An Apple Silicon (MLX) checkpoint of Muse Glimmer, produced as a code-preserving repack of the vendor’s searched dynamic K-quant GGUF muse-glimmer-30B-kquant-dynamic.gguf (per-layer mixed 4/5/6-bit widths chosen by the vendor’s quantization search). Every weight keeps the GGUF’s exact quantization code. The highest-fidelity Muse Glimmer MLX variant. Serve it with SGLang’s MLX backend on a Mac with 48 GB of unified memory or more. Text only — the MLX backend has no vision path. — Source: the vendor GGUF release → ggufweights/muse-glimmer-30B-kquant-dynamic.gguf (sha256 513109c8…), revision aa336138… (the 2026-08-10 «definitive Muse Glimmer 30B GGUFs» re-publish; its quantized payload is byte-identical to the earlier the previous vendor GGUF — this repack’s tensors are bit-identical to the previous revision of this repo, so all evaluation results below are unchanged) — Format: MLX-quantized safetensors, 5 shards (~20 GB); tokenizer / config / chat te
Модальности:
Генерация текста
Области применения:
Диалог / чат
Задача: Генерация текста
Автор: RadixArk
Теги: mlx, muse_glimmer, apple-silicon, sglang, quantized, gguf-mirror, muse-glimmer, conversational
Лайков: 3 | Загрузок: 317
Описание основано на материалах HuggingFace. Перевод выполнен автоматически.
