fix(ppo): 修正日志概率维度与状态张量格式
修复 replay buffer 中 log_probs 的维度错误,从 (buffer_size, action_dim) 改为 buffer_size 修正训练时状态张量格式,从 (N, H, W, C) 转换为 (N, C, H, W) 更新 collect_rollout 返回观测值并修正 log_prob 计算 添加项目配置文件和训练曲线生成脚本
This commit is contained in:
Binary file not shown.
|
After Width: | Height: | Size: 127 KiB |
Reference in New Issue
Block a user