majentik/Qwen3.6-35B-A3B-TurboQuant-MLX-MXFP4 - Каталог нейросетей
Генерация текста

majentik/Qwen3.6-35B-A3B-TurboQuant-MLX-MXFP4

Добавлено:
majentik/Qwen3.6-35B-A3B-TurboQuant-MLX-MXFP4

TurboQuant + MLX-MXFP4 (4-bit) variant of Qwen/Qwen3.6-35B-A3B. Apple Silicon (M1/M2/M3/M4) with TurboQuant structural pre-conditioning and MLX-native MXFP4 layout (E2M1 weights, per-32-element E8M0 (OCP microscaling)). 4.253 bits/weight, ~17 GB on disk, sub-2-s load on M4 Max. Pick this over the affine MLX variants when you want MXFP4 format parity with hardware pipelines while running locally. > [!IMPORTANT] > Text-only extraction. This repo contains only the quantized language tower of > Qwen3.6-35B-A3B. The upstream vision tower (333 tensors) and MTP head are not included, > so image/video input does not work and mlxvlm.load(…) fails with a > Missing … parameters error (the vision tower it expects is absent from the checkpoint). > Load it with mlxlm (recent version with qwen35moe support) as shown above. > For image/video input, use the upstream BF16 model > Qwen/Qwen3.6-35B-A3B on a runtime that supports it. — bf16 — Qwen/Qwen3.6-35B-A3B — FP8 card — majentik/Qwen3.6-35B-A3B-FP8 — TurboQuant MLX-4bit (affine) — majentik/Qwen3.6-35B-A3B-TurboQuant-MLX-4bit — plain MLX-MXFP4 (no rotor/turbo) — majentik/Qwen3.6-35B-A3B-MLX-MXFP4 — Calibration hash: none (mxfp4 is…

Модальности:
Генерация текста

Области применения:
Диалог / чат


Задача: Генерация текста
Автор: majentik
Теги: mlx, qwen3_5_moe, qwen, qwen-3.6, moe, turbo, mxfp4, apple-silicon
Лайков: 4  |  Загрузок: 589

Открыть на HuggingFace →

Описание основано на материалах HuggingFace. Перевод выполнен автоматически.