-
Notifications
You must be signed in to change notification settings - Fork 4
/
Copy pathconfig_model3.py
415 lines (396 loc) · 26.3 KB
/
config_model3.py
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
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
import socket
import numpy as np
import tensorflow as tf
def value(dict, k):
for key in dict:
if k == key or k in key:
return dict[key]
raise NotImplementedError
def n(c, seed):
if c.env_type == 'mujoco':
return 'rrw-%s-%s-bs=%s-c=%s-%s-up' \
'=%s-%s-%s-%s-bs=%s-%s-' \
'ni=%s-%s-ld=%s-max=%s-mlr-%s' \
'-%s-k=%s-gl=%s-cw=%s-sf=%s-tf=%s-mv=%s-rt=%s-sl=%s-sof' \
'=%s-se=%s-om=%s-ol=%s-gr=%s-%s-pos=%s-lr=%s-%s-e=%s-mt=%s' \
'%s%s-w=%s%s%s-%s%s-%sx%s-%s-%s%s-%s' % (
c.xml_name, c.algo, c.bs_per_cpu, c.num_cpus, c.short[c.hostname], c.update_master_interval,
c.pg_pi_hidden_fc_size, c.pg_master_hidden_fc_size, c.pg_bl_hidden_fc_size, c.mini_bs, c.optim_epochs,
c.Tnoise, c.Fnoise, c.short[c.restore_model], c.max_num_timesteps, c.pg_master_pg_lr, c.stop_pg_ts_perc,
c.num_models, str(c.goal).replace(' ', '_'), c.cor_w, c.safe_m_loss, c.short[c.transfer], c.mov_avg,
c.short[c.reset], c.solved_threshold, c.short[c.soft], str(seed), c.short[c.oracle_master],
c.short[c.old_policy], c.goal_r, c.forward_mult, c.short[c.pos_r], c.pg_master_ce_lr_source,
c.pg_master_ce_lr_target, c.short[c.enforced], c.short[c.math], c.short[c.stable_old], c.short[c.l2],
c.short[c.weighted], c.short[c.oracle_master], c.short[c.always_restore],
c.short[c.learn_model_primitive], c.short[c.use_model_primitive], c.pg_mp_n_layers,
c.pg_mp_hidden_fc_size, c.pg_mp_lr, c.short[c.model_primitive_delta], c.short[c.model_primitive_pg],
int(c.success_dist))
elif c.env_type == 'dm_control':
return '13-%s-%s-i=%s-bs=%s-c=%s-%s-up' \
'=%s-%s-%s-%s-en=%s-bs=%s-%s-' \
'ni=%s-%s-l=%s-m=%s-lr-%s' \
'-%s-%s-k=%s-sf=%s-tf=%s-mv=%s-rt=%s-%s-sf=%s' \
'-se=%s-om=%s-ol=%s-lr=%s-%s-e=%s-di=%s-st=%s-b=%s-%s-mh' \
'=%s-ce=%s%s%s-at=%s-mt=%s%s%s-ob=%s%s-%s%s%s%s' % (
c.xml_name, c.algo, c.num_batches, c.bs_per_cpu, c.num_cpus, c.short[c.hostname],
c.update_master_interval, c.pg_pi_hidden_fc_size, c.pg_master_hidden_fc_size, c.pg_bl_hidden_fc_size,
c.ent_coeff, c.mini_bs, c.optim_epochs, c.Tnoise, c.Fnoise, c.short[c.restore_model],
c.max_num_timesteps, c.pg_master_pg_lr, c.stop_pg_ts_perc, c.pg_sub_lr, c.num_models, c.safe_m_loss,
c.short[c.transfer], c.mov_avg, c.short[c.reset], c.solved_threshold, c.short[c.soft], str(seed),
c.short[c.oracle_master], c.short[c.old_policy], c.pg_master_ce_lr_source, c.pg_master_ce_lr_target,
c.short[c.enforced], c.short[c.dist_obv], c.short[c.stage_obv], c.short[c.bounded], c.dist_bound,
c.short[c.manhattan], c.bring_close, c.drop_close, c.drop_width, c.above_target, c.short[c.math],
c.short[c.stable_old], c.short[c.l2], c.short[c.obstacle], c.obstacle_height, c.short[c.redundant],
c.short[c.split], c.repeat, c.short[c.weighted])
class c(object):
task_id = 0
# NOTE: RECORDING
xml_name = 'T3-Ant-v2'
env_name = 'Basic%s' % (xml_name[xml_name.find('-'):])
cor_w = value({'Basic-Humanoid-v2': 7, 'Basic-Ant-v2': 1.5, ('Basic-bring_ball-v0', 'Basic-stack_2-v0'): 1.0},
env_name)
cor_wd = {'E': cor_w * 2, 'N': cor_w * 2, 'W': cor_w * 0.85, 'S': cor_w * 0.85, 'U': None}
paths = np.array(value({'T0-Ant-v2': [[[[0, -8], [0, 0], [8, 0]], [[0, 8], [0, 0], [8, 0]], [[0, 0], [8, 0]]]],
'T1-Ant-v2': [[[[8, 0], [0, 0], [0, 8]], [[-8, 0], [0, 0], [0, 8]], [[0, 0], [0, 8]]]],
'T2-Ant-v2': [[[[0, -8], [0, 0], [-8, 0]], [[0, 8], [0, 0], [-8, 0]], [[0, 0], [-8, 0]]]],
'T3-Ant-v2': [[[[8, 0], [0, 0], [-0, -8]], [[-8, 0], [0, 0], [0, -8]], [[0, 0], [0, -8]]]],
('TS-Ant-v2', 'BS-Ant-v2'): [[[0, 0], [12, 0], [12, 12], [-6, 12], [-6, 0]]],
('BL-Ant-v2', 'TL-Ant-v2', 'TL-Humanoid-v2', 'BL-Humanoid-v2'): [
[[0, 0], [12, 0], [12, 12]]],
('TT-Ant-v2', 'LL-Ant-v2'): [[[0, 0], [12, 0], [12, 12], [24, 12], [24, 24]]], (
'TMaze-Ant-v2', 'RMaze-Ant-v2', 'BMaze-Ant-v2', 'CMaze-Ant-v2', 'VMaze-Ant-v2',
'2Maze-Ant-v2', '5Maze-Ant-v2'): [[[[0, 0], [8, 0], [8, -8], [16, -8], [16, 0], [24, 0]]],
[[[0, 0], [0, 8], [8, 8], [8, 0], [16, 0], [16, -8]]],
[[[0, 0], [-8, 0], [-8, 8], [0, 8], [0, 16], [8, 16]]], [
[[0, 0], [0, -8], [-8, -8], [-8, -16], [-16, -16],
[-16, -8]]], [
[[0, 0], [8, 0], [8, -8], [16, -8], [16, -16],
[24, -16]]], [
[[0, 0], [0, 8], [-8, 8], [-8, 16], [-16, 16],
[-16, 24]]],
[[[0, 0], [8, 0], [8, 8], [0, 8], [0, 16], [-8, 16]]],
[[[0, 0], [0, 8], [8, 8], [8, 16], [0, 16], [0, 24]]],
[[[0, 0], [8, 0], [8, -8], [0, -8], [0, -16], [8, -16]]],
[[[0, 0], [0, -8], [-8, -8], [-8, -16], [-16, -16],
[-16, -24]]]],
# [[[[0, -0.25], [6, -0.25], [6, -4], [13, -4], [13, 3], [6, 3]]],
# [[[0, 0], [0, 6], [6, 6], [6, 0], [10, 0], [10, -5]]],
# [[[0, -0.25], [-5, -0.25], [-5, 7], [1, 7], [1, 3], [8, 3]]],
# [[[0, 0], [0, -5], [-5, -5], [-5, -10], [-11, -10], [-11, -5]]],
# [[[0, 0], [7, 0], [7, -7], [11, -7], [11, -13], [17, -13]]],
# [[[0, 0], [0, 5], [-6, 5], [-6, 12], [-13, 12], [-13, 18]]],
# [[[0, 0], [5, 0], [5, 4], [-2, 4], [-2, 9], [-9, 9]]],
# [[[0, 0], [0, 7], [7, 7], [7, 14], [1, 14], [1, 19]]],
# [[[0, 0], [6, 0], [6, -4], [1, -4], [1, -10], [5, -10]]],
# [[[0, 0], [0, -7], [-7, -7], [-7, -11], [-12, -11], [-12, -18]]]],
('SP-bring_ball-v0', 'SP-stack_2-v0', 'V-stack_2-v0', 'A-stack_2-v0', '2-stack_2-v0'): [
[[0, 0], [1, 0]]],
('SPll-stack_2-v0', 'All-stack_2-v0', '2ll-stack_2-v0', 'Vll-stack_2-v0'): [
[[0, 0], [1, 0]], [[1, 1], [0, 0], [0, 1]], [[1, 0], [1, 1]], [[0, 0], [0, 1], [1, 1]],
[[0, 1], [1, 1], [1, 0]], [[0, 1], [0, 0], [0, 1]], [[1, 0], [0, 0]],
[[1, 1], [0, 1], [0, 0]], ]}, xml_name))
env_type = 'mujoco' if xml_name.endswith('v2') else 'dm_control'
ckpt_path = 'rr-TMaze-Ant-v2-vppo-bs=2048-c=16-tv-up=1-16-64-64-bs=4096-10-ni=0-0.5-ld=F-max=2048-mlr-0-1-k=4-gl' \
'=None-cw=1.5-sf=20.0-tf=T-mv=10-rt=T-sl=0.7-sof=T-se=0-om=F-ol=F-gr=200-50-pos=T-lr=0.001-0.003-e=T' \
'-mt=FFF-w=FFF-FF-2x64-0.0003-TF-3/44'
math = False
stable_old = math
l2 = False
# NOTE: DM_CONTROL
obstacle = True
obstacle_height = 0.09
repeat = 10
redundant = True
bring_close = 1.5
drop_close = 1.5
drop_width = 1.5
split = False
learn_model_primitive = value(
{('T0-Ant-v2', 'T1-Ant-v2', 'T2-Ant-v2', 'T3-Ant-v2'): True, ('BMaze-Ant-v2', "TMaze-Ant-v2"): False}, xml_name)
use_model_primitive = value(
{('T0-Ant-v2', 'T1-Ant-v2', 'T2-Ant-v2', 'T3-Ant-v2', 'BMaze-Ant-v2'): False, "TMaze-Ant-v2": True}, xml_name)
model_primitive_checkpoints = [
'rr-T0-Ant-v2-vppo-bs=2048-c=16-bg-up=1-16-64-64-bs=4096-10-ni=0-0.5-ld=F-max=2048-mlr-0-1-k=1-gl=None-cw=1.5'
'-sf=20.0-tf=T-mv=10-rt=T-sl=1.5-sof=T-se=40-om=F-ol=F-gr=200-50-pos=T-lr=0.001-0.003-e=T-mt=FFF-w=FFF-TF'
'-2x64-0.0003-TF-3/70',
'rr-T1-Ant-v2-vppo-bs=2048-c=16-g1-up=1-16-64-64-bs=4096-10-ni=0-0.5-ld=F-max=2048-mlr-0-1-k=1-gl=None-cw=1.5'
'-sf=20.0-tf=T-mv=10-rt=T-sl=1.5-sof=T-se=40-om=F-ol=F-gr=200-50-pos=T-lr=0.001-0.003-e=T-mt=FFF-w=FFF-TF'
'-2x64-0.0003-TF-3/70',
'rr-T2-Ant-v2-vppo-bs=2048-c=16-oc-up=1-16-64-64-bs=4096-10-ni=0-0.5-ld=F-max=2048-mlr-0-1-k=1-gl=None-cw=1.5'
'-sf=20.0-tf=T-mv=10-rt=T-sl=1.5-sof=T-se=40-om=F-ol=F-gr=200-50-pos=T-lr=0.001-0.003-e=T-mt=FFF-w=FFF-TF'
'-2x64-0.0003-TF-3/70',
'rr-T3-Ant-v2-vppo-bs=2048-c=16-jp-up=1-16-64-64-bs=4096-10-ni=0-0.5-ld=F-max=2048-mlr-0-1-k=1-gl=None-cw=1.5'
'-sf=20.0-tf=T-mv=10-rt=T-sl=1.5-sof=T-se=40-om=F-ol=F-gr=200-50-pos=T-lr=0.001-0.003-e=T-mt=FFF-w=FFF-TF'
'-2x64-0.0003-TF-3/70']
bounded = False
dist_bound = 0.08
dist_obv = True
above_target = 3
stage_obv = True
manhattan = False
model_primitive_pg = False
model_primitive_delta = True
# NOTE: MUJOCO
soft = True
weighted = False
restore_model = False
always_restore = False
oracle_master = False
old_policy = False
enforced = True
reset = value({('SPll-stack_2-v0', 'SP-stack_2-v0'): True,
('All-stack_2-v0', 'A-stack_2-v0', '2ll-stack_2-v0', 'Vll-stack_2-v0'): True, (
'TMaze-Ant-v2', 'CMaze-Ant-v2', 'TS-Ant-v2', 'BS-Ant-v2', 'VMaze-Ant-v2', '2Maze-Ant-v2', 'TL-Ant-v2',
'BL-Ant-v2', '5Maze-Ant-v2', 'T0-Ant-v2', 'T1-Ant-v2', 'T2-Ant-v2', 'T3-Ant-v2'): True, 'BMaze-Ant-v2': False},
xml_name)
transfer = value({('SPll-stack_2-v0', 'SP-stack_2-v0'): False,
('All-stack_2-v0', 'A-stack_2-v0', '2ll-stack_2-v0', 'Vll-stack_2-v0'): True, (
'TMaze-Ant-v2', 'CMaze-Ant-v2', 'TS-Ant-v2', 'BS-Ant-v2', 'VMaze-Ant-v2', '2Maze-Ant-v2', 'TL-Ant-v2',
'BL-Ant-v2', '5Maze-Ant-v2', 'T0-Ant-v2', 'T1-Ant-v2', 'T2-Ant-v2', 'T3-Ant-v2'): True, 'BMaze-Ant-v2': False},
xml_name)
oracle_turn_on = -1
num_tasks = paths.shape[0]
if 'stack_2' in xml_name:
box_to_target = []
for path in paths:
d = {}
for i in range(len(path)):
box = path[i][0]
if box not in d:
target = i
d[box] = target
box_to_target.append(d)
target_height = np.array(value({
('All-stack_2-v0', 'SPll-stack_2-v0', '2ll-stack_2-v0', 'Vll-stack_2-v0'): [[0, 1], [0, 0, 1], [0, 0],
[0, 0, 1], [0, 1, 0], [0, 0, 0],
[0, 1], [0, 1, 0]],
('SP-stack_2-v0', 'A-stack_2-v0', '2-stack_2-v0'): [[0, 1]]}, xml_name))
target_to_drop = [np.array([i[1] for i in path]) for path in paths]
box_of_interest = [np.array([i[0] for i in path]) for path in paths]
domain = value({'Basic-bring_ball-v0': 'manipulator', 'Basic-stack_2-v0': 'stacker', 'Basic-Ant-v2': 'maze'},
env_name)
# NOTE: gamma -> discount factor
gamma = 0.99
record_all_lifelong_tasks = False
stop_pg_ts_perc = 1
survival_reward = 0
# NOTE: algorithm
algo = 'vppo'
record = False
prior = 0.25
# NOTE: # trajectory generation
num_cpus = value({'mujoco': 16, 'dm_control': 24}, env_type)
num_cores = num_cpus
bs_per_cpu = value({'dm_control': 1536, 'mujoco': 2048}, env_type)
max_num_timesteps = bs_per_cpu
bs_per_core = bs_per_cpu * num_cpus // num_cores + 1
total_ts = int(value({'dm_control': 3e7, 'mujoco': 3e7}, env_type))
prev_timestep = int(restore_model)
num_batches = (total_ts - prev_timestep) // num_cpus // bs_per_cpu
decay_steps_factor = 0.5
decay_steps = int(num_batches * decay_steps_factor)
# NOTE: alpha -> learning rate of Q's
alpha_begin = 0.01
alpha_end = 0.005
# NOTE: beta -> learning rate of priors
beta = 0.5
# NOTE: e-greedy params
eps_begin = 1.0
eps_end = 0.01
lam = 0.95
# NOTE: multiple runs
num_models = value(
{('A-stack_2-v0', 'All-stack_2-v0'): 12, '5Maze-Ant-v2': 5, ('Vll-stack_2-v0', 'V-stack_2-v0'): 6,
('VMaze-Ant-v2', '2Maze-Ant-v2', '2ll-stack_2-v0'): 2, (
'Stochastic-Fourrooms-large-v0', 'Stochastic-Fourrooms-small-v0', 'Stochastic-Fourrooms-small-v1',
'SquareNoDist-Ant-v2', 'TMaze-Ant-v2', 'RMaze-Ant-v2', 'L-Ant-v2', 'TS-Ant-v2', 'CMaze-Ant-v2'): 4, (
'TL-Ant-v2', 'L-Swimmer-v2', 'TL-Humanoid-v2', 'T-Humanoid-v2', 'LL-Humanoid-v2', 'L2-Ant-v2',
'BL-Humanoid-v2'): 2,
('BL-Ant-v2', 'BS-Ant-v2', 'BMaze-Ant-v2', 'T0-Ant-v2', 'T1-Ant-v2', 'T2-Ant-v2', 'T3-Ant-v2'): 1,
('SP-bring_ball-v0', 'SP-stack_2-v0', 'SPll-stack_2-v0'): 1}, xml_name)
# NOTE: plot params
mov_avg = 10
obscure_factor = 0
savefig = False
dpi = 300
# NOTE: deep neural network for learn_dynamics
n_layers = 10
hidden_fc_size = 256
num_epochs = 100000
base_lr = 0.0001
# NOTE: PG baseline network
pg_bl_n_layers = 2
pg_bl_hidden_fc_size = value({'dm_control': 64, 'mujoco': 64}, env_type)
pg_bl_lr = 3e-4
# NOTE: Model Primitives
pg_mp_n_layers = 2
pg_mp_hidden_fc_size = value({'dm_control': 64, 'mujoco': 64}, env_type)
pg_mp_lr = 3e-4
# NOTE: PG master network
pg_master_n_layers = 2
pg_master_hidden_fc_size = value({'dm_control': 64, 'mujoco': 64}, env_type)
pg_master_ce_lr_source = value({'dm_control': 3e-2, 'mujoco': 1e-3}, env_type)
pg_master_ce_lr_target = 3e-3
pg_master_pg_lr = value(
{'Basic-Humanoid-v2': 0, 'Basic-Ant-v2': 0, ('Basic-bring_ball-v0', 'Basic-stack_2-v0'): 0, }, env_name)
# NOTE: PG main network
pg_pi_n_layers = 2
pg_pi_hidden_fc_size = value({('SPll-stack_2-v0', 'SP-stack_2-v0', 'BS-Ant-v2', 'BL-Ant-v2', 'BMaze-Ant-v2'): 64,
('All-stack_2-v0', 'A-stack_2-v0', 'Vll-stack_2-v0'): 16, (
'TMaze-Ant-v2', 'CMaze-Ant-v2', 'TS-Ant-v2', 'TL-Ant-v2', 'VMaze-Ant-v2', '2Maze-Ant-v2', '5Maze-Ant-v2',
'T0-Ant-v2', 'T1-Ant-v2', 'T2-Ant-v2', 'T3-Ant-v2'): 16, '2ll-stack_2-v0': 16}, xml_name)
pg_sub_lr = 3e-4
# NOTE: ppo clip
clip_param = 0.2
# NOTE: debug
update_master_interval = 1
execution_stats_interval = 10
activation_fn_str = 'tanh'
activation_fn = {'relu': tf.nn.relu, 'tanh': tf.nn.tanh, 'None': None}[activation_fn_str]
lr_step_interval = 100
validate = False # NOTE: validate: whether to verify logic
log_interval = 1
mini_bs = 256 * num_cpus
ent_coeff = 0
optim_epochs = 10
eval_num_rollouts = num_cpus
hostname = socket.gethostname()
hostname = 'SUNet' if 'SUNet' in hostname else hostname
gpu = True if hostname in ['astoria'] else False
Tnoise = 0
Fnoise = 0.5
# NOTE: joint update params
stop_pg_ts = total_ts * stop_pg_ts_perc
gradclip = 20000.
# NOTE: Tver: 4->3, 2->4; bethpage: current
viewer = True
view_sub = False
# NOTE: GIF CONFIG
line_pad = 15
thickness = 1
gif_s = 512
gif_format = 'gif'
fontscale = 0.5
goal_r = 200
num_eval = 70
safe_m_loss = 20.0
# NOTE: ENV
fix_sub = False
goal = None
viewer_dist = 1.2
success_dist = cor_w * 2
solved_threshold = value(
{('T0-Ant-v2', 'T1-Ant-v2', 'T2-Ant-v2', 'T3-Ant-v2'): 1.5, ("BMaze-Ant-v2", "TMaze-Ant-v2"): 0.7},
xml_name) if env_type == 'mujoco' else 0.75
corridors = None
wl_h = 0.8
wl_w = 0.25
warmup = 0.1
e = 1e-20
max_frames = 1000
forward_mult = 50
pos_r = True
short = {True: 'T', False: 'F', 'SUNet': 'SU', 'sislgc1': 'g1', 'sislgc2': 'g2', 'sislgc3': 'g3', 'sislgc4': 'g4',
'sislgc5': 'g5', 'sislgc6': 'g6', 'sislgc7': 'g7', 'sislgc8': 'g8', 'sislgc9': 'g9', 'sislgc10': 'g10',
'sislgc11': 'g11', 'sislgc12': 'g12', 'sislgc13': 'g13', 'sislgc14': 'g14', 'sislgc15': 'g15',
'sislgc16': 'g16', 'sislgc17': 'g17', 'sislgc18': 'g18', 'sislgc19': 'g19', 'sislgc20': 'g20',
'sislgc21': 'g21', 'sislgc22': 'g22', 'sislgc23': 'g23', 'sislgc24': 'g24', 'sislgc25': 'g25',
'sislgc26': 'g26', 'sislgc27': 'g27', 'sislgc28': 'g28', 'sislgc29': 'g29', 'sislgc30': 'g30', 'bethpage': 'bg',
'cambridge': 'ca', 'tula': 'tl', 'tver': 'tv', 'jodhpur': 'jp', 'astoria': 'as', 'cheonan': 'ch',
'oceanside': 'oc', 'Bohans-MacBook-Pro.local': 'mac', 'dyn-160-39-154-237.dyn.columbia.edu': 'mac', }
dir_to_model = {'E': 0, 'N': 1, 'W': 2, 'S': 3, }
# NOTE: GENERATE MAZES
num_mazes = 20
num_corr = [5] * num_mazes
max_corr_len = 8
min_corr_len = 4
finalize = True
run_stds = [[4.599198, 4.6641273, 1.28923, 2.454618, 0.099999994, 0.099999994, 0.12651926, 0.44238856, 0.24519706,
0.27842945, 0.3273297, 0.3899642, 0.2788939, 0.39544523, 0.2628939, 0.37652737, 0.2529815, 0.39808935,
0.27402195, 1.2319881, 0.6234127, 0.6694361, 1.0196308, 1.0814291, 1.2173401, 2.8830736, 3.0794158,
4.0499864, 1.9910984, 3.6656153, 1.9535313, 4.0345163, 2.2735405, 0.099999994, 0.099999994,
0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.13444856, 0.13409562, 0.099999994, 0.2144815,
0.21478048, 0.2999312, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994,
0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.2986342, 0.30273247,
0.29329726, 0.30751246, 0.30443984, 0.30402762, 0.099999994, 0.099999994, 0.099999994, 0.099999994,
0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994,
0.32695666, 0.3283175, 0.32865548, 0.33718628, 0.3334012, 0.3277009, 0.099999994, 0.099999994,
0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994,
0.099999994, 0.099999994, 0.29600772, 0.29339242, 0.28964946, 0.30079424, 0.30085075, 0.29916844,
0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994,
0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.31006745, 0.31021363, 0.30940148, 0.31883976,
0.316916, 0.31288943],
[1.2312175, 2.5586734, 4.862292, 4.9084835, 0.099999994, 0.099999994, 0.11902989, 0.37661228, 0.19612414,
0.1970121, 0.3508082, 0.41372088, 0.24030799, 0.37279978, 0.2762219, 0.40963152, 0.25021195, 0.356578,
0.29620296, 0.68267214, 1.3554753, 0.762895, 1.0119213, 1.1173822, 1.1998409, 4.2042365, 2.1269264, 2.3674908,
3.1287558, 4.3375616, 2.4620428, 2.987101, 1.4826976, 0.099999994, 0.099999994, 0.099999994, 0.099999994,
0.099999994, 0.099999994, 0.11626974, 0.11602408, 0.099999994, 0.17530695, 0.17600684, 0.22305314, 0.099999994,
0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994,
0.099999994, 0.099999994, 0.099999994, 0.32071027, 0.31924236, 0.3196359, 0.32367983, 0.32657215, 0.31954604,
0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994,
0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.3258689, 0.3229741, 0.31455493, 0.3269703, 0.32847512,
0.32323778, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994,
0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.28302696, 0.2854431, 0.282894, 0.28799674,
0.28955868, 0.2865865, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994,
0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.3557716, 0.3584011, 0.34562215,
0.36439514, 0.3656799, 0.35458568],
[0.27116972, 0.2731105, 0.49103335, 0.35940948, 10.0, 10.0, 8.285499, 3.1395023, 5.3231235, 5.5543966, 2.809905,
2.5080442, 4.685012, 2.5534766, 3.8159177, 2.789227, 3.6269453, 2.5306308, 3.9287062, 0.829414, 1.6191914,
1.1945486, 0.9497294, 0.88029605, 0.71083105, 0.26208684, 0.6023471, 0.22584441, 0.36695793, 0.33044592,
0.31234562, 0.23242629, 0.59796476, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 9.156298, 9.454527, 10.0, 6.133729,
6.050961, 4.6875067, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 3.3002894,
3.3518806, 3.370098, 3.2725973, 3.2803993, 3.3199122, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0,
10.0, 10.0, 10.0, 3.1619124, 3.175854, 3.1968389, 3.0986283, 3.091742, 3.1791952, 10.0, 10.0, 10.0, 10.0, 10.0,
10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 3.7807808, 3.753944, 3.8666487, 3.6851885, 3.694838, 3.7078226, 10.0,
10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 2.9965332, 2.9918234, 2.9949353, 2.9243422,
2.925929, 3.0298653],
[1.2362218, 3.4339106, 3.632699, 3.6123943, 0.099999994, 0.099999994, 0.12220171, 0.3551185, 0.23772441,
0.1933735, 0.347828, 0.35959145, 0.21327308, 0.40731293, 0.2096475, 0.34834144, 0.27511126, 0.38403288,
0.27539626, 0.5728318, 1.1423793, 0.77614814, 1.1704272, 1.0412856, 1.4850398, 2.3779807, 1.7846571, 4.932697,
1.9872103, 2.5577078, 3.2989748, 3.4681187, 2.5812807, 0.099999994, 0.099999994, 0.099999994, 0.099999994,
0.099999994, 0.099999994, 0.115986496, 0.11946792, 0.099999994, 0.18283401, 0.18193267, 0.24615471,
0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994,
0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.2910546, 0.2980322, 0.28877378, 0.29828975, 0.29668167,
0.29561338, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994,
0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.33946756, 0.34012014, 0.3420155, 0.3465538,
0.34908473, 0.3354624, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994,
0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.2563183, 0.25463828,
0.24056835, 0.2594071, 0.26107442, 0.26175717, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994,
0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.099999994, 0.3323459,
0.33377367, 0.32560465, 0.3441525, 0.34077618, 0.3317802],
]
run_std = value({
'Basic-Ant-v2': [0.0001, 0.0001, 0.0001, 0.0001, 11926939, 19020905, 0.12928092, 0.6212888, 0.25102094,
0.22332077, 0.58209974, 0.42657858, 0.2656426, 0.40442517, 0.27392906, 0.40158844, 0.27972764,
0.40863708, 0.30180925, 0.7156728, 0.77419865, 0.65207326, 1.1297017, 1.106495, 1.3923764,
2.8920958, 2.2228127, 3.126648, 2.608846, 3.370323, 2.8734233, 4.168823, 2.3023212, 0.09999999,
0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.14755759, 0.14737378, 0.09999999,
0.21581785, 0.21529765, 0.25864798, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999,
0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.3328616,
0.33258832, 0.3209721, 0.33333594, 0.33337238, 0.3281808, 0.09999999, 0.09999999, 0.09999999,
0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999,
0.09999999, 0.36311337, 0.36590517, 0.35446492, 0.36720052, 0.36716765, 0.35422853, 0.09999999,
0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999,
0.09999999, 0.09999999, 0.09999999, 0.29842412, 0.29977155, 0.29137745, 0.3030151, 0.3025487,
0.2990263, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999,
0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.36081558, 0.36251754, 0.35274118,
0.36504218, 0.3648821, 0.35191768], 'Basic-Humanoid-v2': np.ones(shape=382),
'Basic-bring_ball-v0': np.ones(shape=37), # 'Basic-stack_2-v0': np.ones(
# shape=48 + 14 * dist_obv + (18 if redundant else 8) + repeat * 2)
'Basic-stack_2-v0': [0.59313864, 0.5122037, 1.5283116, 1.3284554, 1.0007584, 0.2044712, 0.159425, 0.20430137,
0.16148166, 0.16000782, 0.09999999, 0.7036024, 0.7099915, 0.19793506, 0.09999999,
0.7012213, 0.7117969, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999,
0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999, 0.09999999,
3.3793998, 2.9180048, 2.562093, 2.7047367, 1.0841713, 0.8094801, 1.0757841, 0.8439478,
0.44067168, 0.33905765, 7.5584087, 0.3428601, 0.3224279, 5.8961425, 1.3366703, 1.8276505,
1.8320743, 1.1618187, 1.1803689, 0.15215792, 0.10663375, 0.17174725, 0.09999999,
0.20094976, 0.10194187, 0.20406862, 0.10068466, 0.25226733, 0.09999999, 0.20357151,
0.09999999, 0.2427386, 0.11436407, 0.36003974, 0.23516357, 0.17657563, 0.47387388,
0.23171127, 0.09999999, 0.47834542, 0.09999999, 0.32406855, 0.18077546, 0.13010821,
0.2439245, 0.31692365, 0.09999999, 0.4783455, 0.09999999, 0.4783455, 0.47834542, 0.4783455,
0.47834542, 0.4783455, 0.47834542, 0.4783455, 0.47834542, 0.4783455, 0.47834542, 0.4783455,
0.47834542, 0.4783455, 0.47834542, 0.4783455, 0.47834542, 0.4783455, 0.47834542,
0.47834548, 0.4783454, 0.47834548, 0.4783454]}, env_name)
debug = True
change_color = True
gold = (0.8 * 255, 0.6 * 255, 0.4 * 255)