Asilarkness commited on
Commit
8cf72af
·
0 Parent(s):

Super-squash branch 'main' using huggingface_hub

Browse files
.gitattributes ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ *.7z filter=lfs diff=lfs merge=lfs -text
2
+ *.arrow filter=lfs diff=lfs merge=lfs -text
3
+ *.bin filter=lfs diff=lfs merge=lfs -text
4
+ *.bz2 filter=lfs diff=lfs merge=lfs -text
5
+ *.ckpt filter=lfs diff=lfs merge=lfs -text
6
+ *.ftz filter=lfs diff=lfs merge=lfs -text
7
+ *.gz filter=lfs diff=lfs merge=lfs -text
8
+ *.h5 filter=lfs diff=lfs merge=lfs -text
9
+ *.joblib filter=lfs diff=lfs merge=lfs -text
10
+ *.lfs.* filter=lfs diff=lfs merge=lfs -text
11
+ *.mlmodel filter=lfs diff=lfs merge=lfs -text
12
+ *.model filter=lfs diff=lfs merge=lfs -text
13
+ *.msgpack filter=lfs diff=lfs merge=lfs -text
14
+ *.npy filter=lfs diff=lfs merge=lfs -text
15
+ *.npz filter=lfs diff=lfs merge=lfs -text
16
+ *.onnx filter=lfs diff=lfs merge=lfs -text
17
+ *.ot filter=lfs diff=lfs merge=lfs -text
18
+ *.parquet filter=lfs diff=lfs merge=lfs -text
19
+ *.pb filter=lfs diff=lfs merge=lfs -text
20
+ *.pickle filter=lfs diff=lfs merge=lfs -text
21
+ *.pkl filter=lfs diff=lfs merge=lfs -text
22
+ *.pt filter=lfs diff=lfs merge=lfs -text
23
+ *.pth filter=lfs diff=lfs merge=lfs -text
24
+ *.rar filter=lfs diff=lfs merge=lfs -text
25
+ *.safetensors filter=lfs diff=lfs merge=lfs -text
26
+ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
27
+ *.tar.* filter=lfs diff=lfs merge=lfs -text
28
+ *.tar filter=lfs diff=lfs merge=lfs -text
29
+ *.tflite filter=lfs diff=lfs merge=lfs -text
30
+ *.tgz filter=lfs diff=lfs merge=lfs -text
31
+ *.wasm filter=lfs diff=lfs merge=lfs -text
32
+ *.xz filter=lfs diff=lfs merge=lfs -text
33
+ *.zip filter=lfs diff=lfs merge=lfs -text
34
+ *.zst filter=lfs diff=lfs merge=lfs -text
35
+ *tfevents* filter=lfs diff=lfs merge=lfs -text
elo_25_b20f.txt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ ckpt round=25 step=24400 pos=97894400
2
+ sf_elo=1320 nodes=60
3
+ games=4 W0 D4 L0 score=0.500 elo=1320
elo_27_b70f.txt ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ ckpt round=27 step=26200 pos=104038400
2
+ sf_elo=1320 nodes=60
3
+ games=2 W0 D2 L0 score=0.500 elo=1320
4
+ why={'repetition': 2}
elo_27_c2ac.txt ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ ckpt round=27 step=26200 pos=104038400
2
+ sf_elo=1320 nodes=60
3
+ games=3 W1 D2 L0 score=0.667 elo=1440
4
+ why={'repetition': 2, 'mate': 1}
elo_28_4987.txt ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ ckpt round=28 step=27200 pos=108134400
2
+ sf_elo=1320 nodes=60
3
+ games=4 W1 D3 L0 score=0.625 elo=1409
4
+ why={'mate': 1, 'repetition': 3}
elo_35_cb92.txt ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ ckpt round=35 step=31400 pos=125542400
2
+ sf_elo=1320 nodes=60
3
+ games=4 W0 D4 L0 score=0.500 elo=1320
4
+ why={'repetition': 4}
elo_47_59da.txt ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ ckpt round=47 step=37000 pos=148275200
2
+ sf_elo=1320 nodes=60
3
+ games=2 W0 D2 L0 score=0.500 elo=1320
4
+ why={'repetition': 2}
elo_80_9bfb.txt ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ ckpt round=80 step=50800 pos=153804800
2
+ sf_elo=1320 nodes=60
3
+ games=3 W2 D1 L0 score=0.833 elo=1600
4
+ why={'mate': 2, 'repetition': 1}
global_AM.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:54d1c9a7a27590019ffc28c7b6ba874e39480e380085f91c9972192f02a976a9
3
+ size 457361611
global_round.txt ADDED
@@ -0,0 +1 @@
 
 
1
+ 144
log_AMa.txt ADDED
@@ -0,0 +1,309 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ cfg AMa d=1280 H=10 tied=1 nl=8 lr=8 T=2 shards=0..250 hours=168.0
2
+ step 21100 loss=0.4227 pos=21606400 dt=94
3
+ step 21200 loss=0.4053 pos=21708800 dt=180
4
+ step 21300 loss=0.4865 pos=21811200 dt=280
5
+ step 21400 loss=0.4350 pos=21913600 dt=367
6
+ step 21500 loss=0.3828 pos=22016000 dt=458
7
+ step 21600 loss=0.4234 pos=22118400 dt=544
8
+ step 21700 loss=0.4910 pos=22220800 dt=643
9
+ step 21800 loss=0.4452 pos=22323200 dt=730
10
+ step 21900 loss=0.3942 pos=22425600 dt=821
11
+ step 22000 loss=0.4022 pos=22528000 dt=907
12
+ EVAL step 22000 pos=22528000 | T=2 loss=0.370 from=0.404 joint=0.683 hard=0.822 | T=16 loss=0.400 from=0.402 joint=0.660 hard=0.793
13
+ step 22100 loss=0.4324 pos=22630400 dt=1049
14
+ step 22200 loss=0.4339 pos=22732800 dt=1136
15
+ step 22300 loss=0.4233 pos=22835200 dt=1227
16
+ step 22400 loss=0.4460 pos=22937600 dt=1314
17
+ step 22500 loss=0.4075 pos=23040000 dt=1415
18
+ step 22600 loss=0.4079 pos=23142400 dt=1501
19
+ step 22700 loss=0.3944 pos=23244800 dt=1593
20
+ step 22800 loss=0.3964 pos=23347200 dt=1679
21
+ step 22900 loss=0.3848 pos=23449600 dt=1770
22
+ step 23000 loss=0.4207 pos=23552000 dt=1857
23
+ EVAL step 23000 pos=23552000 | T=2 loss=0.405 from=0.401 joint=0.676 hard=0.814 | T=16 loss=0.415 from=0.395 joint=0.650 hard=0.781
24
+ step 23100 loss=0.3815 pos=23654400 dt=1993
25
+ step 23200 loss=0.3997 pos=23756800 dt=2080
26
+ step 23300 loss=0.3834 pos=23859200 dt=2171
27
+ step 23400 loss=0.4137 pos=23961600 dt=2258
28
+ step 23500 loss=0.4010 pos=24064000 dt=2349
29
+ step 23600 loss=0.3856 pos=24166400 dt=2436
30
+ step 23700 loss=0.4032 pos=24268800 dt=2534
31
+ step 23800 loss=0.3768 pos=24371200 dt=2620
32
+ step 23900 loss=0.4048 pos=24473600 dt=2712
33
+ step 24000 loss=0.4187 pos=24576000 dt=2798
34
+ EVAL step 24000 pos=24576000 | T=2 loss=0.402 from=0.405 joint=0.658 hard=0.785 | T=16 loss=0.430 from=0.400 joint=0.641 hard=0.766
35
+ step 24100 loss=0.3440 pos=24678400 dt=2936
36
+ step 24200 loss=0.4150 pos=24780800 dt=3023
37
+ step 24300 loss=0.3823 pos=24883200 dt=3114
38
+ step 24400 loss=0.3890 pos=24985600 dt=3200
39
+ step 24500 loss=0.4605 pos=25088000 dt=3291
40
+ step 24600 loss=0.4236 pos=25190400 dt=3378
41
+ step 24700 loss=0.3917 pos=25292800 dt=3469
42
+ step 24800 loss=0.4012 pos=25395200 dt=3556
43
+ step 24900 loss=0.3937 pos=25497600 dt=3653
44
+ step 25000 loss=0.3789 pos=25600000 dt=3740
45
+ EVAL step 25000 pos=25600000 | T=2 loss=0.386 from=0.390 joint=0.687 hard=0.828 | T=16 loss=0.405 from=0.385 joint=0.657 hard=0.792
46
+ step 25100 loss=0.3445 pos=25702400 dt=3877
47
+ step 25200 loss=0.4035 pos=25804800 dt=3963
48
+ step 25300 loss=0.3709 pos=25907200 dt=4054
49
+ step 25400 loss=0.3617 pos=26009600 dt=4141
50
+ step 25500 loss=0.3563 pos=26112000 dt=4237
51
+ step 25600 loss=0.3405 pos=26214400 dt=4323
52
+ step 25700 loss=0.3811 pos=26316800 dt=4415
53
+ step 25800 loss=0.3760 pos=26419200 dt=4501
54
+ step 25900 loss=0.3744 pos=26521600 dt=4592
55
+ step 26000 loss=0.3753 pos=26624000 dt=4678
56
+ EVAL step 26000 pos=26624000 | T=2 loss=0.373 from=0.402 joint=0.689 hard=0.831 | T=16 loss=0.396 from=0.403 joint=0.659 hard=0.791
57
+ step 26100 loss=0.3314 pos=26726400 dt=4816
58
+ step 26200 loss=0.3465 pos=26828800 dt=4902
59
+ step 26300 loss=0.3665 pos=26931200 dt=4993
60
+ step 26400 loss=0.3137 pos=27033600 dt=5079
61
+ step 26500 loss=0.2960 pos=27136000 dt=5176
62
+ step 26600 loss=0.3173 pos=27238400 dt=5262
63
+ step 26700 loss=0.2902 pos=27340800 dt=5352
64
+ step 26800 loss=0.2813 pos=27443200 dt=5438
65
+ step 26900 loss=0.2764 pos=27545600 dt=5527
66
+ step 27000 loss=0.2356 pos=27648000 dt=5613
67
+ EVAL step 27000 pos=27648000 | T=2 loss=0.481 from=0.434 joint=0.583 hard=0.702 | T=16 loss=0.533 from=0.429 joint=0.544 hard=0.655
68
+ step 27100 loss=0.3966 pos=27750400 dt=5749
69
+ step 27200 loss=0.3780 pos=27852800 dt=5836
70
+ step 27300 loss=0.3422 pos=27955200 dt=5927
71
+ step 27400 loss=0.3402 pos=28057600 dt=6014
72
+ step 27500 loss=0.3666 pos=28160000 dt=6107
73
+ step 27600 loss=0.3509 pos=28262400 dt=6194
74
+ step 27700 loss=0.3646 pos=28364800 dt=6285
75
+ step 27800 loss=0.3353 pos=28467200 dt=6372
76
+ step 27900 loss=0.3220 pos=28569600 dt=6464
77
+ step 28000 loss=0.3240 pos=28672000 dt=6551
78
+ EVAL step 28000 pos=28672000 | T=2 loss=0.369 from=0.397 joint=0.694 hard=0.840 | T=16 loss=0.399 from=0.403 joint=0.663 hard=0.796
79
+ step 28100 loss=0.3429 pos=28774400 dt=6699
80
+ step 28200 loss=0.3410 pos=28876800 dt=6786
81
+ step 28300 loss=0.3426 pos=28979200 dt=6877
82
+ step 28400 loss=0.3767 pos=29081600 dt=6964
83
+ step 28500 loss=0.3301 pos=29184000 dt=7065
84
+ step 28600 loss=0.3856 pos=29286400 dt=7152
85
+ step 28700 loss=0.3312 pos=29388800 dt=7243
86
+ step 28800 loss=0.3655 pos=29491200 dt=7330
87
+ step 28900 loss=0.3527 pos=29593600 dt=7431
88
+ step 29000 loss=0.3327 pos=29696000 dt=7518
89
+ EVAL step 29000 pos=29696000 | T=2 loss=0.360 from=0.405 joint=0.705 hard=0.854 | T=16 loss=0.387 from=0.405 joint=0.669 hard=0.805
90
+ step 29100 loss=0.3957 pos=29798400 dt=7653
91
+ step 29200 loss=0.3738 pos=29900800 dt=7740
92
+ step 29300 loss=0.4023 pos=30003200 dt=7831
93
+ step 29400 loss=0.3647 pos=30105600 dt=7917
94
+ step 29500 loss=0.3681 pos=30208000 dt=8021
95
+ step 29600 loss=0.3706 pos=30310400 dt=8107
96
+ step 29700 loss=0.3313 pos=30412800 dt=8199
97
+ step 29800 loss=0.3420 pos=30515200 dt=8286
98
+ step 29900 loss=0.3766 pos=30617600 dt=8381
99
+ step 30000 loss=0.3029 pos=30720000 dt=8467
100
+ EVAL step 30000 pos=30720000 | T=2 loss=0.357 from=0.407 joint=0.700 hard=0.846 | T=16 loss=0.383 from=0.403 joint=0.672 hard=0.810
101
+ step 30100 loss=0.3612 pos=30822400 dt=8604
102
+ step 30200 loss=0.3487 pos=30924800 dt=8691
103
+ step 30300 loss=0.3464 pos=31027200 dt=8788
104
+ step 30400 loss=0.3788 pos=31129600 dt=8875
105
+ step 30500 loss=0.3383 pos=31232000 dt=8966
106
+ step 30600 loss=0.3392 pos=31334400 dt=9052
107
+ step 30700 loss=0.3281 pos=31436800 dt=9149
108
+ step 30800 loss=0.3509 pos=31539200 dt=9235
109
+ step 30900 loss=0.3356 pos=31641600 dt=9326
110
+ step 31000 loss=0.3595 pos=31744000 dt=9412
111
+ EVAL step 31000 pos=31744000 | T=2 loss=0.360 from=0.403 joint=0.692 hard=0.833 | T=16 loss=0.384 from=0.408 joint=0.670 hard=0.802
112
+ step 31100 loss=0.3216 pos=31846400 dt=9548
113
+ step 31200 loss=0.3135 pos=31948800 dt=9635
114
+ step 31300 loss=0.3213 pos=32051200 dt=9725
115
+ step 31400 loss=0.3548 pos=32153600 dt=9812
116
+ step 31500 loss=0.3104 pos=32256000 dt=9907
117
+ step 31600 loss=0.3455 pos=32358400 dt=9993
118
+ step 31700 loss=0.3300 pos=32460800 dt=10084
119
+ step 31800 loss=0.3398 pos=32563200 dt=10170
120
+ step 31900 loss=0.3382 pos=32665600 dt=10269
121
+ step 32000 loss=0.3017 pos=32768000 dt=10355
122
+ EVAL step 32000 pos=32768000 | T=2 loss=0.350 from=0.403 joint=0.717 hard=0.869 | T=16 loss=0.377 from=0.400 joint=0.691 hard=0.836
123
+ step 32100 loss=0.3373 pos=32870400 dt=10490
124
+ step 32200 loss=0.3083 pos=32972800 dt=10576
125
+ step 32300 loss=0.3107 pos=33075200 dt=10671
126
+ step 32400 loss=0.3298 pos=33177600 dt=10757
127
+ step 32500 loss=0.2936 pos=33280000 dt=10847
128
+ step 32600 loss=0.3201 pos=33382400 dt=10933
129
+ step 32700 loss=0.3035 pos=33484800 dt=11027
130
+ step 32800 loss=0.3035 pos=33587200 dt=11113
131
+ step 32900 loss=0.2619 pos=33689600 dt=11204
132
+ step 33000 loss=0.2392 pos=33792000 dt=11289
133
+ EVAL step 33000 pos=33792000 | T=2 loss=0.385 from=0.406 joint=0.632 hard=0.757 | T=16 loss=0.439 from=0.412 joint=0.596 hard=0.707
134
+ step 33100 loss=0.3114 pos=33894400 dt=11424
135
+ step 33200 loss=0.3524 pos=33996800 dt=11510
136
+ step 33300 loss=0.3580 pos=34099200 dt=11606
137
+ step 33400 loss=0.3619 pos=34201600 dt=11693
138
+ step 33500 loss=0.3609 pos=34304000 dt=11784
139
+ step 33600 loss=0.3315 pos=34406400 dt=11871
140
+ step 33700 loss=0.3343 pos=34508800 dt=11968
141
+ step 33800 loss=0.3396 pos=34611200 dt=12055
142
+ step 33900 loss=0.3203 pos=34713600 dt=12148
143
+ step 34000 loss=0.3897 pos=34816000 dt=12234
144
+ EVAL step 34000 pos=34816000 | T=2 loss=0.336 from=0.410 joint=0.723 hard=0.878 | T=16 loss=0.362 from=0.408 joint=0.689 hard=0.832
145
+ step 34100 loss=0.3670 pos=34918400 dt=12376
146
+ step 34200 loss=0.3599 pos=35020800 dt=12462
147
+ step 34300 loss=0.3011 pos=35123200 dt=12554
148
+ step 34400 loss=0.3152 pos=35225600 dt=12641
149
+ step 34500 loss=0.3232 pos=35328000 dt=12738
150
+ step 34600 loss=0.3229 pos=35430400 dt=12824
151
+ step 34700 loss=0.2953 pos=35532800 dt=12916
152
+ step 34800 loss=0.3355 pos=35635200 dt=13002
153
+ step 34900 loss=0.3346 pos=35737600 dt=13098
154
+ step 35000 loss=0.3215 pos=35840000 dt=13185
155
+ EVAL step 35000 pos=35840000 | T=2 loss=0.341 from=0.407 joint=0.710 hard=0.859 | T=16 loss=0.363 from=0.406 joint=0.674 hard=0.812
156
+ step 35100 loss=0.3272 pos=35942400 dt=13324
157
+ step 35200 loss=0.3286 pos=36044800 dt=13411
158
+ step 35300 loss=0.3423 pos=36147200 dt=13507
159
+ step 35400 loss=0.3583 pos=36249600 dt=13594
160
+ step 35500 loss=0.3253 pos=36352000 dt=13686
161
+ step 35600 loss=0.3245 pos=36454400 dt=13772
162
+ step 35700 loss=0.3334 pos=36556800 dt=13863
163
+ step 35800 loss=0.3403 pos=36659200 dt=13949
164
+ step 35900 loss=0.3511 pos=36761600 dt=14046
165
+ step 36000 loss=0.3476 pos=36864000 dt=14133
166
+ EVAL step 36000 pos=36864000 | T=2 loss=0.341 from=0.402 joint=0.717 hard=0.869 | T=16 loss=0.375 from=0.399 joint=0.665 hard=0.798
167
+ step 36100 loss=0.2908 pos=36966400 dt=14271
168
+ step 36200 loss=0.3200 pos=37068800 dt=14357
169
+ step 36300 loss=0.3435 pos=37171200 dt=14455
170
+ step 36400 loss=0.3219 pos=37273600 dt=14541
171
+ step 36500 loss=0.3236 pos=37376000 dt=14632
172
+ step 36600 loss=0.3340 pos=37478400 dt=14719
173
+ step 36700 loss=0.3619 pos=37580800 dt=14817
174
+ step 36800 loss=0.3461 pos=37683200 dt=14903
175
+ step 36900 loss=0.3348 pos=37785600 dt=14995
176
+ step 37000 loss=0.3378 pos=37888000 dt=15082
177
+ EVAL step 37000 pos=37888000 | T=2 loss=0.326 from=0.403 joint=0.718 hard=0.871 | T=16 loss=0.351 from=0.405 joint=0.678 hard=0.819
178
+ step 37100 loss=0.3245 pos=37990400 dt=15224
179
+ step 37200 loss=0.3076 pos=38092800 dt=15311
180
+ step 37300 loss=0.3222 pos=38195200 dt=15403
181
+ step 37400 loss=0.3292 pos=38297600 dt=15489
182
+ step 37500 loss=0.2863 pos=38400000 dt=15585
183
+ step 37600 loss=0.3135 pos=38502400 dt=15672
184
+ step 37700 loss=0.3183 pos=38604800 dt=15767
185
+ step 37800 loss=0.3074 pos=38707200 dt=15854
186
+ step 37900 loss=0.3215 pos=38809600 dt=15952
187
+ step 38000 loss=0.3070 pos=38912000 dt=16038
188
+ EVAL step 38000 pos=38912000 | T=2 loss=0.340 from=0.408 joint=0.717 hard=0.870 | T=16 loss=0.374 from=0.407 joint=0.672 hard=0.809
189
+ step 38100 loss=0.2882 pos=39014400 dt=16174
190
+ step 38200 loss=0.3135 pos=39116800 dt=16260
191
+ step 38300 loss=0.2948 pos=39219200 dt=16358
192
+ step 38400 loss=0.3002 pos=39321600 dt=16445
193
+ step 38500 loss=0.2803 pos=39424000 dt=16537
194
+ step 38600 loss=0.2754 pos=39526400 dt=16623
195
+ step 38700 loss=0.2539 pos=39628800 dt=16714
196
+ step 38800 loss=0.2745 pos=39731200 dt=16799
197
+ step 38900 loss=0.2464 pos=39833600 dt=16898
198
+ step 39000 loss=0.2333 pos=39936000 dt=16983
199
+ EVAL step 39000 pos=39936000 | T=2 loss=0.340 from=0.411 joint=0.652 hard=0.780 | T=16 loss=0.371 from=0.414 joint=0.632 hard=0.756
200
+ step 39100 loss=0.2640 pos=40038400 dt=17120
201
+ step 39200 loss=0.3753 pos=40140800 dt=17207
202
+ step 39300 loss=0.3352 pos=40243200 dt=17305
203
+ step 39400 loss=0.3351 pos=40345600 dt=17392
204
+ step 39500 loss=0.3510 pos=40448000 dt=17486
205
+ step 39600 loss=0.3213 pos=40550400 dt=17572
206
+ step 39700 loss=0.3383 pos=40652800 dt=17667
207
+ step 39800 loss=0.3532 pos=40755200 dt=17754
208
+ step 39900 loss=0.2879 pos=40857600 dt=17845
209
+ step 40000 loss=0.3396 pos=40960000 dt=17932
210
+ EVAL step 40000 pos=40960000 | T=2 loss=0.317 from=0.409 joint=0.719 hard=0.873 | T=16 loss=0.343 from=0.408 joint=0.681 hard=0.821
211
+ step 40100 loss=0.3110 pos=41062400 dt=18073
212
+ step 40200 loss=0.3191 pos=41164800 dt=18160
213
+ step 40300 loss=0.3191 pos=41267200 dt=18250
214
+ step 40400 loss=0.2819 pos=41369600 dt=18337
215
+ step 40500 loss=0.2963 pos=41472000 dt=18438
216
+ step 40600 loss=0.3124 pos=41574400 dt=18524
217
+ step 40700 loss=0.3207 pos=41676800 dt=18615
218
+ step 40800 loss=0.3356 pos=41779200 dt=18702
219
+ step 40900 loss=0.3361 pos=41881600 dt=18799
220
+ step 41000 loss=0.3580 pos=41984000 dt=18886
221
+ EVAL step 41000 pos=41984000 | T=2 loss=0.319 from=0.409 joint=0.730 hard=0.888 | T=16 loss=0.345 from=0.406 joint=0.693 hard=0.837
222
+ step 41100 loss=0.2968 pos=42086400 dt=19024
223
+ step 41200 loss=0.3439 pos=42188800 dt=19110
224
+ step 41300 loss=0.3269 pos=42291200 dt=19201
225
+ step 41400 loss=0.3314 pos=42393600 dt=19288
226
+ step 41500 loss=0.3207 pos=42496000 dt=19382
227
+ step 41600 loss=0.3021 pos=42598400 dt=19469
228
+ step 41700 loss=0.3195 pos=42700800 dt=19562
229
+ step 41800 loss=0.2989 pos=42803200 dt=19648
230
+ step 41900 loss=0.3181 pos=42905600 dt=19749
231
+ step 42000 loss=0.3033 pos=43008000 dt=19836
232
+ EVAL step 42000 pos=43008000 | T=2 loss=0.320 from=0.403 joint=0.727 hard=0.885 | T=16 loss=0.348 from=0.402 joint=0.694 hard=0.841
233
+ step 42100 loss=0.3150 pos=43110400 dt=19974
234
+ step 42200 loss=0.3186 pos=43212800 dt=20061
235
+ step 42300 loss=0.2902 pos=43315200 dt=20159
236
+ step 42400 loss=0.3101 pos=43417600 dt=20246
237
+ step 42500 loss=0.3223 pos=43520000 dt=20339
238
+ step 42600 loss=0.3005 pos=43622400 dt=20426
239
+ step 42700 loss=0.2953 pos=43724800 dt=20523
240
+ step 42800 loss=0.3057 pos=43827200 dt=20610
241
+ step 42900 loss=0.3362 pos=43929600 dt=20702
242
+ step 43000 loss=0.3238 pos=44032000 dt=20788
243
+ EVAL step 43000 pos=44032000 | T=2 loss=0.314 from=0.408 joint=0.724 hard=0.883 | T=16 loss=0.349 from=0.406 joint=0.688 hard=0.834
244
+ step 43100 loss=0.2983 pos=44134400 dt=20931
245
+ step 43200 loss=0.3092 pos=44236800 dt=21017
246
+ step 43300 loss=0.3139 pos=44339200 dt=21108
247
+ step 43400 loss=0.3215 pos=44441600 dt=21194
248
+ step 43500 loss=0.2862 pos=44544000 dt=21290
249
+ step 43600 loss=0.3119 pos=44646400 dt=21376
250
+ step 43700 loss=0.2949 pos=44748800 dt=21467
251
+ step 43800 loss=0.2815 pos=44851200 dt=21553
252
+ step 43900 loss=0.2836 pos=44953600 dt=21648
253
+ step 44000 loss=0.2819 pos=45056000 dt=21734
254
+ EVAL step 44000 pos=45056000 | T=2 loss=0.316 from=0.404 joint=0.726 hard=0.882 | T=16 loss=0.353 from=0.402 joint=0.691 hard=0.832
255
+ step 44100 loss=0.2829 pos=45158400 dt=21870
256
+ step 44200 loss=0.2484 pos=45260800 dt=21956
257
+ step 44300 loss=0.2845 pos=45363200 dt=22054
258
+ step 44400 loss=0.2533 pos=45465600 dt=22140
259
+ step 44500 loss=0.2802 pos=45568000 dt=22230
260
+ step 44600 loss=0.2483 pos=45670400 dt=22316
261
+ step 44700 loss=0.2459 pos=45772800 dt=22407
262
+ step 44800 loss=0.2327 pos=45875200 dt=22493
263
+ step 44900 loss=0.2244 pos=45977600 dt=22587
264
+ step 45000 loss=0.2580 pos=46080000 dt=22672
265
+ EVAL step 45000 pos=46080000 | T=2 loss=0.329 from=0.417 joint=0.670 hard=0.803 | T=16 loss=0.365 from=0.416 joint=0.626 hard=0.741
266
+ step 45100 loss=0.2395 pos=46182400 dt=22806
267
+ step 45200 loss=0.3326 pos=46284800 dt=22892
268
+ step 45300 loss=0.3349 pos=46387200 dt=22990
269
+ step 45400 loss=0.3289 pos=46489600 dt=23077
270
+ step 45500 loss=0.3221 pos=46592000 dt=23170
271
+ step 45600 loss=0.3522 pos=46694400 dt=23257
272
+ step 45700 loss=0.3244 pos=46796800 dt=23353
273
+ step 45800 loss=0.2984 pos=46899200 dt=23439
274
+ step 45900 loss=0.3217 pos=47001600 dt=23530
275
+ step 46000 loss=0.3101 pos=47104000 dt=23617
276
+ EVAL step 46000 pos=47104000 | T=2 loss=0.311 from=0.409 joint=0.734 hard=0.894 | T=16 loss=0.337 from=0.410 joint=0.694 hard=0.838
277
+ step 46100 loss=0.3183 pos=47206400 dt=23757
278
+ step 46200 loss=0.3006 pos=47308800 dt=23843
279
+ step 46300 loss=0.3101 pos=47411200 dt=23934
280
+ step 46400 loss=0.2951 pos=47513600 dt=24021
281
+ step 46500 loss=0.3028 pos=47616000 dt=24116
282
+ step 46600 loss=0.3419 pos=47718400 dt=24202
283
+ step 46700 loss=0.2737 pos=47820800 dt=24293
284
+ step 46800 loss=0.3273 pos=47923200 dt=24380
285
+ step 46900 loss=0.2726 pos=48025600 dt=24474
286
+ step 47000 loss=0.3036 pos=48128000 dt=24561
287
+ EVAL step 47000 pos=48128000 | T=2 loss=0.310 from=0.408 joint=0.729 hard=0.888 | T=16 loss=0.336 from=0.410 joint=0.690 hard=0.835
288
+ step 47100 loss=0.3318 pos=48230400 dt=24696
289
+ step 47200 loss=0.3063 pos=48332800 dt=24782
290
+ step 47300 loss=0.2967 pos=48435200 dt=24876
291
+ step 47400 loss=0.2749 pos=48537600 dt=24963
292
+ step 47500 loss=0.2718 pos=48640000 dt=25054
293
+ step 47600 loss=0.3179 pos=48742400 dt=25140
294
+ step 47700 loss=0.3243 pos=48844800 dt=25236
295
+ step 47800 loss=0.3166 pos=48947200 dt=25322
296
+ step 47900 loss=0.3377 pos=49049600 dt=25414
297
+ step 48000 loss=0.3257 pos=49152000 dt=25501
298
+ EVAL step 48000 pos=49152000 | T=2 loss=0.312 from=0.411 joint=0.717 hard=0.869 | T=16 loss=0.339 from=0.409 joint=0.681 hard=0.817
299
+ step 48100 loss=0.2966 pos=49254400 dt=25643
300
+ step 48200 loss=0.3099 pos=49356800 dt=25730
301
+ step 48300 loss=0.2831 pos=49459200 dt=25823
302
+ step 48400 loss=0.2884 pos=49561600 dt=25910
303
+ step 48500 loss=0.3025 pos=49664000 dt=26015
304
+ step 48600 loss=0.3073 pos=49766400 dt=26101
305
+ step 48700 loss=0.2687 pos=49868800 dt=26194
306
+ step 48800 loss=0.2965 pos=49971200 dt=26281
307
+ step 48900 loss=0.2944 pos=50073600 dt=26376
308
+ step 49000 loss=0.2800 pos=50176000 dt=26462
309
+ EVAL step 49000 pos=50176000 | T=2 loss=0.306 from=0.409 joint=0.728 hard=0.883 | T=16 loss=0.336 from=0.411 joint=0.690 hard=0.832
log_AMb.txt ADDED
@@ -0,0 +1,452 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ cfg AMb d=1280 H=10 tied=1 nl=8 lr=8 T=2 shards=250..500 hours=168.0
2
+ step 13100 loss=0.5495 pos=13414400 dt=92
3
+ step 13200 loss=0.4743 pos=13516800 dt=179
4
+ step 13300 loss=0.6806 pos=13619200 dt=276
5
+ step 13400 loss=0.5667 pos=13721600 dt=362
6
+ step 13500 loss=0.6304 pos=13824000 dt=453
7
+ step 13600 loss=0.5533 pos=13926400 dt=540
8
+ step 13700 loss=0.5601 pos=14028800 dt=635
9
+ step 13800 loss=0.5711 pos=14131200 dt=722
10
+ step 13900 loss=0.5438 pos=14233600 dt=812
11
+ step 14000 loss=0.5214 pos=14336000 dt=899
12
+ EVAL step 14000 pos=14336000 | T=2 loss=0.537 from=0.386 joint=0.625 hard=0.754 | T=16 loss=0.558 from=0.380 joint=0.611 hard=0.731
13
+ step 14100 loss=0.4960 pos=14438400 dt=1038
14
+ step 14200 loss=0.5381 pos=14540800 dt=1125
15
+ step 14300 loss=0.5112 pos=14643200 dt=1216
16
+ step 14400 loss=0.4369 pos=14745600 dt=1303
17
+ step 14500 loss=0.4766 pos=14848000 dt=1398
18
+ step 14600 loss=0.4865 pos=14950400 dt=1485
19
+ step 14700 loss=0.4549 pos=15052800 dt=1577
20
+ step 14800 loss=0.4776 pos=15155200 dt=1663
21
+ step 14900 loss=0.4504 pos=15257600 dt=1761
22
+ step 15000 loss=0.4480 pos=15360000 dt=1848
23
+ EVAL step 15000 pos=15360000 | T=2 loss=0.479 from=0.385 joint=0.627 hard=0.752 | T=16 loss=0.515 from=0.383 joint=0.606 hard=0.723
24
+ step 15100 loss=0.4998 pos=15462400 dt=1984
25
+ step 15200 loss=0.3924 pos=15564800 dt=2070
26
+ step 15300 loss=0.4253 pos=15667200 dt=2164
27
+ step 15400 loss=0.3968 pos=15769600 dt=2250
28
+ step 15500 loss=0.4646 pos=15872000 dt=2341
29
+ step 15600 loss=0.4401 pos=15974400 dt=2427
30
+ step 15700 loss=0.4195 pos=16076800 dt=2523
31
+ step 15800 loss=0.4274 pos=16179200 dt=2609
32
+ step 15900 loss=0.4247 pos=16281600 dt=2700
33
+ step 16000 loss=0.4430 pos=16384000 dt=2787
34
+ EVAL step 16000 pos=16384000 | T=2 loss=0.426 from=0.388 joint=0.663 hard=0.803 | T=16 loss=0.454 from=0.388 joint=0.634 hard=0.766
35
+ step 16100 loss=0.3912 pos=16486400 dt=2929
36
+ step 16200 loss=0.4415 pos=16588800 dt=3015
37
+ step 16300 loss=0.4423 pos=16691200 dt=3106
38
+ step 16400 loss=0.4157 pos=16793600 dt=3193
39
+ step 16500 loss=0.3986 pos=16896000 dt=3289
40
+ step 16600 loss=0.3860 pos=16998400 dt=3376
41
+ step 16700 loss=0.4178 pos=17100800 dt=3467
42
+ step 16800 loss=0.4075 pos=17203200 dt=3554
43
+ step 16900 loss=0.4218 pos=17305600 dt=3650
44
+ step 17000 loss=0.4110 pos=17408000 dt=3736
45
+ EVAL step 17000 pos=17408000 | T=2 loss=0.425 from=0.401 joint=0.666 hard=0.805 | T=16 loss=0.440 from=0.398 joint=0.639 hard=0.766
46
+ step 17100 loss=0.3671 pos=17510400 dt=3877
47
+ step 17200 loss=0.3534 pos=17612800 dt=3963
48
+ step 17300 loss=0.3609 pos=17715200 dt=4060
49
+ step 17400 loss=0.4018 pos=17817600 dt=4146
50
+ step 17500 loss=0.4167 pos=17920000 dt=4239
51
+ step 17600 loss=0.3915 pos=18022400 dt=4325
52
+ step 17700 loss=0.3846 pos=18124800 dt=4422
53
+ step 17800 loss=0.3763 pos=18227200 dt=4508
54
+ step 17900 loss=0.3581 pos=18329600 dt=4599
55
+ step 18000 loss=0.3870 pos=18432000 dt=4685
56
+ EVAL step 18000 pos=18432000 | T=2 loss=0.397 from=0.401 joint=0.683 hard=0.826 | T=16 loss=0.436 from=0.397 joint=0.646 hard=0.778
57
+ step 18100 loss=0.3913 pos=18534400 dt=4835
58
+ step 18200 loss=0.3731 pos=18636800 dt=4921
59
+ step 18300 loss=0.3804 pos=18739200 dt=5012
60
+ step 18400 loss=0.3401 pos=18841600 dt=5098
61
+ step 18500 loss=0.3279 pos=18944000 dt=5195
62
+ step 18600 loss=0.3644 pos=19046400 dt=5280
63
+ step 18700 loss=0.3198 pos=19148800 dt=5371
64
+ step 18800 loss=0.3160 pos=19251200 dt=5457
65
+ step 18900 loss=0.3209 pos=19353600 dt=5552
66
+ step 19000 loss=0.3149 pos=19456000 dt=5638
67
+ EVAL step 19000 pos=19456000 | T=2 loss=0.602 from=0.394 joint=0.565 hard=0.673 | T=16 loss=0.683 from=0.389 joint=0.552 hard=0.661
68
+ step 19100 loss=0.3907 pos=19558400 dt=5776
69
+ step 19200 loss=0.4027 pos=19660800 dt=5863
70
+ step 19300 loss=0.3768 pos=19763200 dt=5965
71
+ step 19400 loss=0.3985 pos=19865600 dt=6052
72
+ step 19500 loss=0.3793 pos=19968000 dt=6145
73
+ step 19600 loss=0.3983 pos=20070400 dt=6232
74
+ step 19700 loss=0.3751 pos=20172800 dt=6330
75
+ step 19800 loss=0.3611 pos=20275200 dt=6416
76
+ step 19900 loss=0.3912 pos=20377600 dt=6507
77
+ step 20000 loss=0.3563 pos=20480000 dt=6594
78
+ EVAL step 20000 pos=20480000 | T=2 loss=0.365 from=0.402 joint=0.695 hard=0.842 | T=16 loss=0.408 from=0.397 joint=0.664 hard=0.800
79
+ step 20100 loss=0.3837 pos=20582400 dt=6737
80
+ step 20200 loss=0.3731 pos=20684800 dt=6824
81
+ step 20300 loss=0.4047 pos=20787200 dt=6915
82
+ step 20400 loss=0.4237 pos=20889600 dt=7001
83
+ step 20500 loss=0.3725 pos=20992000 dt=7093
84
+ step 20600 loss=0.3771 pos=21094400 dt=7180
85
+ step 20700 loss=0.4188 pos=21196800 dt=7279
86
+ step 20800 loss=0.3622 pos=21299200 dt=7366
87
+ step 20900 loss=0.3898 pos=21401600 dt=7460
88
+ step 21000 loss=0.3729 pos=21504000 dt=7546
89
+ EVAL step 21000 pos=21504000 | T=2 loss=0.362 from=0.406 joint=0.687 hard=0.831 | T=16 loss=0.393 from=0.406 joint=0.668 hard=0.806
90
+ step 21100 loss=0.3601 pos=21606400 dt=7688
91
+ step 21200 loss=0.3694 pos=21708800 dt=7775
92
+ step 21300 loss=0.3653 pos=21811200 dt=7866
93
+ step 21400 loss=0.3769 pos=21913600 dt=7953
94
+ step 21500 loss=0.3445 pos=22016000 dt=8044
95
+ step 21600 loss=0.3417 pos=22118400 dt=8131
96
+ step 21700 loss=0.3675 pos=22220800 dt=8223
97
+ step 21800 loss=0.3652 pos=22323200 dt=8310
98
+ step 21900 loss=0.3886 pos=22425600 dt=8401
99
+ step 22000 loss=0.3792 pos=22528000 dt=8488
100
+ EVAL step 22000 pos=22528000 | T=2 loss=0.386 from=0.408 joint=0.679 hard=0.821 | T=16 loss=0.425 from=0.409 joint=0.643 hard=0.771
101
+ step 22100 loss=0.4069 pos=22630400 dt=8631
102
+ step 22200 loss=0.3688 pos=22732800 dt=8717
103
+ step 22300 loss=0.3581 pos=22835200 dt=8809
104
+ step 22400 loss=0.3954 pos=22937600 dt=8895
105
+ step 22500 loss=0.3946 pos=23040000 dt=8988
106
+ step 22600 loss=0.3811 pos=23142400 dt=9075
107
+ step 22700 loss=0.3422 pos=23244800 dt=9172
108
+ step 22800 loss=0.3899 pos=23347200 dt=9258
109
+ step 22900 loss=0.3443 pos=23449600 dt=9349
110
+ step 23000 loss=0.3352 pos=23552000 dt=9435
111
+ EVAL step 23000 pos=23552000 | T=2 loss=0.370 from=0.403 joint=0.685 hard=0.825 | T=16 loss=0.399 from=0.401 joint=0.658 hard=0.790
112
+ step 23100 loss=0.3411 pos=23654400 dt=9579
113
+ step 23200 loss=0.3478 pos=23756800 dt=9666
114
+ step 23300 loss=0.3229 pos=23859200 dt=9757
115
+ step 23400 loss=0.3735 pos=23961600 dt=9844
116
+ step 23500 loss=0.3636 pos=24064000 dt=9944
117
+ step 23600 loss=0.3805 pos=24166400 dt=10031
118
+ step 23700 loss=0.3202 pos=24268800 dt=10121
119
+ step 23800 loss=0.3844 pos=24371200 dt=10208
120
+ step 23900 loss=0.3722 pos=24473600 dt=10300
121
+ step 24000 loss=0.3218 pos=24576000 dt=10387
122
+ EVAL step 24000 pos=24576000 | T=2 loss=0.358 from=0.404 joint=0.705 hard=0.854 | T=16 loss=0.390 from=0.403 joint=0.673 hard=0.812
123
+ step 24100 loss=0.3381 pos=24678400 dt=10525
124
+ step 24200 loss=0.3521 pos=24780800 dt=10612
125
+ step 24300 loss=0.3201 pos=24883200 dt=10705
126
+ step 24400 loss=0.3581 pos=24985600 dt=10791
127
+ step 24500 loss=0.3270 pos=25088000 dt=10883
128
+ step 24600 loss=0.3251 pos=25190400 dt=10970
129
+ step 24700 loss=0.2930 pos=25292800 dt=11066
130
+ step 24800 loss=0.3200 pos=25395200 dt=11152
131
+ step 24900 loss=0.2921 pos=25497600 dt=11243
132
+ step 25000 loss=0.2877 pos=25600000 dt=11329
133
+ EVAL step 25000 pos=25600000 | T=2 loss=0.417 from=0.423 joint=0.612 hard=0.736 | T=16 loss=0.473 from=0.417 joint=0.601 hard=0.720
134
+ step 25100 loss=0.2680 pos=25702400 dt=11467
135
+ step 25200 loss=0.3700 pos=25804800 dt=11554
136
+ step 25300 loss=0.3886 pos=25907200 dt=11645
137
+ step 25400 loss=0.3233 pos=26009600 dt=11733
138
+ step 25500 loss=0.3516 pos=26112000 dt=11824
139
+ step 25600 loss=0.3536 pos=26214400 dt=11912
140
+ step 25700 loss=0.3463 pos=26316800 dt=12132
141
+ step 25800 loss=0.3847 pos=26419200 dt=12219
142
+ step 25900 loss=0.3512 pos=26521600 dt=12319
143
+ step 26000 loss=0.3510 pos=26624000 dt=12406
144
+ EVAL step 26000 pos=26624000 | T=2 loss=0.355 from=0.404 joint=0.713 hard=0.866 | T=16 loss=0.382 from=0.404 joint=0.684 hard=0.827
145
+ step 26100 loss=0.3655 pos=26726400 dt=12543
146
+ step 26200 loss=0.3389 pos=26828800 dt=12630
147
+ step 26300 loss=0.3666 pos=26931200 dt=12726
148
+ step 26400 loss=0.3707 pos=27033600 dt=12813
149
+ step 26500 loss=0.3528 pos=27136000 dt=12904
150
+ step 26600 loss=0.3532 pos=27238400 dt=12991
151
+ step 26700 loss=0.3459 pos=27340800 dt=13083
152
+ step 26800 loss=0.3725 pos=27443200 dt=13170
153
+ step 26900 loss=0.3228 pos=27545600 dt=13262
154
+ step 27000 loss=0.3579 pos=27648000 dt=13349
155
+ EVAL step 27000 pos=27648000 | T=2 loss=0.349 from=0.403 joint=0.716 hard=0.870 | T=16 loss=0.369 from=0.398 joint=0.683 hard=0.827
156
+ step 27100 loss=0.3652 pos=27750400 dt=13485
157
+ step 27200 loss=0.3333 pos=27852800 dt=13572
158
+ step 27300 loss=0.3485 pos=27955200 dt=13673
159
+ step 27400 loss=0.3306 pos=28057600 dt=13760
160
+ step 27500 loss=0.3657 pos=28160000 dt=13852
161
+ step 27600 loss=0.3472 pos=28262400 dt=13939
162
+ step 27700 loss=0.3649 pos=28364800 dt=14030
163
+ step 27800 loss=0.3736 pos=28467200 dt=14116
164
+ step 27900 loss=0.3304 pos=28569600 dt=14207
165
+ step 28000 loss=0.3149 pos=28672000 dt=14293
166
+ EVAL step 28000 pos=28672000 | T=2 loss=0.350 from=0.405 joint=0.705 hard=0.853 | T=16 loss=0.385 from=0.406 joint=0.679 hard=0.818
167
+ step 28100 loss=0.3345 pos=28774400 dt=14430
168
+ step 28200 loss=0.3519 pos=28876800 dt=14516
169
+ step 28300 loss=0.3504 pos=28979200 dt=14607
170
+ step 28400 loss=0.3299 pos=29081600 dt=14694
171
+ step 28500 loss=0.3317 pos=29184000 dt=14786
172
+ step 28600 loss=0.3800 pos=29286400 dt=14873
173
+ step 28700 loss=0.3067 pos=29388800 dt=14966
174
+ step 28800 loss=0.3869 pos=29491200 dt=15052
175
+ step 28900 loss=0.3380 pos=29593600 dt=15143
176
+ step 29000 loss=0.3316 pos=29696000 dt=15230
177
+ EVAL step 29000 pos=29696000 | T=2 loss=0.369 from=0.397 joint=0.691 hard=0.835 | T=16 loss=0.399 from=0.402 joint=0.662 hard=0.796
178
+ step 29100 loss=0.3051 pos=29798400 dt=15378
179
+ step 29200 loss=0.3468 pos=29900800 dt=15464
180
+ step 29300 loss=0.3524 pos=30003200 dt=15556
181
+ step 29400 loss=0.3265 pos=30105600 dt=15642
182
+ step 29500 loss=0.3182 pos=30208000 dt=15742
183
+ step 29600 loss=0.3490 pos=30310400 dt=15828
184
+ step 29700 loss=0.3113 pos=30412800 dt=15919
185
+ step 29800 loss=0.3676 pos=30515200 dt=16005
186
+ step 29900 loss=0.3135 pos=30617600 dt=16106
187
+ step 30000 loss=0.3140 pos=30720000 dt=16192
188
+ EVAL step 30000 pos=30720000 | T=2 loss=0.359 from=0.411 joint=0.699 hard=0.848 | T=16 loss=0.387 from=0.417 joint=0.664 hard=0.801
189
+ step 30100 loss=0.3749 pos=30822400 dt=16327
190
+ step 30200 loss=0.3481 pos=30924800 dt=16414
191
+ step 30300 loss=0.3178 pos=31027200 dt=16516
192
+ step 30400 loss=0.3446 pos=31129600 dt=16603
193
+ step 30500 loss=0.3228 pos=31232000 dt=16694
194
+ step 30600 loss=0.3033 pos=31334400 dt=16780
195
+ step 30700 loss=0.2848 pos=31436800 dt=16876
196
+ step 30800 loss=0.2527 pos=31539200 dt=16961
197
+ step 30900 loss=0.2980 pos=31641600 dt=17051
198
+ step 31000 loss=0.2719 pos=31744000 dt=17137
199
+ EVAL step 31000 pos=31744000 | T=2 loss=0.359 from=0.409 joint=0.676 hard=0.813 | T=16 loss=0.396 from=0.404 joint=0.646 hard=0.772
200
+ step 31100 loss=0.3480 pos=31846400 dt=17277
201
+ step 31200 loss=0.3637 pos=31948800 dt=17364
202
+ step 31300 loss=0.3252 pos=32051200 dt=17454
203
+ step 31400 loss=0.3471 pos=32153600 dt=17541
204
+ step 31500 loss=0.3595 pos=32256000 dt=17636
205
+ step 31600 loss=0.3587 pos=32358400 dt=17723
206
+ step 31700 loss=0.3557 pos=32460800 dt=17814
207
+ step 31800 loss=0.3306 pos=32563200 dt=17901
208
+ step 31900 loss=0.3688 pos=32665600 dt=17991
209
+ step 32000 loss=0.3539 pos=32768000 dt=18078
210
+ EVAL step 32000 pos=32768000 | T=2 loss=0.343 from=0.405 joint=0.712 hard=0.866 | T=16 loss=0.366 from=0.402 joint=0.681 hard=0.820
211
+ step 32100 loss=0.3345 pos=32870400 dt=18214
212
+ step 32200 loss=0.3214 pos=32972800 dt=18301
213
+ step 32300 loss=0.3711 pos=33075200 dt=18393
214
+ step 32400 loss=0.3120 pos=33177600 dt=18479
215
+ step 32500 loss=0.3410 pos=33280000 dt=18574
216
+ step 32600 loss=0.3628 pos=33382400 dt=18661
217
+ step 32700 loss=0.3386 pos=33484800 dt=18752
218
+ step 32800 loss=0.3573 pos=33587200 dt=18839
219
+ step 32900 loss=0.3249 pos=33689600 dt=18934
220
+ step 33000 loss=0.3360 pos=33792000 dt=19021
221
+ EVAL step 33000 pos=33792000 | T=2 loss=0.342 from=0.404 joint=0.718 hard=0.873 | T=16 loss=0.369 from=0.397 joint=0.684 hard=0.823
222
+ step 33100 loss=0.3512 pos=33894400 dt=19157
223
+ step 33200 loss=0.3181 pos=33996800 dt=19243
224
+ step 33300 loss=0.3121 pos=34099200 dt=19339
225
+ step 33400 loss=0.3460 pos=34201600 dt=19425
226
+ step 33500 loss=0.3532 pos=34304000 dt=19524
227
+ step 33600 loss=0.3554 pos=34406400 dt=19610
228
+ step 33700 loss=0.3302 pos=34508800 dt=19707
229
+ step 33800 loss=0.3305 pos=34611200 dt=19793
230
+ step 33900 loss=0.3321 pos=34713600 dt=19885
231
+ step 34000 loss=0.2845 pos=34816000 dt=19971
232
+ EVAL step 34000 pos=34816000 | T=2 loss=0.341 from=0.409 joint=0.711 hard=0.861 | T=16 loss=0.380 from=0.408 joint=0.673 hard=0.811
233
+ step 34100 loss=0.3213 pos=34918400 dt=20114
234
+ step 34200 loss=0.3547 pos=35020800 dt=20200
235
+ step 34300 loss=0.3333 pos=35123200 dt=20293
236
+ step 34400 loss=0.3199 pos=35225600 dt=20379
237
+ step 34500 loss=0.3354 pos=35328000 dt=20489
238
+ step 34600 loss=0.3198 pos=35430400 dt=20575
239
+ step 34700 loss=0.3389 pos=35532800 dt=20667
240
+ step 34800 loss=0.3215 pos=35635200 dt=20754
241
+ step 34900 loss=0.3166 pos=35737600 dt=20850
242
+ step 35000 loss=0.2989 pos=35840000 dt=20936
243
+ EVAL step 35000 pos=35840000 | T=2 loss=0.330 from=0.409 joint=0.724 hard=0.880 | T=16 loss=0.359 from=0.408 joint=0.685 hard=0.828
244
+ step 35100 loss=0.3127 pos=35942400 dt=21075
245
+ step 35200 loss=0.3307 pos=36044800 dt=21162
246
+ step 35300 loss=0.2806 pos=36147200 dt=21260
247
+ step 35400 loss=0.2655 pos=36249600 dt=21346
248
+ step 35500 loss=0.3215 pos=36352000 dt=21437
249
+ step 35600 loss=0.3220 pos=36454400 dt=21524
250
+ step 35700 loss=0.3119 pos=36556800 dt=21623
251
+ step 35800 loss=0.3087 pos=36659200 dt=21709
252
+ step 35900 loss=0.3327 pos=36761600 dt=21800
253
+ step 36000 loss=0.3118 pos=36864000 dt=21886
254
+ EVAL step 36000 pos=36864000 | T=2 loss=0.333 from=0.408 joint=0.711 hard=0.861 | T=16 loss=0.367 from=0.407 joint=0.683 hard=0.823
255
+ step 36100 loss=0.2935 pos=36966400 dt=22023
256
+ step 36200 loss=0.3030 pos=37068800 dt=22109
257
+ step 36300 loss=0.3140 pos=37171200 dt=22204
258
+ step 36400 loss=0.2890 pos=37273600 dt=22290
259
+ step 36500 loss=0.2988 pos=37376000 dt=22381
260
+ step 36600 loss=0.2659 pos=37478400 dt=22467
261
+ step 36700 loss=0.2771 pos=37580800 dt=22565
262
+ step 36800 loss=0.2855 pos=37683200 dt=22650
263
+ step 36900 loss=0.2335 pos=37785600 dt=22740
264
+ step 37000 loss=0.2358 pos=37888000 dt=22826
265
+ EVAL step 37000 pos=37888000 | T=2 loss=0.342 from=0.398 joint=0.697 hard=0.838 | T=16 loss=0.380 from=0.397 joint=0.669 hard=0.803
266
+ step 37100 loss=0.2088 pos=37990400 dt=22969
267
+ step 37200 loss=0.3377 pos=38092800 dt=23055
268
+ step 37300 loss=0.3119 pos=38195200 dt=23146
269
+ step 37400 loss=0.3212 pos=38297600 dt=23233
270
+ step 37500 loss=0.3325 pos=38400000 dt=23330
271
+ step 37600 loss=0.3163 pos=38502400 dt=23416
272
+ step 37700 loss=0.3072 pos=38604800 dt=23510
273
+ step 37800 loss=0.3573 pos=38707200 dt=23597
274
+ step 37900 loss=0.3086 pos=38809600 dt=23696
275
+ step 38000 loss=0.3111 pos=38912000 dt=23783
276
+ EVAL step 38000 pos=38912000 | T=2 loss=0.324 from=0.405 joint=0.720 hard=0.874 | T=16 loss=0.358 from=0.407 joint=0.682 hard=0.824
277
+ step 38100 loss=0.3331 pos=39014400 dt=23951
278
+ step 38200 loss=0.3444 pos=39116800 dt=24038
279
+ step 38300 loss=0.3053 pos=39219200 dt=24133
280
+ step 38400 loss=0.3461 pos=39321600 dt=24219
281
+ step 38500 loss=0.3254 pos=39424000 dt=24310
282
+ step 38600 loss=0.3402 pos=39526400 dt=24397
283
+ step 38700 loss=0.2894 pos=39628800 dt=24496
284
+ step 38800 loss=0.2771 pos=39731200 dt=24582
285
+ step 38900 loss=0.3331 pos=39833600 dt=24674
286
+ step 39000 loss=0.3046 pos=39936000 dt=24760
287
+ EVAL step 39000 pos=39936000 | T=2 loss=0.326 from=0.403 joint=0.720 hard=0.873 | T=16 loss=0.358 from=0.401 joint=0.680 hard=0.818
288
+ step 39100 loss=0.3204 pos=40038400 dt=24897
289
+ step 39200 loss=0.2988 pos=40140800 dt=24984
290
+ step 39300 loss=0.3273 pos=40243200 dt=25081
291
+ step 39400 loss=0.2914 pos=40345600 dt=25167
292
+ step 39500 loss=0.3203 pos=40448000 dt=25258
293
+ step 39600 loss=0.3088 pos=40550400 dt=25345
294
+ step 39700 loss=0.3583 pos=40652800 dt=25443
295
+ step 39800 loss=0.3164 pos=40755200 dt=25530
296
+ step 39900 loss=0.2909 pos=40857600 dt=25622
297
+ step 40000 loss=0.3023 pos=40960000 dt=25709
298
+ EVAL step 40000 pos=40960000 | T=2 loss=0.321 from=0.408 joint=0.705 hard=0.853 | T=16 loss=0.350 from=0.408 joint=0.678 hard=0.817
299
+ step 40100 loss=0.2657 pos=41062400 dt=25854
300
+ step 40200 loss=0.3333 pos=41164800 dt=25940
301
+ step 40300 loss=0.3082 pos=41267200 dt=26034
302
+ step 40400 loss=0.3218 pos=41369600 dt=26120
303
+ step 40500 loss=0.3133 pos=41472000 dt=26217
304
+ step 40600 loss=0.3333 pos=41574400 dt=26304
305
+ step 40700 loss=0.3084 pos=41676800 dt=26395
306
+ step 40800 loss=0.2741 pos=41779200 dt=26481
307
+ step 40900 loss=0.2992 pos=41881600 dt=26578
308
+ step 41000 loss=0.3268 pos=41984000 dt=26664
309
+ EVAL step 41000 pos=41984000 | T=2 loss=0.321 from=0.409 joint=0.729 hard=0.886 | T=16 loss=0.343 from=0.409 joint=0.691 hard=0.834
310
+ step 41100 loss=0.2786 pos=42086400 dt=26801
311
+ step 41200 loss=0.2987 pos=42188800 dt=26887
312
+ step 41300 loss=0.2989 pos=42291200 dt=26986
313
+ step 41400 loss=0.3113 pos=42393600 dt=27073
314
+ step 41500 loss=0.3067 pos=42496000 dt=27163
315
+ step 41600 loss=0.2711 pos=42598400 dt=27250
316
+ step 41700 loss=0.3241 pos=42700800 dt=27347
317
+ step 41800 loss=0.2801 pos=42803200 dt=27433
318
+ step 41900 loss=0.3226 pos=42905600 dt=27524
319
+ step 42000 loss=0.2925 pos=43008000 dt=27611
320
+ EVAL step 42000 pos=43008000 | T=2 loss=0.316 from=0.407 joint=0.728 hard=0.887 | T=16 loss=0.343 from=0.408 joint=0.695 hard=0.839
321
+ step 42100 loss=0.2847 pos=43110400 dt=27748
322
+ step 42200 loss=0.2975 pos=43212800 dt=27835
323
+ step 42300 loss=0.2766 pos=43315200 dt=27930
324
+ step 42400 loss=0.2567 pos=43417600 dt=28016
325
+ step 42500 loss=0.2992 pos=43520000 dt=28108
326
+ step 42600 loss=0.3203 pos=43622400 dt=28194
327
+ step 42700 loss=0.2755 pos=43724800 dt=28289
328
+ step 42800 loss=0.2494 pos=43827200 dt=28375
329
+ step 42900 loss=0.2501 pos=43929600 dt=28466
330
+ step 43000 loss=0.2401 pos=44032000 dt=28551
331
+ EVAL step 43000 pos=44032000 | T=2 loss=0.314 from=0.406 joint=0.709 hard=0.860 | T=16 loss=0.344 from=0.406 joint=0.681 hard=0.820
332
+ step 43100 loss=0.2219 pos=44134400 dt=28692
333
+ step 43200 loss=0.2402 pos=44236800 dt=28778
334
+ step 43300 loss=0.3253 pos=44339200 dt=28871
335
+ step 43400 loss=0.3505 pos=44441600 dt=28957
336
+ step 43500 loss=0.3168 pos=44544000 dt=29053
337
+ step 43600 loss=0.3008 pos=44646400 dt=29139
338
+ step 43700 loss=0.3187 pos=44748800 dt=29231
339
+ step 43800 loss=0.2858 pos=44851200 dt=29317
340
+ step 43900 loss=0.3173 pos=44953600 dt=29415
341
+ step 44000 loss=0.3346 pos=45056000 dt=29501
342
+ EVAL step 44000 pos=45056000 | T=2 loss=0.326 from=0.405 joint=0.716 hard=0.870 | T=16 loss=0.348 from=0.403 joint=0.678 hard=0.817
343
+ step 44100 loss=0.3281 pos=45158400 dt=29639
344
+ step 44200 loss=0.2986 pos=45260800 dt=29726
345
+ step 44300 loss=0.2918 pos=45363200 dt=29824
346
+ step 44400 loss=0.3204 pos=45465600 dt=29910
347
+ step 44500 loss=0.3393 pos=45568000 dt=30003
348
+ step 44600 loss=0.2972 pos=45670400 dt=30089
349
+ step 44700 loss=0.3060 pos=45772800 dt=30185
350
+ step 44800 loss=0.3041 pos=45875200 dt=30272
351
+ step 44900 loss=0.3201 pos=45977600 dt=30363
352
+ step 45000 loss=0.3079 pos=46080000 dt=30449
353
+ EVAL step 45000 pos=46080000 | T=2 loss=0.311 from=0.410 joint=0.729 hard=0.886 | T=16 loss=0.342 from=0.408 joint=0.703 hard=0.851
354
+ step 45100 loss=0.2976 pos=46182400 dt=30584
355
+ step 45200 loss=0.3193 pos=46284800 dt=30671
356
+ step 45300 loss=0.2922 pos=46387200 dt=30767
357
+ step 45400 loss=0.3062 pos=46489600 dt=30854
358
+ step 45500 loss=0.2688 pos=46592000 dt=30945
359
+ step 45600 loss=0.3151 pos=46694400 dt=31032
360
+ step 45700 loss=0.2913 pos=46796800 dt=31127
361
+ step 45800 loss=0.3303 pos=46899200 dt=31213
362
+ step 45900 loss=0.3191 pos=47001600 dt=31308
363
+ step 46000 loss=0.3086 pos=47104000 dt=31395
364
+ EVAL step 46000 pos=47104000 | T=2 loss=0.313 from=0.414 joint=0.700 hard=0.844 | T=16 loss=0.342 from=0.412 joint=0.660 hard=0.787
365
+ step 46100 loss=0.2853 pos=47206400 dt=31536
366
+ step 46200 loss=0.3353 pos=47308800 dt=31623
367
+ step 46300 loss=0.3410 pos=47411200 dt=31714
368
+ step 46400 loss=0.2990 pos=47513600 dt=31801
369
+ step 46500 loss=0.3003 pos=47616000 dt=31896
370
+ step 46600 loss=0.2973 pos=47718400 dt=31983
371
+ step 46700 loss=0.3056 pos=47820800 dt=32073
372
+ step 46800 loss=0.3045 pos=47923200 dt=32160
373
+ step 46900 loss=0.2999 pos=48025600 dt=32255
374
+ step 47000 loss=0.3432 pos=48128000 dt=32342
375
+ EVAL step 47000 pos=48128000 | T=2 loss=0.309 from=0.401 joint=0.730 hard=0.888 | T=16 loss=0.335 from=0.407 joint=0.700 hard=0.848
376
+ step 47100 loss=0.3122 pos=48230400 dt=32479
377
+ step 47200 loss=0.3221 pos=48332800 dt=32566
378
+ step 47300 loss=0.2751 pos=48435200 dt=32662
379
+ step 47400 loss=0.2957 pos=48537600 dt=32748
380
+ step 47500 loss=0.2801 pos=48640000 dt=32839
381
+ step 47600 loss=0.2847 pos=48742400 dt=32925
382
+ step 47700 loss=0.3011 pos=48844800 dt=33019
383
+ step 47800 loss=0.2728 pos=48947200 dt=33105
384
+ step 47900 loss=0.2691 pos=49049600 dt=33196
385
+ step 48000 loss=0.2854 pos=49152000 dt=33282
386
+ EVAL step 48000 pos=49152000 | T=2 loss=0.307 from=0.414 joint=0.734 hard=0.896 | T=16 loss=0.330 from=0.416 joint=0.700 hard=0.849
387
+ step 48100 loss=0.2671 pos=49254400 dt=33423
388
+ step 48200 loss=0.2673 pos=49356800 dt=33510
389
+ step 48300 loss=0.2911 pos=49459200 dt=33601
390
+ step 48400 loss=0.3083 pos=49561600 dt=33687
391
+ step 48500 loss=0.3098 pos=49664000 dt=33784
392
+ step 48600 loss=0.3029 pos=49766400 dt=33870
393
+ step 48700 loss=0.2653 pos=49868800 dt=33962
394
+ step 48800 loss=0.2236 pos=49971200 dt=34047
395
+ step 48900 loss=0.2313 pos=50073600 dt=34145
396
+ step 49000 loss=0.2475 pos=50176000 dt=34230
397
+ EVAL step 49000 pos=50176000 | T=2 loss=0.311 from=0.414 joint=0.701 hard=0.848 | T=16 loss=0.341 from=0.408 joint=0.679 hard=0.818
398
+ step 49100 loss=0.2356 pos=50278400 dt=34367
399
+ step 49200 loss=0.1955 pos=50380800 dt=34452
400
+ step 49300 loss=0.2903 pos=50483200 dt=34546
401
+ step 49400 loss=0.3229 pos=50585600 dt=34633
402
+ step 49500 loss=0.2800 pos=50688000 dt=34726
403
+ step 49600 loss=0.2730 pos=50790400 dt=34813
404
+ step 49700 loss=0.2937 pos=50892800 dt=34913
405
+ step 49800 loss=0.2638 pos=50995200 dt=35000
406
+ step 49900 loss=0.2692 pos=51097600 dt=35092
407
+ step 50000 loss=0.3324 pos=51200000 dt=35179
408
+ EVAL step 50000 pos=51200000 | T=2 loss=0.304 from=0.410 joint=0.723 hard=0.876 | T=16 loss=0.332 from=0.415 joint=0.690 hard=0.831
409
+ step 50100 loss=0.2606 pos=51302400 dt=35320
410
+ step 50200 loss=0.3057 pos=51404800 dt=35406
411
+ step 50300 loss=0.2853 pos=51507200 dt=35498
412
+ step 50400 loss=0.3106 pos=51609600 dt=35584
413
+ step 50500 loss=0.3254 pos=51712000 dt=35683
414
+ step 50600 loss=0.3196 pos=51814400 dt=35770
415
+ step 50700 loss=0.3224 pos=51916800 dt=35864
416
+ step 50800 loss=0.3035 pos=52019200 dt=35950
417
+ step 50900 loss=0.2989 pos=52121600 dt=36042
418
+ step 51000 loss=0.3287 pos=52224000 dt=36129
419
+ EVAL step 51000 pos=52224000 | T=2 loss=0.304 from=0.407 joint=0.735 hard=0.897 | T=16 loss=0.336 from=0.407 joint=0.698 hard=0.845
420
+ step 51100 loss=0.3391 pos=52326400 dt=36268
421
+ step 51200 loss=0.3160 pos=52428800 dt=36355
422
+ step 51300 loss=0.3123 pos=52531200 dt=36448
423
+ step 51400 loss=0.2937 pos=52633600 dt=36534
424
+ step 51500 loss=0.3038 pos=52736000 dt=36628
425
+ step 51600 loss=0.2871 pos=52838400 dt=36714
426
+ step 51700 loss=0.2994 pos=52940800 dt=36809
427
+ step 51800 loss=0.3162 pos=53043200 dt=36895
428
+ step 51900 loss=0.3201 pos=53145600 dt=36986
429
+ step 52000 loss=0.3398 pos=53248000 dt=37073
430
+ EVAL step 52000 pos=53248000 | T=2 loss=0.304 from=0.409 joint=0.729 hard=0.888 | T=16 loss=0.330 from=0.411 joint=0.686 hard=0.825
431
+ step 52100 loss=0.3032 pos=53350400 dt=37212
432
+ step 52200 loss=0.3118 pos=53452800 dt=37298
433
+ step 52300 loss=0.3211 pos=53555200 dt=37391
434
+ step 52400 loss=0.2921 pos=53657600 dt=37477
435
+ step 52500 loss=0.3152 pos=53760000 dt=37572
436
+ step 52600 loss=0.3116 pos=53862400 dt=37659
437
+ step 52700 loss=0.2699 pos=53964800 dt=37749
438
+ step 52800 loss=0.2695 pos=54067200 dt=37836
439
+ step 52900 loss=0.2925 pos=54169600 dt=37928
440
+ step 53000 loss=0.2779 pos=54272000 dt=38014
441
+ EVAL step 53000 pos=54272000 | T=2 loss=0.307 from=0.409 joint=0.738 hard=0.901 | T=16 loss=0.332 from=0.409 joint=0.701 hard=0.849
442
+ step 53100 loss=0.3259 pos=54374400 dt=38149
443
+ step 53200 loss=0.3092 pos=54476800 dt=38236
444
+ step 53300 loss=0.2671 pos=54579200 dt=38328
445
+ step 53400 loss=0.2698 pos=54681600 dt=38415
446
+ step 53500 loss=0.2608 pos=54784000 dt=38505
447
+ step 53600 loss=0.3033 pos=54886400 dt=38592
448
+ step 53700 loss=0.2991 pos=54988800 dt=38684
449
+ step 53800 loss=0.2876 pos=55091200 dt=38770
450
+ step 53900 loss=0.3127 pos=55193600 dt=38861
451
+ step 54000 loss=0.2774 pos=55296000 dt=38947
452
+ EVAL step 54000 pos=55296000 | T=2 loss=0.307 from=0.410 joint=0.739 hard=0.902 | T=16 loss=0.340 from=0.408 joint=0.707 hard=0.857
log_AMc.txt ADDED
@@ -0,0 +1,391 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ cfg AMc d=1280 H=10 tied=1 nl=8 lr=8 T=2 shards=500..750 hours=168.0
2
+ step 23700 loss=0.4073 pos=24268800 dt=91
3
+ step 23800 loss=0.4145 pos=24371200 dt=179
4
+ step 23900 loss=0.4169 pos=24473600 dt=279
5
+ step 24000 loss=0.4036 pos=24576000 dt=366
6
+ EVAL step 24000 pos=24576000 | T=2 loss=0.429 from=0.393 joint=0.649 hard=0.779 | T=16 loss=0.450 from=0.390 joint=0.623 hard=0.747
7
+ step 24100 loss=0.4281 pos=24678400 dt=506
8
+ step 24200 loss=0.4568 pos=24780800 dt=593
9
+ step 24300 loss=0.3895 pos=24883200 dt=694
10
+ step 24400 loss=0.4427 pos=24985600 dt=781
11
+ step 24500 loss=0.4227 pos=25088000 dt=873
12
+ step 24600 loss=0.3977 pos=25190400 dt=960
13
+ step 24700 loss=0.3790 pos=25292800 dt=1059
14
+ step 24800 loss=0.4452 pos=25395200 dt=1146
15
+ step 24900 loss=0.4136 pos=25497600 dt=1238
16
+ step 25000 loss=0.3813 pos=25600000 dt=1325
17
+ EVAL step 25000 pos=25600000 | T=2 loss=0.396 from=0.404 joint=0.682 hard=0.826 | T=16 loss=0.420 from=0.401 joint=0.666 hard=0.806
18
+ step 25100 loss=0.4037 pos=25702400 dt=1465
19
+ step 25200 loss=0.3904 pos=25804800 dt=1552
20
+ step 25300 loss=0.4170 pos=25907200 dt=1644
21
+ step 25400 loss=0.3907 pos=26009600 dt=1731
22
+ step 25500 loss=0.3651 pos=26112000 dt=1823
23
+ step 25600 loss=0.3886 pos=26214400 dt=1910
24
+ step 25700 loss=0.3835 pos=26316800 dt=2013
25
+ step 25800 loss=0.3655 pos=26419200 dt=2100
26
+ step 25900 loss=0.4158 pos=26521600 dt=2192
27
+ step 26000 loss=0.3674 pos=26624000 dt=2279
28
+ EVAL step 26000 pos=26624000 | T=2 loss=0.381 from=0.403 joint=0.677 hard=0.820 | T=16 loss=0.405 from=0.399 joint=0.660 hard=0.797
29
+ step 26100 loss=0.4068 pos=26726400 dt=2416
30
+ step 26200 loss=0.3765 pos=26828800 dt=2503
31
+ step 26300 loss=0.3958 pos=26931200 dt=2594
32
+ step 26400 loss=0.3540 pos=27033600 dt=2681
33
+ step 26500 loss=0.3740 pos=27136000 dt=2773
34
+ step 26600 loss=0.4108 pos=27238400 dt=2860
35
+ step 26700 loss=0.4261 pos=27340800 dt=2957
36
+ step 26800 loss=0.3841 pos=27443200 dt=3044
37
+ step 26900 loss=0.3157 pos=27545600 dt=3135
38
+ step 27000 loss=0.3464 pos=27648000 dt=3222
39
+ EVAL step 27000 pos=27648000 | T=2 loss=0.371 from=0.398 joint=0.686 hard=0.827 | T=16 loss=0.398 from=0.400 joint=0.660 hard=0.792
40
+ step 27100 loss=0.4133 pos=27750400 dt=3359
41
+ step 27200 loss=0.3687 pos=27852800 dt=3446
42
+ step 27300 loss=0.3527 pos=27955200 dt=3543
43
+ step 27400 loss=0.4107 pos=28057600 dt=3630
44
+ step 27500 loss=0.3453 pos=28160000 dt=3724
45
+ step 27600 loss=0.3631 pos=28262400 dt=3811
46
+ step 27700 loss=0.3619 pos=28364800 dt=3918
47
+ step 27800 loss=0.3842 pos=28467200 dt=4005
48
+ step 27900 loss=0.3405 pos=28569600 dt=4098
49
+ step 28000 loss=0.3555 pos=28672000 dt=4186
50
+ EVAL step 28000 pos=28672000 | T=2 loss=0.360 from=0.405 joint=0.697 hard=0.845 | T=16 loss=0.392 from=0.408 joint=0.664 hard=0.800
51
+ step 28100 loss=0.3667 pos=28774400 dt=4331
52
+ step 28200 loss=0.3872 pos=28876800 dt=4418
53
+ step 28300 loss=0.3696 pos=28979200 dt=4511
54
+ step 28400 loss=0.3418 pos=29081600 dt=4598
55
+ step 28500 loss=0.3628 pos=29184000 dt=4690
56
+ step 28600 loss=0.3474 pos=29286400 dt=4777
57
+ step 28700 loss=0.3339 pos=29388800 dt=4875
58
+ step 28800 loss=0.3274 pos=29491200 dt=4962
59
+ step 28900 loss=0.3366 pos=29593600 dt=5055
60
+ step 29000 loss=0.2930 pos=29696000 dt=5141
61
+ EVAL step 29000 pos=29696000 | T=2 loss=0.352 from=0.405 joint=0.701 hard=0.851 | T=16 loss=0.381 from=0.404 joint=0.674 hard=0.813
62
+ step 29100 loss=0.3375 pos=29798400 dt=5283
63
+ step 29200 loss=0.2925 pos=29900800 dt=5370
64
+ step 29300 loss=0.3178 pos=30003200 dt=5462
65
+ step 29400 loss=0.3243 pos=30105600 dt=5548
66
+ step 29500 loss=0.2772 pos=30208000 dt=5645
67
+ step 29600 loss=0.2432 pos=30310400 dt=5731
68
+ step 29700 loss=0.3642 pos=30412800 dt=5822
69
+ step 29800 loss=0.3887 pos=30515200 dt=5909
70
+ step 29900 loss=0.3660 pos=30617600 dt=6006
71
+ step 30000 loss=0.3313 pos=30720000 dt=6093
72
+ EVAL step 30000 pos=30720000 | T=2 loss=0.363 from=0.408 joint=0.696 hard=0.841 | T=16 loss=0.404 from=0.404 joint=0.655 hard=0.787
73
+ step 30100 loss=0.3645 pos=30822400 dt=6233
74
+ step 30200 loss=0.3757 pos=30924800 dt=6320
75
+ step 30300 loss=0.3751 pos=31027200 dt=6411
76
+ step 30400 loss=0.3754 pos=31129600 dt=6498
77
+ step 30500 loss=0.3599 pos=31232000 dt=6590
78
+ step 30600 loss=0.3625 pos=31334400 dt=6677
79
+ step 30700 loss=0.3469 pos=31436800 dt=6772
80
+ step 30800 loss=0.3616 pos=31539200 dt=6859
81
+ step 30900 loss=0.3469 pos=31641600 dt=6950
82
+ step 31000 loss=0.3436 pos=31744000 dt=7037
83
+ EVAL step 31000 pos=31744000 | T=2 loss=0.341 from=0.405 joint=0.702 hard=0.848 | T=16 loss=0.371 from=0.403 joint=0.674 hard=0.810
84
+ step 31100 loss=0.3984 pos=31846400 dt=7180
85
+ step 31200 loss=0.3552 pos=31948800 dt=7267
86
+ step 31300 loss=0.3427 pos=32051200 dt=7359
87
+ step 31400 loss=0.3479 pos=32153600 dt=7446
88
+ step 31500 loss=0.3649 pos=32256000 dt=7544
89
+ step 31600 loss=0.3850 pos=32358400 dt=7632
90
+ step 31700 loss=0.3277 pos=32460800 dt=7723
91
+ step 31800 loss=0.3446 pos=32563200 dt=7810
92
+ step 31900 loss=0.3853 pos=32665600 dt=7906
93
+ step 32000 loss=0.3106 pos=32768000 dt=7993
94
+ EVAL step 32000 pos=32768000 | T=2 loss=0.340 from=0.404 joint=0.716 hard=0.867 | T=16 loss=0.374 from=0.401 joint=0.690 hard=0.833
95
+ step 32100 loss=0.3395 pos=32870400 dt=8129
96
+ step 32200 loss=0.3691 pos=32972800 dt=8217
97
+ step 32300 loss=0.3300 pos=33075200 dt=8315
98
+ step 32400 loss=0.3208 pos=33177600 dt=8403
99
+ step 32500 loss=0.3444 pos=33280000 dt=8496
100
+ step 32600 loss=0.3696 pos=33382400 dt=8583
101
+ step 32700 loss=0.3499 pos=33484800 dt=8675
102
+ step 32800 loss=0.3524 pos=33587200 dt=8762
103
+ step 32900 loss=0.3759 pos=33689600 dt=8862
104
+ step 33000 loss=0.3483 pos=33792000 dt=8949
105
+ EVAL step 33000 pos=33792000 | T=2 loss=0.339 from=0.406 joint=0.714 hard=0.867 | T=16 loss=0.367 from=0.403 joint=0.682 hard=0.825
106
+ step 33100 loss=0.3294 pos=33894400 dt=9087
107
+ step 33200 loss=0.3330 pos=33996800 dt=9174
108
+ step 33300 loss=0.3323 pos=34099200 dt=9270
109
+ step 33400 loss=0.3628 pos=34201600 dt=9357
110
+ step 33500 loss=0.3197 pos=34304000 dt=9449
111
+ step 33600 loss=0.3370 pos=34406400 dt=9536
112
+ step 33700 loss=0.3580 pos=34508800 dt=9632
113
+ step 33800 loss=0.3161 pos=34611200 dt=9719
114
+ step 33900 loss=0.3576 pos=34713600 dt=9810
115
+ step 34000 loss=0.3519 pos=34816000 dt=9897
116
+ EVAL step 34000 pos=34816000 | T=2 loss=0.386 from=0.406 joint=0.719 hard=0.874 | T=16 loss=0.408 from=0.406 joint=0.673 hard=0.809
117
+ step 34100 loss=0.3490 pos=34918400 dt=10040
118
+ step 34200 loss=0.3456 pos=35020800 dt=10126
119
+ step 34300 loss=0.3467 pos=35123200 dt=10217
120
+ step 34400 loss=0.3332 pos=35225600 dt=10304
121
+ step 34500 loss=0.3068 pos=35328000 dt=10399
122
+ step 34600 loss=0.3157 pos=35430400 dt=10486
123
+ step 34700 loss=0.3148 pos=35532800 dt=10578
124
+ step 34800 loss=0.3257 pos=35635200 dt=10665
125
+ step 34900 loss=0.3051 pos=35737600 dt=10762
126
+ step 35000 loss=0.2915 pos=35840000 dt=10849
127
+ EVAL step 35000 pos=35840000 | T=2 loss=0.368 from=0.401 joint=0.693 hard=0.833 | T=16 loss=0.398 from=0.399 joint=0.655 hard=0.782
128
+ step 35100 loss=0.3245 pos=35942400 dt=10985
129
+ step 35200 loss=0.2715 pos=36044800 dt=11071
130
+ step 35300 loss=0.2569 pos=36147200 dt=11168
131
+ step 35400 loss=0.2502 pos=36249600 dt=11254
132
+ step 35500 loss=0.2179 pos=36352000 dt=11344
133
+ step 35600 loss=0.2498 pos=36454400 dt=11430
134
+ step 35700 loss=0.2858 pos=36556800 dt=11528
135
+ step 35800 loss=0.3297 pos=36659200 dt=11615
136
+ step 35900 loss=0.3005 pos=36761600 dt=11707
137
+ step 36000 loss=0.3512 pos=36864000 dt=11794
138
+ EVAL step 36000 pos=36864000 | T=2 loss=0.333 from=0.411 joint=0.714 hard=0.865 | T=16 loss=0.358 from=0.412 joint=0.687 hard=0.828
139
+ step 36100 loss=0.3320 pos=36966400 dt=11931
140
+ step 36200 loss=0.3547 pos=37068800 dt=12018
141
+ step 36300 loss=0.2912 pos=37171200 dt=12118
142
+ step 36400 loss=0.3573 pos=37273600 dt=12205
143
+ step 36500 loss=0.3437 pos=37376000 dt=12298
144
+ step 36600 loss=0.3178 pos=37478400 dt=12385
145
+ step 36700 loss=0.3318 pos=37580800 dt=12484
146
+ step 36800 loss=0.3572 pos=37683200 dt=12571
147
+ step 36900 loss=0.3180 pos=37785600 dt=12665
148
+ step 37000 loss=0.3124 pos=37888000 dt=12752
149
+ EVAL step 37000 pos=37888000 | T=2 loss=0.318 from=0.410 joint=0.732 hard=0.891 | T=16 loss=0.345 from=0.413 joint=0.685 hard=0.827
150
+ step 37100 loss=0.3506 pos=37990400 dt=12892
151
+ step 37200 loss=0.3405 pos=38092800 dt=12980
152
+ step 37300 loss=0.3240 pos=38195200 dt=13073
153
+ step 37400 loss=0.2886 pos=38297600 dt=13160
154
+ step 37500 loss=0.3566 pos=38400000 dt=13259
155
+ step 37600 loss=0.3255 pos=38502400 dt=13346
156
+ step 37700 loss=0.3281 pos=38604800 dt=13437
157
+ step 37800 loss=0.3489 pos=38707200 dt=13524
158
+ step 37900 loss=0.3194 pos=38809600 dt=13620
159
+ step 38000 loss=0.3289 pos=38912000 dt=13707
160
+ EVAL step 38000 pos=38912000 | T=2 loss=0.326 from=0.407 joint=0.716 hard=0.872 | T=16 loss=0.356 from=0.408 joint=0.685 hard=0.829
161
+ step 38100 loss=0.3316 pos=39014400 dt=13851
162
+ step 38200 loss=0.3222 pos=39116800 dt=13938
163
+ step 38300 loss=0.3417 pos=39219200 dt=14034
164
+ step 38400 loss=0.3391 pos=39321600 dt=14121
165
+ step 38500 loss=0.3266 pos=39424000 dt=14213
166
+ step 38600 loss=0.3419 pos=39526400 dt=14300
167
+ step 38700 loss=0.3430 pos=39628800 dt=14398
168
+ step 38800 loss=0.3459 pos=39731200 dt=14485
169
+ step 38900 loss=0.3122 pos=39833600 dt=14581
170
+ step 39000 loss=0.3172 pos=39936000 dt=14668
171
+ EVAL step 39000 pos=39936000 | T=2 loss=0.326 from=0.403 joint=0.722 hard=0.876 | T=16 loss=0.348 from=0.402 joint=0.685 hard=0.824
172
+ step 39100 loss=0.3308 pos=40038400 dt=14805
173
+ step 39200 loss=0.2935 pos=40140800 dt=14892
174
+ step 39300 loss=0.3179 pos=40243200 dt=14988
175
+ step 39400 loss=0.3108 pos=40345600 dt=15075
176
+ step 39500 loss=0.3094 pos=40448000 dt=15169
177
+ step 39600 loss=0.3115 pos=40550400 dt=15256
178
+ step 39700 loss=0.3226 pos=40652800 dt=15352
179
+ step 39800 loss=0.3186 pos=40755200 dt=15439
180
+ step 39900 loss=0.3148 pos=40857600 dt=15534
181
+ step 40000 loss=0.3187 pos=40960000 dt=15621
182
+ EVAL step 40000 pos=40960000 | T=2 loss=0.318 from=0.404 joint=0.729 hard=0.887 | T=16 loss=0.350 from=0.402 joint=0.693 hard=0.839
183
+ step 40100 loss=0.3019 pos=41062400 dt=15765
184
+ step 40200 loss=0.2984 pos=41164800 dt=15852
185
+ step 40300 loss=0.3020 pos=41267200 dt=15943
186
+ step 40400 loss=0.3253 pos=41369600 dt=16030
187
+ step 40500 loss=0.3285 pos=41472000 dt=16128
188
+ step 40600 loss=0.2868 pos=41574400 dt=16214
189
+ step 40700 loss=0.2788 pos=41676800 dt=16308
190
+ step 40800 loss=0.2847 pos=41779200 dt=16395
191
+ step 40900 loss=0.2944 pos=41881600 dt=16490
192
+ step 41000 loss=0.2691 pos=41984000 dt=16577
193
+ EVAL step 41000 pos=41984000 | T=2 loss=0.322 from=0.405 joint=0.714 hard=0.868 | T=16 loss=0.349 from=0.405 joint=0.686 hard=0.829
194
+ step 41100 loss=0.2418 pos=42086400 dt=16714
195
+ step 41200 loss=0.2820 pos=42188800 dt=16801
196
+ step 41300 loss=0.2568 pos=42291200 dt=16895
197
+ step 41400 loss=0.2481 pos=42393600 dt=16982
198
+ step 41500 loss=0.2625 pos=42496000 dt=17073
199
+ step 41600 loss=0.2410 pos=42598400 dt=17159
200
+ step 41700 loss=0.3165 pos=42700800 dt=17256
201
+ step 41800 loss=0.3229 pos=42803200 dt=17343
202
+ step 41900 loss=0.3056 pos=42905600 dt=17436
203
+ step 42000 loss=0.3010 pos=43008000 dt=17523
204
+ EVAL step 42000 pos=43008000 | T=2 loss=0.316 from=0.405 joint=0.725 hard=0.883 | T=16 loss=0.344 from=0.409 joint=0.685 hard=0.826
205
+ step 42100 loss=0.3219 pos=43110400 dt=17659
206
+ step 42200 loss=0.2858 pos=43212800 dt=17746
207
+ step 42300 loss=0.3335 pos=43315200 dt=17842
208
+ step 42400 loss=0.3520 pos=43417600 dt=17929
209
+ step 42500 loss=0.3022 pos=43520000 dt=18021
210
+ step 42600 loss=0.3128 pos=43622400 dt=18108
211
+ step 42700 loss=0.2995 pos=43724800 dt=18208
212
+ step 42800 loss=0.3385 pos=43827200 dt=18295
213
+ step 42900 loss=0.3451 pos=43929600 dt=18388
214
+ step 43000 loss=0.3142 pos=44032000 dt=18475
215
+ EVAL step 43000 pos=44032000 | T=2 loss=0.311 from=0.407 joint=0.728 hard=0.885 | T=16 loss=0.339 from=0.406 joint=0.694 hard=0.835
216
+ step 43100 loss=0.2994 pos=44134400 dt=18619
217
+ step 43200 loss=0.3059 pos=44236800 dt=18706
218
+ step 43300 loss=0.3022 pos=44339200 dt=18798
219
+ step 43400 loss=0.2806 pos=44441600 dt=18885
220
+ step 43500 loss=0.3179 pos=44544000 dt=18983
221
+ step 43600 loss=0.3116 pos=44646400 dt=19070
222
+ step 43700 loss=0.2655 pos=44748800 dt=19162
223
+ step 43800 loss=0.3234 pos=44851200 dt=19249
224
+ step 43900 loss=0.2949 pos=44953600 dt=19344
225
+ step 44000 loss=0.2971 pos=45056000 dt=19431
226
+ EVAL step 44000 pos=45056000 | T=2 loss=0.318 from=0.409 joint=0.708 hard=0.857 | T=16 loss=0.351 from=0.403 joint=0.660 hard=0.788
227
+ step 44100 loss=0.3045 pos=45158400 dt=19570
228
+ step 44200 loss=0.2826 pos=45260800 dt=19657
229
+ step 44300 loss=0.2907 pos=45363200 dt=19752
230
+ step 44400 loss=0.2936 pos=45465600 dt=19839
231
+ step 44500 loss=0.3305 pos=45568000 dt=19931
232
+ step 44600 loss=0.2980 pos=45670400 dt=20018
233
+ step 44700 loss=0.2913 pos=45772800 dt=20113
234
+ step 44800 loss=0.2629 pos=45875200 dt=20200
235
+ step 44900 loss=0.3340 pos=45977600 dt=20291
236
+ step 45000 loss=0.3301 pos=46080000 dt=20379
237
+ EVAL step 45000 pos=46080000 | T=2 loss=0.311 from=0.410 joint=0.735 hard=0.896 | T=16 loss=0.336 from=0.410 joint=0.694 hard=0.838
238
+ step 45100 loss=0.3103 pos=46182400 dt=20522
239
+ step 45200 loss=0.3068 pos=46284800 dt=20609
240
+ step 45300 loss=0.2986 pos=46387200 dt=20700
241
+ step 45400 loss=0.2920 pos=46489600 dt=20787
242
+ step 45500 loss=0.3042 pos=46592000 dt=20882
243
+ step 45600 loss=0.2951 pos=46694400 dt=20969
244
+ step 45700 loss=0.2965 pos=46796800 dt=21060
245
+ step 45800 loss=0.3056 pos=46899200 dt=21146
246
+ step 45900 loss=0.3241 pos=47001600 dt=21243
247
+ step 46000 loss=0.3015 pos=47104000 dt=21329
248
+ EVAL step 46000 pos=47104000 | T=2 loss=0.319 from=0.415 joint=0.725 hard=0.883 | T=16 loss=0.343 from=0.419 joint=0.686 hard=0.827
249
+ step 46100 loss=0.3010 pos=47206400 dt=21469
250
+ step 46200 loss=0.2888 pos=47308800 dt=21556
251
+ step 46300 loss=0.2906 pos=47411200 dt=21653
252
+ step 46400 loss=0.2740 pos=47513600 dt=21740
253
+ step 46500 loss=0.2757 pos=47616000 dt=21831
254
+ step 46600 loss=0.3064 pos=47718400 dt=21918
255
+ step 46700 loss=0.2971 pos=47820800 dt=22017
256
+ step 46800 loss=0.2609 pos=47923200 dt=22104
257
+ step 46900 loss=0.3052 pos=48025600 dt=22195
258
+ step 47000 loss=0.2578 pos=48128000 dt=22282
259
+ EVAL step 47000 pos=48128000 | T=2 loss=0.312 from=0.410 joint=0.714 hard=0.865 | T=16 loss=0.339 from=0.409 joint=0.680 hard=0.813
260
+ step 47100 loss=0.2841 pos=48230400 dt=22426
261
+ step 47200 loss=0.2802 pos=48332800 dt=22513
262
+ step 47300 loss=0.2765 pos=48435200 dt=22604
263
+ step 47400 loss=0.2137 pos=48537600 dt=22691
264
+ step 47500 loss=0.2645 pos=48640000 dt=22787
265
+ step 47600 loss=0.2345 pos=48742400 dt=22872
266
+ step 47700 loss=0.2243 pos=48844800 dt=22963
267
+ step 47800 loss=0.3085 pos=48947200 dt=23049
268
+ step 47900 loss=0.2884 pos=49049600 dt=23145
269
+ step 48000 loss=0.2822 pos=49152000 dt=23232
270
+ EVAL step 48000 pos=49152000 | T=2 loss=0.315 from=0.408 joint=0.730 hard=0.887 | T=16 loss=0.340 from=0.412 joint=0.697 hard=0.844
271
+ step 48100 loss=0.3001 pos=49254400 dt=23371
272
+ step 48200 loss=0.3049 pos=49356800 dt=23458
273
+ step 48300 loss=0.3124 pos=49459200 dt=23556
274
+ step 48400 loss=0.3016 pos=49561600 dt=23643
275
+ step 48500 loss=0.3150 pos=49664000 dt=23735
276
+ step 48600 loss=0.2947 pos=49766400 dt=23822
277
+ step 48700 loss=0.2718 pos=49868800 dt=23921
278
+ step 48800 loss=0.3109 pos=49971200 dt=24008
279
+ step 48900 loss=0.3021 pos=50073600 dt=24101
280
+ step 49000 loss=0.3048 pos=50176000 dt=24188
281
+ EVAL step 49000 pos=50176000 | T=2 loss=0.309 from=0.407 joint=0.737 hard=0.901 | T=16 loss=0.341 from=0.405 joint=0.698 hard=0.846
282
+ step 49100 loss=0.3087 pos=50278400 dt=24326
283
+ step 49200 loss=0.3191 pos=50380800 dt=24412
284
+ step 49300 loss=0.3331 pos=50483200 dt=24504
285
+ step 49400 loss=0.3215 pos=50585600 dt=24591
286
+ step 49500 loss=0.3071 pos=50688000 dt=24685
287
+ step 49600 loss=0.2805 pos=50790400 dt=24772
288
+ step 49700 loss=0.2883 pos=50892800 dt=24865
289
+ step 49800 loss=0.3416 pos=50995200 dt=24952
290
+ step 49900 loss=0.2772 pos=51097600 dt=25045
291
+ step 50000 loss=0.3193 pos=51200000 dt=25132
292
+ EVAL step 50000 pos=51200000 | T=2 loss=0.313 from=0.409 joint=0.736 hard=0.897 | T=16 loss=0.336 from=0.409 joint=0.699 hard=0.844
293
+ step 50100 loss=0.3107 pos=51302400 dt=25270
294
+ step 50200 loss=0.2894 pos=51404800 dt=25357
295
+ step 50300 loss=0.3123 pos=51507200 dt=25449
296
+ step 50400 loss=0.3001 pos=51609600 dt=25536
297
+ step 50500 loss=0.3226 pos=51712000 dt=25629
298
+ step 50600 loss=0.3170 pos=51814400 dt=25716
299
+ step 50700 loss=0.3036 pos=51916800 dt=25811
300
+ step 50800 loss=0.3276 pos=52019200 dt=25898
301
+ step 50900 loss=0.3095 pos=52121600 dt=25990
302
+ step 51000 loss=0.3347 pos=52224000 dt=26077
303
+ EVAL step 51000 pos=52224000 | T=2 loss=0.313 from=0.409 joint=0.727 hard=0.886 | T=16 loss=0.340 from=0.407 joint=0.690 hard=0.835
304
+ step 51100 loss=0.3062 pos=52326400 dt=26214
305
+ step 51200 loss=0.3047 pos=52428800 dt=26301
306
+ step 51300 loss=0.2691 pos=52531200 dt=26393
307
+ step 51400 loss=0.2780 pos=52633600 dt=26480
308
+ step 51500 loss=0.3004 pos=52736000 dt=26571
309
+ step 51600 loss=0.2734 pos=52838400 dt=26658
310
+ step 51700 loss=0.3090 pos=52940800 dt=26749
311
+ step 51800 loss=0.2879 pos=53043200 dt=26836
312
+ step 51900 loss=0.3121 pos=53145600 dt=26927
313
+ step 52000 loss=0.3307 pos=53248000 dt=27014
314
+ EVAL step 52000 pos=53248000 | T=2 loss=0.313 from=0.405 joint=0.724 hard=0.878 | T=16 loss=0.347 from=0.400 joint=0.685 hard=0.824
315
+ step 52100 loss=0.2701 pos=53350400 dt=27151
316
+ step 52200 loss=0.3178 pos=53452800 dt=27238
317
+ step 52300 loss=0.3013 pos=53555200 dt=27329
318
+ step 52400 loss=0.3009 pos=53657600 dt=27416
319
+ step 52500 loss=0.2954 pos=53760000 dt=27508
320
+ step 52600 loss=0.2819 pos=53862400 dt=27595
321
+ step 52700 loss=0.2638 pos=53964800 dt=27687
322
+ step 52800 loss=0.2888 pos=54067200 dt=27774
323
+ step 52900 loss=0.2920 pos=54169600 dt=27866
324
+ step 53000 loss=0.2709 pos=54272000 dt=27952
325
+ EVAL step 53000 pos=54272000 | T=2 loss=0.311 from=0.408 joint=0.727 hard=0.885 | T=16 loss=0.336 from=0.406 joint=0.693 hard=0.838
326
+ step 53100 loss=0.2992 pos=54374400 dt=28088
327
+ step 53200 loss=0.2721 pos=54476800 dt=28174
328
+ step 53300 loss=0.2634 pos=54579200 dt=28267
329
+ step 53400 loss=0.2455 pos=54681600 dt=28353
330
+ step 53500 loss=0.2625 pos=54784000 dt=28443
331
+ step 53600 loss=0.1968 pos=54886400 dt=28529
332
+ step 53700 loss=0.2372 pos=54988800 dt=28619
333
+ step 53800 loss=0.2679 pos=55091200 dt=28705
334
+ step 53900 loss=0.3486 pos=55193600 dt=28797
335
+ step 54000 loss=0.2935 pos=55296000 dt=28883
336
+ EVAL step 54000 pos=55296000 | T=2 loss=0.315 from=0.412 joint=0.724 hard=0.879 | T=16 loss=0.343 from=0.407 joint=0.676 hard=0.815
337
+ step 54100 loss=0.3410 pos=55398400 dt=29020
338
+ step 54200 loss=0.2680 pos=55500800 dt=29107
339
+ step 54300 loss=0.3101 pos=55603200 dt=29200
340
+ step 54400 loss=0.2897 pos=55705600 dt=29287
341
+ step 54500 loss=0.2832 pos=55808000 dt=29381
342
+ step 54600 loss=0.2813 pos=55910400 dt=29468
343
+ step 54700 loss=0.2904 pos=56012800 dt=29560
344
+ step 54800 loss=0.3006 pos=56115200 dt=29647
345
+ step 54900 loss=0.3077 pos=56217600 dt=29740
346
+ step 55000 loss=0.3060 pos=56320000 dt=29827
347
+ EVAL step 55000 pos=56320000 | T=2 loss=0.306 from=0.409 joint=0.735 hard=0.894 | T=16 loss=0.334 from=0.409 joint=0.679 hard=0.817
348
+ step 55100 loss=0.3000 pos=56422400 dt=29968
349
+ step 55200 loss=0.3229 pos=56524800 dt=30055
350
+ step 55300 loss=0.2914 pos=56627200 dt=30147
351
+ step 55400 loss=0.3321 pos=56729600 dt=30234
352
+ step 55500 loss=0.3096 pos=56832000 dt=30328
353
+ step 55600 loss=0.2982 pos=56934400 dt=30415
354
+ step 55700 loss=0.2670 pos=57036800 dt=30509
355
+ step 55800 loss=0.3048 pos=57139200 dt=30596
356
+ step 55900 loss=0.3135 pos=57241600 dt=30688
357
+ step 56000 loss=0.2837 pos=57344000 dt=30775
358
+ EVAL step 56000 pos=57344000 | T=2 loss=0.302 from=0.412 joint=0.731 hard=0.890 | T=16 loss=0.329 from=0.412 joint=0.689 hard=0.834
359
+ step 56100 loss=0.2777 pos=57446400 dt=30915
360
+ step 56200 loss=0.3102 pos=57548800 dt=31002
361
+ step 56300 loss=0.2964 pos=57651200 dt=31094
362
+ step 56400 loss=0.3012 pos=57753600 dt=31181
363
+ step 56500 loss=0.2954 pos=57856000 dt=31273
364
+ step 56600 loss=0.2924 pos=57958400 dt=31360
365
+ step 56700 loss=0.2720 pos=58060800 dt=31455
366
+ step 56800 loss=0.3053 pos=58163200 dt=31542
367
+ step 56900 loss=0.2904 pos=58265600 dt=31634
368
+ step 57000 loss=0.3343 pos=58368000 dt=31721
369
+ EVAL step 57000 pos=58368000 | T=2 loss=0.313 from=0.413 joint=0.734 hard=0.894 | T=16 loss=0.334 from=0.410 joint=0.696 hard=0.841
370
+ step 57100 loss=0.2792 pos=58470400 dt=31859
371
+ step 57200 loss=0.2985 pos=58572800 dt=31946
372
+ step 57300 loss=0.2777 pos=58675200 dt=32039
373
+ step 57400 loss=0.3065 pos=58777600 dt=32126
374
+ step 57500 loss=0.3082 pos=58880000 dt=32218
375
+ step 57600 loss=0.2917 pos=58982400 dt=32304
376
+ step 57700 loss=0.2700 pos=59084800 dt=32399
377
+ step 57800 loss=0.2802 pos=59187200 dt=32486
378
+ step 57900 loss=0.2841 pos=59289600 dt=32583
379
+ step 58000 loss=0.2954 pos=59392000 dt=32670
380
+ EVAL step 58000 pos=59392000 | T=2 loss=0.306 from=0.411 joint=0.735 hard=0.894 | T=16 loss=0.332 from=0.408 joint=0.699 hard=0.846
381
+ step 58100 loss=0.2903 pos=59494400 dt=32807
382
+ step 58200 loss=0.2957 pos=59596800 dt=32894
383
+ step 58300 loss=0.2484 pos=59699200 dt=32986
384
+ step 58400 loss=0.2741 pos=59801600 dt=33073
385
+ step 58500 loss=0.2731 pos=59904000 dt=33165
386
+ step 58600 loss=0.2730 pos=60006400 dt=33252
387
+ step 58700 loss=0.2584 pos=60108800 dt=33344
388
+ step 58800 loss=0.2668 pos=60211200 dt=33431
389
+ step 58900 loss=0.2537 pos=60313600 dt=33522
390
+ step 59000 loss=0.2806 pos=60416000 dt=33609
391
+ EVAL step 59000 pos=60416000 | T=2 loss=0.305 from=0.407 joint=0.742 hard=0.906 | T=16 loss=0.332 from=0.409 joint=0.703 hard=0.852
log_AMd.txt ADDED
@@ -0,0 +1,454 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ cfg AMd d=1280 H=10 tied=1 nl=8 lr=8 T=2 shards=750..1000 hours=168.0
2
+ step 13100 loss=0.6690 pos=13414400 dt=92
3
+ step 13200 loss=0.6455 pos=13516800 dt=179
4
+ step 13300 loss=0.8357 pos=13619200 dt=276
5
+ step 13400 loss=0.5983 pos=13721600 dt=362
6
+ step 13500 loss=0.5179 pos=13824000 dt=453
7
+ step 13600 loss=0.5718 pos=13926400 dt=540
8
+ step 13700 loss=0.5185 pos=14028800 dt=638
9
+ step 13800 loss=0.5214 pos=14131200 dt=725
10
+ step 13900 loss=0.5496 pos=14233600 dt=815
11
+ step 14000 loss=0.5747 pos=14336000 dt=902
12
+ EVAL step 14000 pos=14336000 | T=2 loss=0.537 from=0.386 joint=0.626 hard=0.754 | T=16 loss=0.558 from=0.379 joint=0.609 hard=0.728
13
+ step 14100 loss=0.4980 pos=14438400 dt=1042
14
+ step 14200 loss=0.5183 pos=14540800 dt=1128
15
+ step 14300 loss=0.5227 pos=14643200 dt=1219
16
+ step 14400 loss=0.4609 pos=14745600 dt=1306
17
+ step 14500 loss=0.4842 pos=14848000 dt=1400
18
+ step 14600 loss=0.4401 pos=14950400 dt=1487
19
+ step 14700 loss=0.4346 pos=15052800 dt=1577
20
+ step 14800 loss=0.4672 pos=15155200 dt=1664
21
+ step 14900 loss=0.4273 pos=15257600 dt=1762
22
+ step 15000 loss=0.4218 pos=15360000 dt=1848
23
+ EVAL step 15000 pos=15360000 | T=2 loss=0.448 from=0.394 joint=0.654 hard=0.793 | T=16 loss=0.487 from=0.394 joint=0.638 hard=0.771
24
+ step 15100 loss=0.4706 pos=15462400 dt=1984
25
+ step 15200 loss=0.4059 pos=15564800 dt=2070
26
+ step 15300 loss=0.4546 pos=15667200 dt=2165
27
+ step 15400 loss=0.4403 pos=15769600 dt=2251
28
+ step 15500 loss=0.4088 pos=15872000 dt=2342
29
+ step 15600 loss=0.4573 pos=15974400 dt=2428
30
+ step 15700 loss=0.4372 pos=16076800 dt=2524
31
+ step 15800 loss=0.4211 pos=16179200 dt=2611
32
+ step 15900 loss=0.4309 pos=16281600 dt=2702
33
+ step 16000 loss=0.4511 pos=16384000 dt=2788
34
+ EVAL step 16000 pos=16384000 | T=2 loss=0.426 from=0.388 joint=0.663 hard=0.803 | T=16 loss=0.454 from=0.388 joint=0.637 hard=0.770
35
+ step 16100 loss=0.4267 pos=16486400 dt=2933
36
+ step 16200 loss=0.4061 pos=16588800 dt=3019
37
+ step 16300 loss=0.4038 pos=16691200 dt=3113
38
+ step 16400 loss=0.4574 pos=16793600 dt=3200
39
+ step 16500 loss=0.3902 pos=16896000 dt=3295
40
+ step 16600 loss=0.4072 pos=16998400 dt=3382
41
+ step 16700 loss=0.4142 pos=17100800 dt=3475
42
+ step 16800 loss=0.3669 pos=17203200 dt=3561
43
+ step 16900 loss=0.3875 pos=17305600 dt=3659
44
+ step 17000 loss=0.3737 pos=17408000 dt=3745
45
+ EVAL step 17000 pos=17408000 | T=2 loss=0.414 from=0.399 joint=0.667 hard=0.803 | T=16 loss=0.447 from=0.398 joint=0.649 hard=0.778
46
+ step 17100 loss=0.4335 pos=17510400 dt=3881
47
+ step 17200 loss=0.4248 pos=17612800 dt=3968
48
+ step 17300 loss=0.4133 pos=17715200 dt=4063
49
+ step 17400 loss=0.4308 pos=17817600 dt=4150
50
+ step 17500 loss=0.3980 pos=17920000 dt=4243
51
+ step 17600 loss=0.4081 pos=18022400 dt=4329
52
+ step 17700 loss=0.3854 pos=18124800 dt=4427
53
+ step 17800 loss=0.3604 pos=18227200 dt=4513
54
+ step 17900 loss=0.3788 pos=18329600 dt=4604
55
+ step 18000 loss=0.3858 pos=18432000 dt=4691
56
+ EVAL step 18000 pos=18432000 | T=2 loss=0.397 from=0.401 joint=0.682 hard=0.825 | T=16 loss=0.437 from=0.397 joint=0.646 hard=0.777
57
+ step 18100 loss=0.3722 pos=18534400 dt=4833
58
+ step 18200 loss=0.4146 pos=18636800 dt=4919
59
+ step 18300 loss=0.3327 pos=18739200 dt=5011
60
+ step 18400 loss=0.3499 pos=18841600 dt=5096
61
+ step 18500 loss=0.3457 pos=18944000 dt=5193
62
+ step 18600 loss=0.3319 pos=19046400 dt=5279
63
+ step 18700 loss=0.3270 pos=19148800 dt=5370
64
+ step 18800 loss=0.3479 pos=19251200 dt=5455
65
+ step 18900 loss=0.3048 pos=19353600 dt=5554
66
+ step 19000 loss=0.2930 pos=19456000 dt=5639
67
+ EVAL step 19000 pos=19456000 | T=2 loss=0.566 from=0.404 joint=0.576 hard=0.688 | T=16 loss=0.644 from=0.402 joint=0.552 hard=0.656
68
+ step 19100 loss=0.4574 pos=19558400 dt=5774
69
+ step 19200 loss=0.4525 pos=19660800 dt=5861
70
+ step 19300 loss=0.3792 pos=19763200 dt=5959
71
+ step 19400 loss=0.3864 pos=19865600 dt=6046
72
+ step 19500 loss=0.3596 pos=19968000 dt=6137
73
+ step 19600 loss=0.3895 pos=20070400 dt=6224
74
+ step 19700 loss=0.3906 pos=20172800 dt=6322
75
+ step 19800 loss=0.3617 pos=20275200 dt=6408
76
+ step 19900 loss=0.3467 pos=20377600 dt=6499
77
+ step 20000 loss=0.3969 pos=20480000 dt=6586
78
+ EVAL step 20000 pos=20480000 | T=2 loss=0.369 from=0.403 joint=0.702 hard=0.854 | T=16 loss=0.398 from=0.405 joint=0.668 hard=0.806
79
+ step 20100 loss=0.3712 pos=20582400 dt=6723
80
+ step 20200 loss=0.3677 pos=20684800 dt=6810
81
+ step 20300 loss=0.4230 pos=20787200 dt=6907
82
+ step 20400 loss=0.3997 pos=20889600 dt=6994
83
+ step 20500 loss=0.3925 pos=20992000 dt=7084
84
+ step 20600 loss=0.3858 pos=21094400 dt=7171
85
+ step 20700 loss=0.4115 pos=21196800 dt=7274
86
+ step 20800 loss=0.3627 pos=21299200 dt=7360
87
+ step 20900 loss=0.3733 pos=21401600 dt=7451
88
+ step 21000 loss=0.3806 pos=21504000 dt=7537
89
+ EVAL step 21000 pos=21504000 | T=2 loss=0.362 from=0.406 joint=0.690 hard=0.838 | T=16 loss=0.393 from=0.406 joint=0.667 hard=0.806
90
+ step 21100 loss=0.3704 pos=21606400 dt=7677
91
+ step 21200 loss=0.3855 pos=21708800 dt=7763
92
+ step 21300 loss=0.3751 pos=21811200 dt=7855
93
+ step 21400 loss=0.3738 pos=21913600 dt=7941
94
+ step 21500 loss=0.4106 pos=22016000 dt=8032
95
+ step 21600 loss=0.3450 pos=22118400 dt=8119
96
+ step 21700 loss=0.3578 pos=22220800 dt=8210
97
+ step 21800 loss=0.3864 pos=22323200 dt=8296
98
+ step 21900 loss=0.3669 pos=22425600 dt=8387
99
+ step 22000 loss=0.3536 pos=22528000 dt=8473
100
+ EVAL step 22000 pos=22528000 | T=2 loss=0.386 from=0.407 joint=0.680 hard=0.822 | T=16 loss=0.425 from=0.410 joint=0.644 hard=0.771
101
+ step 22100 loss=0.3652 pos=22630400 dt=8619
102
+ step 22200 loss=0.3794 pos=22732800 dt=8706
103
+ step 22300 loss=0.3811 pos=22835200 dt=8797
104
+ step 22400 loss=0.3722 pos=22937600 dt=8883
105
+ step 22500 loss=0.3704 pos=23040000 dt=8974
106
+ step 22600 loss=0.3505 pos=23142400 dt=9060
107
+ step 22700 loss=0.3659 pos=23244800 dt=9157
108
+ step 22800 loss=0.3765 pos=23347200 dt=9243
109
+ step 22900 loss=0.3522 pos=23449600 dt=9334
110
+ step 23000 loss=0.3496 pos=23552000 dt=9421
111
+ EVAL step 23000 pos=23552000 | T=2 loss=0.370 from=0.404 joint=0.686 hard=0.825 | T=16 loss=0.400 from=0.402 joint=0.658 hard=0.790
112
+ step 23100 loss=0.3415 pos=23654400 dt=9565
113
+ step 23200 loss=0.3571 pos=23756800 dt=9651
114
+ step 23300 loss=0.3397 pos=23859200 dt=9742
115
+ step 23400 loss=0.3499 pos=23961600 dt=9828
116
+ step 23500 loss=0.3709 pos=24064000 dt=9925
117
+ step 23600 loss=0.3452 pos=24166400 dt=10011
118
+ step 23700 loss=0.3455 pos=24268800 dt=10103
119
+ step 23800 loss=0.3619 pos=24371200 dt=10189
120
+ step 23900 loss=0.3339 pos=24473600 dt=10281
121
+ step 24000 loss=0.3285 pos=24576000 dt=10368
122
+ EVAL step 24000 pos=24576000 | T=2 loss=0.370 from=0.405 joint=0.699 hard=0.846 | T=16 loss=0.404 from=0.402 joint=0.673 hard=0.810
123
+ step 24100 loss=0.3256 pos=24678400 dt=10503
124
+ step 24200 loss=0.3261 pos=24780800 dt=10589
125
+ step 24300 loss=0.3348 pos=24883200 dt=10680
126
+ step 24400 loss=0.3160 pos=24985600 dt=10766
127
+ step 24500 loss=0.3353 pos=25088000 dt=10858
128
+ step 24600 loss=0.2983 pos=25190400 dt=10944
129
+ step 24700 loss=0.3167 pos=25292800 dt=11040
130
+ step 24800 loss=0.2744 pos=25395200 dt=11125
131
+ step 24900 loss=0.2835 pos=25497600 dt=11216
132
+ step 25000 loss=0.3023 pos=25600000 dt=11301
133
+ EVAL step 25000 pos=25600000 | T=2 loss=0.411 from=0.412 joint=0.601 hard=0.719 | T=16 loss=0.463 from=0.411 joint=0.583 hard=0.696
134
+ step 25100 loss=0.3401 pos=25702400 dt=11439
135
+ step 25200 loss=0.3923 pos=25804800 dt=11526
136
+ step 25300 loss=0.3761 pos=25907200 dt=11617
137
+ step 25400 loss=0.3500 pos=26009600 dt=11703
138
+ step 25500 loss=0.3566 pos=26112000 dt=11794
139
+ step 25600 loss=0.3504 pos=26214400 dt=11881
140
+ step 25700 loss=0.3856 pos=26316800 dt=11973
141
+ step 25800 loss=0.3686 pos=26419200 dt=12059
142
+ step 25900 loss=0.3798 pos=26521600 dt=12151
143
+ step 26000 loss=0.3733 pos=26624000 dt=12237
144
+ EVAL step 26000 pos=26624000 | T=2 loss=0.357 from=0.402 joint=0.707 hard=0.859 | T=16 loss=0.383 from=0.401 joint=0.672 hard=0.812
145
+ step 26100 loss=0.3826 pos=26726400 dt=12381
146
+ step 26200 loss=0.3666 pos=26828800 dt=12468
147
+ step 26300 loss=0.3724 pos=26931200 dt=12560
148
+ step 26400 loss=0.3540 pos=27033600 dt=12647
149
+ step 26500 loss=0.3538 pos=27136000 dt=12745
150
+ step 26600 loss=0.3748 pos=27238400 dt=12831
151
+ step 26700 loss=0.3977 pos=27340800 dt=12924
152
+ step 26800 loss=0.3873 pos=27443200 dt=13011
153
+ step 26900 loss=0.3685 pos=27545600 dt=13104
154
+ step 27000 loss=0.3692 pos=27648000 dt=13191
155
+ EVAL step 27000 pos=27648000 | T=2 loss=0.351 from=0.404 joint=0.708 hard=0.861 | T=16 loss=0.382 from=0.405 joint=0.672 hard=0.812
156
+ step 27100 loss=0.3902 pos=27750400 dt=13327
157
+ step 27200 loss=0.3762 pos=27852800 dt=13413
158
+ step 27300 loss=0.3254 pos=27955200 dt=13506
159
+ step 27400 loss=0.3708 pos=28057600 dt=13592
160
+ step 27500 loss=0.3134 pos=28160000 dt=13693
161
+ step 27600 loss=0.3661 pos=28262400 dt=13779
162
+ step 27700 loss=0.3554 pos=28364800 dt=13871
163
+ step 27800 loss=0.3354 pos=28467200 dt=13957
164
+ step 27900 loss=0.3324 pos=28569600 dt=14049
165
+ step 28000 loss=0.3385 pos=28672000 dt=14136
166
+ EVAL step 28000 pos=28672000 | T=2 loss=0.343 from=0.408 joint=0.717 hard=0.872 | T=16 loss=0.376 from=0.408 joint=0.690 hard=0.836
167
+ step 28100 loss=0.3524 pos=28774400 dt=14270
168
+ step 28200 loss=0.3088 pos=28876800 dt=14356
169
+ step 28300 loss=0.3164 pos=28979200 dt=14447
170
+ step 28400 loss=0.2921 pos=29081600 dt=14534
171
+ step 28500 loss=0.3280 pos=29184000 dt=14624
172
+ step 28600 loss=0.3377 pos=29286400 dt=14711
173
+ step 28700 loss=0.3448 pos=29388800 dt=14802
174
+ step 28800 loss=0.3477 pos=29491200 dt=14888
175
+ step 28900 loss=0.3815 pos=29593600 dt=14979
176
+ step 29000 loss=0.3362 pos=29696000 dt=15065
177
+ EVAL step 29000 pos=29696000 | T=2 loss=0.345 from=0.404 joint=0.715 hard=0.870 | T=16 loss=0.379 from=0.406 joint=0.687 hard=0.830
178
+ step 29100 loss=0.3422 pos=29798400 dt=15202
179
+ step 29200 loss=0.3052 pos=29900800 dt=15288
180
+ step 29300 loss=0.3248 pos=30003200 dt=15395
181
+ step 29400 loss=0.3125 pos=30105600 dt=15482
182
+ step 29500 loss=0.3004 pos=30208000 dt=15573
183
+ step 29600 loss=0.3027 pos=30310400 dt=15660
184
+ step 29700 loss=0.3377 pos=30412800 dt=15764
185
+ step 29800 loss=0.3233 pos=30515200 dt=15851
186
+ step 29900 loss=0.3285 pos=30617600 dt=15941
187
+ step 30000 loss=0.3201 pos=30720000 dt=16027
188
+ EVAL step 30000 pos=30720000 | T=2 loss=0.360 from=0.406 joint=0.697 hard=0.844 | T=16 loss=0.391 from=0.408 joint=0.664 hard=0.799
189
+ step 30100 loss=0.2964 pos=30822400 dt=16174
190
+ step 30200 loss=0.3494 pos=30924800 dt=16261
191
+ step 30300 loss=0.2793 pos=31027200 dt=16351
192
+ step 30400 loss=0.3397 pos=31129600 dt=16437
193
+ step 30500 loss=0.3026 pos=31232000 dt=16539
194
+ step 30600 loss=0.3200 pos=31334400 dt=16624
195
+ step 30700 loss=0.2781 pos=31436800 dt=16714
196
+ step 30800 loss=0.2641 pos=31539200 dt=16800
197
+ push fail (Request ID: Root=1-6aa049d4-1bae647337213e5b3abf55b7;9c6d8b3f-9059-4a18-a7ca-f9cb4f25ed44)
198
+ step 30900 loss=0.2850 pos=31641600 dt=16894
199
+ step 31000 loss=0.2507 pos=31744000 dt=16979
200
+ EVAL step 31000 pos=31744000 | T=2 loss=0.368 from=0.413 joint=0.638 hard=0.760 | T=16 loss=0.408 from=0.412 joint=0.618 hard=0.739
201
+ step 31100 loss=0.2401 pos=31846400 dt=17115
202
+ step 31200 loss=0.3586 pos=31948800 dt=17202
203
+ step 31300 loss=0.3544 pos=32051200 dt=17299
204
+ step 31400 loss=0.3469 pos=32153600 dt=17385
205
+ step 31500 loss=0.3603 pos=32256000 dt=17476
206
+ step 31600 loss=0.3410 pos=32358400 dt=17562
207
+ step 31700 loss=0.3666 pos=32460800 dt=17658
208
+ step 31800 loss=0.3659 pos=32563200 dt=17744
209
+ step 31900 loss=0.3352 pos=32665600 dt=17836
210
+ step 32000 loss=0.3445 pos=32768000 dt=17923
211
+ EVAL step 32000 pos=32768000 | T=2 loss=0.340 from=0.406 joint=0.713 hard=0.864 | T=16 loss=0.367 from=0.400 joint=0.687 hard=0.829
212
+ step 32100 loss=0.3668 pos=32870400 dt=18059
213
+ step 32200 loss=0.2991 pos=32972800 dt=18146
214
+ step 32300 loss=0.3549 pos=33075200 dt=18236
215
+ step 32400 loss=0.3751 pos=33177600 dt=18323
216
+ step 32500 loss=0.3534 pos=33280000 dt=18418
217
+ step 32600 loss=0.3316 pos=33382400 dt=18505
218
+ step 32700 loss=0.3128 pos=33484800 dt=18597
219
+ step 32800 loss=0.3387 pos=33587200 dt=18684
220
+ step 32900 loss=0.3010 pos=33689600 dt=18775
221
+ step 33000 loss=0.3514 pos=33792000 dt=18862
222
+ EVAL step 33000 pos=33792000 | T=2 loss=0.341 from=0.405 joint=0.702 hard=0.847 | T=16 loss=0.371 from=0.403 joint=0.675 hard=0.810
223
+ step 33100 loss=0.3320 pos=33894400 dt=19002
224
+ step 33200 loss=0.3306 pos=33996800 dt=19089
225
+ step 33300 loss=0.3276 pos=34099200 dt=19180
226
+ step 33400 loss=0.3528 pos=34201600 dt=19267
227
+ step 33500 loss=0.3218 pos=34304000 dt=19363
228
+ step 33600 loss=0.3857 pos=34406400 dt=19450
229
+ step 33700 loss=0.3427 pos=34508800 dt=19542
230
+ step 33800 loss=0.3141 pos=34611200 dt=19629
231
+ step 33900 loss=0.3533 pos=34713600 dt=19724
232
+ step 34000 loss=0.3024 pos=34816000 dt=19811
233
+ EVAL step 34000 pos=34816000 | T=2 loss=0.336 from=0.401 joint=0.721 hard=0.876 | T=16 loss=0.370 from=0.400 joint=0.690 hard=0.833
234
+ step 34100 loss=0.3798 pos=34918400 dt=19947
235
+ step 34200 loss=0.3070 pos=35020800 dt=20034
236
+ step 34300 loss=0.3406 pos=35123200 dt=20130
237
+ step 34400 loss=0.3216 pos=35225600 dt=20216
238
+ step 34500 loss=0.3356 pos=35328000 dt=20307
239
+ step 34600 loss=0.3394 pos=35430400 dt=20393
240
+ step 34700 loss=0.3304 pos=35532800 dt=20492
241
+ step 34800 loss=0.3722 pos=35635200 dt=20579
242
+ step 34900 loss=0.3383 pos=35737600 dt=20672
243
+ step 35000 loss=0.3601 pos=35840000 dt=20759
244
+ EVAL step 35000 pos=35840000 | T=2 loss=0.336 from=0.410 joint=0.724 hard=0.878 | T=16 loss=0.362 from=0.408 joint=0.689 hard=0.834
245
+ step 35100 loss=0.3334 pos=35942400 dt=20900
246
+ step 35200 loss=0.3149 pos=36044800 dt=20987
247
+ step 35300 loss=0.3337 pos=36147200 dt=21077
248
+ step 35400 loss=0.3059 pos=36249600 dt=21163
249
+ step 35500 loss=0.2739 pos=36352000 dt=21261
250
+ step 35600 loss=0.3205 pos=36454400 dt=21347
251
+ step 35700 loss=0.3207 pos=36556800 dt=21438
252
+ step 35800 loss=0.3259 pos=36659200 dt=21524
253
+ step 35900 loss=0.3198 pos=36761600 dt=21620
254
+ step 36000 loss=0.2759 pos=36864000 dt=21707
255
+ EVAL step 36000 pos=36864000 | T=2 loss=0.332 from=0.411 joint=0.729 hard=0.889 | T=16 loss=0.360 from=0.406 joint=0.688 hard=0.830
256
+ step 36100 loss=0.2813 pos=36966400 dt=21841
257
+ step 36200 loss=0.3128 pos=37068800 dt=21927
258
+ step 36300 loss=0.2828 pos=37171200 dt=22026
259
+ step 36400 loss=0.2925 pos=37273600 dt=22112
260
+ step 36500 loss=0.2766 pos=37376000 dt=22203
261
+ step 36600 loss=0.2691 pos=37478400 dt=22289
262
+ step 36700 loss=0.2739 pos=37580800 dt=22379
263
+ step 36800 loss=0.2577 pos=37683200 dt=22465
264
+ step 36900 loss=0.2628 pos=37785600 dt=22561
265
+ step 37000 loss=0.2480 pos=37888000 dt=22646
266
+ EVAL step 37000 pos=37888000 | T=2 loss=0.354 from=0.407 joint=0.677 hard=0.812 | T=16 loss=0.399 from=0.404 joint=0.657 hard=0.789
267
+ step 37100 loss=0.1794 pos=37990400 dt=22780
268
+ step 37200 loss=0.3664 pos=38092800 dt=22866
269
+ step 37300 loss=0.3338 pos=38195200 dt=22961
270
+ step 37400 loss=0.3160 pos=38297600 dt=23047
271
+ step 37500 loss=0.3506 pos=38400000 dt=23139
272
+ step 37600 loss=0.3177 pos=38502400 dt=23226
273
+ step 37700 loss=0.2994 pos=38604800 dt=23323
274
+ step 37800 loss=0.3246 pos=38707200 dt=23410
275
+ step 37900 loss=0.2931 pos=38809600 dt=23502
276
+ step 38000 loss=0.3239 pos=38912000 dt=23588
277
+ EVAL step 38000 pos=38912000 | T=2 loss=0.326 from=0.403 joint=0.718 hard=0.871 | T=16 loss=0.351 from=0.404 joint=0.677 hard=0.816
278
+ step 38100 loss=0.2956 pos=39014400 dt=23730
279
+ step 38200 loss=0.3317 pos=39116800 dt=23817
280
+ step 38300 loss=0.3089 pos=39219200 dt=23908
281
+ step 38400 loss=0.3225 pos=39321600 dt=23994
282
+ step 38500 loss=0.3523 pos=39424000 dt=24091
283
+ step 38600 loss=0.3847 pos=39526400 dt=24178
284
+ step 38700 loss=0.2999 pos=39628800 dt=24269
285
+ step 38800 loss=0.3037 pos=39731200 dt=24356
286
+ step 38900 loss=0.3408 pos=39833600 dt=24451
287
+ step 39000 loss=0.3157 pos=39936000 dt=24538
288
+ EVAL step 39000 pos=39936000 | T=2 loss=0.325 from=0.406 joint=0.725 hard=0.880 | T=16 loss=0.342 from=0.407 joint=0.695 hard=0.840
289
+ step 39100 loss=0.2885 pos=40038400 dt=24674
290
+ step 39200 loss=0.3087 pos=40140800 dt=24761
291
+ step 39300 loss=0.3380 pos=40243200 dt=24858
292
+ step 39400 loss=0.3283 pos=40345600 dt=24945
293
+ step 39500 loss=0.3249 pos=40448000 dt=25036
294
+ step 39600 loss=0.3509 pos=40550400 dt=25122
295
+ step 39700 loss=0.3321 pos=40652800 dt=25213
296
+ step 39800 loss=0.3227 pos=40755200 dt=25299
297
+ step 39900 loss=0.3171 pos=40857600 dt=25396
298
+ step 40000 loss=0.3122 pos=40960000 dt=25483
299
+ EVAL step 40000 pos=40960000 | T=2 loss=0.321 from=0.403 joint=0.729 hard=0.888 | T=16 loss=0.347 from=0.405 joint=0.698 hard=0.845
300
+ step 40100 loss=0.3078 pos=41062400 dt=25622
301
+ step 40200 loss=0.2911 pos=41164800 dt=25708
302
+ step 40300 loss=0.3292 pos=41267200 dt=25806
303
+ step 40400 loss=0.2956 pos=41369600 dt=25892
304
+ step 40500 loss=0.3095 pos=41472000 dt=25990
305
+ step 40600 loss=0.3223 pos=41574400 dt=26076
306
+ step 40700 loss=0.3161 pos=41676800 dt=26177
307
+ step 40800 loss=0.3119 pos=41779200 dt=26264
308
+ step 40900 loss=0.3100 pos=41881600 dt=26356
309
+ step 41000 loss=0.2907 pos=41984000 dt=26442
310
+ EVAL step 41000 pos=41984000 | T=2 loss=0.317 from=0.409 joint=0.718 hard=0.873 | T=16 loss=0.342 from=0.408 joint=0.680 hard=0.818
311
+ step 41100 loss=0.2849 pos=42086400 dt=26582
312
+ step 41200 loss=0.2796 pos=42188800 dt=26669
313
+ step 41300 loss=0.3349 pos=42291200 dt=26759
314
+ step 41400 loss=0.2900 pos=42393600 dt=26846
315
+ push fail Client error '412 Precondition Failed' for url 'https://huggingface.co/api/models/Asilarkness/rdbt-chess/commit/main' (R
316
+ step 41500 loss=0.3071 pos=42496000 dt=26941
317
+ step 41600 loss=0.2584 pos=42598400 dt=27027
318
+ step 41700 loss=0.2939 pos=42700800 dt=27119
319
+ step 41800 loss=0.2868 pos=42803200 dt=27205
320
+ step 41900 loss=0.3136 pos=42905600 dt=27299
321
+ step 42000 loss=0.3254 pos=43008000 dt=27385
322
+ EVAL step 42000 pos=43008000 | T=2 loss=0.318 from=0.403 joint=0.729 hard=0.888 | T=16 loss=0.350 from=0.402 joint=0.694 hard=0.840
323
+ step 42100 loss=0.2919 pos=43110400 dt=27525
324
+ step 42200 loss=0.3016 pos=43212800 dt=27611
325
+ step 42300 loss=0.3004 pos=43315200 dt=27703
326
+ step 42400 loss=0.3011 pos=43417600 dt=27789
327
+ step 42500 loss=0.2804 pos=43520000 dt=27885
328
+ step 42600 loss=0.2516 pos=43622400 dt=27971
329
+ step 42700 loss=0.2802 pos=43724800 dt=28062
330
+ step 42800 loss=0.2355 pos=43827200 dt=28147
331
+ step 42900 loss=0.2728 pos=43929600 dt=28242
332
+ step 43000 loss=0.2251 pos=44032000 dt=28327
333
+ EVAL step 43000 pos=44032000 | T=2 loss=0.323 from=0.412 joint=0.680 hard=0.818 | T=16 loss=0.356 from=0.408 joint=0.661 hard=0.794
334
+ step 43100 loss=0.2265 pos=44134400 dt=28466
335
+ step 43200 loss=0.1697 pos=44236800 dt=28552
336
+ step 43300 loss=0.3377 pos=44339200 dt=28646
337
+ step 43400 loss=0.3269 pos=44441600 dt=28733
338
+ step 43500 loss=0.3117 pos=44544000 dt=28827
339
+ step 43600 loss=0.2831 pos=44646400 dt=28913
340
+ step 43700 loss=0.3152 pos=44748800 dt=29011
341
+ step 43800 loss=0.3072 pos=44851200 dt=29097
342
+ step 43900 loss=0.3322 pos=44953600 dt=29191
343
+ step 44000 loss=0.3073 pos=45056000 dt=29277
344
+ EVAL step 44000 pos=45056000 | T=2 loss=0.314 from=0.408 joint=0.725 hard=0.883 | T=16 loss=0.349 from=0.406 joint=0.690 hard=0.837
345
+ step 44100 loss=0.3063 pos=45158400 dt=29423
346
+ step 44200 loss=0.2686 pos=45260800 dt=29510
347
+ step 44300 loss=0.2881 pos=45363200 dt=29601
348
+ step 44400 loss=0.3028 pos=45465600 dt=29687
349
+ step 44500 loss=0.3521 pos=45568000 dt=29779
350
+ step 44600 loss=0.3068 pos=45670400 dt=29865
351
+ step 44700 loss=0.3363 pos=45772800 dt=29962
352
+ step 44800 loss=0.3412 pos=45875200 dt=30048
353
+ step 44900 loss=0.2783 pos=45977600 dt=30139
354
+ step 45000 loss=0.3092 pos=46080000 dt=30225
355
+ EVAL step 45000 pos=46080000 | T=2 loss=0.311 from=0.408 joint=0.726 hard=0.883 | T=16 loss=0.340 from=0.406 joint=0.698 hard=0.840
356
+ step 45100 loss=0.2641 pos=46182400 dt=30367
357
+ step 45200 loss=0.2897 pos=46284800 dt=30453
358
+ step 45300 loss=0.3084 pos=46387200 dt=30544
359
+ step 45400 loss=0.3160 pos=46489600 dt=30630
360
+ step 45500 loss=0.3478 pos=46592000 dt=30724
361
+ step 45600 loss=0.3141 pos=46694400 dt=30811
362
+ step 45700 loss=0.3198 pos=46796800 dt=30901
363
+ step 45800 loss=0.3049 pos=46899200 dt=30988
364
+ step 45900 loss=0.2778 pos=47001600 dt=31086
365
+ step 46000 loss=0.3369 pos=47104000 dt=31172
366
+ EVAL step 46000 pos=47104000 | T=2 loss=0.314 from=0.404 joint=0.727 hard=0.883 | T=16 loss=0.343 from=0.407 joint=0.696 hard=0.842
367
+ step 46100 loss=0.2910 pos=47206400 dt=31308
368
+ step 46200 loss=0.3340 pos=47308800 dt=31395
369
+ step 46300 loss=0.3069 pos=47411200 dt=31490
370
+ step 46400 loss=0.2937 pos=47513600 dt=31577
371
+ step 46500 loss=0.2925 pos=47616000 dt=31667
372
+ step 46600 loss=0.3249 pos=47718400 dt=31754
373
+ step 46700 loss=0.2829 pos=47820800 dt=31848
374
+ step 46800 loss=0.3298 pos=47923200 dt=31934
375
+ step 46900 loss=0.2987 pos=48025600 dt=32025
376
+ step 47000 loss=0.3265 pos=48128000 dt=32111
377
+ EVAL step 47000 pos=48128000 | T=2 loss=0.311 from=0.410 joint=0.734 hard=0.895 | T=16 loss=0.337 from=0.410 joint=0.694 hard=0.839
378
+ step 47100 loss=0.2936 pos=48230400 dt=32249
379
+ step 47200 loss=0.2856 pos=48332800 dt=32336
380
+ step 47300 loss=0.2912 pos=48435200 dt=32426
381
+ step 47400 loss=0.3078 pos=48537600 dt=32512
382
+ step 47500 loss=0.2639 pos=48640000 dt=32608
383
+ step 47600 loss=0.2946 pos=48742400 dt=32694
384
+ step 47700 loss=0.2825 pos=48844800 dt=32785
385
+ step 47800 loss=0.2783 pos=48947200 dt=32872
386
+ step 47900 loss=0.2913 pos=49049600 dt=32968
387
+ step 48000 loss=0.2583 pos=49152000 dt=33054
388
+ EVAL step 48000 pos=49152000 | T=2 loss=0.310 from=0.410 joint=0.728 hard=0.888 | T=16 loss=0.340 from=0.410 joint=0.692 hard=0.834
389
+ step 48100 loss=0.2526 pos=49254400 dt=33191
390
+ step 48200 loss=0.2785 pos=49356800 dt=33277
391
+ step 48300 loss=0.2863 pos=49459200 dt=33372
392
+ step 48400 loss=0.2716 pos=49561600 dt=33458
393
+ step 48500 loss=0.2790 pos=49664000 dt=33548
394
+ step 48600 loss=0.2825 pos=49766400 dt=33634
395
+ step 48700 loss=0.2418 pos=49868800 dt=33729
396
+ step 48800 loss=0.2513 pos=49971200 dt=33815
397
+ step 48900 loss=0.2267 pos=50073600 dt=33905
398
+ step 49000 loss=0.2331 pos=50176000 dt=33990
399
+ EVAL step 49000 pos=50176000 | T=2 loss=0.312 from=0.410 joint=0.713 hard=0.864 | T=16 loss=0.339 from=0.409 joint=0.681 hard=0.817
400
+ step 49100 loss=0.1878 pos=50278400 dt=34136
401
+ step 49200 loss=0.2768 pos=50380800 dt=34221
402
+ step 49300 loss=0.2977 pos=50483200 dt=34313
403
+ step 49400 loss=0.3423 pos=50585600 dt=34399
404
+ step 49500 loss=0.3297 pos=50688000 dt=34493
405
+ step 49600 loss=0.2741 pos=50790400 dt=34579
406
+ step 49700 loss=0.3071 pos=50892800 dt=34672
407
+ step 49800 loss=0.3064 pos=50995200 dt=34758
408
+ step 49900 loss=0.2602 pos=51097600 dt=34854
409
+ step 50000 loss=0.3163 pos=51200000 dt=34941
410
+ EVAL step 50000 pos=51200000 | T=2 loss=0.302 from=0.414 joint=0.734 hard=0.894 | T=16 loss=0.333 from=0.413 joint=0.698 hard=0.842
411
+ step 50100 loss=0.2965 pos=51302400 dt=35082
412
+ step 50200 loss=0.3184 pos=51404800 dt=35168
413
+ step 50300 loss=0.3000 pos=51507200 dt=35263
414
+ step 50400 loss=0.2834 pos=51609600 dt=35349
415
+ step 50500 loss=0.2885 pos=51712000 dt=35442
416
+ step 50600 loss=0.3278 pos=51814400 dt=35529
417
+ step 50700 loss=0.2870 pos=51916800 dt=35623
418
+ step 50800 loss=0.3141 pos=52019200 dt=35709
419
+ step 50900 loss=0.3158 pos=52121600 dt=35800
420
+ step 51000 loss=0.3219 pos=52224000 dt=35886
421
+ EVAL step 51000 pos=52224000 | T=2 loss=0.302 from=0.412 joint=0.734 hard=0.894 | T=16 loss=0.329 from=0.408 joint=0.696 hard=0.841
422
+ step 51100 loss=0.3232 pos=52326400 dt=36025
423
+ step 51200 loss=0.3193 pos=52428800 dt=36111
424
+ step 51300 loss=0.2773 pos=52531200 dt=36205
425
+ step 51400 loss=0.3379 pos=52633600 dt=36291
426
+ step 51500 loss=0.3107 pos=52736000 dt=36384
427
+ step 51600 loss=0.3337 pos=52838400 dt=36471
428
+ step 51700 loss=0.3173 pos=52940800 dt=36564
429
+ step 51800 loss=0.3289 pos=53043200 dt=36650
430
+ step 51900 loss=0.3023 pos=53145600 dt=36745
431
+ step 52000 loss=0.2666 pos=53248000 dt=36831
432
+ EVAL step 52000 pos=53248000 | T=2 loss=0.306 from=0.412 joint=0.728 hard=0.887 | T=16 loss=0.330 from=0.416 joint=0.693 hard=0.842
433
+ step 52100 loss=0.3092 pos=53350400 dt=36969
434
+ step 52200 loss=0.3052 pos=53452800 dt=37055
435
+ step 52300 loss=0.3289 pos=53555200 dt=37147
436
+ step 52400 loss=0.2715 pos=53657600 dt=37233
437
+ step 52500 loss=0.2813 pos=53760000 dt=37324
438
+ step 52600 loss=0.3015 pos=53862400 dt=37411
439
+ step 52700 loss=0.2746 pos=53964800 dt=37508
440
+ step 52800 loss=0.3042 pos=54067200 dt=37594
441
+ step 52900 loss=0.2999 pos=54169600 dt=37688
442
+ step 53000 loss=0.2719 pos=54272000 dt=37774
443
+ EVAL step 53000 pos=54272000 | T=2 loss=0.308 from=0.412 joint=0.737 hard=0.900 | T=16 loss=0.328 from=0.407 joint=0.705 hard=0.856
444
+ step 53100 loss=0.2866 pos=54374400 dt=37909
445
+ step 53200 loss=0.3316 pos=54476800 dt=37995
446
+ step 53300 loss=0.3192 pos=54579200 dt=38087
447
+ step 53400 loss=0.3025 pos=54681600 dt=38173
448
+ step 53500 loss=0.2843 pos=54784000 dt=38264
449
+ step 53600 loss=0.2782 pos=54886400 dt=38351
450
+ step 53700 loss=0.3111 pos=54988800 dt=38443
451
+ step 53800 loss=0.2593 pos=55091200 dt=38529
452
+ step 53900 loss=0.2716 pos=55193600 dt=38620
453
+ step 54000 loss=0.2899 pos=55296000 dt=38706
454
+ EVAL step 54000 pos=55296000 | T=2 loss=0.312 from=0.411 joint=0.730 hard=0.886 | T=16 loss=0.335 from=0.411 joint=0.696 hard=0.842
log_AMe.txt ADDED
@@ -0,0 +1,114 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ cfg AMe d=1280 H=10 tied=1 nl=8 lr=8 T=2 shards=500..625 hours=168.0
2
+ step 50900 loss=0.4318 pos=153907200 dt=96
3
+ step 51000 loss=0.3471 pos=154009600 dt=185
4
+ EVAL step 51000 pos=154009600 | T=2 loss=0.403 from=0.411 joint=0.684 hard=0.826 | T=16 loss=0.426 from=0.403 joint=0.664 hard=0.800
5
+ step 51100 loss=0.4282 pos=154112000 dt=339
6
+ step 51200 loss=0.4348 pos=154214400 dt=428
7
+ step 51300 loss=0.2957 pos=154316800 dt=529
8
+ step 51400 loss=0.4214 pos=154419200 dt=618
9
+ step 51500 loss=0.3737 pos=154521600 dt=713
10
+ step 51600 loss=0.4121 pos=154624000 dt=802
11
+ step 51700 loss=0.3966 pos=154726400 dt=897
12
+ step 51800 loss=0.3873 pos=154828800 dt=986
13
+ step 51900 loss=0.3556 pos=154931200 dt=1089
14
+ step 52000 loss=0.3665 pos=155033600 dt=1178
15
+ EVAL step 52000 pos=155033600 | T=2 loss=0.360 from=0.406 joint=0.699 hard=0.844 | T=16 loss=0.393 from=0.400 joint=0.666 hard=0.798
16
+ step 52100 loss=0.3446 pos=155136000 dt=1319
17
+ step 52200 loss=0.3437 pos=155238400 dt=1409
18
+ step 52300 loss=0.3867 pos=155340800 dt=1508
19
+ step 52400 loss=0.3051 pos=155443200 dt=1597
20
+ step 52500 loss=0.3331 pos=155545600 dt=1694
21
+ step 52600 loss=0.3568 pos=155648000 dt=1783
22
+ step 52700 loss=0.3343 pos=155750400 dt=1884
23
+ step 52800 loss=0.3297 pos=155852800 dt=1973
24
+ step 52900 loss=0.3900 pos=155955200 dt=2067
25
+ step 53000 loss=0.3662 pos=156057600 dt=2156
26
+ EVAL step 53000 pos=156057600 | T=2 loss=0.350 from=0.398 joint=0.700 hard=0.849 | T=16 loss=0.374 from=0.395 joint=0.671 hard=0.811
27
+ step 53100 loss=0.3171 pos=156160000 dt=2301
28
+ step 53200 loss=0.3551 pos=156262400 dt=2390
29
+ step 53300 loss=0.3552 pos=156364800 dt=2483
30
+ step 53400 loss=0.3530 pos=156467200 dt=2572
31
+ step 53500 loss=0.3492 pos=156569600 dt=2675
32
+ step 53600 loss=0.3476 pos=156672000 dt=2764
33
+ step 53700 loss=0.3373 pos=156774400 dt=2858
34
+ step 53800 loss=0.3499 pos=156876800 dt=2947
35
+ step 53900 loss=0.3002 pos=156979200 dt=3064
36
+ step 54000 loss=0.3212 pos=157081600 dt=3153
37
+ EVAL step 54000 pos=157081600 | T=2 loss=0.341 from=0.406 joint=0.699 hard=0.842 | T=16 loss=0.374 from=0.400 joint=0.669 hard=0.801
38
+ step 54100 loss=0.3174 pos=157184000 dt=3293
39
+ step 54200 loss=0.3705 pos=157286400 dt=3382
40
+ step 54300 loss=0.3583 pos=157388800 dt=3487
41
+ step 54400 loss=0.2905 pos=157491200 dt=3576
42
+ step 54500 loss=0.3486 pos=157593600 dt=3671
43
+ step 54600 loss=0.3805 pos=157696000 dt=3760
44
+ step 54700 loss=0.3155 pos=157798400 dt=3860
45
+ step 54800 loss=0.3181 pos=157900800 dt=3949
46
+ step 54900 loss=0.3745 pos=158003200 dt=4043
47
+ step 55000 loss=0.3271 pos=158105600 dt=4132
48
+ EVAL step 55000 pos=158105600 | T=2 loss=0.340 from=0.402 joint=0.707 hard=0.859 | T=16 loss=0.379 from=0.396 joint=0.666 hard=0.799
49
+ step 55100 loss=0.3242 pos=158208000 dt=4271
50
+ step 55200 loss=0.2840 pos=158310400 dt=4360
51
+ step 55300 loss=0.3068 pos=158412800 dt=4458
52
+ step 55400 loss=0.3233 pos=158515200 dt=4547
53
+ step 55500 loss=0.2911 pos=158617600 dt=4641
54
+ step 55600 loss=0.3217 pos=158720000 dt=4730
55
+ step 55700 loss=0.3481 pos=158822400 dt=4830
56
+ step 55800 loss=0.3119 pos=158924800 dt=4919
57
+ step 55900 loss=0.2761 pos=159027200 dt=5014
58
+ step 56000 loss=0.3078 pos=159129600 dt=5102
59
+ EVAL step 56000 pos=159129600 | T=2 loss=0.333 from=0.408 joint=0.724 hard=0.883 | T=16 loss=0.368 from=0.404 joint=0.689 hard=0.834
60
+ step 56100 loss=0.2633 pos=159232000 dt=5251
61
+ step 56200 loss=0.2946 pos=159334400 dt=5340
62
+ step 56300 loss=0.2884 pos=159436800 dt=5433
63
+ step 56400 loss=0.2713 pos=159539200 dt=5522
64
+ step 56500 loss=0.2877 pos=159641600 dt=5623
65
+ step 56600 loss=0.2625 pos=159744000 dt=5712
66
+ step 56700 loss=0.2509 pos=159846400 dt=5806
67
+ step 56800 loss=0.2871 pos=159948800 dt=5895
68
+ step 56900 loss=0.3035 pos=160051200 dt=5989
69
+ step 57000 loss=0.3334 pos=160153600 dt=6078
70
+ EVAL step 57000 pos=160153600 | T=2 loss=0.331 from=0.406 joint=0.706 hard=0.855 | T=16 loss=0.360 from=0.405 joint=0.654 hard=0.783
71
+ step 57100 loss=0.3239 pos=160256000 dt=6217
72
+ step 57200 loss=0.3225 pos=160358400 dt=6306
73
+ step 57300 loss=0.3635 pos=160460800 dt=6400
74
+ step 57400 loss=0.3444 pos=160563200 dt=6489
75
+ step 57500 loss=0.3467 pos=160665600 dt=6583
76
+ step 57600 loss=0.3465 pos=160768000 dt=6672
77
+ step 57700 loss=0.3020 pos=160870400 dt=6770
78
+ step 57800 loss=0.3349 pos=160972800 dt=6859
79
+ step 57900 loss=0.3118 pos=161075200 dt=6955
80
+ step 58000 loss=0.3337 pos=161177600 dt=7044
81
+ EVAL step 58000 pos=161177600 | T=2 loss=0.325 from=0.402 joint=0.715 hard=0.869 | T=16 loss=0.357 from=0.400 joint=0.666 hard=0.799
82
+ step 58100 loss=0.3066 pos=161280000 dt=7190
83
+ step 58200 loss=0.3014 pos=161382400 dt=7279
84
+ step 58300 loss=0.3602 pos=161484800 dt=7373
85
+ step 58400 loss=0.3113 pos=161587200 dt=7462
86
+ step 58500 loss=0.3189 pos=161689600 dt=7563
87
+ step 58600 loss=0.3803 pos=161792000 dt=7652
88
+ step 58700 loss=0.3052 pos=161894400 dt=7746
89
+ step 58800 loss=0.3131 pos=161996800 dt=7835
90
+ step 58900 loss=0.3119 pos=162099200 dt=7936
91
+ step 59000 loss=0.3291 pos=162201600 dt=8025
92
+ EVAL step 59000 pos=162201600 | T=2 loss=0.323 from=0.404 joint=0.717 hard=0.874 | T=16 loss=0.356 from=0.407 joint=0.663 hard=0.799
93
+ step 59100 loss=0.3078 pos=162304000 dt=8165
94
+ step 59200 loss=0.3205 pos=162406400 dt=8255
95
+ step 59300 loss=0.3282 pos=162508800 dt=8355
96
+ step 59400 loss=0.3378 pos=162611200 dt=8444
97
+ step 59500 loss=0.3341 pos=162713600 dt=8539
98
+ step 59600 loss=0.3068 pos=162816000 dt=8628
99
+ step 59700 loss=0.3132 pos=162918400 dt=8727
100
+ step 59800 loss=0.3302 pos=163020800 dt=8816
101
+ step 59900 loss=0.3377 pos=163123200 dt=8910
102
+ step 60000 loss=0.3658 pos=163225600 dt=8999
103
+ EVAL step 60000 pos=163225600 | T=2 loss=0.323 from=0.403 joint=0.706 hard=0.854 | T=16 loss=0.365 from=0.404 joint=0.655 hard=0.783
104
+ step 60100 loss=0.2784 pos=163328000 dt=9141
105
+ step 60200 loss=0.3256 pos=163430400 dt=9230
106
+ step 60300 loss=0.3117 pos=163532800 dt=9333
107
+ step 60400 loss=0.3290 pos=163635200 dt=9422
108
+ step 60500 loss=0.2947 pos=163737600 dt=9516
109
+ step 60600 loss=0.3316 pos=163840000 dt=9605
110
+ step 60700 loss=0.3222 pos=163942400 dt=9704
111
+ step 60800 loss=0.3434 pos=164044800 dt=9793
112
+ step 60900 loss=0.2965 pos=164147200 dt=9888
113
+ step 61000 loss=0.3143 pos=164249600 dt=9977
114
+ EVAL step 61000 pos=164249600 | T=2 loss=0.314 from=0.407 joint=0.727 hard=0.884 | T=16 loss=0.342 from=0.398 joint=0.670 hard=0.805
log_AMf.txt ADDED
@@ -0,0 +1,455 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ cfg AMf d=1280 H=10 tied=1 nl=8 lr=8 T=2 shards=625..750 hours=168.0
2
+ step 50900 loss=0.3677 pos=153907200 dt=95
3
+ step 51000 loss=0.3665 pos=154009600 dt=185
4
+ EVAL step 51000 pos=154009600 | T=2 loss=0.416 from=0.412 joint=0.664 hard=0.795 | T=16 loss=0.442 from=0.406 joint=0.620 hard=0.737
5
+ step 51100 loss=0.3882 pos=154112000 dt=327
6
+ step 51200 loss=0.3691 pos=154214400 dt=416
7
+ step 51300 loss=0.3422 pos=154316800 dt=510
8
+ step 51400 loss=0.4141 pos=154419200 dt=600
9
+ step 51500 loss=0.3440 pos=154521600 dt=701
10
+ step 51600 loss=0.3763 pos=154624000 dt=790
11
+ step 51700 loss=0.3556 pos=154726400 dt=885
12
+ step 51800 loss=0.4071 pos=154828800 dt=974
13
+ step 51900 loss=0.3256 pos=154931200 dt=1073
14
+ step 52000 loss=0.3288 pos=155033600 dt=1162
15
+ EVAL step 52000 pos=155033600 | T=2 loss=0.366 from=0.409 joint=0.700 hard=0.848 | T=16 loss=0.395 from=0.404 joint=0.648 hard=0.776
16
+ step 52100 loss=0.3370 pos=155136000 dt=1302
17
+ step 52200 loss=0.4058 pos=155238400 dt=1392
18
+ step 52300 loss=0.4160 pos=155340800 dt=1491
19
+ step 52400 loss=0.3640 pos=155443200 dt=1580
20
+ step 52500 loss=0.3697 pos=155545600 dt=1675
21
+ step 52600 loss=0.4113 pos=155648000 dt=1764
22
+ step 52700 loss=0.3384 pos=155750400 dt=1865
23
+ step 52800 loss=0.3358 pos=155852800 dt=1954
24
+ step 52900 loss=0.3459 pos=155955200 dt=2048
25
+ step 53000 loss=0.3294 pos=156057600 dt=2137
26
+ EVAL step 53000 pos=156057600 | T=2 loss=0.350 from=0.398 joint=0.698 hard=0.844 | T=16 loss=0.374 from=0.395 joint=0.672 hard=0.812
27
+ step 53100 loss=0.3822 pos=156160000 dt=2282
28
+ step 53200 loss=0.3149 pos=156262400 dt=2371
29
+ step 53300 loss=0.3484 pos=156364800 dt=2464
30
+ step 53400 loss=0.3441 pos=156467200 dt=2553
31
+ step 53500 loss=0.3055 pos=156569600 dt=2654
32
+ step 53600 loss=0.3404 pos=156672000 dt=2744
33
+ step 53700 loss=0.3319 pos=156774400 dt=2839
34
+ step 53800 loss=0.3074 pos=156876800 dt=2928
35
+ step 53900 loss=0.3306 pos=156979200 dt=3036
36
+ step 54000 loss=0.3043 pos=157081600 dt=3125
37
+ EVAL step 54000 pos=157081600 | T=2 loss=0.353 from=0.408 joint=0.709 hard=0.862 | T=16 loss=0.375 from=0.401 joint=0.662 hard=0.800
38
+ step 54100 loss=0.3485 pos=157184000 dt=3263
39
+ step 54200 loss=0.3473 pos=157286400 dt=3352
40
+ step 54300 loss=0.3554 pos=157388800 dt=3453
41
+ step 54400 loss=0.3306 pos=157491200 dt=3542
42
+ step 54500 loss=0.3141 pos=157593600 dt=3637
43
+ step 54600 loss=0.3031 pos=157696000 dt=3726
44
+ step 54700 loss=0.3762 pos=157798400 dt=3820
45
+ step 54800 loss=0.3539 pos=157900800 dt=3909
46
+ step 54900 loss=0.3219 pos=158003200 dt=4011
47
+ step 55000 loss=0.3233 pos=158105600 dt=4101
48
+ EVAL step 55000 pos=158105600 | T=2 loss=0.342 from=0.411 joint=0.689 hard=0.826 | T=16 loss=0.372 from=0.405 joint=0.642 hard=0.770
49
+ step 55100 loss=0.3326 pos=158208000 dt=4242
50
+ step 55200 loss=0.3402 pos=158310400 dt=4331
51
+ step 55300 loss=0.3417 pos=158412800 dt=4430
52
+ step 55400 loss=0.3412 pos=158515200 dt=4519
53
+ step 55500 loss=0.3258 pos=158617600 dt=4614
54
+ step 55600 loss=0.3238 pos=158720000 dt=4703
55
+ step 55700 loss=0.3261 pos=158822400 dt=4809
56
+ step 55800 loss=0.2807 pos=158924800 dt=4898
57
+ step 55900 loss=0.2980 pos=159027200 dt=4991
58
+ step 56000 loss=0.3137 pos=159129600 dt=5080
59
+ EVAL step 56000 pos=159129600 | T=2 loss=0.333 from=0.408 joint=0.721 hard=0.879 | T=16 loss=0.368 from=0.405 joint=0.689 hard=0.833
60
+ step 56100 loss=0.3287 pos=159232000 dt=5227
61
+ step 56200 loss=0.2485 pos=159334400 dt=5316
62
+ step 56300 loss=0.2626 pos=159436800 dt=5409
63
+ step 56400 loss=0.3160 pos=159539200 dt=5497
64
+ step 56500 loss=0.2488 pos=159641600 dt=5590
65
+ step 56600 loss=0.3053 pos=159744000 dt=5679
66
+ step 56700 loss=0.3446 pos=159846400 dt=5778
67
+ step 56800 loss=0.2219 pos=159948800 dt=5866
68
+ step 56900 loss=0.3715 pos=160051200 dt=5959
69
+ step 57000 loss=0.3263 pos=160153600 dt=6048
70
+ EVAL step 57000 pos=160153600 | T=2 loss=0.331 from=0.407 joint=0.709 hard=0.859 | T=16 loss=0.366 from=0.406 joint=0.658 hard=0.791
71
+ step 57100 loss=0.3393 pos=160256000 dt=6188
72
+ step 57200 loss=0.3355 pos=160358400 dt=6277
73
+ step 57300 loss=0.3120 pos=160460800 dt=6370
74
+ step 57400 loss=0.3510 pos=160563200 dt=6459
75
+ step 57500 loss=0.3396 pos=160665600 dt=6554
76
+ step 57600 loss=0.3149 pos=160768000 dt=6643
77
+ step 57700 loss=0.3291 pos=160870400 dt=6742
78
+ step 57800 loss=0.2995 pos=160972800 dt=6831
79
+ step 57900 loss=0.3380 pos=161075200 dt=6924
80
+ step 58000 loss=0.2870 pos=161177600 dt=7013
81
+ EVAL step 58000 pos=161177600 | T=2 loss=0.325 from=0.403 joint=0.717 hard=0.871 | T=16 loss=0.357 from=0.401 joint=0.668 hard=0.801
82
+ step 58100 loss=0.3026 pos=161280000 dt=7159
83
+ step 58200 loss=0.3649 pos=161382400 dt=7248
84
+ step 58300 loss=0.3067 pos=161484800 dt=7343
85
+ step 58400 loss=0.3340 pos=161587200 dt=7432
86
+ step 58500 loss=0.3290 pos=161689600 dt=7532
87
+ step 58600 loss=0.3431 pos=161792000 dt=7621
88
+ step 58700 loss=0.3453 pos=161894400 dt=7715
89
+ step 58800 loss=0.3159 pos=161996800 dt=7804
90
+ step 58900 loss=0.3453 pos=162099200 dt=7907
91
+ step 59000 loss=0.3077 pos=162201600 dt=7996
92
+ EVAL step 59000 pos=162201600 | T=2 loss=0.331 from=0.396 joint=0.698 hard=0.844 | T=16 loss=0.360 from=0.392 joint=0.645 hard=0.765
93
+ step 59100 loss=0.3378 pos=162304000 dt=8136
94
+ step 59200 loss=0.3038 pos=162406400 dt=8225
95
+ step 59300 loss=0.3258 pos=162508800 dt=8327
96
+ step 59400 loss=0.3292 pos=162611200 dt=8416
97
+ step 59500 loss=0.3394 pos=162713600 dt=8511
98
+ step 59600 loss=0.3569 pos=162816000 dt=8600
99
+ step 59700 loss=0.2958 pos=162918400 dt=8694
100
+ step 59800 loss=0.3176 pos=163020800 dt=8783
101
+ step 59900 loss=0.3286 pos=163123200 dt=8887
102
+ step 60000 loss=0.3302 pos=163225600 dt=8976
103
+ EVAL step 60000 pos=163225600 | T=2 loss=0.321 from=0.405 joint=0.723 hard=0.878 | T=16 loss=0.358 from=0.402 joint=0.658 hard=0.790
104
+ step 60100 loss=0.2812 pos=163328000 dt=9115
105
+ step 60200 loss=0.2981 pos=163430400 dt=9205
106
+ step 60300 loss=0.3221 pos=163532800 dt=9307
107
+ step 60400 loss=0.3013 pos=163635200 dt=9396
108
+ step 60500 loss=0.2853 pos=163737600 dt=9491
109
+ step 60600 loss=0.3298 pos=163840000 dt=9580
110
+ step 60700 loss=0.3238 pos=163942400 dt=9683
111
+ step 60800 loss=0.3158 pos=164044800 dt=9772
112
+ step 60900 loss=0.2841 pos=164147200 dt=9869
113
+ step 61000 loss=0.3151 pos=164249600 dt=9958
114
+ EVAL step 61000 pos=164249600 | T=2 loss=0.324 from=0.413 joint=0.720 hard=0.873 | T=16 loss=0.353 from=0.409 joint=0.672 hard=0.807
115
+ step 61100 loss=0.3306 pos=164352000 dt=10098
116
+ step 61200 loss=0.3047 pos=164454400 dt=10187
117
+ step 61300 loss=0.2845 pos=164556800 dt=10292
118
+ step 61400 loss=0.3151 pos=164659200 dt=10381
119
+ step 61500 loss=0.2754 pos=164761600 dt=10477
120
+ step 61600 loss=0.3234 pos=164864000 dt=10566
121
+ step 61700 loss=0.3088 pos=164966400 dt=10667
122
+ step 61800 loss=0.2812 pos=165068800 dt=10756
123
+ step 61900 loss=0.2528 pos=165171200 dt=10851
124
+ step 62000 loss=0.2779 pos=165273600 dt=10940
125
+ EVAL step 62000 pos=165273600 | T=2 loss=0.313 from=0.407 joint=0.727 hard=0.886 | T=16 loss=0.347 from=0.403 joint=0.686 hard=0.829
126
+ step 62100 loss=0.2833 pos=165376000 dt=11081
127
+ step 62200 loss=0.2875 pos=165478400 dt=11170
128
+ step 62300 loss=0.2581 pos=165580800 dt=11263
129
+ step 62400 loss=0.2422 pos=165683200 dt=11352
130
+ step 62500 loss=0.2354 pos=165785600 dt=11445
131
+ step 62600 loss=0.2349 pos=165888000 dt=11533
132
+ step 62700 loss=0.2351 pos=165990400 dt=11631
133
+ step 62800 loss=0.2104 pos=166092800 dt=11719
134
+ step 62900 loss=0.3503 pos=166195200 dt=11811
135
+ step 63000 loss=0.2939 pos=166297600 dt=11900
136
+ EVAL step 63000 pos=166297600 | T=2 loss=0.312 from=0.413 joint=0.711 hard=0.859 | T=16 loss=0.344 from=0.409 joint=0.671 hard=0.807
137
+ step 63100 loss=0.3266 pos=166400000 dt=12040
138
+ step 63200 loss=0.3114 pos=166502400 dt=12129
139
+ step 63300 loss=0.2907 pos=166604800 dt=12224
140
+ step 63400 loss=0.2886 pos=166707200 dt=12313
141
+ step 63500 loss=0.3033 pos=166809600 dt=12406
142
+ step 63600 loss=0.3529 pos=166912000 dt=12495
143
+ step 63700 loss=0.3322 pos=167014400 dt=12588
144
+ step 63800 loss=0.2886 pos=167116800 dt=12677
145
+ step 63900 loss=0.3162 pos=167219200 dt=12772
146
+ step 64000 loss=0.2963 pos=167321600 dt=12861
147
+ EVAL step 64000 pos=167321600 | T=2 loss=0.311 from=0.410 joint=0.734 hard=0.896 | T=16 loss=0.346 from=0.406 joint=0.693 hard=0.840
148
+ step 64100 loss=0.3149 pos=167424000 dt=13001
149
+ step 64200 loss=0.2916 pos=167526400 dt=13090
150
+ step 64300 loss=0.2967 pos=167628800 dt=13184
151
+ step 64400 loss=0.3141 pos=167731200 dt=13273
152
+ step 64500 loss=0.3023 pos=167833600 dt=13369
153
+ step 64600 loss=0.3124 pos=167936000 dt=13458
154
+ step 64700 loss=0.2994 pos=168038400 dt=13551
155
+ step 64800 loss=0.3316 pos=168140800 dt=13640
156
+ step 64900 loss=0.3270 pos=168243200 dt=13733
157
+ step 65000 loss=0.3106 pos=168345600 dt=13823
158
+ EVAL step 65000 pos=168345600 | T=2 loss=0.317 from=0.408 joint=0.734 hard=0.896 | T=16 loss=0.339 from=0.406 joint=0.684 hard=0.825
159
+ step 65100 loss=0.3058 pos=168448000 dt=13960
160
+ step 65200 loss=0.3146 pos=168550400 dt=14049
161
+ step 65300 loss=0.3121 pos=168652800 dt=14143
162
+ step 65400 loss=0.3223 pos=168755200 dt=14232
163
+ step 65500 loss=0.3025 pos=168857600 dt=14328
164
+ step 65600 loss=0.3020 pos=168960000 dt=14417
165
+ step 65700 loss=0.2979 pos=169062400 dt=14511
166
+ step 65800 loss=0.2873 pos=169164800 dt=14600
167
+ step 65900 loss=0.3004 pos=169267200 dt=14694
168
+ step 66000 loss=0.3229 pos=169369600 dt=14784
169
+ EVAL step 66000 pos=169369600 | T=2 loss=0.308 from=0.405 joint=0.729 hard=0.890 | T=16 loss=0.346 from=0.402 joint=0.670 hard=0.807
170
+ step 66100 loss=0.3185 pos=169472000 dt=14923
171
+ step 66200 loss=0.2892 pos=169574400 dt=15012
172
+ step 66300 loss=0.2523 pos=169676800 dt=15108
173
+ step 66400 loss=0.2840 pos=169779200 dt=15197
174
+ step 66500 loss=0.2815 pos=169881600 dt=15301
175
+ step 66600 loss=0.2856 pos=169984000 dt=15391
176
+ step 66700 loss=0.2774 pos=170086400 dt=15487
177
+ step 66800 loss=0.3139 pos=170188800 dt=15576
178
+ step 66900 loss=0.3182 pos=170291200 dt=15687
179
+ step 67000 loss=0.3036 pos=170393600 dt=15776
180
+ EVAL step 67000 pos=170393600 | T=2 loss=0.311 from=0.407 joint=0.728 hard=0.886 | T=16 loss=0.344 from=0.407 joint=0.672 hard=0.807
181
+ step 67100 loss=0.2917 pos=170496000 dt=15917
182
+ step 67200 loss=0.3161 pos=170598400 dt=16007
183
+ step 67300 loss=0.3043 pos=170700800 dt=16102
184
+ step 67400 loss=0.2939 pos=170803200 dt=16191
185
+ step 67500 loss=0.3030 pos=170905600 dt=16291
186
+ step 67600 loss=0.3114 pos=171008000 dt=16380
187
+ step 67700 loss=0.2902 pos=171110400 dt=16475
188
+ step 67800 loss=0.3186 pos=171212800 dt=16565
189
+ step 67900 loss=0.3033 pos=171315200 dt=16668
190
+ step 68000 loss=0.3057 pos=171417600 dt=16757
191
+ EVAL step 68000 pos=171417600 | T=2 loss=0.316 from=0.406 joint=0.725 hard=0.883 | T=16 loss=0.354 from=0.408 joint=0.683 hard=0.826
192
+ step 68100 loss=0.2924 pos=171520000 dt=16897
193
+ step 68200 loss=0.2762 pos=171622400 dt=16986
194
+ step 68300 loss=0.2498 pos=171724800 dt=17089
195
+ step 68400 loss=0.2901 pos=171827200 dt=17178
196
+ step 68500 loss=0.2697 pos=171929600 dt=17273
197
+ step 68600 loss=0.2666 pos=172032000 dt=17362
198
+ step 68700 loss=0.2163 pos=172134400 dt=17464
199
+ step 68800 loss=0.2513 pos=172236800 dt=17552
200
+ step 68900 loss=0.2116 pos=172339200 dt=17645
201
+ step 69000 loss=0.3325 pos=172441600 dt=17734
202
+ EVAL step 69000 pos=172441600 | T=2 loss=0.325 from=0.408 joint=0.709 hard=0.859 | T=16 loss=0.362 from=0.409 joint=0.658 hard=0.792
203
+ step 69100 loss=0.3221 pos=172544000 dt=17874
204
+ step 69200 loss=0.3646 pos=172646400 dt=17963
205
+ step 69300 loss=0.3304 pos=172748800 dt=18067
206
+ step 69400 loss=0.3364 pos=172851200 dt=18156
207
+ step 69500 loss=0.3088 pos=172953600 dt=18252
208
+ step 69600 loss=0.2819 pos=173056000 dt=18341
209
+ step 69700 loss=0.3219 pos=173158400 dt=18447
210
+ step 69800 loss=0.3111 pos=173260800 dt=18539
211
+ step 69900 loss=0.3291 pos=173363200 dt=18632
212
+ step 70000 loss=0.3224 pos=173465600 dt=18722
213
+ EVAL step 70000 pos=173465600 | T=2 loss=0.342 from=0.403 joint=0.711 hard=0.861 | T=16 loss=0.385 from=0.405 joint=0.666 hard=0.806
214
+ step 70100 loss=0.2964 pos=173568000 dt=18868
215
+ step 70200 loss=0.3389 pos=173670400 dt=18957
216
+ step 70300 loss=0.3016 pos=173772800 dt=19052
217
+ step 70400 loss=0.3005 pos=173875200 dt=19141
218
+ step 70500 loss=0.2978 pos=173977600 dt=19243
219
+ step 70600 loss=0.3220 pos=174080000 dt=19333
220
+ step 70700 loss=0.3227 pos=174182400 dt=19428
221
+ step 70800 loss=0.3229 pos=174284800 dt=19517
222
+ step 70900 loss=0.2918 pos=174387200 dt=19620
223
+ step 71000 loss=0.3138 pos=174489600 dt=19710
224
+ EVAL step 71000 pos=174489600 | T=2 loss=0.313 from=0.415 joint=0.725 hard=0.880 | T=16 loss=0.349 from=0.421 joint=0.673 hard=0.810
225
+ step 71100 loss=0.2925 pos=174592000 dt=19848
226
+ step 71200 loss=0.2584 pos=174694400 dt=19937
227
+ step 71300 loss=0.3324 pos=174796800 dt=20040
228
+ step 71400 loss=0.2999 pos=174899200 dt=20130
229
+ step 71500 loss=0.3007 pos=175001600 dt=20224
230
+ step 71600 loss=0.2879 pos=175104000 dt=20313
231
+ step 71700 loss=0.3539 pos=175206400 dt=20411
232
+ step 71800 loss=0.2767 pos=175308800 dt=20500
233
+ step 71900 loss=0.3139 pos=175411200 dt=20594
234
+ step 72000 loss=0.2930 pos=175513600 dt=20684
235
+ EVAL step 72000 pos=175513600 | T=2 loss=0.306 from=0.411 joint=0.720 hard=0.875 | T=16 loss=0.349 from=0.416 joint=0.652 hard=0.782
236
+ step 72100 loss=0.3083 pos=175616000 dt=20823
237
+ step 72200 loss=0.3242 pos=175718400 dt=20912
238
+ step 72300 loss=0.2853 pos=175820800 dt=21012
239
+ step 72400 loss=0.3020 pos=175923200 dt=21101
240
+ step 72500 loss=0.3178 pos=176025600 dt=21196
241
+ step 72600 loss=0.2630 pos=176128000 dt=21285
242
+ step 72700 loss=0.2680 pos=176230400 dt=21379
243
+ step 72800 loss=0.2903 pos=176332800 dt=21468
244
+ step 72900 loss=0.2655 pos=176435200 dt=21562
245
+ step 73000 loss=0.3046 pos=176537600 dt=21651
246
+ EVAL step 73000 pos=176537600 | T=2 loss=0.306 from=0.409 joint=0.732 hard=0.892 | T=16 loss=0.347 from=0.417 joint=0.673 hard=0.811
247
+ step 73100 loss=0.2956 pos=176640000 dt=21789
248
+ step 73200 loss=0.3052 pos=176742400 dt=21879
249
+ step 73300 loss=0.3042 pos=176844800 dt=21980
250
+ step 73400 loss=0.2607 pos=176947200 dt=22070
251
+ step 73500 loss=0.2871 pos=177049600 dt=22164
252
+ step 73600 loss=0.2770 pos=177152000 dt=22253
253
+ step 73700 loss=0.2765 pos=177254400 dt=22360
254
+ step 73800 loss=0.3036 pos=177356800 dt=22449
255
+ step 73900 loss=0.3009 pos=177459200 dt=22544
256
+ step 74000 loss=0.2659 pos=177561600 dt=22633
257
+ EVAL step 74000 pos=177561600 | T=2 loss=0.337 from=0.409 joint=0.716 hard=0.870 | T=16 loss=0.394 from=0.406 joint=0.646 hard=0.774
258
+ step 74100 loss=0.2758 pos=177664000 dt=22782
259
+ step 74200 loss=0.2717 pos=177766400 dt=22871
260
+ step 74300 loss=0.2648 pos=177868800 dt=22967
261
+ step 74400 loss=0.2579 pos=177971200 dt=23056
262
+ step 74500 loss=0.2194 pos=178073600 dt=23153
263
+ step 74600 loss=0.2124 pos=178176000 dt=23246
264
+ step 74700 loss=0.2296 pos=178278400 dt=23339
265
+ step 74800 loss=0.2169 pos=178380800 dt=23427
266
+ step 74900 loss=0.2343 pos=178483200 dt=23523
267
+ step 75000 loss=0.2973 pos=178585600 dt=23612
268
+ EVAL step 75000 pos=178585600 | T=2 loss=0.314 from=0.410 joint=0.707 hard=0.854 | T=16 loss=0.360 from=0.403 joint=0.646 hard=0.777
269
+ step 75100 loss=0.3203 pos=178688000 dt=23751
270
+ step 75200 loss=0.2799 pos=178790400 dt=23840
271
+ step 75300 loss=0.3047 pos=178892800 dt=23952
272
+ step 75400 loss=0.3052 pos=178995200 dt=24041
273
+ step 75500 loss=0.3239 pos=179097600 dt=24134
274
+ step 75600 loss=0.3351 pos=179200000 dt=24223
275
+ step 75700 loss=0.3010 pos=179302400 dt=24323
276
+ step 75800 loss=0.3583 pos=179404800 dt=24412
277
+ step 75900 loss=0.3000 pos=179507200 dt=24506
278
+ step 76000 loss=0.3211 pos=179609600 dt=24595
279
+ EVAL step 76000 pos=179609600 | T=2 loss=0.342 from=0.410 joint=0.709 hard=0.859 | T=16 loss=0.391 from=0.406 joint=0.651 hard=0.787
280
+ step 76100 loss=0.2952 pos=179712000 dt=24743
281
+ step 76200 loss=0.3266 pos=179814400 dt=24832
282
+ step 76300 loss=0.3031 pos=179916800 dt=24926
283
+ step 76400 loss=0.3164 pos=180019200 dt=25015
284
+ step 76500 loss=0.3165 pos=180121600 dt=25117
285
+ step 76600 loss=0.2719 pos=180224000 dt=25206
286
+ step 76700 loss=0.3013 pos=180326400 dt=25300
287
+ step 76800 loss=0.2516 pos=180428800 dt=25390
288
+ step 76900 loss=0.2853 pos=180531200 dt=25487
289
+ step 77000 loss=0.3087 pos=180633600 dt=25577
290
+ EVAL step 77000 pos=180633600 | T=2 loss=0.306 from=0.408 joint=0.726 hard=0.886 | T=16 loss=0.350 from=0.402 joint=0.667 hard=0.803
291
+ step 77100 loss=0.3282 pos=180736000 dt=25718
292
+ step 77200 loss=0.3090 pos=180838400 dt=25807
293
+ step 77300 loss=0.3044 pos=180940800 dt=25907
294
+ step 77400 loss=0.2792 pos=181043200 dt=25996
295
+ step 77500 loss=0.2897 pos=181145600 dt=26090
296
+ step 77600 loss=0.2792 pos=181248000 dt=26180
297
+ step 77700 loss=0.2772 pos=181350400 dt=26279
298
+ step 77800 loss=0.2844 pos=181452800 dt=26368
299
+ step 77900 loss=0.3217 pos=181555200 dt=26462
300
+ step 78000 loss=0.3133 pos=181657600 dt=26551
301
+ EVAL step 78000 pos=181657600 | T=2 loss=0.325 from=0.409 joint=0.708 hard=0.856 | T=16 loss=0.374 from=0.411 joint=0.641 hard=0.772
302
+ step 78100 loss=0.3093 pos=181760000 dt=26695
303
+ step 78200 loss=0.3057 pos=181862400 dt=26784
304
+ step 78300 loss=0.3213 pos=181964800 dt=26879
305
+ step 78400 loss=0.2714 pos=182067200 dt=26969
306
+ step 78500 loss=0.2830 pos=182169600 dt=27073
307
+ step 78600 loss=0.3302 pos=182272000 dt=27163
308
+ step 78700 loss=0.3304 pos=182374400 dt=27256
309
+ step 78800 loss=0.3014 pos=182476800 dt=27346
310
+ step 78900 loss=0.2785 pos=182579200 dt=27446
311
+ step 79000 loss=0.2917 pos=182681600 dt=27536
312
+ EVAL step 79000 pos=182681600 | T=2 loss=0.305 from=0.408 joint=0.726 hard=0.882 | T=16 loss=0.347 from=0.411 joint=0.660 hard=0.795
313
+ step 79100 loss=0.2496 pos=182784000 dt=27678
314
+ step 79200 loss=0.3047 pos=182886400 dt=27767
315
+ step 79300 loss=0.2939 pos=182988800 dt=27869
316
+ step 79400 loss=0.2937 pos=183091200 dt=27958
317
+ step 79500 loss=0.3101 pos=183193600 dt=28052
318
+ step 79600 loss=0.3153 pos=183296000 dt=28141
319
+ step 79700 loss=0.2751 pos=183398400 dt=28240
320
+ step 79800 loss=0.3086 pos=183500800 dt=28329
321
+ step 79900 loss=0.2792 pos=183603200 dt=28423
322
+ step 80000 loss=0.2843 pos=183705600 dt=28512
323
+ EVAL step 80000 pos=183705600 | T=2 loss=0.303 from=0.410 joint=0.725 hard=0.881 | T=16 loss=0.338 from=0.400 joint=0.657 hard=0.787
324
+ step 80100 loss=0.2689 pos=183808000 dt=28650
325
+ step 80200 loss=0.2598 pos=183910400 dt=28744
326
+ step 80300 loss=0.2597 pos=184012800 dt=28839
327
+ step 80400 loss=0.2338 pos=184115200 dt=28928
328
+ step 80500 loss=0.2892 pos=184217600 dt=29023
329
+ step 80600 loss=0.2108 pos=184320000 dt=29112
330
+ step 80700 loss=0.2055 pos=184422400 dt=29205
331
+ step 80800 loss=0.2166 pos=184524800 dt=29294
332
+ step 80900 loss=0.2015 pos=184627200 dt=29387
333
+ step 81000 loss=0.2603 pos=184729600 dt=29476
334
+ EVAL step 81000 pos=184729600 | T=2 loss=0.308 from=0.408 joint=0.699 hard=0.842 | T=16 loss=0.357 from=0.401 joint=0.608 hard=0.723
335
+ step 81100 loss=0.2725 pos=184832000 dt=29619
336
+ step 81200 loss=0.2938 pos=184934400 dt=29708
337
+ step 81300 loss=0.2810 pos=185036800 dt=29805
338
+ step 81400 loss=0.2738 pos=185139200 dt=29894
339
+ step 81500 loss=0.2622 pos=185241600 dt=29994
340
+ step 81600 loss=0.2983 pos=185344000 dt=30084
341
+ step 81700 loss=0.2945 pos=185446400 dt=30178
342
+ step 81800 loss=0.3172 pos=185548800 dt=30267
343
+ step 81900 loss=0.2850 pos=185651200 dt=30364
344
+ step 82000 loss=0.3351 pos=185753600 dt=30454
345
+ EVAL step 82000 pos=185753600 | T=2 loss=0.297 from=0.408 joint=0.743 hard=0.907 | T=16 loss=0.331 from=0.409 joint=0.680 hard=0.820
346
+ step 82100 loss=0.2545 pos=185856000 dt=30596
347
+ step 82200 loss=0.2952 pos=185958400 dt=30686
348
+ step 82300 loss=0.2792 pos=186060800 dt=30781
349
+ step 82400 loss=0.3096 pos=186163200 dt=30883
350
+ step 82500 loss=0.2961 pos=186265600 dt=30978
351
+ step 82600 loss=0.2790 pos=186368000 dt=31068
352
+ step 82700 loss=0.2888 pos=186470400 dt=31163
353
+ step 82800 loss=0.3356 pos=186572800 dt=31253
354
+ step 82900 loss=0.2839 pos=186675200 dt=31347
355
+ step 83000 loss=0.3061 pos=186777600 dt=31436
356
+ EVAL step 83000 pos=186777600 | T=2 loss=0.292 from=0.410 joint=0.740 hard=0.899 | T=16 loss=0.329 from=0.401 joint=0.690 hard=0.834
357
+ step 83100 loss=0.2923 pos=186880000 dt=31576
358
+ step 83200 loss=0.2676 pos=186982400 dt=31666
359
+ step 83300 loss=0.2816 pos=187084800 dt=31783
360
+ step 83400 loss=0.2924 pos=187187200 dt=31873
361
+ step 83500 loss=0.2905 pos=187289600 dt=31967
362
+ step 83600 loss=0.2931 pos=187392000 dt=32057
363
+ step 83700 loss=0.2943 pos=187494400 dt=32152
364
+ step 83800 loss=0.7399 pos=187596800 dt=32242
365
+ step 83900 loss=0.2943 pos=187699200 dt=32336
366
+ step 84000 loss=0.3100 pos=187801600 dt=32515
367
+ EVAL step 84000 pos=187801600 | T=2 loss=0.296 from=0.408 joint=0.728 hard=0.883 | T=16 loss=0.332 from=0.401 joint=0.665 hard=0.796
368
+ step 84100 loss=0.2976 pos=187904000 dt=32696
369
+ step 84200 loss=0.3016 pos=188006400 dt=32786
370
+ step 84300 loss=0.2637 pos=188108800 dt=32880
371
+ step 84400 loss=0.2626 pos=188211200 dt=32969
372
+ step 84500 loss=0.2608 pos=188313600 dt=33064
373
+ step 84600 loss=0.2698 pos=188416000 dt=33187
374
+ step 84700 loss=0.2426 pos=188518400 dt=33379
375
+ step 84800 loss=0.3028 pos=188620800 dt=33469
376
+ step 84900 loss=0.2713 pos=188723200 dt=33565
377
+ step 85000 loss=0.2902 pos=188825600 dt=33655
378
+ EVAL step 85000 pos=188825600 | T=2 loss=0.304 from=0.414 joint=0.734 hard=0.893 | T=16 loss=0.345 from=0.411 joint=0.673 hard=0.811
379
+ step 85100 loss=0.2871 pos=188928000 dt=33795
380
+ step 85200 loss=0.2961 pos=189030400 dt=33885
381
+ step 85300 loss=0.2567 pos=189132800 dt=33984
382
+ step 85400 loss=0.2853 pos=189235200 dt=34074
383
+ step 85500 loss=0.2855 pos=189337600 dt=34171
384
+ step 85600 loss=0.2776 pos=189440000 dt=34260
385
+ step 85700 loss=0.2640 pos=189542400 dt=34356
386
+ step 85800 loss=0.2793 pos=189644800 dt=34548
387
+ step 85900 loss=0.2870 pos=189747200 dt=34743
388
+ step 86000 loss=0.2450 pos=189849600 dt=34936
389
+ EVAL step 86000 pos=189849600 | T=2 loss=0.293 from=0.410 joint=0.736 hard=0.897 | T=16 loss=0.339 from=0.405 joint=0.678 hard=0.818
390
+ step 86100 loss=0.2813 pos=189952000 dt=35097
391
+ step 86200 loss=0.2595 pos=190054400 dt=35186
392
+ step 86300 loss=0.2783 pos=190156800 dt=35280
393
+ step 86400 loss=0.2426 pos=190259200 dt=35371
394
+ step 86500 loss=0.2430 pos=190361600 dt=35465
395
+ step 86600 loss=0.2195 pos=190464000 dt=35554
396
+ step 86700 loss=0.2269 pos=190566400 dt=35647
397
+ step 86800 loss=0.2233 pos=190668800 dt=35736
398
+ step 86900 loss=0.2231 pos=190771200 dt=35829
399
+ step 87000 loss=0.1537 pos=190873600 dt=35918
400
+ EVAL step 87000 pos=190873600 | T=2 loss=0.351 from=0.415 joint=0.619 hard=0.728 | T=16 loss=0.400 from=0.410 joint=0.588 hard=0.693
401
+ step 87100 loss=0.2620 pos=190976000 dt=36057
402
+ step 87200 loss=0.2922 pos=191078400 dt=36147
403
+ step 87300 loss=0.2782 pos=191180800 dt=36241
404
+ step 87400 loss=0.3187 pos=191283200 dt=36331
405
+ step 87500 loss=0.2976 pos=191385600 dt=36425
406
+ step 87600 loss=0.2706 pos=191488000 dt=36515
407
+ step 87700 loss=0.2902 pos=191590400 dt=36609
408
+ step 87800 loss=0.2586 pos=191692800 dt=36699
409
+ step 87900 loss=0.2837 pos=191795200 dt=36793
410
+ step 88000 loss=0.2773 pos=191897600 dt=36883
411
+ EVAL step 88000 pos=191897600 | T=2 loss=0.286 from=0.413 joint=0.739 hard=0.904 | T=16 loss=0.327 from=0.408 joint=0.678 hard=0.821
412
+ step 88100 loss=0.2832 pos=192000000 dt=37029
413
+ step 88200 loss=0.2719 pos=192102400 dt=37119
414
+ step 88300 loss=0.2987 pos=192204800 dt=37214
415
+ step 88400 loss=0.3102 pos=192307200 dt=37304
416
+ step 88500 loss=0.2950 pos=192409600 dt=37402
417
+ step 88600 loss=0.2659 pos=192512000 dt=37492
418
+ step 88700 loss=0.2759 pos=192614400 dt=37586
419
+ step 88800 loss=0.2500 pos=192716800 dt=37676
420
+ step 88900 loss=0.2725 pos=192819200 dt=37777
421
+ step 89000 loss=0.3067 pos=192921600 dt=37867
422
+ EVAL step 89000 pos=192921600 | T=2 loss=0.288 from=0.412 joint=0.734 hard=0.894 | T=16 loss=0.328 from=0.404 joint=0.689 hard=0.835
423
+ step 89100 loss=0.2822 pos=193024000 dt=38013
424
+ step 89200 loss=0.2774 pos=193126400 dt=38103
425
+ step 89300 loss=0.3006 pos=193228800 dt=38198
426
+ step 89400 loss=0.2901 pos=193331200 dt=38288
427
+ step 89500 loss=0.3306 pos=193433600 dt=38388
428
+ step 89600 loss=0.2849 pos=193536000 dt=38477
429
+ step 89700 loss=0.2943 pos=193638400 dt=38572
430
+ step 89800 loss=0.2707 pos=193740800 dt=38661
431
+ step 89900 loss=0.2791 pos=193843200 dt=38761
432
+ step 90000 loss=0.3008 pos=193945600 dt=38851
433
+ EVAL step 90000 pos=193945600 | T=2 loss=0.289 from=0.410 joint=0.744 hard=0.908 | T=16 loss=0.331 from=0.410 joint=0.673 hard=0.816
434
+ step 90100 loss=0.2816 pos=194048000 dt=38993
435
+ step 90200 loss=0.2776 pos=194150400 dt=39083
436
+ step 90300 loss=0.2890 pos=194252800 dt=39183
437
+ step 90400 loss=0.3030 pos=194355200 dt=39273
438
+ step 90500 loss=0.2989 pos=194457600 dt=39367
439
+ step 90600 loss=0.2663 pos=194560000 dt=39457
440
+ step 90700 loss=0.2892 pos=194662400 dt=39556
441
+ step 90800 loss=0.3460 pos=194764800 dt=39646
442
+ step 90900 loss=0.2882 pos=194867200 dt=39740
443
+ step 91000 loss=0.3255 pos=194969600 dt=39830
444
+ EVAL step 91000 pos=194969600 | T=2 loss=0.296 from=0.412 joint=0.732 hard=0.890 | T=16 loss=0.339 from=0.410 joint=0.666 hard=0.805
445
+ step 91100 loss=0.2831 pos=195072000 dt=39969
446
+ step 91200 loss=0.2560 pos=195174400 dt=40059
447
+ step 91300 loss=0.2664 pos=195276800 dt=40163
448
+ step 91400 loss=0.2633 pos=195379200 dt=40253
449
+ step 91500 loss=0.3141 pos=195481600 dt=40347
450
+ step 91600 loss=0.3052 pos=195584000 dt=40437
451
+ step 91700 loss=0.2864 pos=195686400 dt=40531
452
+ step 91800 loss=0.2736 pos=195788800 dt=40621
453
+ step 91900 loss=0.2739 pos=195891200 dt=40715
454
+ step 92000 loss=0.2026 pos=195993600 dt=40804
455
+ EVAL step 92000 pos=195993600 | T=2 loss=0.287 from=0.413 joint=0.738 hard=0.900 | T=16 loss=0.325 from=0.409 joint=0.688 hard=0.829
log_AMg.txt ADDED
@@ -0,0 +1,312 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ cfg AMg d=1280 H=10 tied=1 nl=8 lr=8 T=2 shards=750..875 hours=168.0
2
+ step 50900 loss=0.3455 pos=153907200 dt=95
3
+ step 51000 loss=0.3859 pos=154009600 dt=185
4
+ EVAL step 51000 pos=154009600 | T=2 loss=0.373 from=0.395 joint=0.697 hard=0.844 | T=16 loss=0.407 from=0.393 joint=0.655 hard=0.787
5
+ step 51100 loss=0.3873 pos=154112000 dt=326
6
+ step 51200 loss=0.5122 pos=154214400 dt=415
7
+ step 51300 loss=0.4222 pos=154316800 dt=510
8
+ step 51400 loss=0.3798 pos=154419200 dt=599
9
+ step 51500 loss=0.3699 pos=154521600 dt=701
10
+ step 51600 loss=0.3686 pos=154624000 dt=790
11
+ step 51700 loss=0.3709 pos=154726400 dt=886
12
+ step 51800 loss=0.5316 pos=154828800 dt=975
13
+ step 51900 loss=0.3579 pos=154931200 dt=1074
14
+ step 52000 loss=0.3183 pos=155033600 dt=1163
15
+ EVAL step 52000 pos=155033600 | T=2 loss=0.353 from=0.405 joint=0.699 hard=0.846 | T=16 loss=0.375 from=0.410 joint=0.666 hard=0.801
16
+ step 52100 loss=0.3923 pos=155136000 dt=1304
17
+ step 52200 loss=0.3347 pos=155238400 dt=1393
18
+ step 52300 loss=0.3384 pos=155340800 dt=1494
19
+ step 52400 loss=0.3458 pos=155443200 dt=1584
20
+ step 52500 loss=0.3352 pos=155545600 dt=1678
21
+ step 52600 loss=0.3707 pos=155648000 dt=1767
22
+ step 52700 loss=0.3548 pos=155750400 dt=1866
23
+ step 52800 loss=0.3183 pos=155852800 dt=1955
24
+ step 52900 loss=0.3420 pos=155955200 dt=2048
25
+ step 53000 loss=0.3473 pos=156057600 dt=2138
26
+ EVAL step 53000 pos=156057600 | T=2 loss=0.350 from=0.398 joint=0.698 hard=0.846 | T=16 loss=0.374 from=0.395 joint=0.673 hard=0.813
27
+ step 53100 loss=0.3488 pos=156160000 dt=2282
28
+ step 53200 loss=0.3715 pos=156262400 dt=2371
29
+ step 53300 loss=0.3356 pos=156364800 dt=2465
30
+ step 53400 loss=0.3700 pos=156467200 dt=2555
31
+ step 53500 loss=0.3402 pos=156569600 dt=2653
32
+ step 53600 loss=0.3021 pos=156672000 dt=2742
33
+ step 53700 loss=0.3601 pos=156774400 dt=2836
34
+ step 53800 loss=0.3021 pos=156876800 dt=2925
35
+ step 53900 loss=0.3337 pos=156979200 dt=3039
36
+ step 54000 loss=0.3362 pos=157081600 dt=3128
37
+ EVAL step 54000 pos=157081600 | T=2 loss=0.355 from=0.403 joint=0.701 hard=0.847 | T=16 loss=0.404 from=0.399 joint=0.666 hard=0.798
38
+ step 54100 loss=0.3180 pos=157184000 dt=3268
39
+ step 54200 loss=0.3772 pos=157286400 dt=3357
40
+ step 54300 loss=0.2990 pos=157388800 dt=3456
41
+ step 54400 loss=0.3505 pos=157491200 dt=3545
42
+ step 54500 loss=0.3309 pos=157593600 dt=3640
43
+ step 54600 loss=0.3116 pos=157696000 dt=3729
44
+ step 54700 loss=0.3573 pos=157798400 dt=3823
45
+ step 54800 loss=0.3212 pos=157900800 dt=3912
46
+ step 54900 loss=0.3135 pos=158003200 dt=4013
47
+ step 55000 loss=0.2937 pos=158105600 dt=4102
48
+ EVAL step 55000 pos=158105600 | T=2 loss=0.334 from=0.409 joint=0.720 hard=0.876 | T=16 loss=0.364 from=0.401 joint=0.676 hard=0.816
49
+ step 55100 loss=0.3505 pos=158208000 dt=4244
50
+ step 55200 loss=0.3125 pos=158310400 dt=4333
51
+ step 55300 loss=0.2967 pos=158412800 dt=4441
52
+ step 55400 loss=0.2855 pos=158515200 dt=4530
53
+ step 55500 loss=0.3086 pos=158617600 dt=4625
54
+ step 55600 loss=0.3155 pos=158720000 dt=4714
55
+ step 55700 loss=0.2980 pos=158822400 dt=4815
56
+ step 55800 loss=0.2732 pos=158924800 dt=4904
57
+ step 55900 loss=0.3243 pos=159027200 dt=4998
58
+ step 56000 loss=0.2923 pos=159129600 dt=5087
59
+ EVAL step 56000 pos=159129600 | T=2 loss=0.333 from=0.408 joint=0.719 hard=0.876 | T=16 loss=0.367 from=0.404 joint=0.688 hard=0.832
60
+ step 56100 loss=0.2929 pos=159232000 dt=5232
61
+ step 56200 loss=0.3009 pos=159334400 dt=5321
62
+ step 56300 loss=0.2745 pos=159436800 dt=5416
63
+ step 56400 loss=0.2592 pos=159539200 dt=5505
64
+ step 56500 loss=0.2464 pos=159641600 dt=5599
65
+ step 56600 loss=0.2572 pos=159744000 dt=5688
66
+ step 56700 loss=0.2984 pos=159846400 dt=5791
67
+ step 56800 loss=0.2496 pos=159948800 dt=5879
68
+ step 56900 loss=0.3373 pos=160051200 dt=5973
69
+ step 57000 loss=0.3523 pos=160153600 dt=6062
70
+ EVAL step 57000 pos=160153600 | T=2 loss=0.359 from=0.407 joint=0.693 hard=0.833 | T=16 loss=0.377 from=0.404 joint=0.633 hard=0.757
71
+ step 57100 loss=0.3306 pos=160256000 dt=6200
72
+ step 57200 loss=0.3375 pos=160358400 dt=6290
73
+ step 57300 loss=0.3273 pos=160460800 dt=6383
74
+ step 57400 loss=0.3295 pos=160563200 dt=6472
75
+ step 57500 loss=0.3201 pos=160665600 dt=6566
76
+ step 57600 loss=0.2959 pos=160768000 dt=6655
77
+ step 57700 loss=0.3583 pos=160870400 dt=6754
78
+ step 57800 loss=0.3119 pos=160972800 dt=6843
79
+ step 57900 loss=0.3089 pos=161075200 dt=6937
80
+ step 58000 loss=0.3423 pos=161177600 dt=7026
81
+ EVAL step 58000 pos=161177600 | T=2 loss=0.325 from=0.403 joint=0.715 hard=0.868 | T=16 loss=0.357 from=0.401 joint=0.669 hard=0.803
82
+ step 58100 loss=0.2785 pos=161280000 dt=7170
83
+ step 58200 loss=0.3320 pos=161382400 dt=7260
84
+ step 58300 loss=0.3319 pos=161484800 dt=7354
85
+ step 58400 loss=0.3366 pos=161587200 dt=7443
86
+ step 58500 loss=0.3028 pos=161689600 dt=7544
87
+ step 58600 loss=0.2973 pos=161792000 dt=7634
88
+ step 58700 loss=0.3364 pos=161894400 dt=7728
89
+ step 58800 loss=0.3290 pos=161996800 dt=7817
90
+ step 58900 loss=0.3539 pos=162099200 dt=7917
91
+ step 59000 loss=0.3133 pos=162201600 dt=8006
92
+ EVAL step 59000 pos=162201600 | T=2 loss=0.322 from=0.410 joint=0.720 hard=0.875 | T=16 loss=0.362 from=0.410 joint=0.668 hard=0.802
93
+ step 59100 loss=0.3379 pos=162304000 dt=8147
94
+ step 59200 loss=0.2862 pos=162406400 dt=8236
95
+ step 59300 loss=0.2964 pos=162508800 dt=8338
96
+ step 59400 loss=0.3252 pos=162611200 dt=8427
97
+ step 59500 loss=0.3120 pos=162713600 dt=8521
98
+ step 59600 loss=0.3469 pos=162816000 dt=8611
99
+ step 59700 loss=0.3139 pos=162918400 dt=8709
100
+ step 59800 loss=0.3020 pos=163020800 dt=8798
101
+ step 59900 loss=0.3343 pos=163123200 dt=8900
102
+ step 60000 loss=0.3571 pos=163225600 dt=8989
103
+ EVAL step 60000 pos=163225600 | T=2 loss=0.353 from=0.407 joint=0.705 hard=0.850 | T=16 loss=0.407 from=0.402 joint=0.664 hard=0.796
104
+ step 60100 loss=0.2885 pos=163328000 dt=9129
105
+ step 60200 loss=0.3069 pos=163430400 dt=9218
106
+ step 60300 loss=0.2805 pos=163532800 dt=9321
107
+ step 60400 loss=0.3389 pos=163635200 dt=9410
108
+ step 60500 loss=0.2874 pos=163737600 dt=9505
109
+ step 60600 loss=0.3215 pos=163840000 dt=9594
110
+ step 60700 loss=0.3228 pos=163942400 dt=9697
111
+ step 60800 loss=0.2968 pos=164044800 dt=9786
112
+ step 60900 loss=0.3235 pos=164147200 dt=9880
113
+ step 61000 loss=0.2983 pos=164249600 dt=9969
114
+ EVAL step 61000 pos=164249600 | T=2 loss=0.314 from=0.408 joint=0.727 hard=0.884 | T=16 loss=0.342 from=0.398 joint=0.670 hard=0.805
115
+ step 61100 loss=0.2871 pos=164352000 dt=10118
116
+ step 61200 loss=0.2932 pos=164454400 dt=10207
117
+ step 61300 loss=0.3276 pos=164556800 dt=10304
118
+ step 61400 loss=0.3244 pos=164659200 dt=10394
119
+ step 61500 loss=0.2881 pos=164761600 dt=10488
120
+ step 61600 loss=0.3275 pos=164864000 dt=10577
121
+ step 61700 loss=0.2766 pos=164966400 dt=10674
122
+ step 61800 loss=0.2801 pos=165068800 dt=10763
123
+ step 61900 loss=0.2925 pos=165171200 dt=10858
124
+ step 62000 loss=0.2620 pos=165273600 dt=10947
125
+ EVAL step 62000 pos=165273600 | T=2 loss=0.322 from=0.405 joint=0.704 hard=0.853 | T=16 loss=0.356 from=0.398 joint=0.660 hard=0.795
126
+ step 62100 loss=0.2842 pos=165376000 dt=11086
127
+ step 62200 loss=0.2718 pos=165478400 dt=11175
128
+ step 62300 loss=0.2740 pos=165580800 dt=11270
129
+ step 62400 loss=0.2932 pos=165683200 dt=11359
130
+ step 62500 loss=0.2663 pos=165785600 dt=11452
131
+ step 62600 loss=0.2431 pos=165888000 dt=11541
132
+ step 62700 loss=0.2484 pos=165990400 dt=11634
133
+ step 62800 loss=0.2669 pos=166092800 dt=11722
134
+ step 62900 loss=0.3503 pos=166195200 dt=11815
135
+ step 63000 loss=0.3086 pos=166297600 dt=11904
136
+ EVAL step 63000 pos=166297600 | T=2 loss=0.311 from=0.404 joint=0.729 hard=0.890 | T=16 loss=0.353 from=0.406 joint=0.676 hard=0.818
137
+ step 63100 loss=0.3293 pos=166400000 dt=12043
138
+ step 63200 loss=0.3000 pos=166502400 dt=12132
139
+ step 63300 loss=0.2928 pos=166604800 dt=12228
140
+ step 63400 loss=0.2868 pos=166707200 dt=12317
141
+ step 63500 loss=0.3365 pos=166809600 dt=12410
142
+ step 63600 loss=0.3082 pos=166912000 dt=12499
143
+ step 63700 loss=0.3170 pos=167014400 dt=12595
144
+ step 63800 loss=0.2989 pos=167116800 dt=12684
145
+ step 63900 loss=0.3115 pos=167219200 dt=12777
146
+ step 64000 loss=0.3408 pos=167321600 dt=12866
147
+ EVAL step 64000 pos=167321600 | T=2 loss=0.314 from=0.410 joint=0.729 hard=0.889 | T=16 loss=0.357 from=0.405 joint=0.678 hard=0.819
148
+ step 64100 loss=0.3295 pos=167424000 dt=13004
149
+ step 64200 loss=0.3372 pos=167526400 dt=13093
150
+ step 64300 loss=0.3181 pos=167628800 dt=13187
151
+ step 64400 loss=0.2638 pos=167731200 dt=13276
152
+ step 64500 loss=0.2787 pos=167833600 dt=13370
153
+ step 64600 loss=0.2826 pos=167936000 dt=13460
154
+ step 64700 loss=0.2877 pos=168038400 dt=13553
155
+ step 64800 loss=0.2918 pos=168140800 dt=13642
156
+ step 64900 loss=0.2945 pos=168243200 dt=13736
157
+ step 65000 loss=0.3254 pos=168345600 dt=13825
158
+ EVAL step 65000 pos=168345600 | T=2 loss=0.301 from=0.410 joint=0.719 hard=0.871 | T=16 loss=0.333 from=0.406 joint=0.674 hard=0.808
159
+ step 65100 loss=0.3116 pos=168448000 dt=13968
160
+ step 65200 loss=0.2845 pos=168550400 dt=14057
161
+ step 65300 loss=0.2905 pos=168652800 dt=14150
162
+ step 65400 loss=0.3087 pos=168755200 dt=14240
163
+ step 65500 loss=0.2969 pos=168857600 dt=14334
164
+ step 65600 loss=0.3009 pos=168960000 dt=14423
165
+ step 65700 loss=0.3069 pos=169062400 dt=14517
166
+ step 65800 loss=0.2889 pos=169164800 dt=14606
167
+ step 65900 loss=0.2886 pos=169267200 dt=14699
168
+ step 66000 loss=0.3217 pos=169369600 dt=14789
169
+ EVAL step 66000 pos=169369600 | T=2 loss=0.305 from=0.409 joint=0.741 hard=0.905 | T=16 loss=0.342 from=0.410 joint=0.689 hard=0.829
170
+ step 66100 loss=0.3174 pos=169472000 dt=14928
171
+ step 66200 loss=0.2710 pos=169574400 dt=15018
172
+ step 66300 loss=0.3128 pos=169676800 dt=15112
173
+ step 66400 loss=0.3211 pos=169779200 dt=15201
174
+ step 66500 loss=0.3254 pos=169881600 dt=15305
175
+ step 66600 loss=0.2763 pos=169984000 dt=15394
176
+ step 66700 loss=0.2760 pos=170086400 dt=15491
177
+ step 66800 loss=0.2709 pos=170188800 dt=15580
178
+ step 66900 loss=0.3125 pos=170291200 dt=15685
179
+ step 67000 loss=0.2877 pos=170393600 dt=15774
180
+ EVAL step 67000 pos=170393600 | T=2 loss=0.308 from=0.403 joint=0.730 hard=0.888 | T=16 loss=0.344 from=0.406 joint=0.689 hard=0.833
181
+ step 67100 loss=0.7565 pos=170496000 dt=15917
182
+ step 67200 loss=0.3055 pos=170598400 dt=16007
183
+ step 67300 loss=0.2939 pos=170700800 dt=16101
184
+ step 67400 loss=0.3126 pos=170803200 dt=16191
185
+ step 67500 loss=0.2649 pos=170905600 dt=16295
186
+ step 67600 loss=0.3084 pos=171008000 dt=16384
187
+ step 67700 loss=0.2830 pos=171110400 dt=16479
188
+ step 67800 loss=0.3052 pos=171212800 dt=16569
189
+ step 67900 loss=0.2964 pos=171315200 dt=16669
190
+ step 68000 loss=0.3032 pos=171417600 dt=16758
191
+ EVAL step 68000 pos=171417600 | T=2 loss=0.316 from=0.403 joint=0.715 hard=0.868 | T=16 loss=0.368 from=0.414 joint=0.662 hard=0.791
192
+ step 68100 loss=0.2724 pos=171520000 dt=16898
193
+ step 68200 loss=0.2780 pos=171622400 dt=16987
194
+ step 68300 loss=0.2696 pos=171724800 dt=17089
195
+ step 68400 loss=0.2721 pos=171827200 dt=17178
196
+ step 68500 loss=0.2276 pos=171929600 dt=17272
197
+ step 68600 loss=0.2598 pos=172032000 dt=17361
198
+ step 68700 loss=0.2820 pos=172134400 dt=17462
199
+ step 68800 loss=0.2405 pos=172236800 dt=17550
200
+ step 68900 loss=0.3340 pos=172339200 dt=17644
201
+ step 69000 loss=0.3388 pos=172441600 dt=17733
202
+ EVAL step 69000 pos=172441600 | T=2 loss=0.335 from=0.411 joint=0.703 hard=0.852 | T=16 loss=0.384 from=0.403 joint=0.652 hard=0.784
203
+ step 69100 loss=0.3413 pos=172544000 dt=17871
204
+ step 69200 loss=0.3231 pos=172646400 dt=17960
205
+ step 69300 loss=0.3070 pos=172748800 dt=18062
206
+ step 69400 loss=0.3158 pos=172851200 dt=18151
207
+ step 69500 loss=0.3350 pos=172953600 dt=18244
208
+ step 69600 loss=0.3128 pos=173056000 dt=18334
209
+ step 69700 loss=0.2841 pos=173158400 dt=18437
210
+ step 69800 loss=0.3293 pos=173260800 dt=18526
211
+ step 69900 loss=0.3224 pos=173363200 dt=18620
212
+ step 70000 loss=0.3215 pos=173465600 dt=18709
213
+ EVAL step 70000 pos=173465600 | T=2 loss=0.342 from=0.402 joint=0.708 hard=0.857 | T=16 loss=0.385 from=0.405 joint=0.668 hard=0.809
214
+ step 70100 loss=0.3121 pos=173568000 dt=18856
215
+ step 70200 loss=0.3035 pos=173670400 dt=18945
216
+ step 70300 loss=0.2804 pos=173772800 dt=19040
217
+ step 70400 loss=0.2988 pos=173875200 dt=19129
218
+ step 70500 loss=0.3203 pos=173977600 dt=19237
219
+ step 70600 loss=0.2822 pos=174080000 dt=19326
220
+ step 70700 loss=0.3290 pos=174182400 dt=19419
221
+ step 70800 loss=0.3019 pos=174284800 dt=19509
222
+ step 70900 loss=0.2899 pos=174387200 dt=19609
223
+ step 71000 loss=0.3075 pos=174489600 dt=19699
224
+ EVAL step 71000 pos=174489600 | T=2 loss=0.308 from=0.413 joint=0.733 hard=0.893 | T=16 loss=0.342 from=0.415 joint=0.686 hard=0.824
225
+ step 71100 loss=0.3048 pos=174592000 dt=19837
226
+ step 71200 loss=0.2740 pos=174694400 dt=19926
227
+ step 71300 loss=0.3167 pos=174796800 dt=20026
228
+ step 71400 loss=0.2867 pos=174899200 dt=20116
229
+ step 71500 loss=0.2961 pos=175001600 dt=20209
230
+ step 71600 loss=0.2980 pos=175104000 dt=20298
231
+ step 71700 loss=0.2658 pos=175206400 dt=20391
232
+ step 71800 loss=0.2747 pos=175308800 dt=20481
233
+ step 71900 loss=0.2865 pos=175411200 dt=20583
234
+ step 72000 loss=0.2995 pos=175513600 dt=20672
235
+ EVAL step 72000 pos=175513600 | T=2 loss=0.310 from=0.407 joint=0.719 hard=0.870 | T=16 loss=0.343 from=0.410 joint=0.672 hard=0.808
236
+ step 72100 loss=0.3041 pos=175616000 dt=20810
237
+ step 72200 loss=0.3482 pos=175718400 dt=20899
238
+ step 72300 loss=0.3280 pos=175820800 dt=21000
239
+ step 72400 loss=0.2606 pos=175923200 dt=21089
240
+ step 72500 loss=0.3199 pos=176025600 dt=21183
241
+ step 72600 loss=0.2956 pos=176128000 dt=21272
242
+ step 72700 loss=0.3228 pos=176230400 dt=21366
243
+ step 72800 loss=0.2793 pos=176332800 dt=21455
244
+ step 72900 loss=0.2817 pos=176435200 dt=21548
245
+ step 73000 loss=0.3171 pos=176537600 dt=21637
246
+ EVAL step 73000 pos=176537600 | T=2 loss=0.304 from=0.410 joint=0.730 hard=0.889 | T=16 loss=0.348 from=0.414 joint=0.678 hard=0.819
247
+ step 73100 loss=0.3010 pos=176640000 dt=21775
248
+ step 73200 loss=0.3352 pos=176742400 dt=21865
249
+ step 73300 loss=0.3355 pos=176844800 dt=21965
250
+ step 73400 loss=0.2756 pos=176947200 dt=22054
251
+ step 73500 loss=0.2899 pos=177049600 dt=22150
252
+ step 73600 loss=0.2806 pos=177152000 dt=22239
253
+ step 73700 loss=0.3270 pos=177254400 dt=22340
254
+ step 73800 loss=0.2653 pos=177356800 dt=22429
255
+ step 73900 loss=0.3124 pos=177459200 dt=22524
256
+ step 74000 loss=0.2801 pos=177561600 dt=22613
257
+ EVAL step 74000 pos=177561600 | T=2 loss=0.337 from=0.409 joint=0.716 hard=0.867 | T=16 loss=0.394 from=0.405 joint=0.645 hard=0.774
258
+ step 74100 loss=0.2710 pos=177664000 dt=22760
259
+ step 74200 loss=0.3140 pos=177766400 dt=22849
260
+ step 74300 loss=0.2752 pos=177868800 dt=22943
261
+ step 74400 loss=0.2736 pos=177971200 dt=23032
262
+ step 74500 loss=0.2286 pos=178073600 dt=23130
263
+ step 74600 loss=0.2323 pos=178176000 dt=23219
264
+ step 74700 loss=0.2313 pos=178278400 dt=23312
265
+ step 74800 loss=0.1640 pos=178380800 dt=23400
266
+ step 74900 loss=0.2166 pos=178483200 dt=23498
267
+ step 75000 loss=0.3110 pos=178585600 dt=23587
268
+ EVAL step 75000 pos=178585600 | T=2 loss=0.316 from=0.405 joint=0.705 hard=0.853 | T=16 loss=0.363 from=0.402 joint=0.628 hard=0.748
269
+ step 75100 loss=0.2911 pos=178688000 dt=23726
270
+ step 75200 loss=0.2973 pos=178790400 dt=23815
271
+ step 75300 loss=0.3175 pos=178892800 dt=23952
272
+ step 75400 loss=0.3106 pos=178995200 dt=24041
273
+ step 75500 loss=0.2859 pos=179097600 dt=24136
274
+ step 75600 loss=0.2991 pos=179200000 dt=24226
275
+ step 75700 loss=0.2723 pos=179302400 dt=24324
276
+ step 75800 loss=0.3307 pos=179404800 dt=24413
277
+ step 75900 loss=0.3115 pos=179507200 dt=24506
278
+ step 76000 loss=0.3192 pos=179609600 dt=24596
279
+ EVAL step 76000 pos=179609600 | T=2 loss=0.342 from=0.411 joint=0.709 hard=0.860 | T=16 loss=0.391 from=0.406 joint=0.648 hard=0.781
280
+ step 76100 loss=0.3118 pos=179712000 dt=24740
281
+ step 76200 loss=0.2924 pos=179814400 dt=24829
282
+ step 76300 loss=0.3070 pos=179916800 dt=24923
283
+ step 76400 loss=0.3022 pos=180019200 dt=25013
284
+ step 76500 loss=0.2884 pos=180121600 dt=25110
285
+ step 76600 loss=0.3243 pos=180224000 dt=25200
286
+ step 76700 loss=0.2870 pos=180326400 dt=25293
287
+ step 76800 loss=0.2730 pos=180428800 dt=25383
288
+ step 76900 loss=0.3352 pos=180531200 dt=25479
289
+ step 77000 loss=0.3038 pos=180633600 dt=25568
290
+ EVAL step 77000 pos=180633600 | T=2 loss=0.302 from=0.409 joint=0.735 hard=0.898 | T=16 loss=0.353 from=0.401 joint=0.678 hard=0.820
291
+ step 77100 loss=0.2807 pos=180736000 dt=25708
292
+ step 77200 loss=0.2927 pos=180838400 dt=25797
293
+ step 77300 loss=0.3105 pos=180940800 dt=25896
294
+ step 77400 loss=0.2940 pos=181043200 dt=25986
295
+ step 77500 loss=0.3039 pos=181145600 dt=26081
296
+ step 77600 loss=0.3129 pos=181248000 dt=26170
297
+ step 77700 loss=0.2600 pos=181350400 dt=26271
298
+ step 77800 loss=0.2751 pos=181452800 dt=26361
299
+ step 77900 loss=0.3169 pos=181555200 dt=26455
300
+ step 78000 loss=0.3028 pos=181657600 dt=26544
301
+ EVAL step 78000 pos=181657600 | T=2 loss=0.325 from=0.410 joint=0.710 hard=0.860 | T=16 loss=0.374 from=0.412 joint=0.641 hard=0.772
302
+ step 78100 loss=0.2929 pos=181760000 dt=26689
303
+ step 78200 loss=0.3035 pos=181862400 dt=26778
304
+ step 78300 loss=0.3082 pos=181964800 dt=26873
305
+ step 78400 loss=0.3180 pos=182067200 dt=26962
306
+ step 78500 loss=0.3069 pos=182169600 dt=27063
307
+ step 78600 loss=0.3248 pos=182272000 dt=27153
308
+ step 78700 loss=0.3135 pos=182374400 dt=27248
309
+ step 78800 loss=0.2854 pos=182476800 dt=27338
310
+ step 78900 loss=0.2731 pos=182579200 dt=27439
311
+ step 79000 loss=0.2966 pos=182681600 dt=27528
312
+ EVAL step 79000 pos=182681600 | T=2 loss=0.298 from=0.411 joint=0.722 hard=0.877 | T=16 loss=0.339 from=0.413 joint=0.669 hard=0.807
log_AMh.txt ADDED
@@ -0,0 +1,137 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ cfg AMh d=1280 H=10 tied=1 nl=8 lr=8 T=2 shards=875..1000 hours=168.0
2
+ step 50900 loss=0.3917 pos=153907200 dt=95
3
+ step 51000 loss=0.4107 pos=154009600 dt=185
4
+ EVAL step 51000 pos=154009600 | T=2 loss=0.445 from=0.405 joint=0.655 hard=0.781 | T=16 loss=0.474 from=0.407 joint=0.635 hard=0.762
5
+ step 51100 loss=0.4135 pos=154112000 dt=326
6
+ step 51200 loss=0.4099 pos=154214400 dt=415
7
+ push fail Client error '412 Precondition Failed' for url 'https://huggingface.co/api/models/Asilarkness/rdbt-chess/commit/main' (R
8
+ step 51300 loss=0.3282 pos=154316800 dt=515
9
+ step 51400 loss=0.3681 pos=154419200 dt=604
10
+ step 51500 loss=0.4218 pos=154521600 dt=699
11
+ step 51600 loss=0.3980 pos=154624000 dt=788
12
+ step 51700 loss=0.3347 pos=154726400 dt=883
13
+ step 51800 loss=0.3547 pos=154828800 dt=972
14
+ step 51900 loss=0.3505 pos=154931200 dt=1073
15
+ step 52000 loss=0.3422 pos=155033600 dt=1162
16
+ EVAL step 52000 pos=155033600 | T=2 loss=0.347 from=0.403 joint=0.705 hard=0.856 | T=16 loss=0.371 from=0.401 joint=0.675 hard=0.816
17
+ step 52100 loss=0.3503 pos=155136000 dt=1303
18
+ step 52200 loss=0.3859 pos=155238400 dt=1392
19
+ step 52300 loss=0.3707 pos=155340800 dt=1492
20
+ step 52400 loss=0.3568 pos=155443200 dt=1581
21
+ step 52500 loss=0.3015 pos=155545600 dt=1677
22
+ step 52600 loss=0.3437 pos=155648000 dt=1766
23
+ step 52700 loss=0.3491 pos=155750400 dt=1866
24
+ step 52800 loss=0.3557 pos=155852800 dt=1955
25
+ step 52900 loss=0.3372 pos=155955200 dt=2049
26
+ step 53000 loss=0.3606 pos=156057600 dt=2139
27
+ EVAL step 53000 pos=156057600 | T=2 loss=0.350 from=0.399 joint=0.697 hard=0.843 | T=16 loss=0.374 from=0.394 joint=0.673 hard=0.815
28
+ step 53100 loss=0.3794 pos=156160000 dt=2284
29
+ step 53200 loss=0.3563 pos=156262400 dt=2373
30
+ step 53300 loss=0.3162 pos=156364800 dt=2467
31
+ step 53400 loss=0.3686 pos=156467200 dt=2556
32
+ step 53500 loss=0.3522 pos=156569600 dt=2659
33
+ step 53600 loss=0.3463 pos=156672000 dt=2749
34
+ step 53700 loss=0.3442 pos=156774400 dt=2843
35
+ step 53800 loss=0.3653 pos=156876800 dt=2933
36
+ step 53900 loss=0.3525 pos=156979200 dt=3041
37
+ step 54000 loss=0.3388 pos=157081600 dt=3131
38
+ EVAL step 54000 pos=157081600 | T=2 loss=0.372 from=0.400 joint=0.698 hard=0.843 | T=16 loss=0.405 from=0.400 joint=0.657 hard=0.789
39
+ step 54100 loss=0.3106 pos=157184000 dt=3269
40
+ step 54200 loss=0.3566 pos=157286400 dt=3358
41
+ step 54300 loss=0.3099 pos=157388800 dt=3457
42
+ step 54400 loss=0.3172 pos=157491200 dt=3546
43
+ step 54500 loss=0.3468 pos=157593600 dt=3642
44
+ step 54600 loss=0.3384 pos=157696000 dt=3731
45
+ step 54700 loss=0.3083 pos=157798400 dt=3825
46
+ step 54800 loss=0.2953 pos=157900800 dt=3914
47
+ step 54900 loss=0.3168 pos=158003200 dt=4014
48
+ step 55000 loss=0.3219 pos=158105600 dt=4103
49
+ EVAL step 55000 pos=158105600 | T=2 loss=0.345 from=0.404 joint=0.719 hard=0.872 | T=16 loss=0.374 from=0.402 joint=0.687 hard=0.832
50
+ step 55100 loss=0.3557 pos=158208000 dt=4242
51
+ step 55200 loss=0.3399 pos=158310400 dt=4332
52
+ step 55300 loss=0.3836 pos=158412800 dt=4432
53
+ step 55400 loss=0.3220 pos=158515200 dt=4521
54
+ step 55500 loss=0.3432 pos=158617600 dt=4615
55
+ step 55600 loss=0.3102 pos=158720000 dt=4705
56
+ step 55700 loss=0.3129 pos=158822400 dt=4808
57
+ step 55800 loss=0.3050 pos=158924800 dt=4897
58
+ step 55900 loss=0.2511 pos=159027200 dt=4992
59
+ step 56000 loss=0.2883 pos=159129600 dt=5081
60
+ EVAL step 56000 pos=159129600 | T=2 loss=0.333 from=0.409 joint=0.721 hard=0.877 | T=16 loss=0.367 from=0.404 joint=0.690 hard=0.834
61
+ step 56100 loss=0.3240 pos=159232000 dt=5228
62
+ step 56200 loss=0.3312 pos=159334400 dt=5317
63
+ step 56300 loss=0.2868 pos=159436800 dt=5411
64
+ step 56400 loss=0.3257 pos=159539200 dt=5500
65
+ step 56500 loss=0.2641 pos=159641600 dt=5593
66
+ step 56600 loss=0.2392 pos=159744000 dt=5682
67
+ step 56700 loss=0.2813 pos=159846400 dt=5783
68
+ step 56800 loss=0.1972 pos=159948800 dt=5871
69
+ step 56900 loss=0.3436 pos=160051200 dt=5967
70
+ step 57000 loss=0.3485 pos=160153600 dt=6056
71
+ EVAL step 57000 pos=160153600 | T=2 loss=0.342 from=0.408 joint=0.694 hard=0.838 | T=16 loss=0.372 from=0.401 joint=0.638 hard=0.763
72
+ step 57100 loss=0.3486 pos=160256000 dt=6195
73
+ step 57200 loss=0.3012 pos=160358400 dt=6284
74
+ step 57300 loss=0.3468 pos=160460800 dt=6379
75
+ step 57400 loss=0.3220 pos=160563200 dt=6468
76
+ step 57500 loss=0.3351 pos=160665600 dt=6562
77
+ step 57600 loss=0.3024 pos=160768000 dt=6651
78
+ step 57700 loss=0.3362 pos=160870400 dt=6754
79
+ step 57800 loss=0.3118 pos=160972800 dt=6843
80
+ step 57900 loss=0.3177 pos=161075200 dt=6938
81
+ step 58000 loss=0.3249 pos=161177600 dt=7027
82
+ EVAL step 58000 pos=161177600 | T=2 loss=0.325 from=0.403 joint=0.715 hard=0.868 | T=16 loss=0.357 from=0.401 joint=0.669 hard=0.802
83
+ step 58100 loss=0.3109 pos=161280000 dt=7172
84
+ step 58200 loss=0.3215 pos=161382400 dt=7261
85
+ step 58300 loss=0.3336 pos=161484800 dt=7360
86
+ step 58400 loss=0.3004 pos=161587200 dt=7449
87
+ step 58500 loss=0.3209 pos=161689600 dt=7553
88
+ step 58600 loss=0.3295 pos=161792000 dt=7642
89
+ step 58700 loss=0.3420 pos=161894400 dt=7738
90
+ step 58800 loss=0.3744 pos=161996800 dt=7827
91
+ step 58900 loss=0.3284 pos=162099200 dt=7929
92
+ step 59000 loss=0.3205 pos=162201600 dt=8018
93
+ EVAL step 59000 pos=162201600 | T=2 loss=0.329 from=0.401 joint=0.719 hard=0.876 | T=16 loss=0.368 from=0.400 joint=0.669 hard=0.801
94
+ step 59100 loss=0.3187 pos=162304000 dt=8160
95
+ step 59200 loss=0.3235 pos=162406400 dt=8249
96
+ step 59300 loss=0.3225 pos=162508800 dt=8346
97
+ step 59400 loss=0.3087 pos=162611200 dt=8435
98
+ step 59500 loss=0.3179 pos=162713600 dt=8529
99
+ step 59600 loss=0.3018 pos=162816000 dt=8618
100
+ step 59700 loss=0.3195 pos=162918400 dt=8714
101
+ step 59800 loss=0.3172 pos=163020800 dt=8803
102
+ step 59900 loss=0.2993 pos=163123200 dt=8904
103
+ step 60000 loss=0.3174 pos=163225600 dt=8994
104
+ EVAL step 60000 pos=163225600 | T=2 loss=0.348 from=0.408 joint=0.719 hard=0.875 | T=16 loss=0.372 from=0.403 joint=0.667 hard=0.803
105
+ step 60100 loss=0.3313 pos=163328000 dt=9132
106
+ step 60200 loss=0.3029 pos=163430400 dt=9222
107
+ step 60300 loss=0.3334 pos=163532800 dt=9326
108
+ step 60400 loss=0.3197 pos=163635200 dt=9415
109
+ step 60500 loss=0.3125 pos=163737600 dt=9511
110
+ step 60600 loss=0.3203 pos=163840000 dt=9600
111
+ step 60700 loss=0.3140 pos=163942400 dt=9703
112
+ step 60800 loss=0.3151 pos=164044800 dt=9792
113
+ step 60900 loss=0.3093 pos=164147200 dt=9888
114
+ step 61000 loss=0.3105 pos=164249600 dt=9977
115
+ EVAL step 61000 pos=164249600 | T=2 loss=0.314 from=0.407 joint=0.724 hard=0.879 | T=16 loss=0.342 from=0.399 joint=0.670 hard=0.805
116
+ step 61100 loss=0.2862 pos=164352000 dt=10121
117
+ step 61200 loss=0.2993 pos=164454400 dt=10210
118
+ step 61300 loss=0.2832 pos=164556800 dt=10303
119
+ step 61400 loss=0.3164 pos=164659200 dt=10393
120
+ step 61500 loss=0.3021 pos=164761600 dt=10489
121
+ step 61600 loss=0.2966 pos=164864000 dt=10578
122
+ step 61700 loss=0.3129 pos=164966400 dt=10680
123
+ step 61800 loss=0.2911 pos=165068800 dt=10769
124
+ step 61900 loss=0.3172 pos=165171200 dt=10865
125
+ step 62000 loss=0.3108 pos=165273600 dt=10954
126
+ EVAL step 62000 pos=165273600 | T=2 loss=0.323 from=0.409 joint=0.718 hard=0.873 | T=16 loss=0.357 from=0.408 joint=0.663 hard=0.799
127
+ step 62100 loss=0.2952 pos=165376000 dt=11094
128
+ step 62200 loss=0.2984 pos=165478400 dt=11183
129
+ step 62300 loss=0.2591 pos=165580800 dt=11277
130
+ step 62400 loss=0.2387 pos=165683200 dt=11366
131
+ step 62500 loss=0.2587 pos=165785600 dt=11462
132
+ step 62600 loss=0.2128 pos=165888000 dt=11551
133
+ step 62700 loss=0.2281 pos=165990400 dt=11647
134
+ step 62800 loss=0.2124 pos=166092800 dt=11735
135
+ step 62900 loss=0.3102 pos=166195200 dt=11829
136
+ step 63000 loss=0.2859 pos=166297600 dt=11918
137
+ EVAL step 63000 pos=166297600 | T=2 loss=0.316 from=0.402 joint=0.709 hard=0.856 | T=16 loss=0.353 from=0.398 joint=0.647 hard=0.773
merge_state.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:92990320db6cc4221739158f9649e4e52fdb50e9857bfda2e727fc14a3b42f19
3
+ size 914722111
sweep_AMa.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:749fcd6249c123988f5364c09cc4e08565ea5a388fd745c87d8897410380a628
3
+ size 457367549
sweep_AMb.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0fdb8a209e195983241d7b93f09e2e6f4e4bb050d829354206f82f5be68c1979
3
+ size 457367549
sweep_AMc.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:05915148901aa197b3c8024e05b994fd5819cfea682be4007032c1e0c164e727
3
+ size 457367549
sweep_AMd.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eac14e854aecfd4a8e3d2022c5811fc7f68f69ad37e30dae101c9564b8f96f2b
3
+ size 457367549
sweep_AMe.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3f1e137b25dd7863c92399a9d2db41a6f48fe926073b45f41b0b62b68cfd59c1
3
+ size 457367549
sweep_AMf.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a3a879209116b504f475c77490725066eb08f366ce3f60d0103eed04f987d472
3
+ size 457367549
sweep_AMg.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f0b8cf633797e6771905b17229ac5cfde0ddf079577a523d64ee2810279de5c4
3
+ size 457367549
sweep_AMh.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:99899b549dbcf8958f40e9530c423aceb514faed5c77a3c9282c22a30f6e21f8
3
+ size 457367549