Skip to content

Infinity v3.0.0-alpha.2

Pre-release
Pre-release

Choose a tag to compare

View all tags
@infinity-collaborator-ci infinity-collaborator-ci released this 14 Oct 20:07
· 1 commit to main since this release
v3.0.0-alpha.2
175ef5a
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Native LibTorch NVIDIA image for Infinity v3.

docker pull ghcr.io/michaelfeil/infinity_v3:nvidia-3.0.0-alpha.2

One image targets SM80, SM86, SM89, SM90, SM100, and SM120, with runtime kernel dispatch. SM75 is unsupported. The image defaults to LibTorch and uses Torch 2.14.1 with CUDA 13.0.

Local H100 checks matched the qualified host build exactly for Laya decisions, ModernBERT reranking, and BGE embeddings. Other SM targets are compile-verified; runtime qualification currently covers H100.

Build locally and run.

Images are published by the tag-triggered GitHub Actions workflow. This prerelease does not replace stable/latest image tags.

Includes repository, native-kernel, SDK, and Cargo dependency notices in the image.

Assets 2
Loading