This model is a compressed version of Qwen/Qwen3-Coder-Next. It is obtained by reducing the number of experts in each MoE layer from 512 to 384 using the REAP baseline method as described in https://bknyaz.github.io/blog/2026/moe/. Compared to other models obtained in this collection, more coding data is used in the calibration data during pruning/merging to better preserve original’s model coding abilities. Specifically, the ratio between c4, math and coding data (see https://bknyaz.github.io/blog/2026/moe/) is 0.0, 0.3, 0.7. The calibration data used here is the same as in Qwen3-Coder-Next-REAM. The compressed model has 60B params (120GB) instead of 80B (160GB) of the original model, reducing storage and GPU memory requirements by roughly 25%. At the same time, the model retains >=96% of the original model’s performance on a variety of benchmarks (see Results section below). Additional efficiency optimization (e.g., quantization) can be added similarly to the original model. Please refer to the license of the original model Qwen/Qwen3-Coder-Next.
Модальности:
Генерация текста
Области применения:
Генерация кода Диалог / чат
Задача: Генерация текста
Автор: SamsungSAILMontreal
Теги: qwen3_next, compression, expert-merging, moe, code, conversational, endpoints_compatible
Лайков: 3 | Загрузок: 35
Описание основано на материалах HuggingFace. Перевод выполнен автоматически.
