Perceptron 开源机器人基础模型 Isaac 0.5
Brilliant open-source robotics foundation-model release from @perceptroninc .
做机器人或具身智能的同学必看，Isaac 0.5 用无动作视频把数据成本打下来，微调即可上手，赶紧去试。
Brilliant open-source robotics foundation-model release from @perceptroninc .
来自@perceptroninc的出色开源机器人基础模型发布。
Isaac 0.5 cuts teleoperation needs by 210X by scaling action-free video.
Isaac 0.5通过扩展无动作视频，将遥操作需求削减了210倍。
The most expensive thing in robotics is still collecting robot data, and this model attacks that problem.
机器人领域最昂贵的部分仍然是收集机器人数据，而这个模型正是针对这一难题。
Isaac 0.5 is just a 36B param sparse model that can take in camera images/video, a human instruction, the robot's current state, and then produce things such as object locations, task progress, visual reasoning, or the robot's next actions.
Isaac 0.5是一个仅36B参数的稀疏模型，能够接收摄像头图像/视频、人类指令、机器人当前状态，并输出物体位置、任务进度、视觉推理或机器人的下一步动作。
So if you have a robot arm, you can take Isaac 0.5, fine-tune it on demonstrations from that arm, and use the resulting policy to control the robot. The same underlying model can also be used only for perception, feeding its visual understanding into another planner or controller.
因此，如果你有机械臂，可以拿Isaac 0.5，在该机械臂的演示数据上进行微调，并用得到的策略来控制机器人。同一个底层模型也可以仅用于感知，将其视觉理解输入到另一个规划器或控制器中。
Perceptron reports faster one-demonstration adaptation: one chess episode cut held-out action loss 7.0× to 10.5×, versus 2.3× to 3.1× for π0.5.
Perceptron报告了更快的单次演示适应能力：一个国际象棋片段将保留动作损失降低了7.0倍至10.5倍，而π0.5则为2.3倍至3.1倍。
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力