Skip to main content

szl-invariants

PyPI OpenSSF Scorecard

SZL Holdings — Orbit v2

Control before action. Evidence after.

Product · Proof · GitHub organization

Software kernel for SZL runtime invariants. Not a model. No weights.

Python lives under torch-ext/szl_invariants/ (__init__.py is the package). Hub kernel: kernels/SZLHOLDINGS/szl-invariants, published from this repository by hf-kernel-publish (card source: hf/kernel-card.md). The model repo SZLHOLDINGS/szl-invariants only points to the kernel repo; it holds no loadable build.

What this is NOT

  • Not trained weights, not a LoRA
  • No MEASURED CUDA benches in this repo
  • Passing selfcheck (if present) is not an eval leaderboard
  • Hub model.joblib is QUARANTINED executable serialization. Do not joblib.load it. GitHub source is the approved path.

Load only a verified publication

A merged GitHub fix is not proof of a Hugging Face release. This repository's hf-kernel-publish workflow is the only writer of the kernel repo. It builds build/torch-cpu and build/torch-universal from torch-ext/ at one reviewed commit, pushes them fast-forward to the kernel repo's main and v1 branches, and then loads them anonymously through get_kernel in a fresh job to check the served bytes against GitHub. Its publish and verify receipts (workflow artifacts) name the kernel-repo commits.

The stable channel is get_kernel("SZLHOLDINGS/szl-invariants", version=1, trust_remote_code=True). For production, pin an exact build: set SZL_INVARIANTS_HF_REVISION to the immutable kernel-repo commit of v1 from a passing verify receipt, not a branch name, version label, or the GitHub source commit. Use kernels>=0.17.2. Do not substitute main when a receipt or revision is unavailable; stop instead.

import os
import re

hf_revision = os.environ.get("SZL_INVARIANTS_HF_REVISION", "")
if re.fullmatch(r"[0-9a-f]{40}", hf_revision) is None:
    raise ValueError("A verified immutable Hugging Face revision is required")

from kernels import get_kernel

kernel = get_kernel(
    "SZLHOLDINGS/szl-invariants",
    revision=hf_revision,
    trust_remote_code=True,
)

The format check above does not verify provenance, hashes, authorization, or runtime compatibility. Those come from the release evidence. Enabling trust_remote_code permits execution of the selected repository's Python; use it only after reviewing that exact revision and its evidence. This example neither unquarantines model.joblib nor establishes acceleration or model quality.

For source-only development, use a separately reviewed, immutable GitHub source revision and run its tests in an isolated environment. Keep that evidence distinct from Hub publication and installed-runtime verification.

Doctrine v11. Λ = Conjecture 1 (advisory, never a theorem). Apache-2.0. Owner: Stephen Lutar / SZL Holdings.

Metadata

Release files for szl-invariants 0.1.2

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

Source distribution (sdist)

Source distribution for szl-invariants 0.1.2
File Size Uploaded Provenance Actions
szl_invariants-0.1.2.tar.gz 31.8 kB With provenance Details for szl_invariants-0.1.2.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for szl-invariants 0.1.2
File Interpreter ABI Platform Actions
szl_invariants-0.1.2-py3-none-any.whl Python 3 none any Details for szl_invariants-0.1.2-py3-none-any.whl

Total release size: 49.7 kB

Release files / szl_invariants-0.1.2.tar.gz

Download URL szl_invariants-0.1.2.tar.gz
Size 31.8 kB
Tags Source
SHA-256 checksum
How to use checksums
c10f0869e8b6a2ac402e889c379cd4bbdf9a0b11b5f77a1e9cedd9bdea42fb18
BLAKE2b-256 checksum
How to use checksums
ce20b178255cd5979e939fbb80bedfb3bb2e43fca06e3e4e67903c77a53b8c5f
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

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 Oct 30, 2026.

Transparency log

Release files / szl_invariants-0.1.2-py3-none-any.whl

Download URL szl_invariants-0.1.2-py3-none-any.whl
Size 17.9 kB
Tags Python 3
SHA-256 checksum
How to use checksums
da73a6c35e2257b399ceef5c789e84ac4486b2574c67a29a610c087b2cd59ba2
BLAKE2b-256 checksum
How to use checksums
8f0610a2210415aa1d4e68ba6fba644d25d0c11d35ce95170fac991d0531c8bb
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

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 Oct 30, 2026.

Transparency log

Security for szl-invariants 0.1.2

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.

Full provenance

Both files in this release have been uploaded with provenance metadata.

Select from 2 files

Attestation for szl_invariants-0.1.2.tar.gz

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 Oct 30, 2026.

Transparency log

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

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