From the source
From the source
A 7B model excelling in both general and spatial reasoning
Ant Ling released M2-Reasoning-7B, a multimodal large language model (MLLM) designed for both general and spatial reasoning.
The model uses a data pipeline generating 294.2K high-quality samples and a dynamic multi-task training strategy with step-wise optimization and task-specific rewards.
It sets new state-of-the-art results across eight benchmarks in general and spatial reasoning domains.
From the source
We introduce M2-Reasoning-7B, a model designed to excel in both general and spatial reasoning. Our approach integrates two key innovations: (1) a novel data pipeline that generates 294.2K high-quality data samples (168K for cold-start fine-tuning and 126.2K for RLVR), which feature logically coherent reasoning trajectories and have undergone comprehensive assessment; and (2) a dynamic multi-task training strategy with step-wise optimization to mitigate conflicts between data, and task-specific rewards for delivering tailored incentive signals.
developer.ant-ling.com
