Skip to content

Releases

Photostacker 0.1.0

Choose a tag to compare

@github-actions github-actions released this 12 Nov 15:25
51cc6c6

Photostacker finds series of geometrically identical photos in a folder (exposure brackets, HDR sets, fast bursts) and stacks each series as layers in one PSD/PSB or layered TIFF. It is an offline, MIT-licensed desktop app and command-line tool. This is its first release.

Downloads

System File
Windows 10 or later, x64 Photostacker_0.1.0_x64-setup.exe (installer), Photostacker_0.1.0_x64_en-US.msi, or Photostacker_0.1.0_x64-portable.zip (nothing to install)
Linux x86_64, glibc 2.35 or later (such as Ubuntu 22.04) Photostacker_0.1.0_amd64.deb or Photostacker_0.1.0_x86_64.AppImage
macOS 11 or later, Apple silicon Photostacker_0.1.0_aarch64.dmg
macOS 11 or later, Intel Photostacker_0.1.0_x64.dmg

Every package holds the desktop app and the command-line tool, photostacker-cli. Installing says how to install each one and what its first start asks: no package is signed by a developer's identity, so Windows and macOS ask for one more step.

Checking a download

  • checksums.txt holds the SHA-256 of every package.
  • Every file has a build-provenance attestation: gh attestation verify <file> --repo riccardogabellone/photostacker.

Good to know

  • Photoshop itself has not opened the files Photostacker writes yet. GIMP, PhotoCraft and Photopea have: see docs/photoshop-checks.md.
  • The macOS packages are built and checked by the project's workflows, but no person has started them on a Mac yet. If you try one, an issue saying how it went is welcome, with your macOS version and the message of the first start.
  • On a Mac with Apple silicon the visual check needs macOS 14 or later. On macOS 11 to 13 the app runs without it and says why.

Full Changelog: https://github.com/riccardogabellone/photostacker/commits/v0.1.0

Model files for the visual check (mirror)

Choose a tag to compare

@riccardogabellone riccardogabellone released this 12 Nov 15:16
51cc6c6

The two models of Photostacker's visual check, kept here as a second place to download them from.

The app and the command-line tool download a model only when asked: first from Hugging Face, and from this release if that fails. Either way they check the file's SHA-256 before installing it (the model registry, crates/core/src/models/registry.rs, holds both addresses, the sizes and the hashes).

File Model Size in bytes SHA-256
dinov2-small.onnx DINOv2 small 88,459,888 83141175ec78b4ff9a2bb58a4c7c264ba0054d1c2e122e5a8114b79a8d4179ea
dinov2-base.onnx DINOv2 base 346,555,570 de9c675d214f0171285f90f1d1d7716bce1c5e280e93826d071d57f7b9314d98

Where they come from, unchanged:

They are Meta's DINOv2 (https://github.com/facebookresearch/dinov2) as Xenova converted them to ONNX on Hugging Face, under the Apache License 2.0. The licence's text is LICENSE-Apache-2.0.txt in this release, as https://www.apache.org/licenses/LICENSE-2.0.txt gives it.

This release is not a version of Photostacker: it holds no program, and it does not change when the app does.