DiffuSpace’s Clover-35B-A3B-Preview (a block-diffusion MoE tuned from Qwen3.6-35B-A3B, 3B active per token) quantized for Kapitsa, the Metal inference engine for Apple silicon. 19.5 GB. Kapitsa downloads this pack with Qwen3.6-35B-A3B’s DFlash2 drafter (incoai/Qwen3.6-35B-A3B-DFlash2) beside it. — MLX affine 4-bit, group 64, for the experts, attention, DeltaNet, embedding and output layer; 8-bit router and shared expert gate; norms, the DeltaNet’s Alog, dtbias and convolutions, and Clover’s own right convolution and softmax controls kept in bf16 as stored. Made from the bf16 checkpoint by tools/convertclover.py in Kapitsa’s repository; conversion.json lists every file’s sha256. — Tensor names under languagemodel. as in MLX’s Qwen3.6 packs. mlx-lm has no clovermoe model, so this pack is for Kapitsa. — config.json, generationconfig.json (Clover’s repetition, length and controls), the tokenizer and chat template, LICENSE and ATTRIBUTIONS.md as DiffuSpace ships them. Prefill: 6,002 tok/s at 4K tokens. Measured with kapitsa bench on a MacBook Pro M5 Max (128 GB), Kapitsa 0.2.24. Apache-2.0, as Clover and Qwen3.6-35B-A3B. Clover is DiffuSpace’s work, derived from Qwen3.6-35B-A3B…
Модальности:
Генерация текста
Области применения:
Диалог / чат
Задача: Генерация текста
Автор: kapitsaprime
Теги: clover_moe, clover, kapitsa, apple-silicon, mixture-of-experts, diffusion, 4-bit, conversational
Лайков: 3 | Загрузок: 874
Описание основано на материалах HuggingFace. Перевод выполнен автоматически.