andrew-healey commited on
Commit
9c23277
·
verified ·
1 Parent(s): d8325ad

Upload folder using huggingface_hub

Browse files
lr1e-4_total_batch_size40960_seq_len128/args.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"hellaswag": true, "attention_kind": "selective", "log_dir": "proxy_model_sweep/lr1e-4_total_batch_size40960_seq_len128", "resume_checkpoint": null, "resume_optimizer": false, "add_a_head": false, "add_head_to_start": true, "new_head_init": "normal", "n_heads": 2, "protect_bos_token": true, "prevent_from_masking_myself": true, "max_steps": null, "group": "proxy_model_sweep", "use_wandb": true, "kill_self_after_run": false, "random_seed": 1337, "memory_penalty_epsilon": 0.1, "selection_head_linear_combo": "none", "selection_head_linear_combo_scale": 1.0, "protection_kind": "none", "leaky_relu_alpha": null, "leaky_relu_bias": null, "use_compile": true, "use_mini_model": false, "upload_to_hf": true, "seq_len": 128, "batch_size": null, "total_batch_size": 40960, "protection_head_scaling_factor": 1.0, "protection_head_bias": 0.0, "n_sliced_masks": null, "n_latent_masks": null, "mask_layernorm": false, "residual_attention_masks": false, "compute_base_shapes": false, "base_shapes_savefile": null, "mup": true, "disable_selection": true, "mup_enable_coord_check_logging": false, "max_lr": 0.0001, "decay_lr": true, "readout_zero_init": false, "query_zero_init": false, "l1_loss": false, "debugpy": false, "n_embd": 128}
lr1e-4_total_batch_size40960_seq_len128/dataloader_00999.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cec289292cf51a5462e2631af59109e64df3eca2d9e62e1fbd1c53a776a0f874
3
+ size 964
lr1e-4_total_batch_size40960_seq_len128/log2.txt ADDED
@@ -0,0 +1,1043 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ max_steps: 1000
2
+ 0 val loss 11.6870
3
+ 0 val perplexity 119016.5000
4
+ 0 train 11.697288 (lr=3.4965e-07) (hash(x)=5534142)
5
+ 1 train 11.705332 (lr=6.9930e-07) (hash(x)=6328012)
6
+ 2 train 11.705779 (lr=1.0490e-06) (hash(x)=5556098)
7
+ 3 train 11.699907 (lr=1.3986e-06) (hash(x)=6219138)
8
+ 4 train 11.687104 (lr=1.7483e-06) (hash(x)=5491972)
9
+ 5 train 11.686815 (lr=2.0979e-06) (hash(x)=5917741)
10
+ 6 train 11.674952 (lr=2.4476e-06) (hash(x)=6112458)
11
+ 7 train 11.660102 (lr=2.7972e-06) (hash(x)=6568379)
12
+ 8 train 11.641198 (lr=3.1469e-06) (hash(x)=7386082)
13
+ 9 train 11.639851 (lr=3.4965e-06) (hash(x)=6151674)
14
+ 10 train 11.649219 (lr=3.8462e-06) (hash(x)=7859060)
15
+ 11 train 11.623875 (lr=4.1958e-06) (hash(x)=6391506)
16
+ 12 train 11.595926 (lr=4.5455e-06) (hash(x)=5985675)
17
+ 13 train 11.571944 (lr=4.8951e-06) (hash(x)=6984772)
18
+ 14 train 11.559936 (lr=5.2448e-06) (hash(x)=6603717)
19
+ 15 train 11.543220 (lr=5.5944e-06) (hash(x)=6190579)
20
+ 16 train 11.523655 (lr=5.9441e-06) (hash(x)=7276303)
21
+ 17 train 11.515402 (lr=6.2937e-06) (hash(x)=6630378)
22
+ 18 train 11.501453 (lr=6.6434e-06) (hash(x)=9989744)
23
+ 19 train 11.448343 (lr=6.9930e-06) (hash(x)=6317894)
24
+ 20 train 11.404428 (lr=7.3427e-06) (hash(x)=7947703)
25
+ 21 train 11.370103 (lr=7.6923e-06) (hash(x)=5777498)
26
+ 22 train 11.356058 (lr=8.0420e-06) (hash(x)=7723226)
27
+ 23 train 11.319388 (lr=8.3916e-06) (hash(x)=6491229)
28
+ 24 train 11.311397 (lr=8.7413e-06) (hash(x)=6019389)
29
+ 25 train 11.273135 (lr=9.0909e-06) (hash(x)=4648609)
30
+ 26 train 11.233518 (lr=9.4406e-06) (hash(x)=5558355)
31
+ 27 train 11.219522 (lr=9.7902e-06) (hash(x)=6369326)
32
+ 28 train 11.159124 (lr=1.0140e-05) (hash(x)=4906853)
33
+ 29 train 11.070101 (lr=1.0490e-05) (hash(x)=6437376)
34
+ 30 train 11.091187 (lr=1.0839e-05) (hash(x)=7084093)
35
+ 31 train 11.025919 (lr=1.1189e-05) (hash(x)=7048804)
36
+ 32 train 11.015018 (lr=1.1538e-05) (hash(x)=5752594)
37
+ 33 train 10.888662 (lr=1.1888e-05) (hash(x)=6393520)
38
+ 34 train 10.979302 (lr=1.2238e-05) (hash(x)=6793550)
39
+ 35 train 10.897839 (lr=1.2587e-05) (hash(x)=6581348)
40
+ 36 train 10.855896 (lr=1.2937e-05) (hash(x)=6235837)
41
+ 37 train 10.872488 (lr=1.3287e-05) (hash(x)=7083021)
42
+ 38 train 10.833597 (lr=1.3636e-05) (hash(x)=6155821)
43
+ 39 train 10.750823 (lr=1.3986e-05) (hash(x)=5657123)
44
+ 40 train 10.650710 (lr=1.4336e-05) (hash(x)=5312267)
45
+ 41 train 10.732980 (lr=1.4685e-05) (hash(x)=6830813)
46
+ 42 train 10.680898 (lr=1.5035e-05) (hash(x)=6424933)
47
+ 43 train 10.639899 (lr=1.5385e-05) (hash(x)=6680023)
48
+ 44 train 10.615314 (lr=1.5734e-05) (hash(x)=7523279)
49
+ 45 train 10.547634 (lr=1.6084e-05) (hash(x)=6635925)
50
+ 46 train 10.577503 (lr=1.6434e-05) (hash(x)=5869441)
51
+ 47 train 10.562243 (lr=1.6783e-05) (hash(x)=6349004)
52
+ 48 train 10.481762 (lr=1.7133e-05) (hash(x)=6008454)
53
+ 49 train 10.500949 (lr=1.7483e-05) (hash(x)=7682741)
54
+ 50 val loss 10.5264
55
+ 50 val perplexity 37288.4062
56
+ 50 train 10.545626 (lr=1.7832e-05) (hash(x)=7794028)
57
+ 51 train 10.516732 (lr=1.8182e-05) (hash(x)=6166062)
58
+ 52 train 10.466324 (lr=1.8531e-05) (hash(x)=4811192)
59
+ 53 train 10.453982 (lr=1.8881e-05) (hash(x)=7822680)
60
+ 54 train 10.423635 (lr=1.9231e-05) (hash(x)=6431409)
61
+ 55 train 10.445662 (lr=1.9580e-05) (hash(x)=5323032)
62
+ 56 train 10.413032 (lr=1.9930e-05) (hash(x)=7363286)
63
+ 57 train 10.427016 (lr=2.0280e-05) (hash(x)=5616075)
64
+ 58 train 10.454243 (lr=2.0629e-05) (hash(x)=5672608)
65
+ 59 train 10.410952 (lr=2.0979e-05) (hash(x)=6839867)
66
+ 60 train 10.373855 (lr=2.1329e-05) (hash(x)=4916928)
67
+ 61 train 10.416348 (lr=2.1678e-05) (hash(x)=8353379)
68
+ 62 train 10.377666 (lr=2.2028e-05) (hash(x)=4838871)
69
+ 63 train 10.308009 (lr=2.2378e-05) (hash(x)=6795665)
70
+ 64 train 10.341059 (lr=2.2727e-05) (hash(x)=6678038)
71
+ 65 train 10.380463 (lr=2.3077e-05) (hash(x)=6112215)
72
+ 66 train 10.354484 (lr=2.3427e-05) (hash(x)=5915361)
73
+ 67 train 10.295052 (lr=2.3776e-05) (hash(x)=5771588)
74
+ 68 train 10.296517 (lr=2.4126e-05) (hash(x)=5688829)
75
+ 69 train 10.249701 (lr=2.4476e-05) (hash(x)=5443305)
76
+ 70 train 10.368075 (lr=2.4825e-05) (hash(x)=6901933)
77
+ 71 train 10.305390 (lr=2.5175e-05) (hash(x)=4294425)
78
+ 72 train 10.266986 (lr=2.5524e-05) (hash(x)=4363941)
79
+ 73 train 10.340864 (lr=2.5874e-05) (hash(x)=5486335)
80
+ 74 train 10.266190 (lr=2.6224e-05) (hash(x)=7295165)
81
+ 75 train 10.302212 (lr=2.6573e-05) (hash(x)=6267424)
82
+ 76 train 10.288569 (lr=2.6923e-05) (hash(x)=5270452)
83
+ 77 train 10.281313 (lr=2.7273e-05) (hash(x)=6305297)
84
+ 78 train 10.316290 (lr=2.7622e-05) (hash(x)=6040077)
85
+ 79 train 10.245375 (lr=2.7972e-05) (hash(x)=7130251)
86
+ 80 train 10.142735 (lr=2.8322e-05) (hash(x)=6508350)
87
+ 81 train 10.266699 (lr=2.8671e-05) (hash(x)=5859576)
88
+ 82 train 10.131472 (lr=2.9021e-05) (hash(x)=4892065)
89
+ 83 train 10.208331 (lr=2.9371e-05) (hash(x)=7317289)
90
+ 84 train 10.222984 (lr=2.9720e-05) (hash(x)=5510218)
91
+ 85 train 10.225883 (lr=3.0070e-05) (hash(x)=6752265)
92
+ 86 train 10.180833 (lr=3.0420e-05) (hash(x)=6530867)
93
+ 87 train 10.181967 (lr=3.0769e-05) (hash(x)=5790078)
94
+ 88 train 10.167474 (lr=3.1119e-05) (hash(x)=7614023)
95
+ 89 train 10.192016 (lr=3.1469e-05) (hash(x)=5245146)
96
+ 90 train 10.143373 (lr=3.1818e-05) (hash(x)=5786972)
97
+ 91 train 10.120044 (lr=3.2168e-05) (hash(x)=6346459)
98
+ 92 train 10.123945 (lr=3.2517e-05) (hash(x)=5209234)
99
+ 93 train 10.069736 (lr=3.2867e-05) (hash(x)=5676768)
100
+ 94 train 10.159283 (lr=3.3217e-05) (hash(x)=5863092)
101
+ 95 train 10.125297 (lr=3.3566e-05) (hash(x)=7019489)
102
+ 96 train 10.048639 (lr=3.3916e-05) (hash(x)=5483019)
103
+ 97 train 10.074350 (lr=3.4266e-05) (hash(x)=5968716)
104
+ 98 train 10.015015 (lr=3.4615e-05) (hash(x)=7780194)
105
+ 99 train 10.036878 (lr=3.4965e-05) (hash(x)=3784321)
106
+ 100 val loss 10.0435
107
+ 100 val perplexity 23006.2656
108
+ 100 train 10.018503 (lr=3.5315e-05) (hash(x)=7774109)
109
+ 101 train 10.071948 (lr=3.5664e-05) (hash(x)=6576749)
110
+ 102 train 10.004955 (lr=3.6014e-05) (hash(x)=7072404)
111
+ 103 train 9.925790 (lr=3.6364e-05) (hash(x)=6556025)
112
+ 104 train 9.979046 (lr=3.6713e-05) (hash(x)=7325027)
113
+ 105 train 9.985698 (lr=3.7063e-05) (hash(x)=6373412)
114
+ 106 train 9.907884 (lr=3.7413e-05) (hash(x)=5262868)
115
+ 107 train 9.821184 (lr=3.7762e-05) (hash(x)=8055563)
116
+ 108 train 9.888244 (lr=3.8112e-05) (hash(x)=5932790)
117
+ 109 train 9.870143 (lr=3.8462e-05) (hash(x)=5680154)
118
+ 110 train 9.889037 (lr=3.8811e-05) (hash(x)=5914217)
119
+ 111 train 9.835287 (lr=3.9161e-05) (hash(x)=6424128)
120
+ 112 train 9.816049 (lr=3.9510e-05) (hash(x)=7804089)
121
+ 113 train 9.910012 (lr=3.9860e-05) (hash(x)=6490149)
122
+ 114 train 9.746980 (lr=4.0210e-05) (hash(x)=6283540)
123
+ 115 train 9.750226 (lr=4.0559e-05) (hash(x)=5728337)
124
+ 116 train 9.777737 (lr=4.0909e-05) (hash(x)=7400272)
125
+ 117 train 9.751750 (lr=4.1259e-05) (hash(x)=6473432)
126
+ 118 train 9.753048 (lr=4.1608e-05) (hash(x)=6922646)
127
+ 119 train 9.689855 (lr=4.1958e-05) (hash(x)=7277141)
128
+ 120 train 9.692665 (lr=4.2308e-05) (hash(x)=7443124)
129
+ 121 train 9.684964 (lr=4.2657e-05) (hash(x)=5632735)
130
+ 122 train 9.567986 (lr=4.3007e-05) (hash(x)=5881015)
131
+ 123 train 9.611203 (lr=4.3357e-05) (hash(x)=5809411)
132
+ 124 train 9.633434 (lr=4.3706e-05) (hash(x)=7232964)
133
+ 125 train 9.567163 (lr=4.4056e-05) (hash(x)=6171829)
134
+ 126 train 9.607805 (lr=4.4406e-05) (hash(x)=5884456)
135
+ 127 train 9.558887 (lr=4.4755e-05) (hash(x)=5077031)
136
+ 128 train 9.572595 (lr=4.5105e-05) (hash(x)=6248895)
137
+ 129 train 9.452215 (lr=4.5455e-05) (hash(x)=6918004)
138
+ 130 train 9.451519 (lr=4.5804e-05) (hash(x)=6482466)
139
+ 131 train 9.404288 (lr=4.6154e-05) (hash(x)=6588573)
140
+ 132 train 9.450846 (lr=4.6503e-05) (hash(x)=6136939)
141
+ 133 train 9.392819 (lr=4.6853e-05) (hash(x)=7040157)
142
+ 134 train 9.384050 (lr=4.7203e-05) (hash(x)=6455595)
143
+ 135 train 9.397627 (lr=4.7552e-05) (hash(x)=4122983)
144
+ 136 train 9.308072 (lr=4.7902e-05) (hash(x)=4724052)
145
+ 137 train 9.347659 (lr=4.8252e-05) (hash(x)=6820797)
146
+ 138 train 9.400518 (lr=4.8601e-05) (hash(x)=7477594)
147
+ 139 train 9.302575 (lr=4.8951e-05) (hash(x)=6146305)
148
+ 140 train 9.217159 (lr=4.9301e-05) (hash(x)=6814831)
149
+ 141 train 9.270012 (lr=4.9650e-05) (hash(x)=5548298)
150
+ 142 train 9.285732 (lr=5.0000e-05) (hash(x)=7609416)
151
+ 143 train 9.249261 (lr=5.0350e-05) (hash(x)=7574342)
152
+ 144 train 9.179375 (lr=5.0699e-05) (hash(x)=5906937)
153
+ 145 train 9.138689 (lr=5.1049e-05) (hash(x)=8013471)
154
+ 146 train 9.114400 (lr=5.1399e-05) (hash(x)=6146213)
155
+ 147 train 9.110011 (lr=5.1748e-05) (hash(x)=6709902)
156
+ 148 train 9.142542 (lr=5.2098e-05) (hash(x)=6402571)
157
+ 149 train 9.043824 (lr=5.2448e-05) (hash(x)=7683614)
158
+ 150 val loss 9.0555
159
+ 150 val perplexity 8565.1602
160
+ 150 train 9.019387 (lr=5.2797e-05) (hash(x)=5453026)
161
+ 151 train 8.964670 (lr=5.3147e-05) (hash(x)=6260017)
162
+ 152 train 8.989517 (lr=5.3497e-05) (hash(x)=6591466)
163
+ 153 train 8.990527 (lr=5.3846e-05) (hash(x)=5501472)
164
+ 154 train 8.986162 (lr=5.4196e-05) (hash(x)=5871698)
165
+ 155 train 8.873516 (lr=5.4545e-05) (hash(x)=4282052)
166
+ 156 train 8.899151 (lr=5.4895e-05) (hash(x)=7824660)
167
+ 157 train 8.860447 (lr=5.5245e-05) (hash(x)=5954950)
168
+ 158 train 8.835849 (lr=5.5594e-05) (hash(x)=5484656)
169
+ 159 train 8.872515 (lr=5.5944e-05) (hash(x)=7031328)
170
+ 160 train 8.800587 (lr=5.6294e-05) (hash(x)=9157069)
171
+ 161 train 8.834757 (lr=5.6643e-05) (hash(x)=5996883)
172
+ 162 train 8.826200 (lr=5.6993e-05) (hash(x)=7645447)
173
+ 163 train 8.906298 (lr=5.7343e-05) (hash(x)=6336164)
174
+ 164 train 8.714073 (lr=5.7692e-05) (hash(x)=5224240)
175
+ 165 train 8.893360 (lr=5.8042e-05) (hash(x)=5447801)
176
+ 166 train 8.699451 (lr=5.8392e-05) (hash(x)=4155598)
177
+ 167 train 8.658438 (lr=5.8741e-05) (hash(x)=7482744)
178
+ 168 train 8.636909 (lr=5.9091e-05) (hash(x)=4733414)
179
+ 169 train 8.670689 (lr=5.9441e-05) (hash(x)=7806284)
180
+ 170 train 8.637245 (lr=5.9790e-05) (hash(x)=7236478)
181
+ 171 train 8.630403 (lr=6.0140e-05) (hash(x)=4656255)
182
+ 172 train 8.656121 (lr=6.0490e-05) (hash(x)=9298325)
183
+ 173 train 8.606151 (lr=6.0839e-05) (hash(x)=6287524)
184
+ 174 train 8.559595 (lr=6.1189e-05) (hash(x)=7895970)
185
+ 175 train 8.580208 (lr=6.1538e-05) (hash(x)=8847992)
186
+ 176 train 8.668791 (lr=6.1888e-05) (hash(x)=6338998)
187
+ 177 train 8.674383 (lr=6.2238e-05) (hash(x)=7104011)
188
+ 178 train 8.506351 (lr=6.2587e-05) (hash(x)=5573214)
189
+ 179 train 8.562291 (lr=6.2937e-05) (hash(x)=6882897)
190
+ 180 train 8.470406 (lr=6.3287e-05) (hash(x)=5854242)
191
+ 181 train 8.388101 (lr=6.3636e-05) (hash(x)=6972494)
192
+ 182 train 8.411957 (lr=6.3986e-05) (hash(x)=6796805)
193
+ 183 train 8.285467 (lr=6.4336e-05) (hash(x)=6139120)
194
+ 184 train 8.368826 (lr=6.4685e-05) (hash(x)=6486796)
195
+ 185 train 8.337702 (lr=6.5035e-05) (hash(x)=5578797)
196
+ 186 train 8.342622 (lr=6.5385e-05) (hash(x)=6895326)
197
+ 187 train 8.378795 (lr=6.5734e-05) (hash(x)=7783160)
198
+ 188 train 8.369094 (lr=6.6084e-05) (hash(x)=4867625)
199
+ 189 train 8.281475 (lr=6.6434e-05) (hash(x)=6631123)
200
+ 190 train 8.160976 (lr=6.6783e-05) (hash(x)=6710965)
201
+ 191 train 8.202462 (lr=6.7133e-05) (hash(x)=6294657)
202
+ 192 train 8.267230 (lr=6.7483e-05) (hash(x)=5443606)
203
+ 193 train 8.269627 (lr=6.7832e-05) (hash(x)=7035770)
204
+ 194 train 8.270896 (lr=6.8182e-05) (hash(x)=4899541)
205
+ 195 train 8.163097 (lr=6.8531e-05) (hash(x)=6227120)
206
+ 196 train 8.220283 (lr=6.8881e-05) (hash(x)=5981648)
207
+ 197 train 8.189032 (lr=6.9231e-05) (hash(x)=5558993)
208
+ 198 train 8.233647 (lr=6.9580e-05) (hash(x)=7015671)
209
+ 199 train 8.146052 (lr=6.9930e-05) (hash(x)=5931234)
210
+ 200 val loss 8.1800
211
+ 200 val perplexity 3569.0156
212
+ 200 train 8.143770 (lr=7.0280e-05) (hash(x)=4582159)
213
+ 201 train 8.113088 (lr=7.0629e-05) (hash(x)=5029472)
214
+ 202 train 8.093628 (lr=7.0979e-05) (hash(x)=5434102)
215
+ 203 train 8.280756 (lr=7.1329e-05) (hash(x)=6810638)
216
+ 204 train 8.099703 (lr=7.1678e-05) (hash(x)=6521562)
217
+ 205 train 8.187403 (lr=7.2028e-05) (hash(x)=7226932)
218
+ 206 train 8.066390 (lr=7.2378e-05) (hash(x)=6957849)
219
+ 207 train 8.108775 (lr=7.2727e-05) (hash(x)=5334281)
220
+ 208 train 8.147187 (lr=7.3077e-05) (hash(x)=5887983)
221
+ 209 train 8.072449 (lr=7.3427e-05) (hash(x)=6393576)
222
+ 210 train 8.212346 (lr=7.3776e-05) (hash(x)=5057959)
223
+ 211 train 8.254953 (lr=7.4126e-05) (hash(x)=8595271)
224
+ 212 train 8.106679 (lr=7.4476e-05) (hash(x)=8395167)
225
+ 213 train 8.140000 (lr=7.4825e-05) (hash(x)=7634417)
226
+ 214 train 7.963014 (lr=7.5175e-05) (hash(x)=5257771)
227
+ 215 train 8.034513 (lr=7.5524e-05) (hash(x)=7674238)
228
+ 216 train 8.135171 (lr=7.5874e-05) (hash(x)=5300555)
229
+ 217 train 7.906672 (lr=7.6224e-05) (hash(x)=6066139)
230
+ 218 train 8.012554 (lr=7.6573e-05) (hash(x)=6688212)
231
+ 219 train 8.150175 (lr=7.6923e-05) (hash(x)=6642404)
232
+ 220 train 7.964653 (lr=7.7273e-05) (hash(x)=4961570)
233
+ 221 train 7.859112 (lr=7.7622e-05) (hash(x)=5105190)
234
+ 222 train 8.080304 (lr=7.7972e-05) (hash(x)=7155991)
235
+ 223 train 7.941610 (lr=7.8322e-05) (hash(x)=5132972)
236
+ 224 train 7.943127 (lr=7.8671e-05) (hash(x)=6269946)
237
+ 225 train 7.946702 (lr=7.9021e-05) (hash(x)=5396133)
238
+ 226 train 7.957126 (lr=7.9371e-05) (hash(x)=6394083)
239
+ 227 train 7.933005 (lr=7.9720e-05) (hash(x)=5237804)
240
+ 228 train 7.948494 (lr=8.0070e-05) (hash(x)=5216638)
241
+ 229 train 7.884851 (lr=8.0420e-05) (hash(x)=6735565)
242
+ 230 train 7.990614 (lr=8.0769e-05) (hash(x)=5983575)
243
+ 231 train 7.943805 (lr=8.1119e-05) (hash(x)=5831619)
244
+ 232 train 7.807380 (lr=8.1469e-05) (hash(x)=9109083)
245
+ 233 train 7.954245 (lr=8.1818e-05) (hash(x)=6912967)
246
+ 234 train 8.091221 (lr=8.2168e-05) (hash(x)=5965674)
247
+ 235 train 7.952698 (lr=8.2517e-05) (hash(x)=5822533)
248
+ 236 train 7.882913 (lr=8.2867e-05) (hash(x)=5268216)
249
+ 237 train 7.788689 (lr=8.3217e-05) (hash(x)=6531900)
250
+ 238 train 7.794518 (lr=8.3566e-05) (hash(x)=5426499)
251
+ 239 train 8.186525 (lr=8.3916e-05) (hash(x)=7455372)
252
+ 240 train 8.421332 (lr=8.4266e-05) (hash(x)=8319119)
253
+ 241 train 8.411531 (lr=8.4615e-05) (hash(x)=7592128)
254
+ 242 train 8.447161 (lr=8.4965e-05) (hash(x)=8704170)
255
+ 243 train 8.044080 (lr=8.5315e-05) (hash(x)=7165037)
256
+ 244 train 7.922281 (lr=8.5664e-05) (hash(x)=7678000)
257
+ 245 train 7.803015 (lr=8.6014e-05) (hash(x)=7414788)
258
+ 246 train 7.842107 (lr=8.6364e-05) (hash(x)=7326216)
259
+ 247 train 7.921700 (lr=8.6713e-05) (hash(x)=5857162)
260
+ 248 train 8.317278 (lr=8.7063e-05) (hash(x)=14715669)
261
+ 249 train 10.722971 (lr=8.7413e-05) (hash(x)=16950560)
262
+ 250 val loss 7.8102
263
+ 250 val perplexity 2465.6804
264
+ 250 train 7.842124 (lr=8.7762e-05) (hash(x)=6159125)
265
+ 251 train 7.839992 (lr=8.8112e-05) (hash(x)=5908442)
266
+ 252 train 7.711533 (lr=8.8462e-05) (hash(x)=6668887)
267
+ 253 train 7.799224 (lr=8.8811e-05) (hash(x)=7662975)
268
+ 254 train 7.777843 (lr=8.9161e-05) (hash(x)=6669539)
269
+ 255 train 7.765620 (lr=8.9510e-05) (hash(x)=5015998)
270
+ 256 train 7.717095 (lr=8.9860e-05) (hash(x)=5458386)
271
+ 257 train 7.822394 (lr=9.0210e-05) (hash(x)=4849847)
272
+ 258 train 7.850494 (lr=9.0559e-05) (hash(x)=5125562)
273
+ 259 train 7.863416 (lr=9.0909e-05) (hash(x)=6012047)
274
+ 260 train 7.769378 (lr=9.1259e-05) (hash(x)=6079364)
275
+ 261 train 7.768021 (lr=9.1608e-05) (hash(x)=6030395)
276
+ 262 train 7.748400 (lr=9.1958e-05) (hash(x)=6781762)
277
+ 263 train 7.779303 (lr=9.2308e-05) (hash(x)=5238325)
278
+ 264 train 7.842370 (lr=9.2657e-05) (hash(x)=8611144)
279
+ 265 train 7.718645 (lr=9.3007e-05) (hash(x)=6901605)
280
+ 266 train 7.746768 (lr=9.3357e-05) (hash(x)=5499337)
281
+ 267 train 7.748135 (lr=9.3706e-05) (hash(x)=7134874)
282
+ 268 train 7.726998 (lr=9.4056e-05) (hash(x)=6149038)
283
+ 269 train 7.738696 (lr=9.4406e-05) (hash(x)=6609760)
284
+ 270 train 7.839108 (lr=9.4755e-05) (hash(x)=6652627)
285
+ 271 train 7.795169 (lr=9.5105e-05) (hash(x)=8618579)
286
+ 272 train 7.696238 (lr=9.5455e-05) (hash(x)=5896892)
287
+ 273 train 7.797287 (lr=9.5804e-05) (hash(x)=8945854)
288
+ 274 train 7.654253 (lr=9.6154e-05) (hash(x)=6775103)
289
+ 275 train 7.698669 (lr=9.6503e-05) (hash(x)=6200024)
290
+ 276 train 7.613345 (lr=9.6853e-05) (hash(x)=5238139)
291
+ 277 train 7.733175 (lr=9.7203e-05) (hash(x)=5780948)
292
+ 278 train 7.718573 (lr=9.7552e-05) (hash(x)=6168407)
293
+ 279 train 7.690348 (lr=9.7902e-05) (hash(x)=3775881)
294
+ 280 train 7.673783 (lr=9.8252e-05) (hash(x)=6848709)
295
+ 281 train 7.989607 (lr=9.8601e-05) (hash(x)=6478113)
296
+ 282 train 7.709716 (lr=9.8951e-05) (hash(x)=6483914)
297
+ 283 train 7.677688 (lr=9.9301e-05) (hash(x)=5927795)
298
+ 284 train 7.688281 (lr=9.9650e-05) (hash(x)=8438033)
299
+ 285 train 7.720707 (lr=1.0000e-04) (hash(x)=5684788)
300
+ 286 train 7.564610 (lr=1.0000e-04) (hash(x)=5710182)
301
+ 287 train 7.647551 (lr=1.0000e-04) (hash(x)=6160063)
302
+ 288 train 7.519366 (lr=9.9998e-05) (hash(x)=5955100)
303
+ 289 train 7.449736 (lr=9.9996e-05) (hash(x)=6672309)
304
+ 290 train 7.806431 (lr=9.9993e-05) (hash(x)=9001564)
305
+ 291 train 8.120126 (lr=9.9989e-05) (hash(x)=5454654)
306
+ 292 train 7.552334 (lr=9.9984e-05) (hash(x)=5174062)
307
+ 293 train 7.666497 (lr=9.9979e-05) (hash(x)=6199203)
308
+ 294 train 7.591342 (lr=9.9972e-05) (hash(x)=8352747)
309
+ 295 train 7.711527 (lr=9.9965e-05) (hash(x)=4547106)
310
+ 296 train 7.694947 (lr=9.9956e-05) (hash(x)=6433099)
311
+ 297 train 7.550387 (lr=9.9947e-05) (hash(x)=6424428)
312
+ 298 train 7.615989 (lr=9.9937e-05) (hash(x)=5975712)
313
+ 299 train 7.579180 (lr=9.9926e-05) (hash(x)=5458532)
314
+ 300 val loss 7.6401
315
+ 300 val perplexity 2079.9033
316
+ 300 train 7.770287 (lr=9.9915e-05) (hash(x)=7002774)
317
+ 301 train 7.894138 (lr=9.9902e-05) (hash(x)=8206173)
318
+ 302 train 7.953984 (lr=9.9889e-05) (hash(x)=7784971)
319
+ 303 train 7.917474 (lr=9.9874e-05) (hash(x)=6001843)
320
+ 304 train 7.710042 (lr=9.9859e-05) (hash(x)=5698642)
321
+ 305 train 7.629550 (lr=9.9843e-05) (hash(x)=6600751)
322
+ 306 train 7.594560 (lr=9.9826e-05) (hash(x)=5804610)
323
+ 307 train 7.552759 (lr=9.9808e-05) (hash(x)=5162652)
324
+ 308 train 7.552864 (lr=9.9789e-05) (hash(x)=5297006)
325
+ 309 train 7.561480 (lr=9.9770e-05) (hash(x)=6728323)
326
+ 310 train 7.559034 (lr=9.9749e-05) (hash(x)=3964882)
327
+ 311 train 7.555285 (lr=9.9728e-05) (hash(x)=6735993)
328
+ 312 train 7.646221 (lr=9.9706e-05) (hash(x)=5150316)
329
+ 313 train 7.512674 (lr=9.9683e-05) (hash(x)=5220448)
330
+ 314 train 7.659996 (lr=9.9659e-05) (hash(x)=6971289)
331
+ 315 train 7.722139 (lr=9.9634e-05) (hash(x)=5012069)
332
+ 316 train 7.583620 (lr=9.9609e-05) (hash(x)=6576136)
333
+ 317 train 7.573574 (lr=9.9582e-05) (hash(x)=6077647)
334
+ 318 train 7.673189 (lr=9.9555e-05) (hash(x)=6032381)
335
+ 319 train 7.535773 (lr=9.9526e-05) (hash(x)=8410716)
336
+ 320 train 7.576756 (lr=9.9497e-05) (hash(x)=6212516)
337
+ 321 train 7.556870 (lr=9.9467e-05) (hash(x)=5298545)
338
+ 322 train 7.640959 (lr=9.9437e-05) (hash(x)=5779073)
339
+ 323 train 7.529157 (lr=9.9405e-05) (hash(x)=6340947)
340
+ 324 train 7.317072 (lr=9.9372e-05) (hash(x)=6530441)
341
+ 325 train 7.622580 (lr=9.9339e-05) (hash(x)=7818007)
342
+ 326 train 7.500859 (lr=9.9305e-05) (hash(x)=6277435)
343
+ 327 train 7.511684 (lr=9.9270e-05) (hash(x)=6226320)
344
+ 328 train 7.580492 (lr=9.9234e-05) (hash(x)=5463751)
345
+ 329 train 7.506289 (lr=9.9197e-05) (hash(x)=6579329)
346
+ 330 train 7.604564 (lr=9.9159e-05) (hash(x)=6660540)
347
+ 331 train 7.561738 (lr=9.9121e-05) (hash(x)=6491272)
348
+ 332 train 7.554421 (lr=9.9081e-05) (hash(x)=6766947)
349
+ 333 train 7.793908 (lr=9.9041e-05) (hash(x)=3721218)
350
+ 334 train 7.660347 (lr=9.9000e-05) (hash(x)=5229898)
351
+ 335 train 7.476605 (lr=9.8958e-05) (hash(x)=6054428)
352
+ 336 train 7.631429 (lr=9.8915e-05) (hash(x)=6290464)
353
+ 337 train 7.603313 (lr=9.8872e-05) (hash(x)=6904498)
354
+ 338 train 7.626131 (lr=9.8827e-05) (hash(x)=7765805)
355
+ 339 train 7.591843 (lr=9.8782e-05) (hash(x)=6543764)
356
+ 340 train 7.596447 (lr=9.8736e-05) (hash(x)=7204955)
357
+ 341 train 7.533058 (lr=9.8689e-05) (hash(x)=4423711)
358
+ 342 train 7.485778 (lr=9.8641e-05) (hash(x)=5343477)
359
+ 343 train 7.623853 (lr=9.8592e-05) (hash(x)=6595856)
360
+ 344 train 7.460827 (lr=9.8543e-05) (hash(x)=7229015)
361
+ 345 train 7.528174 (lr=9.8492e-05) (hash(x)=5916034)
362
+ 346 train 7.655439 (lr=9.8441e-05) (hash(x)=6939521)
363
+ 347 train 7.528080 (lr=9.8389e-05) (hash(x)=5492185)
364
+ 348 train 7.452679 (lr=9.8336e-05) (hash(x)=6898006)
365
+ 349 train 7.369509 (lr=9.8282e-05) (hash(x)=5462008)
366
+ 350 val loss 7.5809
367
+ 350 val perplexity 1960.3555
368
+ 350 train 7.462333 (lr=9.8228e-05) (hash(x)=7821902)
369
+ 351 train 7.544786 (lr=9.8172e-05) (hash(x)=6608108)
370
+ 352 train 7.598093 (lr=9.8116e-05) (hash(x)=6295394)
371
+ 353 train 7.478810 (lr=9.8059e-05) (hash(x)=6803449)
372
+ 354 train 7.473893 (lr=9.8001e-05) (hash(x)=7444134)
373
+ 355 train 7.503682 (lr=9.7942e-05) (hash(x)=5361923)
374
+ 356 train 7.503118 (lr=9.7882e-05) (hash(x)=7803039)
375
+ 357 train 7.627037 (lr=9.7822e-05) (hash(x)=7042578)
376
+ 358 train 7.517677 (lr=9.7761e-05) (hash(x)=6647007)
377
+ 359 train 7.539265 (lr=9.7699e-05) (hash(x)=6303658)
378
+ 360 train 7.635437 (lr=9.7636e-05) (hash(x)=6158927)
379
+ 361 train 7.475334 (lr=9.7572e-05) (hash(x)=6419656)
380
+ 362 train 7.383346 (lr=9.7507e-05) (hash(x)=7891271)
381
+ 363 train 7.582181 (lr=9.7442e-05) (hash(x)=7099888)
382
+ 364 train 7.543154 (lr=9.7376e-05) (hash(x)=6352424)
383
+ 365 train 7.480652 (lr=9.7309e-05) (hash(x)=5863145)
384
+ 366 train 7.502373 (lr=9.7241e-05) (hash(x)=6844597)
385
+ 367 train 7.561295 (lr=9.7172e-05) (hash(x)=6009861)
386
+ 368 train 7.458214 (lr=9.7103e-05) (hash(x)=5374957)
387
+ 369 train 7.511323 (lr=9.7032e-05) (hash(x)=3923225)
388
+ 370 train 7.453778 (lr=9.6961e-05) (hash(x)=5848862)
389
+ 371 train 7.472786 (lr=9.6889e-05) (hash(x)=4440634)
390
+ 372 train 7.429233 (lr=9.6817e-05) (hash(x)=4788998)
391
+ 373 train 7.404116 (lr=9.6743e-05) (hash(x)=7378165)
392
+ 374 train 7.441351 (lr=9.6669e-05) (hash(x)=4771243)
393
+ 375 train 7.434455 (lr=9.6593e-05) (hash(x)=5716089)
394
+ 376 train 7.466268 (lr=9.6518e-05) (hash(x)=7991009)
395
+ 377 train 7.490191 (lr=9.6441e-05) (hash(x)=5348607)
396
+ 378 train 7.564111 (lr=9.6363e-05) (hash(x)=6036146)
397
+ 379 train 7.442731 (lr=9.6285e-05) (hash(x)=6658718)
398
+ 380 train 7.391121 (lr=9.6206e-05) (hash(x)=6176253)
399
+ 381 train 7.409175 (lr=9.6126e-05) (hash(x)=4542895)
400
+ 382 train 7.303351 (lr=9.6045e-05) (hash(x)=5549073)
401
+ 383 train 7.605895 (lr=9.5963e-05) (hash(x)=5342783)
402
+ 384 train 7.643357 (lr=9.5881e-05) (hash(x)=6549546)
403
+ 385 train 7.398177 (lr=9.5798e-05) (hash(x)=6170599)
404
+ 386 train 7.717548 (lr=9.5714e-05) (hash(x)=8483134)
405
+ 387 train 7.811637 (lr=9.5629e-05) (hash(x)=4865036)
406
+ 388 train 7.421105 (lr=9.5544e-05) (hash(x)=4433890)
407
+ 389 train 7.967146 (lr=9.5457e-05) (hash(x)=6738597)
408
+ 390 train 7.553675 (lr=9.5370e-05) (hash(x)=5459620)
409
+ 391 train 7.559793 (lr=9.5282e-05) (hash(x)=4604074)
410
+ 392 train 7.564392 (lr=9.5194e-05) (hash(x)=8045501)
411
+ 393 train 7.669625 (lr=9.5104e-05) (hash(x)=6176464)
412
+ 394 train 7.490872 (lr=9.5014e-05) (hash(x)=6127317)
413
+ 395 train 7.520163 (lr=9.4923e-05) (hash(x)=5571462)
414
+ 396 train 7.523367 (lr=9.4831e-05) (hash(x)=6640755)
415
+ 397 train 7.583597 (lr=9.4739e-05) (hash(x)=6147295)
416
+ 398 train 7.507103 (lr=9.4646e-05) (hash(x)=6120461)
417
+ 399 train 7.472795 (lr=9.4551e-05) (hash(x)=5123928)
418
+ 400 val loss 7.5337
419
+ 400 val perplexity 1870.0198
420
+ 400 train 7.501031 (lr=9.4457e-05) (hash(x)=5428813)
421
+ 401 train 7.472812 (lr=9.4361e-05) (hash(x)=6448586)
422
+ 402 train 7.613503 (lr=9.4265e-05) (hash(x)=8116987)
423
+ 403 train 7.647737 (lr=9.4168e-05) (hash(x)=5974342)
424
+ 404 train 7.666876 (lr=9.4070e-05) (hash(x)=5261336)
425
+ 405 train 7.464823 (lr=9.3971e-05) (hash(x)=5307056)
426
+ 406 train 7.381677 (lr=9.3872e-05) (hash(x)=5865119)
427
+ 407 train 7.422967 (lr=9.3772e-05) (hash(x)=6672380)
428
+ 408 train 7.478955 (lr=9.3671e-05) (hash(x)=6886810)
429
+ 409 train 7.608885 (lr=9.3569e-05) (hash(x)=6564071)
430
+ 410 train 7.359358 (lr=9.3467e-05) (hash(x)=7705089)
431
+ 411 train 7.379630 (lr=9.3364e-05) (hash(x)=5773073)
432
+ 412 train 7.384902 (lr=9.3260e-05) (hash(x)=5905976)
433
+ 413 train 8.177303 (lr=9.3155e-05) (hash(x)=6936653)
434
+ 414 train 7.458583 (lr=9.3050e-05) (hash(x)=6592773)
435
+ 415 train 7.403864 (lr=9.2944e-05) (hash(x)=6796551)
436
+ 416 train 7.456775 (lr=9.2837e-05) (hash(x)=6151589)
437
+ 417 train 7.715071 (lr=9.2729e-05) (hash(x)=7045164)
438
+ 418 train 7.491668 (lr=9.2621e-05) (hash(x)=6650506)
439
+ 419 train 7.418640 (lr=9.2512e-05) (hash(x)=7839748)
440
+ 420 train 7.657086 (lr=9.2402e-05) (hash(x)=6028671)
441
+ 421 train 7.500286 (lr=9.2292e-05) (hash(x)=6118408)
442
+ 422 train 7.449217 (lr=9.2181e-05) (hash(x)=5757526)
443
+ 423 train 7.602367 (lr=9.2069e-05) (hash(x)=6030172)
444
+ 424 train 7.627944 (lr=9.1956e-05) (hash(x)=6719882)
445
+ 425 train 7.491516 (lr=9.1843e-05) (hash(x)=5028050)
446
+ 426 train 7.479056 (lr=9.1729e-05) (hash(x)=5682910)
447
+ 427 train 7.576201 (lr=9.1614e-05) (hash(x)=5747891)
448
+ 428 train 7.359124 (lr=9.1499e-05) (hash(x)=5734208)
449
+ 429 train 7.441398 (lr=9.1382e-05) (hash(x)=7024381)
450
+ 430 train 7.714160 (lr=9.1266e-05) (hash(x)=7814361)
451
+ 431 train 7.548769 (lr=9.1148e-05) (hash(x)=5765893)
452
+ 432 train 7.375860 (lr=9.1030e-05) (hash(x)=5572048)
453
+ 433 train 7.543426 (lr=9.0911e-05) (hash(x)=5610058)
454
+ 434 train 7.403123 (lr=9.0791e-05) (hash(x)=4459483)
455
+ 435 train 7.356299 (lr=9.0671e-05) (hash(x)=6758411)
456
+ 436 train 7.455046 (lr=9.0550e-05) (hash(x)=6701381)
457
+ 437 train 7.634153 (lr=9.0428e-05) (hash(x)=5416400)
458
+ 438 train 7.346602 (lr=9.0306e-05) (hash(x)=5919569)
459
+ 439 train 7.502546 (lr=9.0182e-05) (hash(x)=5282870)
460
+ 440 train 7.351941 (lr=9.0059e-05) (hash(x)=8072181)
461
+ 441 train 7.298467 (lr=8.9934e-05) (hash(x)=7996770)
462
+ 442 train 7.373286 (lr=8.9809e-05) (hash(x)=5093519)
463
+ 443 train 7.397471 (lr=8.9683e-05) (hash(x)=6270917)
464
+ 444 train 7.290353 (lr=8.9557e-05) (hash(x)=5852780)
465
+ 445 train 7.188518 (lr=8.9430e-05) (hash(x)=5327728)
466
+ 446 train 7.416437 (lr=8.9302e-05) (hash(x)=6260765)
467
+ 447 train 7.498930 (lr=8.9173e-05) (hash(x)=6212894)
468
+ 448 train 7.394323 (lr=8.9044e-05) (hash(x)=6724010)
469
+ 449 train 7.487379 (lr=8.8914e-05) (hash(x)=6059934)
470
+ 450 val loss 7.5099
471
+ 450 val perplexity 1826.0381
472
+ 450 train 7.241916 (lr=8.8784e-05) (hash(x)=5587382)
473
+ 451 train 7.376804 (lr=8.8653e-05) (hash(x)=7274108)
474
+ 452 train 7.453814 (lr=8.8521e-05) (hash(x)=6085396)
475
+ 453 train 7.341969 (lr=8.8388e-05) (hash(x)=4784666)
476
+ 454 train 7.442936 (lr=8.8255e-05) (hash(x)=6288638)
477
+ 455 train 7.467599 (lr=8.8122e-05) (hash(x)=7714288)
478
+ 456 train 7.522188 (lr=8.7987e-05) (hash(x)=5534748)
479
+ 457 train 7.433039 (lr=8.7852e-05) (hash(x)=6361270)
480
+ 458 train 7.431294 (lr=8.7717e-05) (hash(x)=4682956)
481
+ 459 train 7.400281 (lr=8.7580e-05) (hash(x)=7568935)
482
+ 460 train 7.473555 (lr=8.7444e-05) (hash(x)=7973311)
483
+ 461 train 7.330730 (lr=8.7306e-05) (hash(x)=6987352)
484
+ 462 train 7.351486 (lr=8.7168e-05) (hash(x)=5413904)
485
+ 463 train 7.539089 (lr=8.7029e-05) (hash(x)=6732924)
486
+ 464 train 7.470128 (lr=8.6890e-05) (hash(x)=6219754)
487
+ 465 train 7.404897 (lr=8.6750e-05) (hash(x)=6839070)
488
+ 466 train 7.336129 (lr=8.6609e-05) (hash(x)=6090120)
489
+ 467 train 7.786535 (lr=8.6468e-05) (hash(x)=7115176)
490
+ 468 train 7.489489 (lr=8.6326e-05) (hash(x)=4989299)
491
+ 469 train 7.379863 (lr=8.6184e-05) (hash(x)=7497590)
492
+ 470 train 7.491033 (lr=8.6041e-05) (hash(x)=3977557)
493
+ 471 train 7.432912 (lr=8.5897e-05) (hash(x)=5524165)
494
+ 472 train 7.450377 (lr=8.5753e-05) (hash(x)=5124936)
495
+ 473 train 7.440199 (lr=8.5608e-05) (hash(x)=5522119)
496
+ 474 train 7.346241 (lr=8.5462e-05) (hash(x)=5030865)
497
+ 475 train 7.414797 (lr=8.5316e-05) (hash(x)=6411783)
498
+ 476 train 7.352341 (lr=8.5170e-05) (hash(x)=6581937)
499
+ 477 train 7.400636 (lr=8.5022e-05) (hash(x)=5116910)
500
+ 478 train 7.200799 (lr=8.4875e-05) (hash(x)=7353459)
501
+ 479 train 7.243629 (lr=8.4726e-05) (hash(x)=6344535)
502
+ 480 train 7.615621 (lr=8.4577e-05) (hash(x)=5840169)
503
+ 481 train 7.374841 (lr=8.4428e-05) (hash(x)=6754290)
504
+ 482 train 7.609682 (lr=8.4278e-05) (hash(x)=6484573)
505
+ 483 train 7.241075 (lr=8.4127e-05) (hash(x)=5260125)
506
+ 484 train 7.289880 (lr=8.3976e-05) (hash(x)=4881179)
507
+ 485 train 7.382818 (lr=8.3824e-05) (hash(x)=5001634)
508
+ 486 train 7.317724 (lr=8.3672e-05) (hash(x)=7424810)
509
+ 487 train 7.188832 (lr=8.3519e-05) (hash(x)=5176431)
510
+ 488 train 7.296174 (lr=8.3366e-05) (hash(x)=5421624)
511
+ 489 train 7.348325 (lr=8.3212e-05) (hash(x)=5674439)
512
+ 490 train 7.263091 (lr=8.3057e-05) (hash(x)=5819157)
513
+ 491 train 7.437231 (lr=8.2902e-05) (hash(x)=6506856)
514
+ 492 train 7.261283 (lr=8.2746e-05) (hash(x)=6119124)
515
+ 493 train 7.514084 (lr=8.2590e-05) (hash(x)=6241669)
516
+ 494 train 7.106044 (lr=8.2434e-05) (hash(x)=4992756)
517
+ 495 train 7.583076 (lr=8.2276e-05) (hash(x)=7277927)
518
+ 496 train 7.512190 (lr=8.2119e-05) (hash(x)=6622097)
519
+ 497 train 7.614405 (lr=8.1960e-05) (hash(x)=7026582)
520
+ 498 train 7.494966 (lr=8.1801e-05) (hash(x)=5840987)
521
+ 499 train 7.196585 (lr=8.1642e-05) (hash(x)=4351074)
522
+ 500 val loss 7.4676
523
+ 500 val perplexity 1750.3236
524
+ 500 train 7.340311 (lr=8.1482e-05) (hash(x)=7301477)
525
+ 501 train 7.564056 (lr=8.1322e-05) (hash(x)=6264167)
526
+ 502 train 7.281240 (lr=8.1161e-05) (hash(x)=4673363)
527
+ 503 train 7.579064 (lr=8.1000e-05) (hash(x)=6332240)
528
+ 504 train 7.447380 (lr=8.0838e-05) (hash(x)=5093790)
529
+ 505 train 7.582738 (lr=8.0676e-05) (hash(x)=6890713)
530
+ 506 train 7.318880 (lr=8.0513e-05) (hash(x)=6042691)
531
+ 507 train 7.503478 (lr=8.0349e-05) (hash(x)=7080746)
532
+ 508 train 7.494229 (lr=8.0186e-05) (hash(x)=7609207)
533
+ 509 train 7.290571 (lr=8.0021e-05) (hash(x)=4591273)
534
+ 510 train 7.464218 (lr=7.9856e-05) (hash(x)=8183111)
535
+ 511 train 7.741896 (lr=7.9691e-05) (hash(x)=5893095)
536
+ 512 train 7.493413 (lr=7.9525e-05) (hash(x)=5709077)
537
+ 513 train 7.444020 (lr=7.9359e-05) (hash(x)=6218715)
538
+ 514 train 7.326029 (lr=7.9192e-05) (hash(x)=4722188)
539
+ 515 train 7.222136 (lr=7.9025e-05) (hash(x)=6313422)
540
+ 516 train 7.364163 (lr=7.8858e-05) (hash(x)=6350660)
541
+ 517 train 7.441318 (lr=7.8689e-05) (hash(x)=5371913)
542
+ 518 train 7.549626 (lr=7.8521e-05) (hash(x)=6283595)
543
+ 519 train 7.541395 (lr=7.8352e-05) (hash(x)=6179167)
544
+ 520 train 7.386250 (lr=7.8182e-05) (hash(x)=7865194)
545
+ 521 train 7.588134 (lr=7.8012e-05) (hash(x)=4883760)
546
+ 522 train 7.382503 (lr=7.7842e-05) (hash(x)=5769286)
547
+ 523 train 7.362512 (lr=7.7671e-05) (hash(x)=6309912)
548
+ 524 train 7.353332 (lr=7.7500e-05) (hash(x)=7490849)
549
+ 525 train 7.323593 (lr=7.7328e-05) (hash(x)=5631134)
550
+ 526 train 7.326139 (lr=7.7156e-05) (hash(x)=4754560)
551
+ 527 train 7.260207 (lr=7.6984e-05) (hash(x)=5448702)
552
+ 528 train 7.475715 (lr=7.6811e-05) (hash(x)=7168842)
553
+ 529 train 7.447642 (lr=7.6637e-05) (hash(x)=7469195)
554
+ 530 train 7.273111 (lr=7.6463e-05) (hash(x)=5479500)
555
+ 531 train 7.469433 (lr=7.6289e-05) (hash(x)=5634177)
556
+ 532 train 7.382682 (lr=7.6115e-05) (hash(x)=9779982)
557
+ 533 train 7.252612 (lr=7.5940e-05) (hash(x)=3641890)
558
+ 534 train 7.371501 (lr=7.5764e-05) (hash(x)=7994103)
559
+ 535 train 7.344176 (lr=7.5588e-05) (hash(x)=5358547)
560
+ 536 train 7.378058 (lr=7.5412e-05) (hash(x)=6797424)
561
+ 537 train 7.437099 (lr=7.5235e-05) (hash(x)=4876580)
562
+ 538 train 7.215506 (lr=7.5058e-05) (hash(x)=4867629)
563
+ 539 train 7.347958 (lr=7.4881e-05) (hash(x)=6601967)
564
+ 540 train 7.295176 (lr=7.4703e-05) (hash(x)=5480657)
565
+ 541 train 7.342244 (lr=7.4525e-05) (hash(x)=7127346)
566
+ 542 train 7.495181 (lr=7.4346e-05) (hash(x)=7224794)
567
+ 543 train 7.329207 (lr=7.4167e-05) (hash(x)=5967252)
568
+ 544 train 7.416029 (lr=7.3988e-05) (hash(x)=7415262)
569
+ 545 train 7.308474 (lr=7.3808e-05) (hash(x)=6522544)
570
+ 546 train 7.338857 (lr=7.3628e-05) (hash(x)=5181640)
571
+ 547 train 7.249084 (lr=7.3448e-05) (hash(x)=5916880)
572
+ 548 train 7.248053 (lr=7.3267e-05) (hash(x)=5620397)
573
+ 549 train 7.321142 (lr=7.3086e-05) (hash(x)=5994970)
574
+ 550 val loss 7.4364
575
+ 550 val perplexity 1696.6321
576
+ 550 train 7.383569 (lr=7.2904e-05) (hash(x)=6532717)
577
+ 551 train 7.408372 (lr=7.2723e-05) (hash(x)=6262762)
578
+ 552 train 7.193246 (lr=7.2540e-05) (hash(x)=5495326)
579
+ 553 train 7.560027 (lr=7.2358e-05) (hash(x)=5927868)
580
+ 554 train 7.313498 (lr=7.2175e-05) (hash(x)=5485433)
581
+ 555 train 7.293728 (lr=7.1992e-05) (hash(x)=6462425)
582
+ 556 train 7.306563 (lr=7.1808e-05) (hash(x)=6299518)
583
+ 557 train 7.400600 (lr=7.1624e-05) (hash(x)=5950376)
584
+ 558 train 7.511989 (lr=7.1440e-05) (hash(x)=6296440)
585
+ 559 train 7.651197 (lr=7.1256e-05) (hash(x)=6974102)
586
+ 560 train 7.423055 (lr=7.1071e-05) (hash(x)=6228679)
587
+ 561 train 7.300619 (lr=7.0886e-05) (hash(x)=6120319)
588
+ 562 train 7.575762 (lr=7.0701e-05) (hash(x)=6289494)
589
+ 563 train 7.355860 (lr=7.0515e-05) (hash(x)=6877818)
590
+ 564 train 7.318218 (lr=7.0329e-05) (hash(x)=6490976)
591
+ 565 train 7.458638 (lr=7.0143e-05) (hash(x)=6419450)
592
+ 566 train 7.410298 (lr=6.9956e-05) (hash(x)=6840857)
593
+ 567 train 7.347164 (lr=6.9769e-05) (hash(x)=4837686)
594
+ 568 train 7.403899 (lr=6.9582e-05) (hash(x)=5561493)
595
+ 569 train 7.296288 (lr=6.9394e-05) (hash(x)=5831242)
596
+ 570 train 7.419981 (lr=6.9207e-05) (hash(x)=6851511)
597
+ 571 train 7.488548 (lr=6.9019e-05) (hash(x)=6169616)
598
+ 572 train 7.410009 (lr=6.8830e-05) (hash(x)=5357151)
599
+ 573 train 7.230301 (lr=6.8642e-05) (hash(x)=5471028)
600
+ 574 train 7.304619 (lr=6.8453e-05) (hash(x)=5051894)
601
+ 575 train 7.302619 (lr=6.8264e-05) (hash(x)=4930450)
602
+ 576 train 7.421576 (lr=6.8075e-05) (hash(x)=6537527)
603
+ 577 train 7.389394 (lr=6.7885e-05) (hash(x)=6330572)
604
+ 578 train 7.521694 (lr=6.7695e-05) (hash(x)=5315155)
605
+ 579 train 7.331930 (lr=6.7505e-05) (hash(x)=5416079)
606
+ 580 train 7.287281 (lr=6.7315e-05) (hash(x)=6255865)
607
+ 581 train 7.575086 (lr=6.7124e-05) (hash(x)=6723036)
608
+ 582 train 7.276083 (lr=6.6933e-05) (hash(x)=7107190)
609
+ 583 train 7.380904 (lr=6.6742e-05) (hash(x)=5868958)
610
+ 584 train 7.498897 (lr=6.6551e-05) (hash(x)=7385923)
611
+ 585 train 7.296333 (lr=6.6360e-05) (hash(x)=5373053)
612
+ 586 train 7.304144 (lr=6.6168e-05) (hash(x)=5371748)
613
+ 587 train 7.480772 (lr=6.5976e-05) (hash(x)=5783717)
614
+ 588 train 7.206968 (lr=6.5784e-05) (hash(x)=6388848)
615
+ 589 train 7.228001 (lr=6.5592e-05) (hash(x)=4865068)
616
+ 590 train 7.190323 (lr=6.5399e-05) (hash(x)=5844961)
617
+ 591 train 7.277395 (lr=6.5206e-05) (hash(x)=5989273)
618
+ 592 train 7.191666 (lr=6.5013e-05) (hash(x)=5136202)
619
+ 593 train 7.347380 (lr=6.4820e-05) (hash(x)=5811889)
620
+ 594 train 7.271703 (lr=6.4627e-05) (hash(x)=6554415)
621
+ 595 train 7.312141 (lr=6.4433e-05) (hash(x)=5969272)
622
+ 596 train 7.304579 (lr=6.4240e-05) (hash(x)=7121518)
623
+ 597 train 7.847871 (lr=6.4046e-05) (hash(x)=6363768)
624
+ 598 train 7.366454 (lr=6.3852e-05) (hash(x)=7030284)
625
+ 599 train 7.495286 (lr=6.3658e-05) (hash(x)=7243339)
626
+ 600 val loss 7.4092
627
+ 600 val perplexity 1651.1241
628
+ 600 train 7.583957 (lr=6.3463e-05) (hash(x)=7199493)
629
+ 601 train 7.625485 (lr=6.3269e-05) (hash(x)=5819281)
630
+ 602 train 7.328205 (lr=6.3074e-05) (hash(x)=7285909)
631
+ 603 train 7.360774 (lr=6.2879e-05) (hash(x)=6730878)
632
+ 604 train 7.411400 (lr=6.2684e-05) (hash(x)=5200141)
633
+ 605 train 7.496822 (lr=6.2489e-05) (hash(x)=5208573)
634
+ 606 train 7.603688 (lr=6.2294e-05) (hash(x)=5976779)
635
+ 607 train 7.407602 (lr=6.2098e-05) (hash(x)=5921767)
636
+ 608 train 7.471373 (lr=6.1903e-05) (hash(x)=6596032)
637
+ 609 train 7.492226 (lr=6.1707e-05) (hash(x)=5428058)
638
+ 610 train 7.425014 (lr=6.1511e-05) (hash(x)=5748050)
639
+ 611 train 7.428777 (lr=6.1315e-05) (hash(x)=6397421)
640
+ 612 train 7.511092 (lr=6.1119e-05) (hash(x)=5638644)
641
+ 613 train 7.381983 (lr=6.0923e-05) (hash(x)=7317623)
642
+ 614 train 7.466084 (lr=6.0726e-05) (hash(x)=5254932)
643
+ 615 train 7.525526 (lr=6.0530e-05) (hash(x)=4945141)
644
+ 616 train 7.482324 (lr=6.0333e-05) (hash(x)=7093994)
645
+ 617 train 7.399872 (lr=6.0137e-05) (hash(x)=7511894)
646
+ 618 train 7.380695 (lr=5.9940e-05) (hash(x)=6706832)
647
+ 619 train 7.364240 (lr=5.9743e-05) (hash(x)=6175371)
648
+ 620 train 7.402891 (lr=5.9546e-05) (hash(x)=7089897)
649
+ 621 train 7.404323 (lr=5.9349e-05) (hash(x)=6828202)
650
+ 622 train 7.740974 (lr=5.9152e-05) (hash(x)=7095301)
651
+ 623 train 7.410327 (lr=5.8955e-05) (hash(x)=5652191)
652
+ 624 train 7.457762 (lr=5.8758e-05) (hash(x)=7336431)
653
+ 625 train 7.468296 (lr=5.8560e-05) (hash(x)=6426339)
654
+ 626 train 7.559300 (lr=5.8363e-05) (hash(x)=5874982)
655
+ 627 train 7.498997 (lr=5.8165e-05) (hash(x)=6978359)
656
+ 628 train 7.250139 (lr=5.7968e-05) (hash(x)=4438173)
657
+ 629 train 7.328531 (lr=5.7770e-05) (hash(x)=6475356)
658
+ 630 train 7.558687 (lr=5.7573e-05) (hash(x)=5030069)
659
+ 631 train 7.303802 (lr=5.7375e-05) (hash(x)=6507939)
660
+ 632 train 7.314055 (lr=5.7177e-05) (hash(x)=6373406)
661
+ 633 train 7.477476 (lr=5.6979e-05) (hash(x)=6623130)
662
+ 634 train 7.372625 (lr=5.6782e-05) (hash(x)=6276741)
663
+ 635 train 7.284429 (lr=5.6584e-05) (hash(x)=7169874)
664
+ 636 train 7.358021 (lr=5.6386e-05) (hash(x)=4947348)
665
+ 637 train 7.457016 (lr=5.6188e-05) (hash(x)=7407715)
666
+ 638 train 7.603267 (lr=5.5990e-05) (hash(x)=5453341)
667
+ 639 train 7.322425 (lr=5.5792e-05) (hash(x)=6520161)
668
+ 640 train 7.424628 (lr=5.5594e-05) (hash(x)=6011882)
669
+ 641 train 7.381405 (lr=5.5396e-05) (hash(x)=5573372)
670
+ 642 train 7.556123 (lr=5.5198e-05) (hash(x)=5344827)
671
+ 643 train 7.636243 (lr=5.5000e-05) (hash(x)=6127887)
672
+ 644 train 7.339408 (lr=5.4802e-05) (hash(x)=6052304)
673
+ 645 train 7.358655 (lr=5.4604e-05) (hash(x)=7039733)
674
+ 646 train 7.443398 (lr=5.4406e-05) (hash(x)=7746451)
675
+ 647 train 7.647322 (lr=5.4208e-05) (hash(x)=7694373)
676
+ 648 train 7.501451 (lr=5.4010e-05) (hash(x)=6911004)
677
+ 649 train 7.432812 (lr=5.3812e-05) (hash(x)=8055224)
678
+ 650 val loss 7.3872
679
+ 650 val perplexity 1615.1925
680
+ 650 train 7.463547 (lr=5.3614e-05) (hash(x)=7060476)
681
+ 651 train 7.454604 (lr=5.3416e-05) (hash(x)=6002317)
682
+ 652 train 7.484139 (lr=5.3218e-05) (hash(x)=6524390)
683
+ 653 train 7.450332 (lr=5.3021e-05) (hash(x)=8331083)
684
+ 654 train 7.483687 (lr=5.2823e-05) (hash(x)=7084852)
685
+ 655 train 7.421829 (lr=5.2625e-05) (hash(x)=6190174)
686
+ 656 train 7.377133 (lr=5.2427e-05) (hash(x)=5601780)
687
+ 657 train 7.528090 (lr=5.2230e-05) (hash(x)=4764034)
688
+ 658 train 7.400972 (lr=5.2032e-05) (hash(x)=6361606)
689
+ 659 train 7.387288 (lr=5.1835e-05) (hash(x)=5697877)
690
+ 660 train 7.438275 (lr=5.1637e-05) (hash(x)=6249793)
691
+ 661 train 7.486716 (lr=5.1440e-05) (hash(x)=7424709)
692
+ 662 train 7.504687 (lr=5.1242e-05) (hash(x)=8203076)
693
+ 663 train 7.229175 (lr=5.1045e-05) (hash(x)=5290019)
694
+ 664 train 7.313414 (lr=5.0848e-05) (hash(x)=6278399)
695
+ 665 train 7.382229 (lr=5.0651e-05) (hash(x)=5709192)
696
+ 666 train 7.477844 (lr=5.0454e-05) (hash(x)=5660292)
697
+ 667 train 7.444237 (lr=5.0257e-05) (hash(x)=5563697)
698
+ 668 train 7.441769 (lr=5.0060e-05) (hash(x)=6325245)
699
+ 669 train 7.418172 (lr=4.9863e-05) (hash(x)=6259741)
700
+ 670 train 7.530769 (lr=4.9667e-05) (hash(x)=6238355)
701
+ 671 train 7.509060 (lr=4.9470e-05) (hash(x)=7778802)
702
+ 672 train 7.429945 (lr=4.9274e-05) (hash(x)=6283496)
703
+ 673 train 7.510829 (lr=4.9077e-05) (hash(x)=5325614)
704
+ 674 train 7.354575 (lr=4.8881e-05) (hash(x)=5376057)
705
+ 675 train 7.332964 (lr=4.8685e-05) (hash(x)=5203215)
706
+ 676 train 7.439693 (lr=4.8489e-05) (hash(x)=4893030)
707
+ 677 train 7.382489 (lr=4.8293e-05) (hash(x)=7595573)
708
+ 678 train 7.445487 (lr=4.8097e-05) (hash(x)=7188575)
709
+ 679 train 7.493494 (lr=4.7902e-05) (hash(x)=6858739)
710
+ 680 train 7.348853 (lr=4.7706e-05) (hash(x)=7157032)
711
+ 681 train 7.475469 (lr=4.7511e-05) (hash(x)=5871492)
712
+ 682 train 7.334995 (lr=4.7316e-05) (hash(x)=5636927)
713
+ 683 train 7.314777 (lr=4.7121e-05) (hash(x)=6135391)
714
+ 684 train 7.405115 (lr=4.6926e-05) (hash(x)=3351432)
715
+ 685 train 7.387613 (lr=4.6731e-05) (hash(x)=6781160)
716
+ 686 train 7.432035 (lr=4.6537e-05) (hash(x)=5824767)
717
+ 687 train 7.409390 (lr=4.6342e-05) (hash(x)=4512723)
718
+ 688 train 7.355200 (lr=4.6148e-05) (hash(x)=4600884)
719
+ 689 train 7.261007 (lr=4.5954e-05) (hash(x)=5079210)
720
+ 690 train 7.317743 (lr=4.5760e-05) (hash(x)=4671963)
721
+ 691 train 7.192114 (lr=4.5567e-05) (hash(x)=4169508)
722
+ 692 train 7.483804 (lr=4.5373e-05) (hash(x)=5341676)
723
+ 693 train 7.260602 (lr=4.5180e-05) (hash(x)=7593391)
724
+ 694 train 7.373847 (lr=4.4987e-05) (hash(x)=6723106)
725
+ 695 train 7.481277 (lr=4.4794e-05) (hash(x)=5136994)
726
+ 696 train 7.434025 (lr=4.4601e-05) (hash(x)=6571636)
727
+ 697 train 7.446311 (lr=4.4408e-05) (hash(x)=6975234)
728
+ 698 train 7.259545 (lr=4.4216e-05) (hash(x)=4868479)
729
+ 699 train 7.585925 (lr=4.4024e-05) (hash(x)=7086593)
730
+ 700 val loss 7.3623
731
+ 700 val perplexity 1575.4070
732
+ 700 train 7.430545 (lr=4.3832e-05) (hash(x)=6870808)
733
+ 701 train 7.728683 (lr=4.3640e-05) (hash(x)=7186398)
734
+ 702 train 7.504910 (lr=4.3449e-05) (hash(x)=4907570)
735
+ 703 train 7.331970 (lr=4.3258e-05) (hash(x)=5812649)
736
+ 704 train 7.390525 (lr=4.3067e-05) (hash(x)=6577447)
737
+ 705 train 7.342052 (lr=4.2876e-05) (hash(x)=5923453)
738
+ 706 train 7.608717 (lr=4.2685e-05) (hash(x)=7282962)
739
+ 707 train 7.270675 (lr=4.2495e-05) (hash(x)=4767725)
740
+ 708 train 7.341816 (lr=4.2305e-05) (hash(x)=6894202)
741
+ 709 train 7.866931 (lr=4.2115e-05) (hash(x)=4904377)
742
+ 710 train 7.335125 (lr=4.1925e-05) (hash(x)=5810443)
743
+ 711 train 7.435909 (lr=4.1736e-05) (hash(x)=5193944)
744
+ 712 train 7.337388 (lr=4.1547e-05) (hash(x)=5188803)
745
+ 713 train 7.315445 (lr=4.1358e-05) (hash(x)=7235549)
746
+ 714 train 7.525253 (lr=4.1170e-05) (hash(x)=5692576)
747
+ 715 train 7.438269 (lr=4.0981e-05) (hash(x)=5103770)
748
+ 716 train 7.374381 (lr=4.0793e-05) (hash(x)=6084786)
749
+ 717 train 7.302500 (lr=4.0606e-05) (hash(x)=5949126)
750
+ 718 train 7.388938 (lr=4.0418e-05) (hash(x)=6321938)
751
+ 719 train 7.422428 (lr=4.0231e-05) (hash(x)=6227413)
752
+ 720 train 7.485180 (lr=4.0044e-05) (hash(x)=6795372)
753
+ 721 train 7.345835 (lr=3.9857e-05) (hash(x)=6506400)
754
+ 722 train 7.360313 (lr=3.9671e-05) (hash(x)=4595742)
755
+ 723 train 7.489711 (lr=3.9485e-05) (hash(x)=7342185)
756
+ 724 train 7.517817 (lr=3.9299e-05) (hash(x)=5762071)
757
+ 725 train 7.338600 (lr=3.9114e-05) (hash(x)=6337670)
758
+ 726 train 7.381857 (lr=3.8929e-05) (hash(x)=6831170)
759
+ 727 train 7.444521 (lr=3.8744e-05) (hash(x)=7080870)
760
+ 728 train 7.362189 (lr=3.8560e-05) (hash(x)=6449970)
761
+ 729 train 7.528371 (lr=3.8376e-05) (hash(x)=5588514)
762
+ 730 train 7.279354 (lr=3.8192e-05) (hash(x)=5379746)
763
+ 731 train 7.474344 (lr=3.8008e-05) (hash(x)=8049901)
764
+ 732 train 7.552766 (lr=3.7825e-05) (hash(x)=8398209)
765
+ 733 train 7.429967 (lr=3.7642e-05) (hash(x)=6152707)
766
+ 734 train 7.505499 (lr=3.7460e-05) (hash(x)=7682617)
767
+ 735 train 7.376218 (lr=3.7277e-05) (hash(x)=7053352)
768
+ 736 train 7.445709 (lr=3.7096e-05) (hash(x)=8491505)
769
+ 737 train 7.420139 (lr=3.6914e-05) (hash(x)=6102853)
770
+ 738 train 7.496243 (lr=3.6733e-05) (hash(x)=4166591)
771
+ 739 train 7.380925 (lr=3.6552e-05) (hash(x)=7501152)
772
+ 740 train 7.316469 (lr=3.6372e-05) (hash(x)=8231949)
773
+ 741 train 7.378488 (lr=3.6192e-05) (hash(x)=5741690)
774
+ 742 train 7.511323 (lr=3.6012e-05) (hash(x)=5420426)
775
+ 743 train 7.319014 (lr=3.5833e-05) (hash(x)=5808766)
776
+ 744 train 7.282396 (lr=3.5654e-05) (hash(x)=7533322)
777
+ 745 train 7.320943 (lr=3.5475e-05) (hash(x)=6350197)
778
+ 746 train 7.465116 (lr=3.5297e-05) (hash(x)=6207170)
779
+ 747 train 7.442750 (lr=3.5119e-05) (hash(x)=5811552)
780
+ 748 train 7.352421 (lr=3.4942e-05) (hash(x)=4540607)
781
+ 749 train 7.203982 (lr=3.4765e-05) (hash(x)=5599396)
782
+ 750 val loss 7.3458
783
+ 750 val perplexity 1549.6556
784
+ 750 train 7.210698 (lr=3.4588e-05) (hash(x)=6180984)
785
+ 751 train 7.367527 (lr=3.4412e-05) (hash(x)=6519152)
786
+ 752 train 7.377176 (lr=3.4236e-05) (hash(x)=9018585)
787
+ 753 train 7.402687 (lr=3.4060e-05) (hash(x)=5988202)
788
+ 754 train 7.459009 (lr=3.3885e-05) (hash(x)=7243449)
789
+ 755 train 7.245204 (lr=3.3711e-05) (hash(x)=4377360)
790
+ 756 train 7.328814 (lr=3.3537e-05) (hash(x)=5928295)
791
+ 757 train 7.258076 (lr=3.3363e-05) (hash(x)=4996189)
792
+ 758 train 7.302147 (lr=3.3189e-05) (hash(x)=6829435)
793
+ 759 train 7.388505 (lr=3.3016e-05) (hash(x)=6929283)
794
+ 760 train 7.495571 (lr=3.2844e-05) (hash(x)=7733662)
795
+ 761 train 7.410961 (lr=3.2672e-05) (hash(x)=5975918)
796
+ 762 train 7.295754 (lr=3.2500e-05) (hash(x)=5566111)
797
+ 763 train 7.382801 (lr=3.2329e-05) (hash(x)=7407511)
798
+ 764 train 7.301639 (lr=3.2158e-05) (hash(x)=5896879)
799
+ 765 train 7.242899 (lr=3.1988e-05) (hash(x)=5569196)
800
+ 766 train 7.482467 (lr=3.1818e-05) (hash(x)=6235218)
801
+ 767 train 7.396662 (lr=3.1648e-05) (hash(x)=7467630)
802
+ 768 train 7.385614 (lr=3.1479e-05) (hash(x)=5268563)
803
+ 769 train 7.591837 (lr=3.1311e-05) (hash(x)=5915215)
804
+ 770 train 7.395827 (lr=3.1142e-05) (hash(x)=5439067)
805
+ 771 train 7.347938 (lr=3.0975e-05) (hash(x)=6167203)
806
+ 772 train 7.407609 (lr=3.0808e-05) (hash(x)=4940526)
807
+ 773 train 7.457906 (lr=3.0641e-05) (hash(x)=5897107)
808
+ 774 train 7.306570 (lr=3.0475e-05) (hash(x)=6780557)
809
+ 775 train 7.439525 (lr=3.0309e-05) (hash(x)=7566733)
810
+ 776 train 7.558398 (lr=3.0144e-05) (hash(x)=7939982)
811
+ 777 train 7.270278 (lr=2.9979e-05) (hash(x)=6447254)
812
+ 778 train 7.272952 (lr=2.9814e-05) (hash(x)=5525831)
813
+ 779 train 7.213912 (lr=2.9651e-05) (hash(x)=5059170)
814
+ 780 train 7.310160 (lr=2.9487e-05) (hash(x)=7485229)
815
+ 781 train 7.452171 (lr=2.9324e-05) (hash(x)=5955219)
816
+ 782 train 7.290710 (lr=2.9162e-05) (hash(x)=6625876)
817
+ 783 train 7.624340 (lr=2.9000e-05) (hash(x)=7141085)
818
+ 784 train 7.535786 (lr=2.8839e-05) (hash(x)=5777157)
819
+ 785 train 7.242429 (lr=2.8678e-05) (hash(x)=6236468)
820
+ 786 train 7.092711 (lr=2.8518e-05) (hash(x)=5381032)
821
+ 787 train 7.265240 (lr=2.8358e-05) (hash(x)=6733716)
822
+ 788 train 7.542943 (lr=2.8199e-05) (hash(x)=7073122)
823
+ 789 train 7.500874 (lr=2.8040e-05) (hash(x)=7622163)
824
+ 790 train 7.573501 (lr=2.7881e-05) (hash(x)=6866834)
825
+ 791 train 7.491663 (lr=2.7724e-05) (hash(x)=7008739)
826
+ 792 train 7.409664 (lr=2.7566e-05) (hash(x)=7839835)
827
+ 793 train 7.273984 (lr=2.7410e-05) (hash(x)=5313637)
828
+ 794 train 7.476710 (lr=2.7254e-05) (hash(x)=6410359)
829
+ 795 train 7.286410 (lr=2.7098e-05) (hash(x)=5721008)
830
+ 796 train 7.315864 (lr=2.6943e-05) (hash(x)=6137936)
831
+ 797 train 7.154595 (lr=2.6788e-05) (hash(x)=6881406)
832
+ 798 train 7.387875 (lr=2.6634e-05) (hash(x)=6088931)
833
+ 799 train 7.411271 (lr=2.6481e-05) (hash(x)=5689728)
834
+ 800 val loss 7.3276
835
+ 800 val perplexity 1521.7565
836
+ 800 train 7.459803 (lr=2.6328e-05) (hash(x)=5800778)
837
+ 801 train 7.349700 (lr=2.6176e-05) (hash(x)=4419021)
838
+ 802 train 7.301584 (lr=2.6024e-05) (hash(x)=5544294)
839
+ 803 train 7.359852 (lr=2.5873e-05) (hash(x)=5663946)
840
+ 804 train 7.294063 (lr=2.5722e-05) (hash(x)=4573911)
841
+ 805 train 7.211605 (lr=2.5572e-05) (hash(x)=5879465)
842
+ 806 train 7.334810 (lr=2.5423e-05) (hash(x)=5641517)
843
+ 807 train 7.510048 (lr=2.5274e-05) (hash(x)=6390336)
844
+ 808 train 7.365402 (lr=2.5125e-05) (hash(x)=8412819)
845
+ 809 train 7.505763 (lr=2.4978e-05) (hash(x)=7005536)
846
+ 810 train 7.221760 (lr=2.4830e-05) (hash(x)=6630048)
847
+ 811 train 7.355828 (lr=2.4684e-05) (hash(x)=7267424)
848
+ 812 train 7.239872 (lr=2.4538e-05) (hash(x)=6497645)
849
+ 813 train 7.210892 (lr=2.4392e-05) (hash(x)=5498697)
850
+ 814 train 7.305768 (lr=2.4247e-05) (hash(x)=6323524)
851
+ 815 train 7.237651 (lr=2.4103e-05) (hash(x)=6529460)
852
+ 816 train 7.133391 (lr=2.3959e-05) (hash(x)=5498587)
853
+ 817 train 7.280392 (lr=2.3816e-05) (hash(x)=6828893)
854
+ 818 train 7.439272 (lr=2.3674e-05) (hash(x)=6591055)
855
+ 819 train 7.301411 (lr=2.3532e-05) (hash(x)=6739572)
856
+ 820 train 7.295975 (lr=2.3391e-05) (hash(x)=5614231)
857
+ 821 train 7.331375 (lr=2.3250e-05) (hash(x)=5462474)
858
+ 822 train 7.317928 (lr=2.3110e-05) (hash(x)=5906742)
859
+ 823 train 7.425270 (lr=2.2971e-05) (hash(x)=8249268)
860
+ 824 train 7.417471 (lr=2.2832e-05) (hash(x)=4886590)
861
+ 825 train 7.155892 (lr=2.2694e-05) (hash(x)=6736150)
862
+ 826 train 7.286769 (lr=2.2556e-05) (hash(x)=6601014)
863
+ 827 train 7.350247 (lr=2.2420e-05) (hash(x)=8283020)
864
+ 828 train 7.302823 (lr=2.2283e-05) (hash(x)=6526568)
865
+ 829 train 7.378872 (lr=2.2148e-05) (hash(x)=6609259)
866
+ 830 train 7.274437 (lr=2.2013e-05) (hash(x)=6079648)
867
+ 831 train 7.241787 (lr=2.1878e-05) (hash(x)=6607178)
868
+ 832 train 7.336267 (lr=2.1745e-05) (hash(x)=6815798)
869
+ 833 train 7.609237 (lr=2.1612e-05) (hash(x)=7277017)
870
+ 834 train 7.242269 (lr=2.1479e-05) (hash(x)=5326415)
871
+ 835 train 7.344043 (lr=2.1347e-05) (hash(x)=4215944)
872
+ 836 train 7.413520 (lr=2.1216e-05) (hash(x)=4689010)
873
+ 837 train 7.343024 (lr=2.1086e-05) (hash(x)=5690469)
874
+ 838 train 7.264034 (lr=2.0956e-05) (hash(x)=6699212)
875
+ 839 train 7.197639 (lr=2.0827e-05) (hash(x)=5494876)
876
+ 840 train 7.662753 (lr=2.0698e-05) (hash(x)=10756054)
877
+ 841 train 7.272571 (lr=2.0570e-05) (hash(x)=6553219)
878
+ 842 train 7.125119 (lr=2.0443e-05) (hash(x)=7323893)
879
+ 843 train 7.629426 (lr=2.0317e-05) (hash(x)=5481970)
880
+ 844 train 7.380283 (lr=2.0191e-05) (hash(x)=6102855)
881
+ 845 train 7.357008 (lr=2.0066e-05) (hash(x)=6889406)
882
+ 846 train 7.395590 (lr=1.9941e-05) (hash(x)=4468013)
883
+ 847 train 7.191195 (lr=1.9818e-05) (hash(x)=6621321)
884
+ 848 train 7.358621 (lr=1.9694e-05) (hash(x)=6185717)
885
+ 849 train 7.233253 (lr=1.9572e-05) (hash(x)=5594828)
886
+ 850 val loss 7.3202
887
+ 850 val perplexity 1510.5052
888
+ 850 train 7.365483 (lr=1.9450e-05) (hash(x)=6530853)
889
+ 851 train 7.265092 (lr=1.9329e-05) (hash(x)=5688551)
890
+ 852 train 7.418205 (lr=1.9209e-05) (hash(x)=6979027)
891
+ 853 train 7.272730 (lr=1.9089e-05) (hash(x)=6065040)
892
+ 854 train 7.218517 (lr=1.8970e-05) (hash(x)=6836310)
893
+ 855 train 7.317423 (lr=1.8852e-05) (hash(x)=4871742)
894
+ 856 train 7.217106 (lr=1.8734e-05) (hash(x)=7461300)
895
+ 857 train 7.381676 (lr=1.8618e-05) (hash(x)=5073610)
896
+ 858 train 7.302594 (lr=1.8501e-05) (hash(x)=6725326)
897
+ 859 train 7.416997 (lr=1.8386e-05) (hash(x)=6114900)
898
+ 860 train 7.383035 (lr=1.8271e-05) (hash(x)=5249155)
899
+ 861 train 7.412766 (lr=1.8157e-05) (hash(x)=8179785)
900
+ 862 train 7.294283 (lr=1.8044e-05) (hash(x)=5588627)
901
+ 863 train 7.333791 (lr=1.7931e-05) (hash(x)=7084451)
902
+ 864 train 7.532066 (lr=1.7819e-05) (hash(x)=7478960)
903
+ 865 train 7.370672 (lr=1.7708e-05) (hash(x)=5609947)
904
+ 866 train 7.355056 (lr=1.7598e-05) (hash(x)=5741827)
905
+ 867 train 7.309427 (lr=1.7488e-05) (hash(x)=3359718)
906
+ 868 train 7.191789 (lr=1.7379e-05) (hash(x)=5480813)
907
+ 869 train 7.296831 (lr=1.7271e-05) (hash(x)=4774534)
908
+ 870 train 7.398872 (lr=1.7163e-05) (hash(x)=5895917)
909
+ 871 train 7.248298 (lr=1.7056e-05) (hash(x)=4902539)
910
+ 872 train 7.325887 (lr=1.6950e-05) (hash(x)=5734618)
911
+ 873 train 7.266691 (lr=1.6845e-05) (hash(x)=7086920)
912
+ 874 train 7.534166 (lr=1.6740e-05) (hash(x)=5338668)
913
+ 875 train 7.686028 (lr=1.6636e-05) (hash(x)=5238889)
914
+ 876 train 7.203884 (lr=1.6533e-05) (hash(x)=6244468)
915
+ 877 train 7.255644 (lr=1.6431e-05) (hash(x)=6715860)
916
+ 878 train 7.380940 (lr=1.6329e-05) (hash(x)=6216047)
917
+ 879 train 7.389113 (lr=1.6228e-05) (hash(x)=6565285)
918
+ 880 train 7.317013 (lr=1.6128e-05) (hash(x)=5257335)
919
+ 881 train 7.318216 (lr=1.6029e-05) (hash(x)=6474071)
920
+ 882 train 7.343215 (lr=1.5930e-05) (hash(x)=5705620)
921
+ 883 train 7.220967 (lr=1.5832e-05) (hash(x)=7193697)
922
+ 884 train 7.324241 (lr=1.5735e-05) (hash(x)=6264039)
923
+ 885 train 7.374564 (lr=1.5639e-05) (hash(x)=6303643)
924
+ 886 train 7.279689 (lr=1.5543e-05) (hash(x)=6120402)
925
+ 887 train 7.279310 (lr=1.5449e-05) (hash(x)=4158764)
926
+ 888 train 7.036485 (lr=1.5354e-05) (hash(x)=6173916)
927
+ 889 train 7.299476 (lr=1.5261e-05) (hash(x)=6211683)
928
+ 890 train 7.344223 (lr=1.5169e-05) (hash(x)=4950548)
929
+ 891 train 7.158161 (lr=1.5077e-05) (hash(x)=4885446)
930
+ 892 train 7.319761 (lr=1.4986e-05) (hash(x)=8117411)
931
+ 893 train 7.341642 (lr=1.4896e-05) (hash(x)=7446434)
932
+ 894 train 7.230519 (lr=1.4806e-05) (hash(x)=5371870)
933
+ 895 train 7.182268 (lr=1.4718e-05) (hash(x)=5022800)
934
+ 896 train 7.187124 (lr=1.4630e-05) (hash(x)=5666758)
935
+ 897 train 7.216234 (lr=1.4543e-05) (hash(x)=5872751)
936
+ 898 train 7.668936 (lr=1.4456e-05) (hash(x)=8428026)
937
+ 899 train 7.455252 (lr=1.4371e-05) (hash(x)=5222699)
938
+ 900 val loss 7.3157
939
+ 900 val perplexity 1503.7671
940
+ 900 train 7.301188 (lr=1.4286e-05) (hash(x)=6048200)
941
+ 901 train 7.168407 (lr=1.4202e-05) (hash(x)=5125556)
942
+ 902 train 7.172472 (lr=1.4119e-05) (hash(x)=6198583)
943
+ 903 train 7.150073 (lr=1.4037e-05) (hash(x)=5112805)
944
+ 904 train 7.235983 (lr=1.3955e-05) (hash(x)=6212803)
945
+ 905 train 7.269233 (lr=1.3874e-05) (hash(x)=5642978)
946
+ 906 train 7.288847 (lr=1.3794e-05) (hash(x)=7307168)
947
+ 907 train 7.215481 (lr=1.3715e-05) (hash(x)=6115590)
948
+ 908 train 7.080295 (lr=1.3637e-05) (hash(x)=7014937)
949
+ 909 train 7.507735 (lr=1.3559e-05) (hash(x)=6374852)
950
+ 910 train 7.216549 (lr=1.3482e-05) (hash(x)=5780051)
951
+ 911 train 7.208725 (lr=1.3407e-05) (hash(x)=4762818)
952
+ 912 train 7.090237 (lr=1.3331e-05) (hash(x)=6585420)
953
+ 913 train 7.413379 (lr=1.3257e-05) (hash(x)=6291590)
954
+ 914 train 7.482652 (lr=1.3183e-05) (hash(x)=5941561)
955
+ 915 train 7.252762 (lr=1.3111e-05) (hash(x)=6824617)
956
+ 916 train 7.357590 (lr=1.3039e-05) (hash(x)=5763341)
957
+ 917 train 7.373316 (lr=1.2968e-05) (hash(x)=6030150)
958
+ 918 train 7.289693 (lr=1.2897e-05) (hash(x)=5325490)
959
+ 919 train 7.311139 (lr=1.2828e-05) (hash(x)=6373236)
960
+ 920 train 7.329778 (lr=1.2759e-05) (hash(x)=6189151)
961
+ 921 train 7.268406 (lr=1.2691e-05) (hash(x)=6645157)
962
+ 922 train 7.280089 (lr=1.2624e-05) (hash(x)=6478394)
963
+ 923 train 7.437422 (lr=1.2558e-05) (hash(x)=5615918)
964
+ 924 train 7.308346 (lr=1.2493e-05) (hash(x)=6376988)
965
+ 925 train 7.349792 (lr=1.2428e-05) (hash(x)=5456350)
966
+ 926 train 7.395862 (lr=1.2364e-05) (hash(x)=5491217)
967
+ 927 train 7.179298 (lr=1.2301e-05) (hash(x)=6871163)
968
+ 928 train 7.199232 (lr=1.2239e-05) (hash(x)=7003708)
969
+ 929 train 7.337853 (lr=1.2178e-05) (hash(x)=7271154)
970
+ 930 train 7.275816 (lr=1.2118e-05) (hash(x)=5378966)
971
+ 931 train 7.292715 (lr=1.2058e-05) (hash(x)=5912082)
972
+ 932 train 7.171347 (lr=1.1999e-05) (hash(x)=5199606)
973
+ 933 train 7.213825 (lr=1.1941e-05) (hash(x)=5195464)
974
+ 934 train 7.151923 (lr=1.1884e-05) (hash(x)=7534052)
975
+ 935 train 7.161028 (lr=1.1828e-05) (hash(x)=6594135)
976
+ 936 train 7.281318 (lr=1.1772e-05) (hash(x)=6782051)
977
+ 937 train 7.401269 (lr=1.1718e-05) (hash(x)=6485577)
978
+ 938 train 7.316434 (lr=1.1664e-05) (hash(x)=6356813)
979
+ 939 train 7.209404 (lr=1.1611e-05) (hash(x)=6903068)
980
+ 940 train 7.261940 (lr=1.1559e-05) (hash(x)=6200862)
981
+ 941 train 7.300323 (lr=1.1508e-05) (hash(x)=5873929)
982
+ 942 train 7.294487 (lr=1.1457e-05) (hash(x)=5624440)
983
+ 943 train 7.265259 (lr=1.1408e-05) (hash(x)=7166972)
984
+ 944 train 7.420985 (lr=1.1359e-05) (hash(x)=6973234)
985
+ 945 train 7.202220 (lr=1.1311e-05) (hash(x)=6626956)
986
+ 946 train 7.339950 (lr=1.1264e-05) (hash(x)=5265592)
987
+ 947 train 7.035192 (lr=1.1218e-05) (hash(x)=5569439)
988
+ 948 train 7.182633 (lr=1.1173e-05) (hash(x)=4981059)
989
+ 949 train 7.208423 (lr=1.1128e-05) (hash(x)=7488974)
990
+ 950 val loss 7.3021
991
+ 950 val perplexity 1483.4410
992
+ 950 train 7.280080 (lr=1.1085e-05) (hash(x)=5160526)
993
+ 951 train 7.216691 (lr=1.1042e-05) (hash(x)=5340877)
994
+ 952 train 7.074330 (lr=1.1000e-05) (hash(x)=6333269)
995
+ 953 train 7.270759 (lr=1.0959e-05) (hash(x)=7205151)
996
+ 954 train 7.352469 (lr=1.0919e-05) (hash(x)=7256545)
997
+ 955 train 7.207728 (lr=1.0879e-05) (hash(x)=6374904)
998
+ 956 train 7.264159 (lr=1.0841e-05) (hash(x)=7124927)
999
+ 957 train 7.221771 (lr=1.0803e-05) (hash(x)=5263132)
1000
+ 958 train 7.408572 (lr=1.0766e-05) (hash(x)=8427208)
1001
+ 959 train 7.166051 (lr=1.0730e-05) (hash(x)=6884232)
1002
+ 960 train 7.178084 (lr=1.0695e-05) (hash(x)=7416977)
1003
+ 961 train 7.312136 (lr=1.0661e-05) (hash(x)=6733050)
1004
+ 962 train 7.201138 (lr=1.0628e-05) (hash(x)=5251404)
1005
+ 963 train 7.331251 (lr=1.0595e-05) (hash(x)=8162997)
1006
+ 964 train 7.504494 (lr=1.0563e-05) (hash(x)=7234382)
1007
+ 965 train 7.385792 (lr=1.0533e-05) (hash(x)=6255275)
1008
+ 966 train 7.361961 (lr=1.0503e-05) (hash(x)=4874742)
1009
+ 967 train 7.253449 (lr=1.0474e-05) (hash(x)=6303037)
1010
+ 968 train 7.227972 (lr=1.0445e-05) (hash(x)=5002696)
1011
+ 969 train 7.258224 (lr=1.0418e-05) (hash(x)=7636268)
1012
+ 970 train 7.255203 (lr=1.0391e-05) (hash(x)=5843590)
1013
+ 971 train 7.314649 (lr=1.0366e-05) (hash(x)=5122289)
1014
+ 972 train 7.191046 (lr=1.0341e-05) (hash(x)=5785819)
1015
+ 973 train 7.262475 (lr=1.0317e-05) (hash(x)=7350343)
1016
+ 974 train 7.051187 (lr=1.0294e-05) (hash(x)=7612943)
1017
+ 975 train 7.423087 (lr=1.0272e-05) (hash(x)=7955067)
1018
+ 976 train 7.249758 (lr=1.0251e-05) (hash(x)=5520886)
1019
+ 977 train 7.139700 (lr=1.0230e-05) (hash(x)=5299200)
1020
+ 978 train 7.160234 (lr=1.0211e-05) (hash(x)=7820550)
1021
+ 979 train 7.139097 (lr=1.0192e-05) (hash(x)=3895312)
1022
+ 980 train 7.272741 (lr=1.0174e-05) (hash(x)=4726896)
1023
+ 981 train 7.051008 (lr=1.0157e-05) (hash(x)=5061167)
1024
+ 982 train 7.214511 (lr=1.0141e-05) (hash(x)=7006804)
1025
+ 983 train 7.241957 (lr=1.0126e-05) (hash(x)=7392600)
1026
+ 984 train 7.461648 (lr=1.0111e-05) (hash(x)=7717172)
1027
+ 985 train 7.564340 (lr=1.0098e-05) (hash(x)=5800587)
1028
+ 986 train 7.260957 (lr=1.0085e-05) (hash(x)=5813784)
1029
+ 987 train 7.213284 (lr=1.0074e-05) (hash(x)=5513267)
1030
+ 988 train 7.208223 (lr=1.0063e-05) (hash(x)=7679326)
1031
+ 989 train 7.278224 (lr=1.0053e-05) (hash(x)=4820278)
1032
+ 990 train 7.312238 (lr=1.0044e-05) (hash(x)=5950845)
1033
+ 991 train 7.222884 (lr=1.0035e-05) (hash(x)=6988067)
1034
+ 992 train 7.143286 (lr=1.0028e-05) (hash(x)=4536923)
1035
+ 993 train 7.327156 (lr=1.0021e-05) (hash(x)=6874134)
1036
+ 994 train 7.602639 (lr=1.0016e-05) (hash(x)=7252013)
1037
+ 995 train 7.156361 (lr=1.0011e-05) (hash(x)=6180446)
1038
+ 996 train 7.433685 (lr=1.0007e-05) (hash(x)=5936769)
1039
+ 997 train 7.359143 (lr=1.0004e-05) (hash(x)=5627313)
1040
+ 998 train 7.266846 (lr=1.0002e-05) (hash(x)=7541743)
1041
+ 999 val loss 7.2946
1042
+ 999 val perplexity 1472.3033
1043
+ 999 train 7.659437 (lr=1.0000e-05) (hash(x)=9812956)
lr1e-4_total_batch_size40960_seq_len128/model_00999.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d2c64aa6571109f1d158c7fdb91fa4e15d0e277536e0466f6f25fffacfb85e24
3
+ size 36163010
lr1e-4_total_batch_size40960_seq_len128/optimizer_00999.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:25a7752c6b4d8a3a1342311fa6c17a7bc0a5e474f08384d8e488156f11621144
3
+ size 70764358