-
Notifications
You must be signed in to change notification settings - Fork 28
/
training_run_notes_costar.txt
2393 lines (1800 loc) · 266 KB
/
training_run_notes_costar.txt
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
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
saving plot: 2020-01-28-13-38-32_Sim-Stack-Trial-Reward-Training-Sim-Stack-Trial-Reward-Training_success_plot.png
export CUDA_VISIBLE_DEVICES="1" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --tcp_port 19980 --place --future_reward_discount 0.65 --nn densenet
STACKING COMMON SENSE DENSENET, trial reward, check_z_height
-------------------------------------------------------------
costar@costar-desktop|~/src/real_good_robot on fast_sim_thread!?
export CUDA_VISIBLE_DEVICES="0" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65
GPU 0, Tab 0, port 19990, commit 2353c4a9ca39438eca18855b8da68d64a7258706
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-01-30-13-41-13_Sim-Stack-Trial-Reward-Common-Sense-Training
STACKING, DENSENET, NO COMMON SENSE, trial reward, check_z_height
-------------------------------------------------------------
GPU 1, Tab 1, port 19998, commit 2353c4a9ca39438eca18855b8da68d64a7258706
export CUDA_VISIBLE_DEVICES="1" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --check_z_height --tcp_port 19998 --place --future_reward_discount 0.65
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-01-30-13-42-19_Sim-Stack-Trial-Rewa
rd-Training
STACKING COMMON SENSE DENSENET, trial reward, check_z_height
-------------------------------------------------------------
costar@costar-desktop|~/src/real_good_robot on fast_sim_thread!?
export CUDA_VISIBLE_DEVICES="0" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65
GPU 0, Tab 0, port 19990, commit 0e0a4749fc5560d64e3129d1f269fc5fc7e0dc32
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-01-30-13-41-13_Sim-Stack-Trial-Reward-Common-Sense-Training
Experience replay update, now alternates training on past success and failure for the current action
=====================================================================================================
note: alternating slowed down learning and has thus been removed
************************************************************
STACKING, DENSENET, NO COMMON SENSE, trial reward, check_z_height -- ABSOLUTE BEST RUN AS OF 2020-02-02
-------------------------------------------------------------
GPU 1, Tab 1, port 19998, commit b3661e21bf715f93f23833583e6ee5e9ffb607aa
export CUDA_VISIBLE_DEVICES="1" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --check_z_height --tcp_port 19998 --place --future_reward_discount 0.65
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-01-18-56-14_Sim-Stack-Trial-Reward-Training
stats for training:
Max grasp success rate: 0.9424603174603174, at action iteration: 13453. (total of 15651 actions, max excludes first 1000 actions)
Max place success rate: 0.79957805907173, at action iteration: 14487. (total of 15652 actions, max excludes first 1000 actions)
Max action efficiency: 0.582, at action iteration: 12444. (total of 15652 actions, max excludes first 1000 actions)
Max trial success rate: 0.7904761904761904, at action iteration: 12989. (total of 15651 actions, max excludes first 1000 actions)
************************************************************
STACKING COMMON SENSE DENSENET, trial reward, check_z_height -- This Did very well but not quite as well as above
-------------------------------------------------------------
GPU 0, Tab 0, port 19990, commit b3661e21bf715f93f23833583e6ee5e9ffb607aa
export CUDA_VISIBLE_DEVICES="0" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-01-18-38-22_Sim-Stack-Trial-Reward-Common-Sense-Training
STACKING COMMON SENSE DENSENET, trial reward, check_z_height -- COMMON SENSE TRAINS ARGMAX VALUE 0 WHEN IT IS A GEOMETRICALLY KNOWN FAILURE
-------------------------------------------------------------
GPU 0, Tab 0, port 19990, commit bfeaf0326812af89093a72c97e2e43506cb9ef4c, "main.py trainer.py utils_torch.py COMMON SENSE TRAIN GEOMETRIC ARGMAX FAILURE"
export CUDA_VISIBLE_DEVICES="0" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-03-11-49-57_Sim-Stack-Trial-Reward-Common-Sense-Training
SPOT REWARD UPDATE, spot reward now gives full double credit to final time step of a successful trial, with no explore_rate_decay
=================================================================================================================================
STACKING, DENSENET, NO COMMON SENSE, trial reward, check_z_height, IMPORTANT GOES IN PAPER
-------------------------------------------------------------
GPU 1, Tab 1, port 19998, commit 786a5fc256a8c9eb1b1edf39f6a4f3ce274dd455, "trainer.py MAJOR SPOT REWARD CHANGE, 2X FINAL REWARD ON LAST TIMESTEP
export CUDA_VISIBLE_DEVICES="1" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --check_z_height --tcp_port 19998 --place --future_reward_discount 0.65
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-03-16-58-06_Sim-Stack-Trial-Reward-Training
> TESTING RUN Random Arrangements
> Commit: e6583b8e7ed093887b8f08261683a2220c374bdd
> export CUDA_VISIBLE_DEVICES="1" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --check_z_height --tcp_port 19998 --place --future_reward_discount 0.65 --is_testing --random_seed 1238 --snapshot_file '/home/costar/src/real_good_robot/logs/2020-02-03-16-58-06_Sim-Stack-Trial-Reward-Training/models/snapshot.reinforcement-best-stack-rate.pth'
> Pre-trained model snapshot loaded from: /home/costar/src/real_good_robot/logs/2020-02-03-16-58-06_Sim-Stack-Trial-Reward-Training/models/snapshot.reinforcement-best-stack-rate.pth
> Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-20-18-00-10_Sim-Stack-SPOT-Trial-Reward-Testing
> Video: recording_2020_02_20-17_59-55.avi
> Results: {'trial_success_rate_best_value': 0.97, 'trial_success_rate_best_index': 1567, 'grasp_success_rate_best_value': 0.8146002317497103, 'grasp_success_rate_best_index': 1567, 'place_success_rate_best_value': 0.7886524822695036, 'place_success_rate_best_index': 1567, 'action_efficiency_best_value': 0.3752393107849394, 'action_efficiency_best_index': 1569}
STACKING COMMON SENSE DENSENET, trial reward, check_z_height -- COMMON SENSE TRAINS ARGMAX VALUE 0 WHEN IT IS A GEOMETRICALLY KNOWN FAILURE + double credit
-------------------------------------------------------------
GPU 0, Tab 0, port 19990, commit 786a5fc256a8c9eb1b1edf39f6a4f3ce274dd455, "main.py trainer.py utils_torch.py COMMON SENSE TRAIN GEOMETRIC ARGMAX FAILURE" + "trainer.py MAJOR SPOT REWARD CHANGE, 2X FINAL REWARD ON LAST TIMESTEP
export CUDA_VISIBLE_DEVICES="0" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-03-16-57-28_Sim-Stack-Trial-Reward-Common-Sense-Training
> TESTING RUN
> export CUDA_VISIBLE_DEVICES="0" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --tcp_port 19990 --place --future_reward_discount 0.65 --nn densenet --check_z_height --is_testing --random_seed 1238 --snapshot_file '/home/costar/src/real_good_robot/logs/2020-02-03-16-57-28_Sim-Stack-Trial-Reward-Common-Sense-Training/models/snapshot-backup.reinforcement-best-stack-rate.pth'
> Commit: bfea389e37b7205dc54bf5dd2357eb658a0c3527
> Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-16-22-38-38_Sim-Stack-SPOT-Trial-Reward-Testing
> VIDEO: recording_2020_02_16-22_40-23.avi
> GPU 0, Tab 0, port 19990
PUSHING AND GRASPING WITH ALL FEATURES & STUCK OBJECT FIXES - Feb 7
--------------------------------------
± export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/toys --num_obj 10 --push_rewards --experience_replay --explore_rate_decay --common_sense --trial_reward --save_visualizations --future_reward_discount 0.65 --tcp_port 19998
Commit: 5c78490ae6f25cc257ac5fa2030118bc0644e9e8
logging session: logs/2020-02-07-14-43-44_Sim-Push-and-Grasp-Trial-Reward-Common-Sense-Training
GPU 0, Tab 0, port 19998, right v-rep window
> Preset testing run
> Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-11-15-59-07_Sim-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Testing
> export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir 'objects/toys' --num_obj 10 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --future_reward_discount 0.65 --tcp_port 19990 --is_testing --random_seed 1238 --snapshot_file '/home/costar/src/real_good_robot/logs/2020-02-07-14-43-44_Sim-Push-and-Grasp-Trial-Reward-Common-Sense-Training/models/snapshot.reinforcement.pth' --max_test_trials 10 --test_preset_cases
> Commit: 7b6c54ad615d592d86e71d90ea36c6478193a456
> GPU 1, tab 13, port 19999, left v-rep window
STACKING COMMON SENSE DENSENET WITH RANDOM PLACEMENTS OF OBJECTS STUCK TO GRIPPER DUE TO SIMULATOR BUGS
----------------------------------------
± export CUDA_VISIBLE_DEVICES="1" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65
commit: 5c78490ae6f25cc257ac5fa2030118bc0644e9e8
Comment: we manually uncommented the PixelNet() call in trainer.py and commented reinforcement_net()
RESUME with no two step backprop: ± export CUDA_VISIBLE_DEVICES="2" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_z_height --place --tcp_port 20000 --future_reward_discount 0.65 --max_train_actions 10000 --nn efficientnet --disable_two_step_backprop --random_actions --resume '/home/ahundt/src/real_good_robot/logs/2020-05-04-12-08-15_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Training'
RESUME commit: 3ba0b91c5accac6387345c62d5a4e8b7ff9769cd
logging session: logs/2020-02-07-14-35-59_Sim-Stack-Trial-Reward-Common-Sense
GPU 1, Tab 1, port 19990, left v-rep window
SIM STACKING, ANY BLOCK, COMMON SENSE, DENSENET, SPOT TRIAL REWARD
---------------------------------------------------------
export CUDA_VISIBLE_DEVICES="0" && python main.py --is_sim --obj_mesh_dir objects/toys --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --tcp_port 19998 --place --future_reward_discount 0.65 --nn densenet --common_sense --check_z_height
commit: 7b6c54ad615d592d86e71d90ea36c6478193a456
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-10-19-09-09_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Training
GPU 0, Tab 9, port 19998, right v-rep window
> Random Testing Any block
> export CUDA_VISIBLE_DEVICES="0" && python main.py --is_sim --obj_mesh_dir objects/toys --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --tcp_port 19990 --place --future_reward_discount 0.65 --nn densenet --common_sense --check_z_height --random_seed 1238 --is_testing --max_test_trials 50 --snapshot_file '/media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-10-19-09-09_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Training/models/snapshot.reinforcement-best-stack-rate.pth'
> Commit: e6583b8e7ed093887b8f08261683a2220c374bdd
> Video: recording_2020_02_20-19_14-02.avi
> Pre-trained model snapshot loaded from: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-10-19-09-09_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Training/models/snapshot.reinforcement-best-stack-rate.pth
> Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-20-19-14-08_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Testing
> Testing Results: {'trial_success_rate_best_value': 0.24, 'trial_success_rate_best_index': 2680, 'grasp_success_rate_best_value': 0.554670528602462, 'grasp_success_rate_best_index': 2680, 'place_success_rate_best
_value': 0.4846153846153846, 'place_success_rate_best_index': 2680, 'action_efficiency_best_value': 0.03805970149253731, 'action_efficiency_best_index': 2680}
> GPU 0, Tab 8, port 19990, left v-rep window,
PUSHING AND GRASPING WITH ALL FEATURES & SAVE ALL MODELS ACCORDING TO BEST STATS - Feb 12
--------------------------------------
export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir objects/toys --num_obj 10 --push_rewards --experience_replay --explore_rate_decay --common_sense --trial_reward --save_visualizations --future_reward_discount 0.65 --tcp_port 19990
Commit: 22f63b9eea28bbeaaf31930e9731cc7b17b43c35
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-12-17-58-04_Sim-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training
GPU 1, Tab 14, port 19990, left v-rep window
REAL ROBOT PUSHING AND GRASPING COMMON SENSE SPOT - NICK aborted FEB 13
---------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="1" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --future_reward_discount 0.65
commit: 5d59f747024e92918d3e8403ee816e9f86d5352b
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-13-15-16-38_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training
GPU 1: Tab 0, port N/A, Real Robot
^^^ NOTE: not sure what went wrong. things stopped showing up in the camera. will come back later to check.
the visualizations folder is still there.
PUSHING AND GRASPING WITH ALL FEATURES & SAVE ALL MODELS ACCORDING TO BEST STATS - Feb 13
--------------------------------------
export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/toys --num_obj 10 --push_rewards --experience_replay --explore_rate_decay --common_sense --trial_reward --save_visualizations --future_reward_discount 0.65 --tcp_port 19990
Commit: 2b55d4b48c2c6fa1959e52947691b26355aa4180
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-13-18-38-34_Sim-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training
GPU 0, Tab 0, port 19990, left v-rep window
SIM STACKING, COMMON SENSE, DENSENET, SPOT TRIAL REWARD - Feb 13 - Critical bugfix to place experience replay
---------------------------------------------------------
export CUDA_VISIBLE_DEVICES="1" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --tcp_port 19998 --place --future_reward_discount 0.65 --nn densenet --common_sense --check_z_height
commit: 2b55d4b48c2c6fa1959e52947691b26355aa4180
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-13-19-02-32_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Training
GPU 1, Tab 1, port 19998, right v-rep window
REAL ROBOT PUSHING AND GRASPING COMMON SENSE, SPOT - FEB 14 - LONG AND GOOD RUN
=================================================================================
export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --future_reward_discount 0.65
commit: d5e28bcac0dc41d3d41e7f7d538f91bab73c69f8
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-14-13-08-21_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training
Resume command: export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --future_reward_discount 0.65 --resume '/home/costar/src/real_good_robot/logs/2020-02-14-13-08-21_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training'
GPU 1, Tab 0, port N/A, Real Robot
SIM STACKING, COMMON SENSE, DENSENET, SPOT TRIAL REWARD - Feb 14 - Critical bugfix to place experience replay, plotting
---------------------------------------------------------
commit: d5e28bcac0dc41d3d41e7f7d538f91bab73c69f8
export CUDA_VISIBLE_DEVICES="1" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --tcp_port 19998 --place --future_reward_discount 0.65 --nn densenet --common_sense --check_z_height
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-14-19-47-20_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Training
GPU 1, Tab 1, port 19998, right v-rep window
REAL STACKING, COMMON SENSE, SPOT - FEB 15 - JUNK DO NOT USE
=================================================================================
export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --future_reward_discount 0.65 --place --check_z_height
RESUME COMMAND: export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --future_reward_discount 0.65 --place --check_z_height --resume '/home/costar/src/real_good_robot/logs/2020-02-15-15-55-40_Real-Stack-SPOT-Trial-Reward-Common-Sense-Training'
commit: ea6b6d90967aaadc0d3ef8620f1d3a590cff0757
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-15-14-12-00_Real-Stack-SPOT-Trial-Reward-Common-Sense-Training
GPU 0, Tab 0, port N/A, Real Robot
REAL PUSHING AND GRASPING - SUPER BASIC RUN - FEB 18 - FOR FINAL PAPER RESULTS!!!!!! IMPORTANT - planning on 1000 actions.
====================================================
export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --save_visualizations
Commit: 656625133ed3c7d750f99c22b44c82e288c7e6be
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-18-18-58-15_Real-Push-and-Grasp-Two-Step-Reward-Training
GPU 0, Tab 0, port N/A, Real Robot
REAL, PUSHING AND GRASPING, COMMON SENSE, SPOT - FEB 19 - LONG AND GOOD RUN
=================================================================================
export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --future_reward_discount 0.65
Commit: 67b792c6a08309c8406de30804d1fe147c9d967f
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-19-15-33-05_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training
> OUTDATED
> Commit: bb5ae93d373a3bbc40786cc23f76eed0ae2ad233
> OUTDATED DUE TO PAUSE BEFORE FIRST TRIAL IS OVER BUG: Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-19-14-28-23_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training
> OUTDATED DUE TO PAUSE BEFORE FIRST TRIAL IS OVER BUG: Resume command: export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --future_reward_discount 0.65 --resume /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-19-14-28-23_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training
GPU 0, Tab 0, port N/A, Real Robot
REAL, STACKING, COMMON SENSE, SPOT - FEB 09 (multi day) - FOR FINAL PAPER RESULTS!!!
====================================================================================
https://github.com/jhu-lcsr/real_good_robot/releases/tag/v0.14.0
Commit: 8e01a12758f25ab3e4535b861bdbb140d8415ce9
> Final Testing Run, 10 trials
> export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --check_z_height --place --future_reward_discount 0.65 --is_testing --max_test_trials 10 --snapshot_file '/home/costar/src/real_good_robot/logs/2020-02-09-11-02-57_Real-Stack-SPOT-Trial-Reward-Common-Sense-Training/models/snapshot.reinforcement_trial_success_rate_best_value.pth' --random_seed 1238
> Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-19-23-21-59_Real-Stack-SPOT-Trial-Reward-Common-Sense-Testing
SIM ROWS DENSENET - OLD ALGORITHM - WITH NO COMMON SENSE, NO TRIAL REWARD
-------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="1" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --save_visualizations --tcp_port 19998 --place --check_row --max_train_actions 10000
Commit: 8b6937f3597815e3cf0c62294d2235ea14c26aec
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-21-20-33-17_Sim-Rows-Two-Step-Reward-Training
GPU 1, Tab 1, port 19998, right v-rep window
SIM STACK DENSENET - OLD ALGORITHM - WITH NO COMMON SENSE, NO TRIAL REWARD (TODO RESUME ME!!!!!)
-------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="0" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --save_visualizations --tcp_port 19998 --place --check_z_height --max_train_actions 10000
RESUME: export CUDA_VISIBLE_DEVICES="0" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --save_visualizations --tcp_port 19998 --place --check_z_height --max_train_actions 10000 --resume /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-21-20-33-47_Sim-Stack-Two-Step-Reward-Training
Commit: 8b6937f3597815e3cf0c62294d2235ea14c26aec
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-21-20-33-47_Sim-Stack-Two-Step-Reward-Training
GPU 0, Tab 0, port 19990, left v-rep window
SIM TO REAL STACKING TRAINING COMMAND, Common Sense, SPOT Trial Reward
====================================================================
/media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-22-Sim-to-Real-2020-02-03-16-57-28_Sim-Stack-Trial-Reward-Common-Sense-Training
export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --check_z_height --place --future_reward_discount 0.65 --resume /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-22-Sim-to-Real-2020-02-03-16-57-28_Sim-Stack-Trial-Reward-Common-Sense-Training
export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --check_z_height --place --future_reward_discount 0.65 --is_testing --random_seed 1238 --max_test_trials 10 --load_snapshot '/home/costar/src/real_good_robot/logs/2020-02-22-Sim-to-Real-2020-02-03-16-57-28_Sim-Stack-Trial-Reward-Common-Sense-Training/models/snapshot.reinforcement.pth'
Commit: 42f0fc09a2ed776c7089ea346d14509957dd0f5c
GPU 0, Tab 0, port N/A, Real Robot
SIM TO REAL TESTING STACKING - 9 of 10 stack successes.
===================
costar@costar-desktop|/media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot on revert_pixelnet! [0/95245]
export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --check_z_height --place --future_reward_discount 0.65 --is_testing --random_seed 1238 --max_test_trials 10 --snapshot_file '/home/costar/src/real_good_robot/logs/2020-02-22-Sim-to-Real-2020-02-03-16-57-28_Sim-Stack-Trial-Reward-Common-Sense-Training/models/snapshot.reinforcement-best-stack-rate.pth'
Pre-trained model snapshot loaded from: /home/costar/src/real_good_robot/logs/2020-02-22-Sim-to-Real-2020-02-03-16-57-28_Sim-Stack-Trial-Reward-Common-Sense-Training/models/snapshot.reinforcement-best-stack-rate.pth
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-22-17-52-17_Real-Stack-SPOT-Trial-Reward-Common-Sense-Testing
2020-02-22-17-52-17_Real-Stack-SPOT-Trial-Reward-Common-Sense-Testing
{"action_efficiency_best_index": 183, "action_efficiency_best_value": 0.29508196721311475, "grasp_success_rate_best_index": 183, "grasp_success_rate_best_value": 0.4263565891472868, "place_success_rate_best_index": 183, "place_success_rate_best_value": 0.7818181818181819, "trial_success_rate_best_index": null, "trial_success_rate_best_value": -Infinity}
TEST-V2, we applied WD-40 to the gripper, but LR still too low
export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --check_z_height --place --future_reward_discount 0.65 --is_testing --random_seed 1238 --max_test_trials 10 --snapshot_file '/home/costar/src/real_good_robot/logs/2020-02-22-Sim-to-Real-2020-02-03-16-57-28_Sim-Stack-Trial-Reward-Common-Sense-Training/models/snapshot.reinforcement-best-stack-rate.pth'
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-22-19-54-28_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Testing
TEST-V3
{'trial_success_rate_best_value': 0.9, 'trial_success_rate_best_index': 101, 'grasp_success_rate_best_value': 0.8035714285714286, 'grasp_success_rate_best_index': 101, 'place_success_rate_best_value': 0.8043478260869565, 'place_success_rate_best_index': 101, 'action_efficiency_best_value': 0.594059405940594, 'action_efficiency_best_index': 103}
SIM TO REAL TESTING Pushing and Grasping
========================================
export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --future_reward_discount 0.65 --is_testing --random_seed 1238 --max_test_trials 1 --snapshot_file '/home/costar/Downloads/snapshot.reinforcement_grasp_action_efficiency_best_value.pth'
from femur: 2020-02-16-21-33-59_Sim-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training
Training Complete! Dir: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-22-19-54-28_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Testing
Grasp Count: 52, grasp success rate: 0.34615384615384615
V2 testing 2020-02-24-0001
export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --future_reward_discount 0.65 --is_testing --random_seed 1238 --max_test_trials 1 --snapshot_file '/home/costar/Downloads/snapshot.reinforcement_grasp_action_efficiency_best_value.pth'
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-24-01-03-39_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Testing
Max grasp success rate: 0.21621621621621623, at action iteration: 104. (total of 106 actions, max excludes first 104 actions)
Max grasp action efficiency: 0.15384615384615385, at action iteration: 104. (total of 107 actions, max excludes first 104 actions)
saving plot: 2020-02-24-01-16-21_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Testing-Sim-to-Real-Pushing-And-Grasping-SPOT-Q_success_plot.png
saving best stats to: /home/costar/src/real_good_robot/logs/2020-02-24-01-16-21_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Testing/data/best_stats.json
saving best stats to: /home/costar/src/real_good_robot/logs/2020-02-24-01-16-21_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Testing/best_stats.json
{"grasp_action_efficiency_best_index": 104, "grasp_action_efficiency_best_value": 0.15384615384615385, "grasp_success_rate_best_index": 104, "grasp_success_rate_best_value": 0.21621621621621623}
(this first run ended early, not sure what happened. - Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-24-01-03-39_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Testing)
SIM TO REAL TESTING ROWS
========================
export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --check_z_height --place --future_reward_discount 0.65 --is_testing --random_seed 1238 --max_test_trials 10 --check_row --num_obj 4 --snapshot_file '/home/costar/Downloads/2020-02-10-18-38-rows-snapshot.reinforcement-best-stack-rate.pth'
REAL, PUSHING AND GRASPING, COMMON SENSE, SPOT - FEB 23 - LONG AND GOOD RUN - in paper
=================================================================================
export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --future_reward_discount 0.65
'/home/costar/src/real_good_robot/logs/2020-02-23-11-43-55_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training'
RESUME: export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --future_reward_discount 0.65 --resume '/home/costar/src/real_good_robot/logs/2020-02-23-11-43-55_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training'
Max trial success rate: 1.0, at action iteration: 682. (total of 1032 actions, max excludes first 500 actions)
Max grasp success rate: 0.6054421768707483, at action iteration: 774. (total of 1032 actions, max excludes first 500 actions)
Max grasp action efficiency: 0.534, at action iteration: 774. (total of 1033 actions, max excludes first 500 actions)
saving plot: 2020-02-23-11-43-55_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training-Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training_success_plot.png
saving best stats to: /home/costar/src/real_good_robot/logs/2020-02-23-11-43-55_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training/data/best_stats.json
saving best stats to: /home/costar/src/real_good_robot/logs/2020-02-23-11-43-55_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training/best_stats.json
Training Complete! Dir: /home/costar/src/real_good_robot/logs/2020-02-23-11-43-55_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training
Training results:
{'trial_success_rate_best_value': 1.0, 'trial_success_rate_best_index': 682, 'grasp_success_rate_best_value': 0.6054421768707483, 'grasp_success_rate_best_index': 774, 'grasp_action_efficiency_best_value': 0.534, 'grasp_action_efficiency_best_index': 774}
TESTING push and grasp
export CUDA_VISIBLE_DEVICES="0" && python main.py --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --future_reward_discount 0.65 --is_testing --random_seed 1238 --max_test_trials 1 --snapshot_file '/home/costar/src/real_good_robot/logs/2020-02-23-11-43-55_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Training/models/snapshot.reinforcement_grasp_success_rate_best_value.pth'
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-02-23-18-51-58_Real-Push-and-Grasp-SPOT-Trial-Reward-Common-Sense-Testing
=============================================================
2020-04 and 2020-05
=============================================================
Tab 7: ~/src/CoppeliaSim_Edu_V4_0_0_Ubuntu18_04/coppeliaSim.sh -gREMOTEAPISERVERSERVICE_19990_FALSE_TRUE -s ~/src/real_good_robot/simulation/simulation.ttt
Tab 8: ~/src/CoppeliaSim_Edu_V4_0_0_Ubuntu18_04/coppeliaSim.sh -gREMOTEAPISERVERSERVICE_19998_FALSE_TRUE -s ~/src/real_good_robot/simulation/simulation.ttt
SIM STACK DENSENET - OLD ALGORITHM - WITH NO COMMON SENSE, NO TRIAL REWARD - NO HEIGHT REWARD - 2020-04-25-19-59-01
----------------------------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="0" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --save_visualizations --tcp_port 19990 --place --check_z_height --max_train_actions 10000 --no_height_reward --disable_situation_removal
RESUME:
Commit: de5f639ae814bcb1870abe3d8190bebf84abe1ec
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-04-25-21-41-04_Sim-Stack-Two-Step-Reward-Training
IGNORE, FORGOT TO DISABLE SITUATION REMOVAL: Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-04-25-19-59-01_Sim-Stack-Two-Step-Reward-Training
± cat logs/2020-04-25-21-41-04_Sim-Stack-Two-Step-Reward-Training/2020-04-27-17-44-27_Sim-Stack-Two-Step-Reward-Testing/best_stats.json
{"action_efficiency_best_index": 3991, "action_efficiency_best_value": 0.019543973941368076, "grasp_success_rate_best_index": 3991, "grasp_success_rate_best_value": 0.9404958677685951, "place_success_rate_best_index": 3991, "place_success_rate_best_value": 0.5837526959022286, "trial_success_rate_best_index": 3991, "trial_success_rate_best_value": 0.13}%
GPU 0, Tab 4, port 19990, left v-rep window
SIM ROW DENSENET - OLD ALGORITHM - WITH NO COMMON SENSE, NO TRIAL REWARD - NO HEIGHT REWARD - 2020-04-25-20-00-41
--------------------------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="1" && python main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --save_visualizations --tcp_port 19998 --place --check_row --max_train_actions 10000 --no_height_reward --disable_situation_removal
Commit: de5f639ae814bcb1870abe3d8190bebf84abe1ec
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-04-25-21-41-35_Sim-Rows-Two-Step-Reward-Training
IGNORE, FORGOT TO DISABLE SITUATION REMOVAL: Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-04-25-20-00-41_Sim-Rows-Two-Step-Reward-Training
± cat logs/2020-04-25-21-41-35_Sim-Rows-Two-Step-Reward-Training/2020-04-27-17-25-31_Sim-Rows-Two-Step-Reward-Testing/best_stats.json
{"action_efficiency_best_index": 2124, "action_efficiency_best_value": 0.00847457627118644, "grasp_success_rate_best_index": 2124, "grasp_success_rate_best_value": 0.5886075949367089, "place_success_rate_best_index": 2124, "place_success_rate_best_value": 0.2507204610951009, "trial_success_rate_best_index": 2124, "trial_success_rate_best_value": 0.13}
GPU 1, Tab 5, port 19998, right v-rep window
XXXX IGNORE XXXX SIM STACK - COMMON SENSE - TRIAL REWARD - FULL FEATURED RUN - costar 2020-04-28-16-15-22
----------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --max_train_actions 10000
RUN HAD PROBLEMS: Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-04-28-16-15-22_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Training
Commit: cf8fdeb86eed278fe9cb9b863662e2eaa327ebea
GPU 0, Tab 0, port 19990, left v-rep window, v-rep tab 7
XXXX IGNORE XXXX SIM ROW - COMMON SENSE - TRIAL REWARD - FULL FEATURED RUN - costar 2020-04-28-16-16-15
--------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --trial_reward --save_visualizations --common_sense --check_row --tcp_port 19998 --place --future_reward_discount 0.65 --max_train_actions 10000
RUN HAD PROBLEMS: Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-04-28-16-16-15_Sim-Rows-SPOT-Trial-Reward-Common-Sense-Training
Commit: cf8fdeb86eed278fe9cb9b863662e2eaa327ebea
GPU 1, Tab 1, port 19998, right v-rep window, v-rep tab 8
SIM STACK - COMMON SENSE - TRIAL REWARD - FULL FEATURED RUN - costar 2020-05-01-21-47-56
----------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --max_train_actions 10000
RESUME 20k: ± export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --max_train_actions 20000 --resume /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-01-21-47-56_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Training
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-01-21-47-56_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Training
RESUME: ± export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --max_train_actions 10000 --resume /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-01-21-47-56_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Training
Commit (crash): dae67d0f89fba1917e4fb89fc82f8f6171330f1f
Commit (resume): 2f9f569f0c9bfd00df480a9dbce1dba8d43b5020
GPU 0, Tab 0, port 19990, left v-rep window, v-rep tab 7
SIM ROW - COMMON SENSE - TRIAL REWARD - FULL FEATURED RUN - costar 2020-05-01-21-48-39
--------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_row --tcp_port 19998 --place --future_reward_discount 0.65 --max_train_actions 10000
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-01-21-48-39_Sim-Rows-SPOT-Trial-Reward-Common-Sense-Training
RESUME 20k: ± export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_row --tcp_port 19998 --place --future_reward_discount 0.65 --max_train_actions 20000 --resume /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-01-21-48-39_Sim-Rows-SPOT-Trial-Reward-Common-Sense-Training
RESUME: ± export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_row --tcp_port 19998 --place --future_reward_discount 0.65 --max_train_actions 10000 --resume /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-01-21-48-39_Sim-Rows-SPOT-Trial-Reward-Common-Sense-Training
RESUME2: ± export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_row --tcp_port 19999 --place --future_reward_discount 0.65 --max_train_actions 10000 --random_actions --resume '/home/ahundt/src/real_good_robot/logs/2020-05-03-20-04-47_Sim-Rows-SPOT-Trial-Reward-Common-Sense-Training'
Commit (crash): dae67d0f89fba1917e4fb89fc82f8f6171330f1f
Commit (resume): 2f9f569f0c9bfd00df480a9dbce1dba8d43b5020
Commit (resume2): 3ba0b91c5accac6387345c62d5a4e8b7ff9769cd
GPU 1, Tab 1, port 19998, right v-rep window, v-rep tab 8
SIM STACK - COMMON SENSE - TRIAL REWARD - FULL FEATURED RUN - NO TWO STEP BACKPROP - SORT TRIAL REWARD - costar
----------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --max_train_actions 10000 --disable_two_step_backprop
2020-05-05-14-26-12_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Training
Commit: 3ba0b91c5accac6387345c62d5a4e8b7ff9769cd
GPU 0, Tab 0, port 19990, left v-rep window, v-rep tab 7
SIM STACK - COMMON SENSE - TRIAL REWARD - FULL FEATURED RUN - NO TWO STEP BACKPROP - SORT TRIAL REWARD - RANDOM ACTIONS - costar
----------------------------------------------------------------------------------------
± export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --max_train_actions 10000 --disable_two_step_backprop --random_actions
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-05-16-44-34_Sim-Stack-SPOT-Trial-Reward-Commo
Commit: 3ba0b91c5accac6387345c62d5a4e8b7ff9769cd
GPU 0, Tab 0, port 19990, left v-rep window, v-rep tab 7
SIM STACK - COMMON SENSE - TRIAL REWARD - FULL FEATURED RUN - SORT TRIAL REWARD - REWARD SCHEDULE 0.25, 1, 1 - costar 2020-05-06
----------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions
'/home/costar/src/real_good_robot/logs/2020-05-06-10-03-58_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Training'
RESUME: export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions --resume '/home/costar/src/real_good_robot/logs/2020-05-06-10-03-58_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Training'
Commit: d4e776ffb89f6d916ca7ff96ebaf717bfdd45db5
Commit (resume): 7dbec777fd08d9e66b53ec72564880cebdb452e1
GPU 0, Tab 0, port 19990, left v-rep window, v-rep tab 7
Max trial success rate: 0.89, at action iteration: 4489. (total of 4491 actions, max excludes first 4489 actions)
Max grasp success rate: 0.6816311260755705, at action iteration: 4489. (total of 4491 actions, max excludes first 4489 actions)
Max place success rate: 0.6521739130434783, at action iteration: 4489. (total of 4492 actions, max excludes first 4489 actions)
Max action efficiency: 0.12029405212742258, at action iteration: 4491. (total of 4492 actions, max excludes first 4489 actions)
saving plot: 2020-05-10-01-20-13_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Testing-Sim-Stack-SPOT-Trial-Reward-Common-Sense-Testing_success_plot.png
saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-10-01-20-13_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Testing/data/best_stats.json
saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-10-01-20-13_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Testing/best_stats.json
Random Testing Complete! Dir: /home/costar/src/real_good_robot/logs/2020-05-06-10-03-58_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Training/2020-05-10-01-20-13_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Testing
Random Testing results:
{'trial_success_rate_best_value': 0.89, 'trial_success_rate_best_index': 4489, 'grasp_success_rate_best_value': 0.6816311260755705, 'grasp_success_rate_best_index': 4489, 'place_success_rate_best_value': 0.6521739130434783, 'place_success_rate_best_index': 4489, 'action_efficiency_best_value': 0.12029405212742258, 'action_efficiency_best_index': 4491}
Training Complete! Dir: /home/costar/src/real_good_robot/logs/2020-05-06-10-03-58_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Training
Training results:
{'trial_success_rate_best_value': 0.7692307692307693, 'trial_success_rate_best_index': 8827, 'grasp_success_rate_best_value': 0.8487084870848709, 'grasp_success_rate_best_index': 7472, 'place_success_rate_best_value': 0.8325991189427313, 'place_success_rate_best_index': 11194, 'action_efficiency_best_value': 0.552, 'action_efficiency_best_index': 11183}
MANUAL TESTING RUN ON action_efficiency_best_index
> export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --random_actions --snapshot_file '/home/costar/src/real_good_robot/logs/2020-05-06-10-03-58_Sim-Stack-SPOT-Trial-Reward-Common-Sense-Training/models/snapshot.reinforcement_action_efficiency_best_index.pth' --is_testing --save_visualizations --max_test_trials 100 --random_seed 1238
> Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-12-17-48-13_Sim-Stack-SPOT-Trial-Reward-Masked-Testing
> Commit: 13068e53c269b01b1385a3b185d38b006eca762b
> TODO(ahundt) move the testing directory into the training directory once complete
> TODO(ahundt) rerun this, the simulation became unstable because of placing out of arm workspace
>
>
> Testing iteration: 1665
> prev_height: 0.0 max_z: 0.0511079157217398 goal_success: True needed to reset: False max_workspace_height: -0.02 <<<<<<<<<<<
> Current count of pixels with stuff: 5131.0 threshold below which the scene is considered empty: 10
> Change detected: True (value: 5239)
> Trainer.get_label_value(): Current reward: 4.139246 Current reward multiplier: 4.139246 Predicted Future reward: 5.071979 Expected reward: 4.139246 + 0.650000 x 5.071979 = 7.436033
> trial_complete_indices: [ 6. 29. 41. 59. 67. 85. 91. 101. 109. 135. 143. 169.
> 179. 187. 192. 208. 214. 226. 265. 280. 291. 313. 323. 329.
> 342. 351. 361. 383. 387. 420. 439. 450. 477. 485. 506. 524.
> 533. 546. 576. 624. 647. 673. 681. 691. 701. 711. 729. 741.
> 752. 777. 793. 842. 872. 893. 908. 918. 930. 936. 948. 976.
> 993. 1001. 1016. 1041. 1049. 1072. 1078. 1084. 1094. 1102. 1133. 1141.
> 1149. 1168. 1174. 1184. 1210. 1221. 1245. 1255. 1271. 1278. 1293. 1297.
> 1311. 1325. 1331. 1342. 1352. 1360. 1369. 1376. 1386. 1390. 1496. 1512.
> 1522. 1541. 1545. 1636. 1665.]
> Max trial success rate: 0.98, at action iteration: 1662. (total of 1664 actions, max excludes first 1662 actions)
> Max grasp success rate: 0.5967117988394585, at action iteration: 1662. (total of 1664 actions, max excludes first 1662 actions)
> Max place success rate: 0.7615262321144675, at action iteration: 1662. (total of 1665 actions, max excludes first 1662 actions)
> Max action efficiency: 0.36101083032490977, at action iteration: 1664. (total of 1665 actions, max excludes first 1662 actions)
> saving plot: 2020-05-12-17-48-13_Sim-Stack-SPOT-Trial-Reward-Masked-Testing-Sim-Stack-SPOT-Trial-Reward-Masked-Testing_success_plot.png
> saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-12-17-48-13_Sim-Stack-SPOT-Trial-Reward-Masked-Testing/data/best_stats.json
> saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-12-17-48-13_Sim-Stack-SPOT-Trial-Reward-Masked-Testing/best_stats.json
> Trial logging complete: 101 --------------------------------------------------------------
SIM ROW - COMMON SENSE - TRIAL REWARD - FULL FEATURED RUN - SORT TRIAL REWARD - RANDOM ACTIONS - REWARD SCHEDULE 0.25, 1, 1 - costar 2020-05-06
----------------------------------------------------------------------------------------
± export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_row --tcp_port 19998 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions
'/home/costar/src/real_good_robot/logs/2020-05-06-09-59-31_Sim-Rows-SPOT-Trial-Reward-Common-Sense-Training'
RESUME: ± export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_row --tcp_port 19998 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions --resume /home/costar/src/real_good_robot/logs/2020-05-06-09-59-31_Sim-Rows-SPOT-Trial-Reward-Common-Sense-Training
Commit: d4e776ffb89f6d916ca7ff96ebaf717bfdd45db5
Commit (resume): 7dbec777fd08d9e66b53ec72564880cebdb452e1
Commit (resume2): 67bf4b2a56a4aac72a460d5d8598d38a2daac0fd
Commit (resume3 - check for full row on every place): c6c4b401fe719aae89966adaf9ed5ca24cf95fde
GPU 1, Tab 1, port 19998, right v-rep window, v-rep tab 8
Max trial success rate: 0.67, at action iteration: 2523. (total of 2525 actions, max excludes first 2523 actions)
Max grasp success rate: 0.63566388710712, at action iteration: 2523. (total of 2525 actions, max excludes first 2523 actions)
/media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/plot.py:178: RuntimeWarning: Mean of empty slice.
success_rate[i] = successes.mean()
/home/costar/.local/lib/python3.6/site-packages/numpy/core/_methods.py:161: RuntimeWarning: invalid value encountered in double_scalars
ret = ret.dtype.type(ret / rcount)
/media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/plot.py:180: RuntimeWarning: invalid value encountered in double_scalars
var = np.sqrt(success_rate[i] * (1 - success_rate[i]) / successes.shape[0])
Max action efficiency: 0.26634958382877527, at action iteration: 2525. (total of 2526 actions, max excludes first 2523 actions)
saving plot: 2020-05-10-00-30-05_Sim-Rows-SPOT-Trial-Reward-Common-Sense-Testing-Sim-Rows-SPOT-Trial-Reward-Common-Sense-Testing_success_plot.png
saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-10-00-30-05_Sim-Rows-SPOT-Trial-Reward-Common-Sense-Testing/data/best_stats.json
saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-10-00-30-05_Sim-Rows-SPOT-Trial-Reward-Common-Sense-Testing/best_stats.json
Random Testing Complete! Dir: /home/costar/src/real_good_robot/logs/2020-05-06-09-59-31_Sim-Rows-SPOT-Trial-Reward-Common-Sense-Training/2020-05-10-00-30-05_Sim-Rows-SPOT-Trial-Reward-Common-Sense-Testing
Random Testing results:
{'trial_success_rate_best_value': 0.67, 'trial_success_rate_best_index': 2523, 'grasp_success_rate_best_value': 0.63566388710712, 'grasp_success_rate_best_index': 2523, 'place_success_rate_best_value': -inf, 'place_success_rate_best_index': None, 'action_efficiency_best_value': 0.26634958382877527, 'action_efficiency_best_index': 2525}
Training Complete! Dir: /home/costar/src/real_good_robot/logs/2020-05-06-09-59-31_Sim-Rows-SPOT-Trial-Reward-Common-Sense-Training
Training results:
{'trial_success_rate_best_value': 0.5045871559633027, 'trial_success_rate_best_index': 12648, 'grasp_success_rate_best_value': 0.7074829931972789, 'grasp_success_rate_best_index': 12492, 'place_success_rate_best_value': 0.8138297872340425, 'place_success_rate_best_index': 11971, 'action_efficiency_best_value': 0.696, 'action_efficiency_best_index': 12624}
MANUAL TESTING RUN ON action_efficiency_best_index
> export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_row --tcp_port 19998 --place --future_reward_discount 0.65 --is_testing --max_test_trials 100 --random_seed 1238 --snapshot_file '/home/costar/src/real_good_robot/logs/2020-05-06-09-59-31_Sim-Rows-SPOT-Trial-Reward-Common-Sense-Training/models/snapshot.reinforcement_action_efficiency_best_index.pth'
> Commit: 13068e53c269b01b1385a3b185d38b006eca762b
> Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-12-18-04-15_Sim-Rows-SPOT-Trial-Reward-Masked-Testing
>
> TRIAL 100 SUCCESS!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
> prev_height: 0.0 max_z: 0.05112534729294889 goal_success: True needed to reset: False max_workspace_height: -0.02 <<<<<<<<<<<
> STACK: trial: 101 actions/partial: 6.153191489361702 actions/full stack: 16.431818181818183 (lower is better) Grasp Count: 866, grasp success rate: 0.6812933025404158 place_on_stack_rate: 0.4051724137931034 place_attempts: 580 partial_stack_successes: 235 stack_successes: 88 trial_success_r
> ate: 0.8712871287128713 stack goal: [2 1 3] current_height: 4
> Time elapsed: 26.519145
> Trainer iteration: 1445 complete
>
> Testing iteration: 1446
> prev_height: 0.0 max_z: 0.05111105777395382 goal_success: True needed to reset: False max_workspace_height: -0.02 <<<<<<<<<<<
> Current count of pixels with stuff: 2593.0 threshold below which the scene is considered empty: 900
> Change detected: True (value: 4472)
> Trainer.get_label_value(): Current reward: 4.000000 Current reward multiplier: 4.000000 Predicted Future reward: 6.469117 Expected reward: 4.000000 + 0.650000 x 6.469117 = 8.204926
> trial_complete_indices: [ 8. 12. 19. 24. 28. 32. 34. 53. 65. 71. 77. 102.
> 107. 113. 127. 185. 189. 193. 197. 199. 226. 230. 233. 239.
> 255. 268. 296. 300. 305. 308. 317. 327. 331. 335. 410. 424.
> 430. 448. 463. 467. 489. 550. 555. 559. 568. 572. 578. 588.
> 724. 746. 750. 765. 771. 780. 825. 836. 853. 855. 862. 866.
> 878. 921. 925. 927. 931. 935. 971. 984. 990. 1016. 1074. 1107.
> 1115. 1133. 1138. 1146. 1150. 1156. 1166. 1172. 1186. 1199. 1203. 1259.
> 1274. 1280. 1286. 1290. 1292. 1298. 1318. 1331. 1338. 1342. 1353. 1361.
> 1365. 1389. 1419. 1442. 1446.]
> Max trial success rate: 0.86, at action iteration: 1443. (total of 1445 actions, max excludes first 1443 actions)
> Max grasp success rate: 0.6809248554913295, at action iteration: 1443. (total of 1445 actions, max excludes first 1443 actions)
> Max place success rate: 0.7582037996545768, at action iteration: 1445. (total of 1446 actions, max excludes first 1443 actions) Max action efficiency: 0.3700623700623701, at action iteration: 1445. (total of 1446 actions, max excludes first 1443 actions)
> saving plot: 2020-05-12-18-04-15_Sim-Rows-SPOT-Trial-Reward-Masked-Testing-Sim-Rows-SPOT-Trial-Reward-Masked-Testing_success_plot.png
> saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-12-18-04-15_Sim-Rows-SPOT-Trial-Reward-Masked-Testing/data/best_stats.json
> saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-12-18-04-15_Sim-Rows-SPOT-Trial-Reward-Masked-Testing/best_stats.json
> Trial logging complete: 101 --------------------------------------------------------------
SIM STACK - SPOT-Q-MASKED - COMMON SENSE - TRIAL REWARD - FULL FEATURED RUN - SORT TRIAL REWARD - REWARD SCHEDULE 0.1, 1, 1 - costar 2020-05-13
----------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions
RESUME: export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions --resume /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-13-12-51-39_Sim-Stack-SPOT-Trial-Reward-Masked-Training
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-13-12-51-39_Sim-Stack-SPOT-Trial-Reward-Masked-Training
Commit: ccf30348e265a471080b3ee906065e059f6e8573
Commmit (resume, for testing, training complete): 41d2eaff3dc0f3572cecf43805de8582d62d9b31
GPU 0, Tab 0, port 19990, left v-rep window, v-rep tab 7
> Testing run prioritize trial success
> STACK: trial: 101 actions/partial: 4.525547445255475 actions/full stack: 18.979591836734695 (lower is better) Grasp Count: 1090, grasp success rate: 0.7018348623853211 place_on_stack_rate: 0.5379581151832461 place_attempts: 764 partial_stack_successes: 411 stack_successes: 98 trial_success_rate: 0.9702970297029703 stack goal: None current_height: 2.0596242913479084
> trial_complete_indices: [ 23. 29. 83. 87. 93. 98. 122. 154. 165. 188. 205. 213.
> 227. 233. 239. 279. 285. 294. 300. 334. 350. 359. 367. 437.
> 441. 447. 453. 642. 651. 669. 676. 685. 690. 696. 703. 727.
> 733. 750. 763. 769. 778. 784. 790. 805. 861. 911. 924. 931.
> 965. 972. 980. 996. 1009. 1015. 1023. 1108. 1115. 1127. 1142. 1173.
> 1179. 1186. 1208. 1268. 1274. 1294. 1307. 1313. 1325. 1329. 1335. 1377.
> 1381. 1415. 1453. 1472. 1484. 1503. 1511. 1517. 1526. 1537. 1546. 1554.
> 1605. 1612. 1622. 1630. 1676. 1707. 1716. 1740. 1747. 1753. 1775. 1803.
> 1814. 1822. 1836. 1851. 1859.]
> Max trial success rate: 0.97, at action iteration: 1856. (total of 1858 actions, max excludes first 1856 actions)
> Max grasp success rate: 0.7022058823529411, at action iteration: 1856. (total of 1858 actions, max excludes first 1856 actions)
> Max place success rate: 0.7451235370611183, at action iteration: 1856. (total of 1859 actions, max excludes first 1856 actions)
> Max action efficiency: 0.32004310344827586, at action iteration: 1858. (total of 1859 actions, max excludes first 1856 actions)
> saving plot: 2020-05-17-13-07-19_Sim-Stack-SPOT-Trial-Reward-Masked-Testing-Sim-Stack-SPOT-Trial-Reward-Masked-Testing_success_plot.png
> saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-17-13-07-19_Sim-Stack-SPOT-Trial-Reward-Masked-Testing/data/best_stats.json
> saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-17-13-07-19_Sim-Stack-SPOT-Trial-Reward-Masked-Testing/best_stats.json
> Random Testing Complete! Dir: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-13-12-51-39_Sim-Stack-SPOT-Trial-Reward-Masked-Training/2020-05-17-13-07-19_Sim-Stack-SPOT-Trial-Reward-Masked-Testing
> Random Testing results:
> {'trial_success_rate_best_value': 0.97, 'trial_success_rate_best_index': 1856, 'grasp_success_rate_best_value': 0.7022058823529411, 'grasp_success_rate_best_index': 1856, 'place_success_rate_best_value': 0.7451235370611183, 'place_success_rate_best_index': 1856, 'action_efficiency_best_value': 0.32004310344827586, 'action_efficiency_best_index': 1858}
> Training Complete! Dir: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-13-12-51-39_Sim-Stack-SPOT-Trial-Reward-Masked-Training
> Training results:
> {'action_efficiency_best_index': 10008, 'action_efficiency_best_value': 0.684, 'grasp_success_rate_best_index': 15783, 'grasp_success_rate_best_value': 0.8834586466165414, 'place_success_rate_best_index': 17570, 'place_success_rate_best_value': 0.8616071428571429, 'trial_success_rate_best_index': 10011, 'trial_success_rate_best_value': 0.8507462686567164}
> *********** 100% trial success testing **********
> Testing run prioritizing action efficiency:
> {"action_efficiency_best_index": 1325, "action_efficiency_best_value": 0.4580498866213152, "grasp_success_rate_best_index": 1323, "grasp_success_rate_best_value": 0.7697456492637216, "place_success_rate_best_index": 1325, "place_success_rate_best_value": 0.7885615251299827, "trial_success_rate_best_index": 1323, "trial_success_rate_best_value": 1.0}
> *********** 100% trial success testing **********
SIM ROW - SPOT-Q-MASKED - COMMON SENSE - TRIAL REWARD - FULL FEATURED RUN - SORT TRIAL REWARD - RANDOM ACTIONS - REWARD SCHEDULE 0.1, 1, 1 - costar 2020-05-13
----------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_row --tcp_port 19998 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions
RESUME: export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_row --tcp_port 19998 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions --resume /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-13-12-21-00_Sim-Rows-SPOT-Trial-Reward-Masked-Training
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-13-12-21-00_Sim-Rows-SPOT-Trial-Reward-Masked-Training
Commit: ccf30348e265a471080b3ee906065e059f6e8573
Commmit (resume, for testing, training complete): 41d2eaff3dc0f3572cecf43805de8582d62d9b31
GPU 1, Tab 1, port 19998, right v-rep window, v-rep tab 8
> Testing run prioritize trial successes
> STACK: trial: 101 actions/partial: 6.809885931558935 actions/full stack: 18.852631578947367 (lower is better) Grasp Count: 1085, grasp success rate: 0.6552995391705069 place_on_stack_rate: 0.37252124645892354 place_attempts: 706 partial_stack_successes: 263 stack_successes: 95 trial_success_rate: 0.9405940594059405 stack goal: [2] current_height: 1
> trial_complete_indices: [ 6. 10. 18. 28. 58. 62. 69. 87. 94. 102. 108. 115.
> 121. 175. 181. 186. 188. 196. 201. 208. 217. 219. 228. 258.
> 269. 296. 340. 346. 350. 354. 366. 469. 474. 478. 490. 494.
> 505. 521. 547. 557. 599. 628. 648. 676. 680. 684. 740. 772.
> 783. 794. 799. 812. 817. 908. 916. 925. 931. 940. 947. 953.
> 966. 981. 991. 1000. 1066. 1146. 1154. 1188. 1196. 1200. 1294. 1308.
> 1312. 1320. 1324. 1338. 1344. 1381. 1389. 1397. 1401. 1417. 1441. 1481.
> 1485. 1491. 1546. 1582. 1588. 1595. 1601. 1648. 1664. 1706. 1720. 1727.
> 1729. 1767. 1769. 1785. 1790.]
> Max trial success rate: 0.94, at action iteration: 1787. (total of 1789 actions, max excludes first 1787 actions)
> Max grasp success rate: 0.6555863342566943, at action iteration: 1787. (total of 1789 actions, max excludes first 1787 actions)
> Max action efficiency: 0.37604924454392835, at action iteration: 1789. (total of 1790 actions, max excludes first 1787 actions)
> saving plot: 2020-05-17-13-08-59_Sim-Rows-SPOT-Trial-Reward-Masked-Testing-Sim-Rows-SPOT-Trial-Reward-Masked-Testing_success_plot.png
> saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-17-13-08-59_Sim-Rows-SPOT-Trial-Reward-Masked-Testing/data/best_stats.json
> saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-17-13-08-59_Sim-Rows-SPOT-Trial-Reward-Masked-Testing/best_stats.json
> Random Testing Complete! Dir: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-13-12-21-00_Sim-Rows-SPOT-Trial-Reward-Masked-Training/2020-05-17-13-08-59_Sim-Rows-SPOT-Trial-Reward-Masked-Testing
> Random Testing results:
> Random testing results after manual bugfix: {"action_efficiency_best_index": 1789, "action_efficiency_best_value": 0.3764705882352941, "grasp_success_rate_best_index": 1785, "grasp_success_rate_best_value": 0.6561922365988909, "place_success_rate_best_index": 1785, "place_success_rate_best_value": 0.7634560906515581, "trial_success_rate_best_index": 1787, "trial_success_rate_best_value": 0.94}
> XXX SEE correction above {'trial_success_rate_best_value': 0.94, 'trial_success_rate_best_index': 1787, 'grasp_success_rate_best_value': 0.6555863342566943, 'grasp_success_rate_best_index': 1787, 'place_success_rate_best_value': -inf, 'place_success_rate_best_index': None, 'action_efficiency_best_value': 0.37604924454392835, 'action_efficiency_best_index': 1789}
> Training Complete! Dir: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-13-12-21-00_Sim-Rows-SPOT-Trial-Reward-Masked-Training
> Training results:
> {'action_efficiency_best_index': 12997, 'action_efficiency_best_value': 0.672, 'grasp_success_rate_best_index': 17482, 'grasp_success_rate_best_value': 0.6845637583892618, 'place_success_rate_best_index': 12935, 'place_success_rate_best_value': 0.8296703296703297, 'trial_success_rate_best_index': 12367, 'trial_success_rate_best_value': 0.5662650602409639}
> Testing run prioritizing action efficiency
> STACK: trial: 101 actions/partial: 6.560975609756097 actions/full stack: 17.543478260869566 (lower is better) Grasp Count: 1018, grasp success rate: 0.5943025540275049 place_on_stack_rate: 0.412751677852349 place_attempts: 596 partial_stack_successes: 246 stack_successes: 92 trial_success_rate: 0.9108910891089109 stack goal: [0 2] current_height: 2
> trial_complete_indices: [ 16. 23. 31. 58. 76. 81. 100. 104. 157. 168. 172. 252.
> 262. 266. 271. 360. 393. 397. 405. 411. 415. 423. 432. 437.
> 445. 460. 473. 571. 579. 588. 592. 599. 619. 626. 632. 747.
> 753. 769. 782. 788. 794. 816. 877. 887. 897. 903. 905. 974.
> 978. 986. 994. 1005. 1012. 1016. 1020. 1026. 1039. 1043. 1088. 1137.
> 1139. 1148. 1155. 1164. 1172. 1183. 1233. 1242. 1253. 1257. 1264. 1275.
> 1287. 1290. 1297. 1309. 1313. 1317. 1322. 1342. 1350. 1354. 1359. 1402.
> 1443. 1452. 1458. 1462. 1469. 1473. 1482. 1489. 1495. 1504. 1506. 1518.
> 1523. 1532. 1541. 1569. 1613.]
> Max trial success rate: 0.92, at action iteration: 1610. (total of 1612 actions, max excludes first 1610 actions)
> Max grasp success rate: 0.5950738916256157, at action iteration: 1610. (total of 1612 actions, max excludes first 1610 actions)
> /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/plot.py:178: RuntimeWarning: Mean of empty slice.
> success_rate[i] = successes.mean()
> /home/costar/.local/lib/python3.6/site-packages/numpy/core/_methods.py:161: RuntimeWarning: invalid value encountered in double_scalars
> ret = ret.dtype.type(ret / rcount)
> /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/plot.py:180: RuntimeWarning: invalid value encountered in double_scalars
> var = np.sqrt(success_rate[i] * (1 - success_rate[i]) / successes.shape[0])
> Max action efficiency: 0.3875776397515528, at action iteration: 1610. (total of 1613 actions, max excludes first 1610 actions)
> saving plot: 2020-05-17-22-05-52_Sim-Rows-SPOT-Trial-Reward-Masked-Testing-Sim-Rows-SPOT-Trial-Reward-Masked-Testing_success_plot.png
> saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-17-22-05-52_Sim-Rows-SPOT-Trial-Reward-Masked-Testing/data/best_stats.json
> saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-17-22-05-52_Sim-Rows-SPOT-Trial-Reward-Masked-Testing/best_stats.json
> Random Testing Complete! Dir: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-13-12-21-00_Sim-Rows-SPOT-Trial-Reward-Masked-Training/2020-05-17-22-05-52_Sim-Rows-SPOT-Trial-Reward-Masked-Testing
> Random Testing results:
> {'trial_success_rate_best_value': 0.92, 'trial_success_rate_best_index': 1610, 'grasp_success_rate_best_value': 0.5950738916256157, 'grasp_success_rate_best_index': 1610, 'place_success_rate_best_value': -inf, 'place_success_rate_best_index': None, 'action_efficiency_best_value': 0.3875776397515528, 'action_efficiency_best_index': 1610}
> Training Complete! Dir: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-13-12-21-00_Sim-Rows-SPOT-Trial-Reward-Masked-Training
> Training results:
> {'action_efficiency_best_index': 12997, 'action_efficiency_best_value': 0.672, 'grasp_success_rate_best_index': 17482, 'grasp_success_rate_best_value': 0.6845637583892618, 'place_success_rate_best_index': 12935, 'place_success_rate_best_value':
> 0.8296703296703297, 'trial_success_rate_best_index': 12367, 'trial_success_rate_best_value': 0.5662650602409639}
SIM STACK - SPOT STANDARD - TRIAL REWARD - RANDOM ACTIONS - SORT TRIAL REWARD - REWARD SCHEDULE 0.1, 1, 1 - costar 2020-05-18
----------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions
RESUME: export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions --resume /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-18-19-56-49_Sim-Stack-SPOT-Trial-Reward-Training
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-18-19-56-49_Sim-Stack-SPOT-Trial-Reward-Training
Commit: e99391ae3c0921bd95b5b5d2a7d6e992efa69d63
GPU 0, Tab 0, port 19990, left v-rep window, v-rep tab 7
> Test run
> {"action_efficiency_best_index": 1966, "action_efficiency_best_value": 0.29633401221995925, "grasp_success_rate_best_index": 1964, "grasp_success_rate_best_value": 0.689328743545611, "place_success_rate_best_index": 1964, "place_success_rate_best_value": 0.6961394769613948, "trial_success_rate_best_index": 1964, "trial_success_rate_best_value": 0.95}
> '/home/costar/src/real_good_robot/logs/2020-05-18-19-56-49_Sim-Stack-SPOT-Trial-Reward-Training/2020-05-22-12-55-27_Sim-Stack-SPOT-Trial-Reward-Testing/best_stats.json'
SIM ROW - SPOT STANDARD - TRIAL REWARD - RANDOM ACTIONS - SORT TRIAL REWARD - REWARD SCHEDULE 0.1, 1, 1 - costar 2020-05-18
----------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --trial_reward --check_row --tcp_port 19998 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-18-19-57-17_Sim-Rows-SPOT-Trial-Reward-Training
Commit: e99391ae3c0921bd95b5b5d2a7d6e992efa69d63
GPU 1, Tab 1, port 19998, right v-rep window, v-rep tab 8
> testing run prioritizing action efficiency
> export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --trial_reward --check_row --tcp_port 19998 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions --resume '/home/costar/src/real_good_robot/logs/2020-05-18-19-57-17_Sim-Rows-SPOT-Trial-Reward-Training'
> Trial logging complete: 101 --------------------------------------------------------------
> Running two step backprop()
> Primitive confidence scores: 0.783435 (push), 4.346649 (grasp), 8.664590 (place)
> Action: grasp at (8, 87, 152)
> Training loss: 0.981009
> Executing: grasp at (-0.420000, -0.050000, 0.050996) orientation: 3.141593
> gripper position: 0.029739439487457275
> gripper position: 0.02550262212753296
> gripper position: 0.004024624824523926
> gripper position: 0.003876298666000366
> Grasp successful: True
> prev_height: 0.0 max_z: 0.051110193888699765 goal_success: True needed to reset: False max_workspace_height: -0.02 <<<<<<<<<<<
> check_row: True | row_size: 2 | blocks: ['blue' 'yellow']
> check_stack() stack_height: 2 stack matches current goal: True partial_stack_success: True Does the code think a reset is needed: False
> STACK: trial: 101 actions/partial: 13.72289156626506 actions/full stack: 39.275862068965516 (lower is better) Grasp Count: 1899, grasp success rate: 0.8072669826224329 place_on_stack_rate: 0.16403162055335968 place_attempts: 1518 partial_stack_successes: 249 stack_successes: 87 trial_success_rate: 0.8613861386138614 stack goal: [0 1] current_height: 2
> trial_complete_indices: [ 143. 210. 268. 274. 280. 305. 307. 319. 323. 388. 450. 454.
> 475. 477. 624. 643. 649. 656. 667. 671. 675. 776. 788. 817.
> 821. 846. 850. 927. 929. 942. 981. 994. 1003. 1007. 1016. 1059.
> 1224. 1234. 1240. 1251. 1257. 1261. 1263. 1307. 1310. 1321. 1325. 1334.
> 1342. 1351. 1361. 1371. 1391. 1665. 1670. 1674. 1680. 1686. 1693. 1701.
> 1717. 1723. 1742. 1759. 1763. 1769. 1823. 1831. 1848. 1854. 2335. 2597.
> 2605. 2618. 2624. 2632. 2636. 2744. 2758. 2766. 2772. 2959. 2961. 2967.
> 2981. 2983. 2985. 3006. 3010. 3019. 3170. 3182. 3206. 3214. 3216. 3220.
> 3224. 3235. 3283. 3367. 3416.]
> Max trial success rate: 0.87, at action iteration: 3413. (total of 3415 actions, max excludes first 3413 actions)
> Max grasp success rate: 0.8074894514767933, at action iteration: 3413. (total of 3415 actions, max excludes first 3413 actions)
> Max place success rate: 0.6021080368906456, at action iteration: 3413. (total of 3414 actions, max excludes first 3413 actions)
> Max action efficiency: 0.1652505127453853, at action iteration: 3413. (total of 3416 actions, max excludes first 3413 actions)
> saving plot: 2020-05-23-16-29-39_Sim-Rows-SPOT-Trial-Reward-Testing-Sim-Rows-SPOT-Trial-Reward-Testing_success_plot.png
> saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-23-16-29-39_Sim-Rows-SPOT-Trial-Reward-Testing/data/best_stats.json
> saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-23-16-29-39_Sim-Rows-SPOT-Trial-Reward-Testing/best_stats.json
> Choosing a snapshot from the following options:{'action_efficiency_best_index': 19725, 'action_efficiency_best_value': 0.576, 'grasp_success_rate_best_index': 17982, 'grasp_success_rate_best_value': 0.9609375, 'place_success_rate_best_index': 1949, 'place_success_rate_best_value': 0.8333333333333334, 'trial_success_rate_best_index': 18012, 'trial_success_rate_best_value': 0.5714285714285714}
> Evaluating trial_success_rate_best_value
> The trial_success_rate_best_value is fantastic at 0.5714285714285714, so we will look for the best action_efficiency_best_value.
> Shapshot chosen: /home/costar/src/real_good_robot/logs/2020-05-18-19-57-17_Sim-Rows-SPOT-Trial-Reward-Training/models/snapshot.reinforcement_action_efficiency_best_value.pth
> Random Testing Complete! Dir: /home/costar/src/real_good_robot/logs/2020-05-18-19-57-17_Sim-Rows-SPOT-Trial-Reward-Training/2020-05-23-16-29-39_Sim-Rows-SPOT-Trial-Reward-Testing
> Random Testing results:
> {'trial_success_rate_best_value': 0.87, 'trial_success_rate_best_index': 3413, 'grasp_success_rate_best_value': 0.8074894514767933, 'grasp_success_rate_best_index': 3413, 'place_success_rate_best_value': 0.6021080368906456, 'place_success_rate_best_index': 3413, 'action_efficiency_best_value': 0.1652505127453853, 'action_efficiency_best_index': 3413}
> Training Complete! Dir: /home/costar/src/real_good_robot/logs/2020-05-18-19-57-17_Sim-Rows-SPOT-Trial-Reward-Training
> Training results:
> {'action_efficiency_best_index': 19725, 'action_efficiency_best_value': 0.576, 'grasp_success_rate_best_index': 17982, 'grasp_success_rate_best_value': 0.9609375, 'place_success_rate_best_index': 1949, 'place_success_rate_best_value': 0.8333333333333334, 'trial_success_rate_best_index': 18012, 'trial_success_rate_best_value': 0.5714285714285714}
SIM STACK - SPOT Masked no SPOT-Q (alg 1 if statement mask backprop is disabled) - TRIAL REWARD - RANDOM ACTIONS - SORT TRIAL REWARD - REWARD SCHEDULE 0.1, 1, 1 - costar 2020-05-23
----------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions --trial_reward --common_sense --no_common_sense_backprop
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-23-14-31-09_Sim-Stack-SPOT-Trial-Reward-Masked-Training
Commit: a534735959ec2747c3b134a6d3067135a5c7bd75 release tag:v0.16.0
GPU 0, Tab 0, port 19990, left v-rep window, v-rep tab 7
> Trial logging complete: 101 --------------------------------------------------------------
> Running two step backprop()
> Primitive confidence scores: 1.434610 (push), 1.922926 (grasp), 2.361718 (place)
> Action: grasp at (4, 151, 135)
> Training loss: 3.844077
> Executing: grasp at (-0.454000, 0.078000, 0.001002) orientation: 1.570796
> gripper position: 0.030432865023612976
> gripper position: 0.026735419407486916
> gripper position: 0.0015385448932647705
> gripper position: -0.02276727557182312
> gripper position: -0.042291462421417236
> Grasp successful: False
> prev_height: 0.0 max_z: 0.10307922494011586 goal_success: True needed to reset: False max_workspace_height: -0.02 <<<<<<<<<<<
> prev_height: 1.0 max_z: 2.061584498802317 goal_success: True needed to reset: False max_workspace_height: 0.6 <<<<<<<<<<<
> check_stack() stack_height: 2.061584498802317 stack matches current goal: True partial_stack_success: True Does the code think a reset is needed: False
> STACK: trial: 101 actions/partial: 3.2612612612612613 actions/full stack: 10.86 (lower is better) Grasp Count: 593, grasp success rate: 0.8145025295109612 place_on_stack_rate: 0.6894409937888198 place_attempts: 483 partial_stack_successes: 333 stack_successes: 100 trial_success_rate: 0.9900990099009901 stack goal: None current_height: 2.061584498802317
> trial_complete_indices: [ 6. 18. 24. 30. 76. 82. 86. 95. 130. 136. 140. 161.
> 167. 177. 183. 199. 212. 224. 233. 239. 245. 251. 260. 272.
> 278. 284. 294. 300. 304. 317. 327. 335. 343. 347. 353. 360.
> 371. 391. 399. 405. 413. 417. 425. 431. 437. 445. 451. 455.
> 476. 486. 497. 518. 524. 549. 559. 565. 577. 582. 616. 625.
> 631. 639. 649. 655. 671. 685. 694. 698. 704. 724. 731. 742.
> 762. 827. 833. 841. 847. 854. 860. 867. 873. 908. 914. 934.
> 940. 946. 952. 963. 969. 977. 993. 999. 1007. 1016. 1027. 1038.
> 1044. 1053. 1065. 1071. 1085.]
> Max trial success rate: 0.99, at action iteration: 1082. (total of 1084 actions, max excludes first 1082 actions)
> Max grasp success rate: 0.8155668358714044, at action iteration: 1082. (total of 1084 actions, max excludes first 1082 actions)
> Max place success rate: 0.790650406504065, at action iteration: 1082. (total of 1085 actions, max excludes first 1082 actions)
> Max action efficiency: 0.5545286506469501, at action iteration: 1084. (total of 1085 actions, max excludes first 1082 actions)
> saving plot: 2020-05-27-04-58-39_Sim-Stack-SPOT-Trial-Reward-Masked-Testing-Sim-Stack-SPOT-Trial-Reward-Masked-Testing_success_plot.png
> saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-27-04-58-39_Sim-Stack-SPOT-Trial-Reward-Masked-Testing/data/best_stats.json
> saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-27-04-58-39_Sim-Stack-SPOT-Trial-Reward-Masked-Testing/best_stats.json
> Choosing a snapshot from the following options:{'trial_success_rate_best_value': 0.8913043478260869, 'trial_success_rate_best_index': 16675, 'grasp_success_rate_best_value': 0.8388278388278388, 'grasp_success_rate_best_index': 19892, 'place_success_rate_best_value': 0.8356164383561644, 'place_success_rate_best_index': 15066, 'action_efficiency_best_value': 0.576, 'action_efficiency_best_index': 18579}
> Evaluating trial_success_rate_best_value
> The trial_success_rate_best_value is fantastic at 0.8913043478260869, so we will look for the best action_efficiency_best_value.
> Shapshot chosen: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-23-14-31-09_Sim-Stack-SPOT-Trial-Reward-Masked-Training/models/snapshot.reinforcement_action_efficiency_best_value.pth
> Random Testing Complete! Dir: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-23-14-31-09_Sim-Stack-SPOT-Trial-Reward-Masked-Training/2020-05-27-04-58-39_Sim-Stack-SPOT-Trial-Reward-Masked-Testing
> Random Testing results:
> {'trial_success_rate_best_value': 0.99, 'trial_success_rate_best_index': 1082, 'grasp_success_rate_best_value': 0.8155668358714044, 'grasp_success_rate_best_index': 1082, 'place_success_rate_best_value': 0.790650406504065, 'place_success_rate_best_index': 1082, 'action_efficiency_best_value': 0.5545286506469501, 'action_efficiency_best_index': 1084}
> Training Complete! Dir: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-23-14-31-09_Sim-Stack-SPOT-Trial-Reward-Masked-Training
> Training results:
> {'trial_success_rate_best_value': 0.8913043478260869, 'trial_success_rate_best_index': 16675, 'grasp_success_rate_best_value': 0.8388278388278388, 'grasp_success_rate_best_index': 19892, 'place_success_rate_best_value': 0.8356164383561644, 'place_success_rate_best_index': 15066, 'action_efficiency_best_value': 0.576, 'action_efficiency_best_index': 18579}
SIM ROW - SPOT Masked no SPOT-Q (alg 1 if statement mask backprop is disabled) - TRIAL REWARD - RANDOM ACTIONS - SORT TRIAL REWARD - REWARD SCHEDULE 0.1, 1, 1 - costar 2020-05-23
----------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --trial_reward --check_row --tcp_port 19998 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions --trial_reward --common_sense --no_common_sense_backprop
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-24-09-36-39_Sim-Rows-SPOT-Trial-Reward-Masked-Training
Commit: a534735959ec2747c3b134a6d3067135a5c7bd75 release tag:v0.16.0
GPU 1, Tab 1, port 19998, right v-rep window, v-rep tab 8
> /home/costar/src/real_good_robot/logs/2020-05-24-09-36-39_Sim-Rows-SPOT-Trial-Reward-Masked-Training/2020-05-28-03-27-32_Sim-Rows-SPOT-Trial-Reward-Masked-Testing/best_stats.json
> {"action_efficiency_best_index": 1189, "action_efficiency_best_value": 0.4894869638351556, "grasp_success_rate_best_index": 1189, "grasp_success_rate_best_value": 0.7434402332361516, "place_success_rate_best_index": 1189, "place_success_rate_best_value": 0.8452380952380952, "trial_success_rate_best_index": 1189, "trial_success_rate_best_value": 0.93}
SIM STACK - SPOT STANDARD progress TRIAL aka rtrial - TRIAL REWARD - RANDOM ACTIONS - SORT TRIAL REWARD - REWARD SCHEDULE 0.1, 1, 1 - costar 2020-05-27
----------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions
RESUME: ± export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions --resume /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-27-12-35-08_Sim-Stack-SPOT-Trial-Reward-Training
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-27-12-35-08_Sim-Stack-SPOT-Trial-Reward-Training
Commit: a534735959ec2747c3b134a6d3067135a5c7bd75 release tag:v0.16.0
GPU 0, Tab 0, port 19990, left v-rep window, v-rep tab 7
> Trial logging complete: 101 --------------------------------------------------------------
> Running two step backprop()
> Primitive confidence scores: 0.926641 (push), 2.734785 (grasp), 6.851398 (place)
> Action: grasp at (0, 103, 151)
> Training loss: 0.528587
> Executing: grasp at (-0.422000, -0.018000, 0.001003) orientation: 0.000000
> gripper position: 0.03009691834449768
> gripper position: 0.0258101224899292
> gripper position: 0.0006317198276519775
> gripper position: -0.02364581823348999
> gripper position: -0.04264447093009949
> Grasp successful: False
> prev_height: 0.0 max_z: 0.051131368098522104 goal_success: True needed to reset: False max_workspace_height: -0.02 <<<<<<<<<<<
> prev_height: 1.0 max_z: 1.022627361970442 goal_success: False needed to reset: False max_workspace_height: 0.6 <<<<<<<<<<<
> check_stack() stack_height: 1.022627361970442 stack matches current goal: False partial_stack_success: False Does the code think a reset is needed: False
> STACK: trial: 101 actions/partial: 4.434782608695652 actions/full stack: 18.545454545454547 (lower is better) Grasp Count: 1082, grasp success rate: 0.6977818853974121 place_on_stack_rate: 0.5490716180371353 place_attempts: 754 partial_stack_successes: 414 stack_successes: 99 trial_success_rate: 0.9801980198019802 stack goal: None current_height: 1.022627361970442
> trial_complete_indices: [ 10. 16. 26. 66. 72. 118. 124. 128. 136. 186. 235. 241.
> 252. 263. 382. 411. 444. 454. 458. 491. 529. 566. 579. 586.
> 671. 680. 781. 787. 795. 827. 875. 883. 891. 919. 931. 942.
> 957. 974. 982. 997. 1001. 1009. 1019. 1031. 1044. 1066. 1107. 1134.
> 1155. 1163. 1169. 1192. 1201. 1219. 1225. 1248. 1254. 1281. 1301. 1311.
> 1317. 1345. 1351. 1390. 1394. 1404. 1415. 1421. 1431. 1452. 1458. 1469.
> 1475. 1485. 1522. 1542. 1549. 1563. 1578. 1585. 1609. 1620. 1642. 1646.
> 1658. 1669. 1679. 1705. 1711. 1717. 1725. 1736. 1742. 1748. 1762. 1785.
> 1793. 1806. 1813. 1818. 1835.]
> Max trial success rate: 0.98, at action iteration: 1832. (total of 1834 actions, max excludes first 1832 actions)
> Max grasp success rate: 0.6981481481481482, at action iteration: 1832. (total of 1834 actions, max excludes first 1832 actions)
> Max place success rate: 0.7861885790172642, at action iteration: 1834. (total of 1835 actions, max excludes first 1832 actions)
> Max action efficiency: 0.324235807860262, at action iteration: 1834. (total of 1835 actions, max excludes first 1832 actions)
> saving plot: 2020-05-31-05-18-07_Sim-Stack-SPOT-Trial-Reward-Testing-Sim-Stack-SPOT-Trial-Reward-Testing_success_plot.png
> saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-31-05-18-07_Sim-Stack-SPOT-Trial-Reward-Testing/data/best_stats.json
> saving best stats to: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-31-05-18-07_Sim-Stack-SPOT-Trial-Reward-Testing/best_stats.json
> Choosing a snapshot from the following options:{'trial_success_rate_best_value': 0.8157894736842105, 'trial_success_rate_best_index': 10807, 'grasp_success_rate_best_value': 0.8550185873605948, 'grasp_success_rate_best_index': 10825, 'place_success_rate_best_value': 0.7741935483870968, 'place_success_rate_best_index': 13745, 'action_efficiency_best_value': 0.384, 'action_efficiency_best_index': 10746}
> Evaluating trial_success_rate_best_value
> Shapshot chosen: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-27-12-35-08_Sim-Stack-SPOT-Trial-Reward-Training/models/snapshot.reinforcement_trial_success_rate_best_value.pth
> Random Testing Complete! Dir: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-27-12-35-08_Sim-Stack-SPOT-Trial-Reward-Training/2020-05-31-05-18-07_Sim-Stack-SPOT-Trial-Reward-Testing
> Random Testing results:
> {'trial_success_rate_best_value': 0.98, 'trial_success_rate_best_index': 1832, 'grasp_success_rate_best_value': 0.6981481481481482, 'grasp_success_rate_best_index': 1832, 'place_success_rate_best_value': 0.7861885790172642, 'place_success_rate_best_index': 1834, 'action_efficiency_best_value': 0.324235807860262, 'action_efficiency_best_index': 1834}
> Training Complete! Dir: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-27-12-35-08_Sim-Stack-SPOT-Trial-Reward-Training
> Training results:
> {'trial_success_rate_best_value': 0.8157894736842105, 'trial_success_rate_best_index': 10807, 'grasp_success_rate_best_value': 0.8550185873605948, 'grasp_success_rate_best_index': 10825, 'place_success_rate_best_value': 0.7741935483870968, 'place_success_rate_best_index': 13745, 'action_efficiency_best_value': 0.384, 'action_efficiency_best_index': 10746}
XXXX BAD RUN XXXX - SIM ROW - SPOT STANDARD progress TRIAL aka rtrial - TRIAL REWARD - RANDOM ACTIONS - SORT TRIAL REWARD - REWARD SCHEDULE 0.1, 1, 1 - costar 2020-05-23
----------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --trial_reward --check_row --tcp_port 19998 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions
RESUME: export CUDA_VISIBLE_DEVICES="1" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 4 --push_rewards --experience_replay --explore_rate_decay --trial_reward --check_row --tcp_port 19998 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions --resume /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-28-10-46-31_Sim-Rows-SPOT-Trial-Reward-Training
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-28-10-46-31_Sim-Rows-SPOT-Trial-Reward-Training
Commit: a534735959ec2747c3b134a6d3067135a5c7bd75 release tag:v0.16.0
GPU 1, Tab 1, port 19998, right v-rep window, v-rep tab 8
> The simulator state went bad in this run (not the training algorithm), probably the sim causing the robot arm to go elbow down and the row detector to see false positives, and the good models were thus overwritten. This run cannot be used in results.
XXXX The trial_success_rate_best_value is fantastic at 1.0, so we will look for the best grasp_success_rate_best_value.
XXXX The trial_success_rate_best_value is fantastic at 1.0, so we will look for the best action_efficiency_best_value. Shapshot chosen: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-28-10-46-31_Sim-Rows-SPOT-Trial-Reward-Training/models/snapshot.reinforcement_action_efficiency_best_value.pth
XXXX Random Testing Complete! Dir: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-28-10-46-31_Sim-Rows-SPOT-Trial-Reward-Training/2020-06-01-00-55-27_Sim-Rows-SPOT-Trial-Reward-Testing
XXXX Random Testing results:
XXXX {'trial_success_rate_best_value': 0.74, 'trial_success_rate_best_index': 2430, 'grasp_success_rate_best_value': 0.6106719367588933, 'grasp_success_rate_best_index': 2430, 'place_success_rate_best_value' : 0.7897042716319824, 'place_success_rate_best_index': 2430, 'action_efficiency_best_value': 0.2, 'action_efficiency_best_index': 2430}
XXXX Training Complete! Dir: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-28-10-46-31_Sim-Rows-SPOT-Trial-Reward-Training
XXXX Training results:
XXXX {'trial_success_rate_best_value': 1.0, 'trial_success_rate_best_index': 15761, 'grasp_success_rate_best_value': 0.9960159362549801, 'grasp_success_rate_best_index': 15763, 'place_success_rate_best_value': 0.7959183673469388, 'place_success_rate_best_index': 12706, 'action_efficiency_best_value': 0.588, 'action_efficiency_best_index': 10820}
SIM STACK - SPOT-Q-MASKED SPOT FRAMEWORK - COMMON SENSE - TRIAL REWARD - FULL FEATURED RUN - SORT TRIAL REWARD - REWARD SCHEDULE 0.1, 1, 1 - costar 2020-05-31
----------------------------------------------------------------------------------------
export CUDA_VISIBLE_DEVICES="0" && python3 main.py --is_sim --obj_mesh_dir objects/blocks --num_obj 8 --push_rewards --experience_replay --explore_rate_decay --trial_reward --common_sense --check_z_height --tcp_port 19990 --place --future_reward_discount 0.65 --max_train_actions 20000 --random_actions
Creating data logging session: /media/costar/f5f1f858-3666-4832-beea-b743127f1030/real_good_robot/logs/2020-05-31-17-25-41_Sim-Stack-SPOT-Trial-Reward-Masked-Training
Commit: 12d9481717486342dbfcaff191ddb1428f102406 release tag:v0.16.1
GPU 0, Tab 0, port 19990, left v-rep window, v-rep tab 7
> Trial logging complete: 101 --------------------------------------------------------------
> Running two step backprop()
> Primitive confidence scores: 0.580019 (push), 4.184012 (grasp), 8.394482 (place)
> Action: grasp at (0, 73, 119)
> Training loss: 0.248444
> Executing: grasp at (-0.486000, -0.078000, 0.001000) orientation: 0.000000
> gripper position: 0.03083541989326477
> gripper position: 0.026231884956359863
> gripper position: 0.0011520087718963623
> gripper position: -0.023060262203216553
> gripper position: -0.04178208112716675
> gripper position: -0.044988662004470825
> Grasp successful: False
> prev_height: 0.0 max_z: 0.05113248210487194 goal_success: True needed to reset: False max_workspace_height: -0.02 <<<<<<<<<<<
> prev_height: 1.0 max_z: 1.0226496420974387 goal_success: False needed to reset: False max_workspace_height: 0.6 <<<<<<<<<<<
> check_stack() stack_height: 1.0226496420974387 stack matches current goal: False partial_stack_success: False Does the code think a reset is needed: False
> STACK: trial: 101 actions/partial: 3.913793103448276 actions/full stack: 13.485148514851485 (lower is better) Grasp Count: 801, grasp success rate: 0.700374531835206 place_on_stack_rate: 0.6203208556149733 place_attempts: 561 partial_stack_successes: 348 stack_successes: 101 trial_success_rate: 1.0 stack goal: None current_height: 1.0226496420974387
> trial_complete_indices: [ 9. 38. 51. 64. 79. 87. 94. 98. 102. 108. 122. 162.
> 171. 190. 198. 210. 219. 225. 229. 235. 241. 259. 273. 295.
> 305. 311. 331. 354. 360. 368. 380. 386. 404. 419. 427. 451.
> 469. 475. 511. 522. 528. 538. 561. 575. 579. 587. 612. 622.
> 642. 648. 660. 678. 689. 696. 708. 723. 768. 777. 798. 812.
> 821. 827. 833. 839. 850. 858. 885. 895. 911. 931. 937. 958.
> 966. 980. 991. 995. 1002. 1073. 1090. 1103. 1115. 1170. 1189. 1199.
> 1209. 1217. 1225. 1231. 1244. 1261. 1272. 1280. 1287. 1291. 1300. 1309.
> 1329. 1341. 1348. 1354. 1361.]
> Max trial success rate: 1.0, at action iteration: 1358. (total of 1360 actions, max excludes first 1358 actions)