search

andyzeng / visual-pushing-grasping

Train robotic agents to learn to plan pushing and grasping actions for manipulation with deep reinforcement learning.
http://vpg.cs.princeton.edu/
BSD 2-Clause "Simplified" License
898 stars 315 forks source link

How to train your own dataset using the train.py file #34

Open TONG369 opened 5 years ago

TONG369 commented 5 years ago

Hello， Looked at your train.py file, the training data makes the model use the snapshot_file file. In addition, the files called in #def preload(self, transitions_directory): in the file are the results obtained after running the program. If I want to test with my own data, what needs to be corrected for this file? Thank you~ Looking for your reply,thankyou!!!

jayraj3 commented 5 years ago

If you want to train model using your using own objects, create new directory in "objects" containing your own objects. Then specify the "obj_mesh_dir" in argument

Refer the following python main.py --is_sim --obj_mesh_dir 'objects/YOUR-OBJECT-DIRECTORY' --num_obj 10 --push_rewards --experience_replay --explore_rate_decay --save_visualizations

TONG369 commented 5 years ago

I am very happy to receive your email,thanks your reply.Good luck for you!!!

At 2019-08-12 16:26:45, "Jayraj Parmar" notifications@github.com wrote:

If you want to train model using your using own objects, create new directory in "objects" containing your own objects. Then specify the "obj_mesh_dir" in argument

Refer the following python main.py --is_sim --obj_mesh_dir 'objects/YOUR-OBJECT-DIRECTORY' --num_obj 10 --push_rewards --experience_replay --explore_rate_decay --save_visualizations

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

sevenandseven commented 4 years ago

excuse.i want to know how to run this code,and the simulate is our simulate plane?can you tell me ,thankyou.