mlx-community/Muse-Glimmer-30B-OptiQ-4bit

Update kv_config.json: re-measured with `optiq kv-cache` float32 sensitivity (the previous config was ranked on bfloat16 rounding noise). 8-bit layers [3, 5, 14, 16, 21, 30, 39] -> [7, 8, 14, 17, 19, 21, 23].

Add kv_config.json: per-layer KV cache precision from `optiq kv-cache` (target 4.5 bits). Use with `optiq serve --kv-config kv_config.json`.

Add files using upload-large-folder tool

initial commit