Repository navigation
release: 0.60.0 - #1350
Merged
Merged
release: 0.60.0#1350
Conversation
The 0.59.0 changelog scheduled the SKEEP-006 compatibility surface for ERROR in 0.60 and removal in 0.61. This release moves it: the `sk.ainet.io.tokenizer.*` aliases, `sk.ainet.io.model.ModelFormat`, `sk.ainet.io.weights.NameMap` / `RoleTableNameMap` / `TransformerNameMaps`, and in skainet-io-gguf `ModelArchitecture.fromGguf`, `ModelLoaderRegistry`, `LlamaLoaderDescriptor` and `Gemma3nLoaderDescriptor` are errors with their `ReplaceWith` pointing at the new homes. In-repo users (the GGUF and ONNX parsers, the io-core tokenizer tests, the GGUF name-map fixture test) import the new homes. `TokenizerFactory` and `UnsupportedTokenizerException` stay real classes for Java callers; `TransformerNameMaps.Gguf.forArchitecture` stays at WARNING.
|
📖 Documentation Preview The documentation has been built successfully for this PR. Generated Files:
Artifacts:
This comment will be updated automatically when the PR is updated. |
This was referenced Oct 16, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Release 0.60.0, the same sequence as 0.59.0 (#1347), with the scheduled SKEEP-006 deprecation step in front:
refactor(io)!: SKEEP-006 compatibility aliases to DeprecationLevel.ERROR— thesk.ainet.io.tokenizer.*aliases,sk.ainet.io.model.ModelFormat,sk.ainet.io.weights.NameMap/RoleTableNameMap/TransformerNameMaps, and the GGUFModelArchitecture.fromGguf,ModelLoaderRegistryand its two descriptors are errors since this release (removed in 0.61), as the 0.59.0 changelog scheduled. Every annotation keeps itsReplaceWith; in-repo tests moved to the new homes.TokenizerFactory/UnsupportedTokenizerExceptionstay real classes for Java callers;TransformerNameMaps.Gguf.forArchitecturestays at WARNING.docs(changelog): 0.60.0 entry— the[Unreleased]blocks from feat(ops)!: key-padding lengths and window metadata for scaledDotProductAttention (#1343, part 1) #1348 and feat(compile): n-ary composite seam and the skainet.sdpa composite (#1343, part 2) #1349 plus the deprecation step become one[0.60.0] - 2026-10-15entry with a headline and version note.docs(readme): 0.60.0 highlights and BOM coordinate— What's New, BOM0.60.0, contributors, BibTeX note.docs: bump version references to 0.60.0—CITATION.cff,docs/antora.yml, Android tutorial coordinates, the input-frontends reference page's alias schedule.release: 0.60.0—gradle.properties.docs: regenerate operator and kernel matrices at 0.60.0— version lines only.Contents since 0.59.0: #1348 (
scaledDotProductAttention(keyLengths, window, rightContext),AttentionMask, mask rules as op metadata, lengths-based StableHLO lowering, the batched dense-mask fix) and #1349 (n-ary composite seam,skainet.sdpacomposite), both parts of #1343.Verified against the consumer before cutting: SKaiNET-transformers built and tested against a local
0.60.0-SNAPSHOTof this develop head, with padded batches for the BERT and Laya encoders (#490) on top; the Laya multilingual reference fixtures reproduce through the batched path and batched vs single logits agree within 2e-6. That consumer change follows as the next SKaiNET-transformers release.Test plan
apiCheckgenerateDocs/generateKernelMatrixat 0.60.0 (only the version lines changed)workflow_dispatchonrelease/0.60.0) green before tagging0.60.0, tag-triggered publish,mainfast-forwarded, Central resolvessk.ainet:skainet-bom:0.60.0