Rofunc: The Full Process Python Package for Robot Learning from Demonstration and Robot Manipulation
A pre-trained soft object manipulation skill learning model, namely SoftGPT, that is trained using large amounts of exploration data, consisting of a three-dimensional heterogeneous graph representation and a GPT-based dynamics model.
pip install rofunc
import rofunc as rf
import numpy as np
from isaacgym import gymutil
from importlib_resources import files
# Demo
raw_demo_l = np.load(files('rofunc.data.RAW_DEMO').joinpath('taichi_raw_l.npy'))
raw_demo_r = np.load(files('rofunc.data.RAW_DEMO').joinpath('taichi_raw_r.npy'))
demos_x_l = [raw_demo_l[300:435, :], raw_demo_l[435:570, :], raw_demo_l[570:705, :]]
demos_x_r = [raw_demo_r[300:435, :], raw_demo_r[435:570, :], raw_demo_r[570:705, :]]
rf.lqt.plot_3d_bi(demos_x_l, demos_x_r, ori=False, save=False)
# TP-GMM
show_demo_idx = 1
_, _, gmm_rep_l, gmm_rep_r = rf.tpgmm.bi(demos_x_l, demos_x_r, show_demo_idx=show_demo_idx, plot=True)
🖥 Github: https://github.com/skylark0924/rofunc
A pre-trained soft object manipulation skill learning model, namely SoftGPT, that is trained using large amounts of exploration data, consisting of a three-dimensional heterogeneous graph representation and a GPT-based dynamics model.
pip install rofunc
import rofunc as rf
import numpy as np
from isaacgym import gymutil
from importlib_resources import files
# Demo
raw_demo_l = np.load(files('rofunc.data.RAW_DEMO').joinpath('taichi_raw_l.npy'))
raw_demo_r = np.load(files('rofunc.data.RAW_DEMO').joinpath('taichi_raw_r.npy'))
demos_x_l = [raw_demo_l[300:435, :], raw_demo_l[435:570, :], raw_demo_l[570:705, :]]
demos_x_r = [raw_demo_r[300:435, :], raw_demo_r[435:570, :], raw_demo_r[570:705, :]]
rf.lqt.plot_3d_bi(demos_x_l, demos_x_r, ori=False, save=False)
# TP-GMM
show_demo_idx = 1
_, _, gmm_rep_l, gmm_rep_r = rf.tpgmm.bi(demos_x_l, demos_x_r, show_demo_idx=show_demo_idx, plot=True)
🖥 Github: https://github.com/skylark0924/rofunc