,CTest 除单 rank 外,还会根据 MPIEXEC_MAX_NUMPROCS 自动注册最多 2/4/8 ranks 的 spin 回归;在 Slurm 上应先申请足够的计算资源,再运行 ctest -L mpi --output-on-failure。
  • qNEP 的 CUDA batch API 可用,但 LAMMPS Kokkos 需要独立的 ghost/charge 数据流,因此当前会明确拒绝该组合,不会静默降级。
  • virial 在不同公共入口有不同顺序;调用前请查 Python 接口指南 或 公共 API 契约。

项目结构

  • include/nep_adapters/:公共 C/C++ API 和数据约定;
  • src/:core 注册、调度和错误处理;
  • engines/cpu/、engines/cuda/:计算后端;
  • frontends/python/、frontends/lammps/:用户前端;
  • tests/:契约、fixture、oracle 和集成测试;
  • benchmarks/:性能与扩展性入口。

LAMMPS 和 Python 都是前端,不是计算后端;算法实现只保留在 engine 层。

科研引用

NEPAdapters 目前没有独立的软件论文。论文方法部分可以记录所用的 NEPAdapters 版本和仓库地址,但这不能替代对 NEP 原始工作的引用。如果 NEPAdapters 用于科研工作,请至少引用 NEP 的基础论文:

  • Z. Fan, Z. Zeng, C. Zhang, Y. Wang, K. Song, H. Dong, Y. Chen, and T. Ala-Nissila, “Neuroevolution machine learning potentials: Combining high accuracy and low cost in atomistic simulations and application to heat transport,” Physical Review B 104, 104309 (2021). DOI: 10.1103/PhysRevB.104.104309

如果 NepTrainKit 参与了数据准备、主动学习、训练管理或可视化,也请引用:

  • C. Chen, Y. Li, R. Zhao, Z. Liu, Z. Fan, G. Tang, and Z. Wang, “NepTrain and NepTrainKit: Automated active learning and visualization toolkit for neuroevolution potentials,” Computer Physics Communications 317, 109859 (2025). DOI: 10.1016/j.cpc.2025.109859

涉及 GPUMD 的 NEP 训练、实现或分子动力学工作时,建议同时引用 GPUMD 软件论文:

  • Z. Fan et al., “GPUMD: A package for constructing accurate machine-learned potentials and performing highly efficient atomistic simulations,” The Journal of Chemical Physics 157, 114801 (2022). DOI: 10.1063/5.0106617

许可证

NEPAdapters 以 GNU GPL v3 或更高版本 发布。

Metadata

Release files for nep-adapters 1.0.3

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for nep-adapters 1.0.3
File Size Uploaded
nep_adapters-1.0.3.tar.gz 2.0 MB Details

Built distributions (wheels)

Table of built distributions (wheels) for nep-adapters 1.0.3
File
nep_adapters-1.0.3-cp314-cp314-win_amd64.whl CPython 3.14 CPython 3.14 Windows x86-64 Details
nep_adapters-1.0.3-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl CPython 3.14 CPython 3.14 Linux glibc 2.27+ x86-64, Linux glibc 2.28+ x86-64 Details
nep_adapters-1.0.3-cp314-cp314-macosx_14_0_x86_64.whl CPython 3.14 CPython 3.14 macOS 14.0+ x86-64 Details
nep_adapters-1.0.3-cp314-cp314-macosx_14_0_arm64.whl CPython 3.14 CPython 3.14 macOS 14.0+ ARM64 Details
nep_adapters-1.0.3-cp313-cp313-win_amd64.whl CPython 3.13 CPython 3.13 Windows x86-64 Details
nep_adapters-1.0.3-cp313-cp313-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl CPython 3.13 CPython 3.13 Linux glibc 2.27+ x86-64, Linux glibc 2.28+ x86-64 Details
nep_adapters-1.0.3-cp313-cp313-macosx_14_0_x86_64.whl CPython 3.13 CPython 3.13 macOS 14.0+ x86-64 Details
nep_adapters-1.0.3-cp313-cp313-macosx_14_0_arm64.whl CPython 3.13 CPython 3.13 macOS 14.0+ ARM64 Details
nep_adapters-1.0.3-cp312-cp312-win_amd64.whl CPython 3.12 CPython 3.12 Windows x86-64 Details
nep_adapters-1.0.3-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl CPython 3.12 CPython 3.12 Linux glibc 2.28+ x86-64, Linux glibc 2.27+ x86-64 Details
nep_adapters-1.0.3-cp312-cp312-macosx_14_0_x86_64.whl CPython 3.12 CPython 3.12 macOS 14.0+ x86-64 Details
nep_adapters-1.0.3-cp312-cp312-macosx_14_0_arm64.whl CPython 3.12 CPython 3.12 macOS 14.0+ ARM64 Details
nep_adapters-1.0.3-cp311-cp311-win_amd64.whl CPython 3.11 CPython 3.11 Windows x86-64 Details
nep_adapters-1.0.3-cp311-cp311-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl CPython 3.11 CPython 3.11 Linux glibc 2.28+ x86-64, Linux glibc 2.27+ x86-64 Details
nep_adapters-1.0.3-cp311-cp311-macosx_14_0_x86_64.whl CPython 3.11 CPython 3.11 macOS 14.0+ x86-64 Details
nep_adapters-1.0.3-cp311-cp311-macosx_14_0_arm64.whl CPython 3.11 CPython 3.11 macOS 14.0+ ARM64 Details
nep_adapters-1.0.3-cp310-cp310-win_amd64.whl CPython 3.10 CPython 3.10 Windows x86-64 Details
nep_adapters-1.0.3-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl CPython 3.10 CPython 3.10 Linux glibc 2.28+ x86-64, Linux glibc 2.27+ x86-64 Details
nep_adapters-1.0.3-cp310-cp310-macosx_14_0_x86_64.whl CPython 3.10 CPython 3.10 macOS 14.0+ x86-64 Details
nep_adapters-1.0.3-cp310-cp310-macosx_14_0_arm64.whl CPython 3.10 CPython 3.10 macOS 14.0+ ARM64 Details

Total release size: 228.6 MB

Release files / nep_adapters-1.0.3.tar.gz

Download URL nep_adapters-1.0.3.tar.gz
Size 2.0 MB
Tags Source
SHA-256 checksum
How to use checksums
76bbc83a04c04888ea75c1b53587b63e805fc55326302747abd23d69a38c3d6b
BLAKE2b-256 checksum
How to use checksums
e32e9cfb39187d78ebe18773e155a480f50f01741c980adb2e693855414313e6
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp314-cp314-win_amd64.whl

Download URL nep_adapters-1.0.3-cp314-cp314-win_amd64.whl
Size 21.6 MB
Tags CPython 3.14 Windows x86-64
SHA-256 checksum
How to use checksums
304ff80226a34d394867204a3e864cce751a561ad779661197c98c4eac8fc646
BLAKE2b-256 checksum
How to use checksums
0cd9a05d7c149f53f6590f60513ef5a0b22dff314b71b9522e70bd5b45849701
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl

Download URL nep_adapters-1.0.3-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl
Size 22.0 MB
Tags CPython 3.14 Linux glibc 2.27+ x86-64 Linux glibc 2.28+ x86-64
SHA-256 checksum
How to use checksums
4a53034dbf8925856f5557f87aacffc99373f42206b1f14c6a05395d9a76cd7d
BLAKE2b-256 checksum
How to use checksums
41bc8eb63f2ea7b91c5738fd6fa577154afccb516b93c37f856ef5d354b63067
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp314-cp314-macosx_14_0_x86_64.whl

Download URL nep_adapters-1.0.3-cp314-cp314-macosx_14_0_x86_64.whl
Size 974.1 kB
Tags CPython 3.14 macOS 14.0+ x86-64
SHA-256 checksum
How to use checksums
89cde8c31de5026299d4fd12f88702da2d0465379e35881b6ac2917e3d423060
BLAKE2b-256 checksum
How to use checksums
6b52086c206624bc8496241362769db85851535f78d65aee175001f7cb21e22f
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp314-cp314-macosx_14_0_arm64.whl

Download URL nep_adapters-1.0.3-cp314-cp314-macosx_14_0_arm64.whl
Size 882.5 kB
Tags CPython 3.14 macOS 14.0+ ARM64
SHA-256 checksum
How to use checksums
9744940f43a18ed571d72db44b1754507c7fdb26b05389d8b517599e84c0883d
BLAKE2b-256 checksum
How to use checksums
a6997babf6674962ee2fc7a7edabfc76abccead1d0db87bfc92fb137fabf3a4b
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp313-cp313-win_amd64.whl

Download URL nep_adapters-1.0.3-cp313-cp313-win_amd64.whl
Size 21.5 MB
Tags CPython 3.13 Windows x86-64
SHA-256 checksum
How to use checksums
bc5e28d2c0287c333fbf2af3072d8fdf9602f350710aaa6b2c7e656bc08a712e
BLAKE2b-256 checksum
How to use checksums
c7092a3b372083d1ab444d4148bf3bf078c1d59c47170426ced8de980aac8c31
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp313-cp313-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl

Download URL nep_adapters-1.0.3-cp313-cp313-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl
Size 22.0 MB
Tags CPython 3.13 Linux glibc 2.27+ x86-64 Linux glibc 2.28+ x86-64
SHA-256 checksum
How to use checksums
7ceea532f795e5f195e828143fe967865c82fb85be7db353f6eaf0f614239f21
BLAKE2b-256 checksum
How to use checksums
4f1207c5f65ac65e8b210ca2d070b419fb0d7bcc00b425cd53a5366b5e1ee3e6
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp313-cp313-macosx_14_0_x86_64.whl

Download URL nep_adapters-1.0.3-cp313-cp313-macosx_14_0_x86_64.whl
Size 973.9 kB
Tags CPython 3.13 macOS 14.0+ x86-64
SHA-256 checksum
How to use checksums
a57e013beef6a5987e2e9d8203db8fb2c2f3b244e1df2ff3ea6fab77b37caae8
BLAKE2b-256 checksum
How to use checksums
00e81d5fb73e59be6b4e5ccde6e7118153d2ef6c111d4de0c37353f789cabef7
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp313-cp313-macosx_14_0_arm64.whl

Download URL nep_adapters-1.0.3-cp313-cp313-macosx_14_0_arm64.whl
Size 882.2 kB
Tags CPython 3.13 macOS 14.0+ ARM64
SHA-256 checksum
How to use checksums
8e5696c14c83b8ca711abe6cf639f1862e836542d03c80931ec59f2d7b7bfcf0
BLAKE2b-256 checksum
How to use checksums
d2e474837b19b06e849298915892a7e2b77498fc25233eaea17c3ebb169655a9
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp312-cp312-win_amd64.whl

Download URL nep_adapters-1.0.3-cp312-cp312-win_amd64.whl
Size 21.5 MB
Tags CPython 3.12 Windows x86-64
SHA-256 checksum
How to use checksums
af8e92ad7429e04242ec6804d6c13dbd98bac2bda32589f33a6c74c8ebea42a1
BLAKE2b-256 checksum
How to use checksums
fc4636f84aafb4f5728a0e0d48564bd2acba045d70a8e0ab4535a1f5124d5795
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl

Download URL nep_adapters-1.0.3-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl
Size 22.0 MB
Tags CPython 3.12 Linux glibc 2.27+ x86-64 Linux glibc 2.28+ x86-64
SHA-256 checksum
How to use checksums
0ae05b15850fa1d44032ab1f43faef4aa4a7921d36e36eb89692d2a1d204ad37
BLAKE2b-256 checksum
How to use checksums
594678e91e6690d9209bab4f59e02d9bccc021bb5d0ad17695cfdb63f12dfd21
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp312-cp312-macosx_14_0_x86_64.whl

Download URL nep_adapters-1.0.3-cp312-cp312-macosx_14_0_x86_64.whl
Size 973.7 kB
Tags CPython 3.12 macOS 14.0+ x86-64
SHA-256 checksum
How to use checksums
2bdc34539070f3aa5c1fcd9a07ba20c0eeeab79db5c6337b4030218b176fea0b
BLAKE2b-256 checksum
How to use checksums
d41d55b5aa1ac74a65e0c729dd4be33b1a1695f0ae245a5d67e70ac947f97b3e
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp312-cp312-macosx_14_0_arm64.whl

Download URL nep_adapters-1.0.3-cp312-cp312-macosx_14_0_arm64.whl
Size 882.1 kB
Tags CPython 3.12 macOS 14.0+ ARM64
SHA-256 checksum
How to use checksums
04539d55d43745dcfd0c75a4977d92864ff1713089b3bba55dbade16fd69d9f8
BLAKE2b-256 checksum
How to use checksums
75504bfd4855f5905349fe051e73df382d1035a55e9e387a5942891716987c1d
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp311-cp311-win_amd64.whl

Download URL nep_adapters-1.0.3-cp311-cp311-win_amd64.whl
Size 21.5 MB
Tags CPython 3.11 Windows x86-64
SHA-256 checksum
How to use checksums
9a909d4b9d723be6b297d598b142c8e93b24cc0571aed6ba2a1a3fd946e836be
BLAKE2b-256 checksum
How to use checksums
34976da5a35666288e9f58d4635710e427c56db22ba166a5fdbdec27f251ec90
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp311-cp311-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl

Download URL nep_adapters-1.0.3-cp311-cp311-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl
Size 21.9 MB
Tags CPython 3.11 Linux glibc 2.27+ x86-64 Linux glibc 2.28+ x86-64
SHA-256 checksum
How to use checksums
76110930f9475fc4b5cdac926d66a103cef22871d36cbd022ebccc2da7b9a452
BLAKE2b-256 checksum
How to use checksums
7fc3736faf16507dd10cbbbb236e0a01e3b2f2e3f194c2d40fc42beb70e2edc8
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp311-cp311-macosx_14_0_x86_64.whl

Download URL nep_adapters-1.0.3-cp311-cp311-macosx_14_0_x86_64.whl
Size 971.5 kB
Tags CPython 3.11 macOS 14.0+ x86-64
SHA-256 checksum
How to use checksums
bb295dafddd0ce176f9574862930dafd4a88b8edae5db9fe2045c75cfac075ef
BLAKE2b-256 checksum
How to use checksums
c1cb7bdd707e614b1da463ab547c7bbf1777d0779d9d6b739cd87e7493b08318
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp311-cp311-macosx_14_0_arm64.whl

Download URL nep_adapters-1.0.3-cp311-cp311-macosx_14_0_arm64.whl
Size 880.6 kB
Tags CPython 3.11 macOS 14.0+ ARM64
SHA-256 checksum
How to use checksums
9c52e56c25a48197eab8e55ed0b2840c26a39e5b80e00e8a44e9385e4b74a006
BLAKE2b-256 checksum
How to use checksums
f9430a5ff33c18c7d16ba7f8744707b7c7adb8416d01f22b07b61f3e270c0add
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp310-cp310-win_amd64.whl

Download URL nep_adapters-1.0.3-cp310-cp310-win_amd64.whl
Size 21.5 MB
Tags CPython 3.10 Windows x86-64
SHA-256 checksum
How to use checksums
9bca5cb23259540e59fd362d01c23d06aa3994795364f37ce68b88b2f8e48e44
BLAKE2b-256 checksum
How to use checksums
0597008ea75f94c682693b5dd1f2c7dcad860d308570349e610eab07bde3ac3e
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl

Download URL nep_adapters-1.0.3-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl
Size 21.9 MB
Tags CPython 3.10 Linux glibc 2.27+ x86-64 Linux glibc 2.28+ x86-64
SHA-256 checksum
How to use checksums
f1239628b63244467aa8a8dd9d9710a98ed74e25e14a56c6b13c9514e72aba3f
BLAKE2b-256 checksum
How to use checksums
9a437cb6bb0c8fee2f6af9e6d23f646fa1097b2993b1300b94ef55ee7bd18b1a
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp310-cp310-macosx_14_0_x86_64.whl

Download URL nep_adapters-1.0.3-cp310-cp310-macosx_14_0_x86_64.whl
Size 970.1 kB
Tags CPython 3.10 macOS 14.0+ x86-64
SHA-256 checksum
How to use checksums
a08423be7076e348bff44ed32aa73a134fb0e23fda8a7cabe9451866bf552a5c
BLAKE2b-256 checksum
How to use checksums
df21b4eef8ad4c1d125870f9d3d2a1072cc958b1b4c8570c5bbeee90b3b90628
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release files / nep_adapters-1.0.3-cp310-cp310-macosx_14_0_arm64.whl

Download URL nep_adapters-1.0.3-cp310-cp310-macosx_14_0_arm64.whl
Size 879.3 kB
Tags CPython 3.10 macOS 14.0+ ARM64
SHA-256 checksum
How to use checksums
bdcd1878255140e70658fd41fb4f33686ea47d507a34cb152893e9e48899e598
BLAKE2b-256 checksum
How to use checksums
5d0ebb044f9cea869f76a1a6acad75168f67c8afc46c9197834e7ddb362e9021
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 11, 2026.

Transparency log

Release history Release notifications | RSS feed

This release

1.0.3 This release

21 release files

1.0.2

21 release files

1.0.1

21 release files

1.0.0

21 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page