Repository navigation
Update llama.swift to llama.cpp 1.0.0 - #428
Merged
Merged
Conversation
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This PR moves the
Llamatrait to llama.swift 3.0.0, which wraps llama.cpp v1.0.0 (nightlyb12359). The release removed the deprecated C API and added a trailingmtmd_helper_init_optargument tomtmd_helper_bitmap_init_from_buf, whichLlamaLanguageModelcalls to decode image input. The backend now passes the default options; they only configure video decoding, and the backend rejects video input.One behavior changes on macOS: with ffmpeg on
PATH, llama.cpp now decodes WebP image data as an image, where the previous build treated it as video and the backend threwunsupportedFeature.Package.resolveddoesn't change, because it has no pin for llama.swift, which only theLlamatrait uses.The Swift 6.4 CI jobs now also remove
SwiftExplicitPrecompiledModulesafter restoring the cache: the new XCFramework's headers keep their release's timestamps, so Swift Build reused the module built from the old headers and failed to compileLlamaLanguageModel.