Thanks for this great contribution! I was able to train my own model from scratch for image subgoal synthesis. However, as my dataset is in sim and much smaller, the performance could be improved. Do you have any recommendations on how to proceed if I were to finetune your checkpoint rather than instructpix2pix?
Thanks for this great contribution! I was able to train my own model from scratch for image subgoal synthesis. However, as my dataset is in sim and much smaller, the performance could be improved. Do you have any recommendations on how to proceed if I were to finetune your checkpoint rather than instructpix2pix?