safetensors

safetensors Permissive license Active ★ 3,852 v0.8.0 released 2026-06-09

Install

safetensors on PyPI

pip

pip install safetensors

uv

uv add safetensors

poetry

poetry add safetensors

Package facts

License	not declared (permissive)
Python support	supports the current Python release (>=3.10)
Install friction	medium — platform-specific wheel
Runtime dependencies	none
Maintenance	actively maintained — 65 days since the last release
Last repo commit	2026-08-04
First released	2022-09-22
Popularity	one of the top 1,000 most-downloaded packages on PyPI (30-day window, as of 2026-08-13)
Known vulnerabilities	none known (OSV.dev, checked 2026-08-13)

Evidence: safetensors-0.8.0-cp310-abi3-macosx_10_12_x86_64.whl; safetensors-0.8.0-cp310-abi3-macosx_11_0_arm64.whl; safetensors-0.8.0-cp310-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl; safetensors-0.8.0-cp310-abi3-manylinux_2_17_armv7l.manylinux2014_armv7l.whl; safetensors-0.8.0-cp310-abi3-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl; safetensors-0.8.0-cp310-abi3-manylinux_2_17_s390x.manylinux2014_s390x.whl; safetensors-0.8.0-cp310-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl; safetensors-0.8.0-cp310-abi3-manylinux_2_31_riscv64.whl; safetensors-0.8.0-cp310-abi3-manylinux_2_5_i686.manylinux1_i686.whl; safetensors-0.8.0-cp310-abi3-musllinux_1_2_aarch64.whl; safetensors-0.8.0-cp310-abi3-musllinux_1_2_armv7l.whl; safetensors-0.8.0-cp310-abi3-musllinux_1_2_i686.whl; safetensors-0.8.0-cp310-abi3-musllinux_1_2_x86_64.whl; safetensors-0.8.0-cp310-abi3-win32.whl; safetensors-0.8.0-cp310-abi3-win_amd64.whl; safetensors-0.8.0-cp310-abi3-win_arm64.whl

Development Status :: 5 - Production/StableIntended Audience :: DevelopersIntended Audience :: EducationIntended Audience :: Science/ResearchLicense :: OSI Approved :: Apache Software LicenseOperating System :: OS IndependentProgramming Language :: Python :: 3Programming Language :: Python :: 3.10Programming Language :: Python :: 3.11Programming Language :: Python :: 3.12Programming Language :: Python :: 3.13Programming Language :: Python :: 3.14Topic :: Scientific/Engineering :: Artificial IntelligenceTyping :: Typed

About safetensors

from the package's own PyPI description — quoted content, verbatim

Installation

pip install safetensors

Usage

Numpy

from safetensors.numpy import save_file, load_file
import numpy as np

tensors = {
   "a": np.zeros((2, 2)),
   "b": np.zeros((2, 3), dtype=np.uint8)
}

save_file(tensors, "./model.safetensors")


# Now loading
loaded = load_file("./model.safetensors")

Torch

from safetensors.torch import save_file, load_file
import torch

tensors = {
   "a": torch.zeros((2, 2)),
   "b": torch.zeros((2, 3), dtype=torch.uint8)
}

save_file(tensors, "./model.safetensors")


# Now loading
loaded = load_file("./model.safetensors")

Developing

# inside ./safetensors/bindings/python
pip install .[dev]

Should be enough to install this library locally.

Testing

# inside ./safetensors/bindings/python
pip install .[dev]
pytest -sv tests/

Read as markdown · JSON record · Source repository · Homepage

AI interpretation — verify before relying

AI-generated interpretation of the package facts above; every digit, version, license, or vulnerability id it cites is grounded in the facts already shown on this page

Serializes and deserializes tensors to a safe, language-agnostic binary format designed for secure model storage and transfer across frameworks.

Medium install friction due to compiled wheels across many platforms, but pre-built binaries are available for Python 3.10+. Active maintenance with recent commits and no runtime dependencies simplifies deployment.

Licensed under Apache Software License (permissive), allowing commercial and private use with minimal restrictions.

Usage

pip install safetensors

from safetensors.numpy import save_file, load_file

tensors = {"a": [[0, 0], [0, 0]]}
save_file(tensors, "model.safetensors")
loaded = load_file("model.safetensors")

Requires Python 3.10 or later; compiled wheels available for common platforms but may require build tools on unsupported architectures.

Verdict: A production-stable, actively maintained serialization library with no known vulnerabilities, zero runtime dependencies, and broad platform support. Medium install friction is offset by pre-built wheels and permissive licensing, making it a low-risk choice for model persistence.

Needs verification

Performance characteristics (serialization speed, file size overhead) relative to alternatives.
Compatibility guarantees across safetensors versions for forward/backward reading of saved files.
Framework-specific integration details beyond the documented usage patterns.

tensor serialization formatsafe model checkpoint storagemachine learning model weights formatcross-framework tensor persistencesecure tensor file formatmodel weights binary storage