ToPo-ToPo/Qwen3.6-27B-mlx-8bit

MLX 8bit conversion of Qwen/Qwen3.6-27B for Apple Silicon (mlx-vlm).

Provenance (self-converted from official weights)

• Source: Qwen/Qwen3.6-27B (license: apache-2.0)
• Tool: mlx-vlm 0.6.3 — mlx_vlm.convert --hf-path Qwen/Qwen3.6-27B --mlx-path . -q --q-bits 8 --q-group-size 64
• Effective: 8.627 bits/weight
• Validation: reproduced geometrically exact CAD output in an agentic CAD+FEM pipeline (volumes match the reference mlx-community conversion).

Usage

from mlx_vlm import load, generate
model, processor = load("ToPo-ToPo/Qwen3.6-27B-mlx-8bit")