Generalist GEN-1.5：单次演示成功率59%，少样本达83%
Experiments across 10 diverse tasks show 59% average success with one-shot physical prompting, straight from pretraining. With few-shot learning, performance ri
Generalist 的 GEN-1.5 在 10 项多样化物理任务上，仅凭预训练后的单次物理提示（one-shot physical prompting）平均成功率为 59%；经少样本学习，每任务用 5 分钟数据做 10 步梯度更新后升至 83%。该模型可在 1-10 步梯度、1-5 分钟数据（约 10-50 次演示）内适配新物理任务，且上下文学习可零样本跨过 sim-to-real 差距。GEN-1.5 已连续训练超过 8 个月。
正文
当前显示原文，中文全文译文尚未就绪。
Experiments across 10 diverse tasks show 59% average success with one-shot physical prompting, straight from pretraining. With few-shot learning, performance rises to 83% via 10 gradient steps on 5 minutes of data per task.
Although the tasks are simple and success rates are modest, it’s the first model we know of that exhibits the general ability to learn a wide range of dexterous closed-loop physical tasks from just one or few demonstrations. This accelerates reaching a base level of competence for new skills that can be subsequently refined towards mastery.