
Connecting to preview…正在连接预览…
Training return训练回报
Training iteration训练轮次
Movement动作表现
Training iteration训练轮次
Reward components分项奖励
Training iteration训练轮次
Showing the last available samples正在显示上次读取的样本
MakerDuck assistantMakerDuck 助手
Training school训练学校
AI is not configured yetAI 尚未配置
An administrator can configure the model in AI assistant settings.管理员可在 AI 助手设置中配置模型。
AI assistant settings打开 AI 助手设置Device and logs设备与日志No device selected尚未选择设备
Connect your duck in setup or the debug workbench, then read its current state here.在初始化或调试工作台连接鸭子后,可在这里读取当前状态。
Import a log from this computer从电脑导入日志JSON, up to 8 MB and 20,000 records. The selected log is sent to the website assistant for analysis.JSON 格式,最大 8 MB、20,000 条记录。所选日志会发送到网站助手进行分析。
Log summary日志摘要
Choose a training record and describe what you want the duck to do differently.选一条训练记录,告诉助手你想让鸭子的动作有什么变化。
Review training plan核对训练方案
| Parameter参数 | Previous上次 | Proposed本次 |
|---|
Apply these settings, then start the run in Training School.应用参数后,可在训练学校左侧开始训练。
Stop this training run?停止这次训练?
Latest saved model最近保存的模型 ·
The run will end and its latest complete checkpoint will be kept.本轮会结束,保留最近一次完整保存的模型。