simpler-env
How to use my own dataset in evaluating a finetuned octo model
Thanks for your great work!
When I wanted to use simpler env to evaluate my own finetuned octo model with my data, I found some missing parameters: robot_init_x, robot_init_y, object_init_x, and object_init_y. Are these parameters determined by my own data? Or could you please tell me how can I define them?
Moreover, how should I define the dataset_id parameter in my own way? And it seems that there is no explicit place for me to modify the code to load my own dataset?
Looking forward to your reply!
Thank you very much.
The robot and object poses (parameters) should be the same regardless of the model you evaluate. You can do custom eval by modifying the robot and object poses in eval script. See scripts in https://github.com/simpler-env/SimplerEnv/blob/main/scripts/
