R2R数据集中的obj特征

CrystalSixone / DSRG

Code for A Dual Semantic-Aware Recurrent Global-Adaptive Network For Vision-and-Language Navigation

MIT License

15 stars 0 forks source link

R2R数据集中的obj特征 #8

Closed jasscia18 closed 6 months ago

jasscia18 commented 6 months ago

你好，请问在R2R数据集中，是没有用到obj特征吗？在代码vilmodel.py中的ImageEmbeddings函数中，发现bs, view_len, obj_len = traj_obj_img_fts.shape[0], traj_obj_img_fts.shape[1], traj_obj_img_fts.shape[2] 的，traj_obj_img_fts特征为None。

CrystalSixone commented 6 months ago

Thank you for the reminder! The error occurred due to the cleaning code not properly updating the parameter names, an issue that I have now rectified.