File size: 2,088 Bytes
e5214b3 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 | """
归一化常量 - point2_v30
从 data/each-point-V30/point2_v30/meta/stats.json 自动生成
"""
import numpy as np
# ====== 归一化统计量（从 point2_v30 数据集提取）======
# State 归一化参数 (observation.state) - 14D
STATE_MEAN = np.array([
-0.2105864585016499,
1.3310560174962338,
-1.1387206377712409,
0.06850543825862183,
0.7681250187706506,
-0.08478798886625565,
0.05835128650106446,
0.02370642884928235,
-0.031137483014934606,
-0.15235544200127704,
-0.006249059893535669,
0.5941734570462274,
-0.05018620472882003,
7.000005561990393e-05
], dtype=np.float32)
STATE_STD = np.array([
0.2400051179695121,
0.9237869478542775,
0.7118947011828334,
0.08144195556664052,
0.23071447286401414,
0.179722397158341,
0.03285219262761961,
4.953138735344596e-10,
8.388984945120932e-05,
0.0004753077244397484,
0.0008577420475323502,
0.002150086383523521,
0.00010284846023543875,
3.132144638762018e-08
], dtype=np.float32)
# Action 反归一化参数 - 14D (left_arm 7D + right_arm 7D)
ACTION_MEAN = np.array([
-0.2109139475252697,
1.3108079862689859,
-1.1673434242464766,
0.07370919146763717,
0.7044337684964903,
-0.0844366189408058,
0.057783793058325136,
0.02365400559923624,
-0.03171325462574329,
-0.17187520877501722,
-0.005509715998848856,
0.5043942625478558,
-0.05123314912850634,
-0.02110003694254407
], dtype=np.float32)
ACTION_STD = np.array([
0.24125223557585568,
0.9046525674814566,
0.7146504447920251,
0.09142734038550057,
0.25411883130066965,
0.18318653299101498,
0.037955622978664776,
2.7160945227379783e-05,
5.362838404012091e-06,
0.0004134820247080275,
0.001425757586551424,
2.979843662272417e-08,
9.064657920470249e-09,
1.5702117327601945e-05
], dtype=np.float32)
# Image 归一化参数（ImageNet 标准）
IMAGE_MEAN = np.array([0.485, 0.456, 0.406], dtype=np.float32)
IMAGE_STD = np.array([0.229, 0.224, 0.225], dtype=np.float32)
|