Repository navigation
chore(kernelrace): SKaiNET 0.60.0 / transformers 0.61.0 - #40
Merged
Merged
Conversation
The vendored QuantPolicy is gone from transformers; the loaders' default WeightForm (null) keeps quantized tensors packed and mapped, which is what NATIVE_OPTIMIZED meant here. DecoderGgufWeightLoader moved to llm-core and Tokenizer.decode(Int) became decodeToken(Int); both deprecations fixed so the sample is ready for 0.62.
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Moves KernelRace from SKaiNET 0.40.1 / transformers 0.40.2 to SKaiNET 0.60.0 / transformers 0.61.0.
Source changes, all in
shared:QuantPolicyno longer exists in transformers. The loaders' defaultWeightForm(null) keeps quantized tensors packed in their stored block encoding with mapped residency, which is whatNATIVE_OPTIMIZEDselected here, so the argument is simply dropped in the common fallback and in the JVM, Android and iOS builders. Comments updated.DecoderGgufWeightLoaderis imported from its new homesk.ainet.lang.nn.dsl.decoder(thesk.ainet.models.llamaalias is deprecated and goes away).Tokenizer.decode(Int)becamedecodeToken(Int)(SKEEP-006); the two call sites inLlmEngineare switched so the sample compiles without warnings and stays ready for 0.62, where that legacy surface becomes an error.Verified against Maven Central:
:shared:jvmTest(25 tests),:shared:testAndroidHostTest,:composeApp:compileKotlinJvm, and the wasm compiles ofshared,composeAppandwebWorkerare green with no deprecation warnings.:androidApp:assembleDebugis left to CI. The iOS builder received the same mechanical change but cannot be compiled on this machine.