MLX Studio — the only app that natively supports JANG models with reasoning Nemotron-Cascade-2-30B-A3B — JANG_2L (2.3-bit, 8-bit attention) — Reasoning JANG — Jang Adaptive N-bit Grading | The GGUF Equivalent for MLX JANG2L is the only quantization that fits 16 GB Macs while delivering 88% MMLU with reasoning. JANG4M beats MLX 4-bit (93.0% vs 92.5%) at the same 17 GB size. Also see: JANG_4M (17 GB) — same size as MLX 4-bit, higher reasoning score. — Apple Silicon Mac with 16+ GB unified memory — MLX Studio or pip install «jang[mlx]>=2.1.5» — Mamba-2 SSM: Most layers use state-space models, enabling efficient long-context with minimal KV cache. — Only 6 attention layers: KV cache is tiny (0.2 GB at 32K). Most models use 25-100% attention layers. — nemotronh architecture: Requires JANG loader for proper weight mapping. Standard mlx-lm has incomplete support. — IMO Gold Medal**: This model achieves competition-level mathematical reasoning at 30B scale. JANG — Created by Jinho Jang (eric@jangq.ai) · @dealignai GitHub · PyPI · HuggingFace
Модальности:
Генерация текста
Области применения:
Логика и рассуждение Диалог / чат
Задача: Генерация текста
Автор: JANGQ-AI
Теги: mlx, nemotron_h, jang, quantized, mixed-precision, apple-silicon, reasoning, thinking
Лайков: 3 | Загрузок: 150
Описание основано на материалах HuggingFace. Перевод выполнен автоматически.
