The latest llama.cpp release adds Metis-1.6 model support and offers downloadable builds for several operating systems and processor types.
NoteBy OossaPublished by Oossa: 1 min read
The ggml-org team’s llama.cpp release b12089 adds support for Metis-1.6 models. It also provides downloadable builds for platforms including macOS, iOS and Linux, with some Linux builds tailored to different processors. The release notes do not specify which Metis-1.6 models are supported or describe performance changes.
People running Metis-1.6 models may be able to use them with llama.cpp, but the release notes do not say which models or configurations are covered.
Models
OpenAI says its model helped design a talk-to-Codex device, from circuit boards to firmware. Prototypes still needed several rounds of revisions.
Models
An arXiv study finds several single-pass models match or beat some text-generating judges on evaluation benchmarks, but results vary by task and model.
Models
OpenAI says ChatGPT handles one-fifth as many search queries as Google, using its own definition of a search. The company did not explain how it got Google’s figure.
Oossa · Newsletter
The week in AI, explained
Every Monday: the stories worth knowing, in plain language. Free, no spam.
