Autoforge

Latest version: v1.6.4

Safety actively analyzes 723217 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 1 of 5

1.6.3

Just a small change to fix a bug with the initialization getting stuck sometimes

**Full Changelog**: https://github.com/hvoss-techfak/AutoForge/compare/v1.6.2...V1.6.3

1.6.2

Added a patch based smothness loss using a laplacian kernel. In my tests this gives better smoothness with more defined edges.

**Full Changelog**: https://github.com/hvoss-techfak/AutoForge/compare/v1.6.1...v1.6.2

1.6.1

We ran a hyperparameter search on a lot of parameters and found better default values.
Also the "firefly" issue was pretty bad in the last version, so we now multiply the learned height map with 1e-8 to only change layers with bad clustering results. This gives far better results than in the previous versions and the noisy pixel (firefily) issue seems to be fixed.

**Full Changelog**: https://github.com/hvoss-techfak/AutoForge/compare/v1.6.0...v1.6.1

1.6.0

We now learn the height map again, which is so much better for image quality.
Additionally STLs should now be produced almost instantly.

Also I squashed a lot of annoying bugs and cleaned the code a bit.

**Full Changelog**: https://github.com/hvoss-techfak/AutoForge/compare/V1.5.5...v1.6.0

1.5.5

Just a small pruning bugfix

**Full Changelog**: https://github.com/hvoss-techfak/AutoForge/compare/v1.5.4...V1.5.5

1.5.4

I just realized that kmeans is maybe not quite the brightest idea for image color clustering and quantization for images already exists.

**Full Changelog**: https://github.com/hvoss-techfak/AutoForge/compare/V1.5.3...v1.5.4

Page 1 of 5

© 2025 Safety CLI Cybersecurity Inc. All Rights Reserved.