chore(deps): bump skainet to 0.23.0#101
Merged
michalharakal merged 1 commit intodevelopfrom May 2, 2026
Merged
Conversation
Picks up: - The K/N `createRandomAccessSource` POSIX-pread implementation (SKaiNET PR #591), so K/N consumers can now load GGUFs above the ~2 GiB Kotlin `ByteArray` ceiling. Combined with this repo's existing `randomAccessProvider` factory overloads on `*NetworkLoader.fromGguf(...)` and the `fromRandomAccessSource(...)` variant on `GGUFTokenizer`, downstream native CLIs / consumers can switch to the streaming path mechanically. - Lazy zero-init for parameter placeholders (SKaiNET PR #588), which ends real-model load-time OOMs in the DSL builders that this repo's loaders sit on top of. No source changes — purely a version pin bump. Verified by compiling :llm-runtime:kllama (JVM + macosArm64), :llm-agent (JVM), :llm-inference:llama (JVM), :llm-inference:qwen (JVM) against the new dep set. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Picks up two SKaiNET 0.23.0 changes that benefit this repo's consumers without any source changes here:
createRandomAccessSourcenow has a POSIX-preadimplementation (SKaiNET#591). K/N consumers can now load GGUFs above the ~2 GiB KotlinByteArrayceiling. TherandomAccessProvideroverloads on*NetworkLoader.fromGguf(...)and thefromRandomAccessSource(...)variant onGGUFTokenizeralready exist in this repo — they just couldn't actually open a file on macOS / Linux / iOS native targets before 0.23.0. Verified end-to-end againstQwen3-1.7B-Q8_0.gguf(1.8 GiB) on macOS arm64: previously OOMed at construction; now loads.Test plan
./gradlew :llm-runtime:kllama:compileKotlinJvm :llm-runtime:kllama:compileKotlinMacosArm64./gradlew :llm-agent:compileKotlinJvm./gradlew :llm-inference:llama:compileKotlinJvm :llm-inference:qwen:compileKotlinJvmAll clean (only pre-existing
LlamaRuntimedeprecation warnings, unrelated). No source changes — pure version bump.Follow-ups (separate PRs)
kllamaetc.) fromLlamaIngestion.load { source }+GGUFTokenizer.fromSource(...)to the streaming siblings (loadStreaming(randomAccessProvider),fromRandomAccessSource(...)) so they can actually open larger GGUFs on K/N.