dljzx

May I ask how much loss and val_loss have you got? I have got a val_loss of 3.736 which I thought is not good enough.

Bad caption

1

I down load your weight and model in order reduce time of training. But when I test it Generated Caption was totally bad. For example I tested beach.jpg and described...

bad result on Antmaze enviornment

1

It works well on mujoco environments, but not on antmze environment .It did not work even if I changed the parameters according to the paper(expectile=0.9, temperature=10). Can you help me...

Loss json in download.sh

6

captions ./data/MSCOCO/MSCOCO/mscoco_caption_train2014_cn_processed.json file is not downloaded by the shell command file.Or it could be instead by captions_train2014_cn_translation.json ?

Thanks for your work. Could your code behave well in Antmaze environments?

parameter initialize error for clip model

1

Thanks for your code. While Implementing clip rl Model by changing TRANSFER_TYPE to "clip_vit_b32", code faced an error like this: Traceback (most recent call last): File "/nvme/.../anaconda3/envs/instructrl/lib/python3.9/runpy.py", line 197, in...

Action loss is not declining on LISA training

9

Thanks for your work! While i was training LISA model on lorel dataset with 50000 episodes, I found out that the action loss did not efficiently decline while commitment loss...