misc/py-immich-model 0.2.0 FreeBSD

Runtime for the machine learning models used by Immich

System: FreeBSD
Repository: FreeBSD ports
Maintainer: [email protected]
Port added: 2026-10-24
Last update: 2026-10-24
Commit: c40cf98ecc
License: AGPLv3
Categories: misc, python

Description

This is the model runtime that the Immich machine learning service loads at inference time. It holds the ONNX graph lowerings for Immich's CLIP and face recognition models, the per-execution-provider variants of them, and the model catalogue that maps a model name to its files.

The exporter and benchmarking tools that live in the same source tree are not installed; they pull in PyTorch and are only needed to regenerate the models.

Version history 1

0.2.0 2026-10-24 c40cf98ecc

Upstream's version as each commit to the port's Makefile set it, or as this site first saw it in the system's packages; the tree's own revisions are not versions here.

Install

Binary package:

pkg install py312-immich-model

From the ports tree:

cd /usr/ports/misc/py-immich-model/ && make install clean

To depend on this port in another port's Makefile:

RUN_DEPENDS=	py312-immich-model>0:misc/py-immich-model

Packages

The version each package repository carries. “-” means the repository builds no package of it, “n/a” that there is no such repository.

Repository

aarch64

amd64

armv6

armv7

i386

FreeBSD:14:latest

FreeBSD:14:quarterly

n/a

FreeBSD:15:latest

n/a

FreeBSD:15:quarterly

n/a

FreeBSD:16:latest

n/a

FreeBSD:16:quarterly

n/a

Dependency graph

Ports that need it on the left, what it needs to run on the right; at most 8 of each. "Dependencies" below lists them all.

graph LR
  me["misc/py-immich-model"]
  u0["www/immich-ml"] --> me
  me --> d0["archivers/brotli"]
  me --> d1["archivers/libdeflate"]
  me --> d2["archivers/liblz4"]
  me --> d3["archivers/lzo2"]
  me --> d4["archivers/minizip"]
  me --> d5["archivers/py-brotli"]
  me --> d6["archivers/py-zopfli"]
  me --> d7["archivers/zopfli"]
  me -.-> dmore(["200 more"])
  classDef this fill:#ab2b28,color:#ffffff,stroke:#ab2b28
  class me this

Dependencies

As this port's Makefile names them.

Build 1

${PYTHON_PKGNAMEPREFIX}hatchling>0 : devel/py-hatchling

Runtime 5

${PYTHON_PKGNAMEPREFIX}numpy>=1.26 : math/py-numpy
${PYTHON_PKGNAMEPREFIX}onnx>=1.16.1 : misc/py-onnx
${PYTHON_PKGNAMEPREFIX}onnx-ir>=0.2.1 : misc/py-onnx-ir
${PYTHON_PKGNAMEPREFIX}onnxscript>=0.7.0 : misc/py-onnxscript
${PYTHON_PKGNAMEPREFIX}safetensors>=0.7 : misc/py-safetensors

The full lists below include those, and their dependencies in turn.

Build dependencies 19

Run dependencies 208

The first 100 are listed.

Required by 1

www/immich-ml

Configuration

No options to configure.

USES: python
Service scripts: none

Files

distinfo 1

File

Size

SHA256

immich-app-ml-models-0.2.0-8bf3fd1b44b0_GH0.tar.gz

1.784907e+06

56a1352b8125a9afb2b1fc2f70d9a5092e97c45215cb00cbd8cb423c7ee4e18a

Commit history 1

Date Commit Credits Log message

2026-10-24

c40cf98ecc

Jochen Neumeister

misc/py-immich-model: Add New Port

The model runtime that Immich's machine learning service loads at inference time: the ONNX graph lowerings for the CLIP and face recognition models, the per execution provider variants of them, and the catalogue that maps a model name to its files. The exporter and benchmarking tools from the same source tree are not installed; they pull in PyTorch and are only needed to regenerate the models. Sponsored by: Netzkommune GmbH