Upload folder using huggingface_hub

n_heads12_lr14e-4_total_batch_size61440_seed1339/args.json

@@ -0,0 +1 @@
+{"hellaswag": true, "attention_kind": "selective", "log_dir": "wider_is_better_11/n_heads12_lr14e-4_total_batch_size61440_seed1339", "resume_checkpoint": null, "resume_optimizer": false, "add_a_head": false, "add_head_to_start": true, "new_head_init": "normal", "n_heads": 12, "protect_bos_token": true, "prevent_from_masking_myself": true, "max_steps": 8750, "warmup_steps": 500, "group": "wider_is_better_11", "use_wandb": true, "kill_self_after_run": false, "random_seed": 1339, "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": 256, "batch_size": 60, "total_batch_size": 61440, "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": false, "mup_enable_coord_check_logging": false, "max_lr": 0.0014, "decay_lr": true, "readout_zero_init": false, "query_zero_init": false, "l1_loss": false, "debugpy": false, "key": "n_heads12_14e-4_61440", "n_embd": 768}

n_heads12_lr14e-4_total_batch_size61440_seed1339/dataloader_08749.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9bbf5b7eef643b850a13a0d3881f9b995af25e67f722e39e1853dc09a06ad1dc
+size 964

n_heads12_lr14e-4_total_batch_size61440_seed1339/log2.txt

@@ -0,0 +1,1054 @@
+max_steps: 8750
+0 val loss 10.9491
+0 val perplexity 56902.2969
+0 train 10.947037 (lr=2.8000e-06) (hash(x)=84111305)
+10 train 9.696257 (lr=3.0800e-05) (hash(x)=76061209)
+20 train 9.286982 (lr=5.8800e-05) (hash(x)=72815142)
+30 train 8.850385 (lr=8.6800e-05) (hash(x)=74281158)
+40 train 8.175850 (lr=1.1480e-04) (hash(x)=70759787)
+50 train 7.791744 (lr=1.4280e-04) (hash(x)=67498044)
+60 train 7.606891 (lr=1.7080e-04) (hash(x)=74288138)
+70 train 7.538900 (lr=1.9880e-04) (hash(x)=72712110)
+80 train 7.527149 (lr=2.2680e-04) (hash(x)=76893978)
+90 train 7.290340 (lr=2.5480e-04) (hash(x)=69479499)
+100 val loss 7.2023
+100 val perplexity 1342.5669
+100 train 7.169267 (lr=2.8280e-04) (hash(x)=78805238)
+110 train 7.054847 (lr=3.1080e-04) (hash(x)=72379905)
+120 train 6.960956 (lr=3.3880e-04) (hash(x)=84941914)
+130 train 6.761159 (lr=3.6680e-04) (hash(x)=74520609)
+140 train 6.816748 (lr=3.9480e-04) (hash(x)=78600114)
+150 train 6.496477 (lr=4.2280e-04) (hash(x)=71364632)
+160 train 6.498137 (lr=4.5080e-04) (hash(x)=88692902)
+170 train 6.547095 (lr=4.7880e-04) (hash(x)=76813970)
+180 train 6.405209 (lr=5.0680e-04) (hash(x)=78277546)
+190 train 6.384841 (lr=5.3480e-04) (hash(x)=69274625)
+200 val loss 6.3754
+200 val perplexity 587.2024
+200 train 6.300391 (lr=5.6280e-04) (hash(x)=74752080)
+210 train 6.195320 (lr=5.9080e-04) (hash(x)=73366585)
+220 train 6.166953 (lr=6.1880e-04) (hash(x)=82441512)
+230 train 5.945963 (lr=6.4680e-04) (hash(x)=70370492)
+240 train 6.007533 (lr=6.7480e-04) (hash(x)=76271528)
+250 train 5.924657 (lr=7.0280e-04) (hash(x)=70391215)
+260 train 5.843722 (lr=7.3080e-04) (hash(x)=70932307)
+270 train 5.785228 (lr=7.5880e-04) (hash(x)=73903274)
+280 train 5.817907 (lr=7.8680e-04) (hash(x)=72833455)
+290 train 5.877813 (lr=8.1480e-04) (hash(x)=86446336)
+300 val loss 5.9359
+300 val perplexity 378.3824
+300 train 5.664558 (lr=8.4280e-04) (hash(x)=72231879)
+310 train 5.775199 (lr=8.7080e-04) (hash(x)=83289406)
+320 train 5.853682 (lr=8.9880e-04) (hash(x)=79292084)
+330 train 6.013921 (lr=9.2680e-04) (hash(x)=81963141)
+340 train 5.893492 (lr=9.5480e-04) (hash(x)=76688193)
+350 train 5.684438 (lr=9.8280e-04) (hash(x)=67974208)
+360 train 5.863403 (lr=1.0108e-03) (hash(x)=71759924)
+370 train 5.766744 (lr=1.0388e-03) (hash(x)=82538335)
+380 train 5.639980 (lr=1.0668e-03) (hash(x)=73719341)
+390 train 5.784976 (lr=1.0948e-03) (hash(x)=81731118)
+400 val loss 5.6350
+400 val perplexity 280.0556
+400 train 5.590763 (lr=1.1228e-03) (hash(x)=65538392)
+410 train 5.556056 (lr=1.1508e-03) (hash(x)=73037722)
+420 train 5.471239 (lr=1.1788e-03) (hash(x)=71698939)
+430 train 5.673790 (lr=1.2068e-03) (hash(x)=89649867)
+440 train 5.715183 (lr=1.2348e-03) (hash(x)=59262658)
+450 train 5.401849 (lr=1.2628e-03) (hash(x)=79057900)
+460 train 5.423159 (lr=1.2908e-03) (hash(x)=68890702)
+470 train 5.472520 (lr=1.3188e-03) (hash(x)=69474321)
+480 train 5.378677 (lr=1.3468e-03) (hash(x)=74715332)
+490 train 5.496259 (lr=1.3748e-03) (hash(x)=81929495)
+500 val loss 5.3864
+500 val perplexity 218.4182
+500 train 5.544104 (lr=1.4000e-03) (hash(x)=81405581)
+510 train 5.144229 (lr=1.4000e-03) (hash(x)=66596183)
+520 train 5.474530 (lr=1.4000e-03) (hash(x)=75298362)
+530 train 5.280549 (lr=1.4000e-03) (hash(x)=78449774)
+540 train 5.258121 (lr=1.3999e-03) (hash(x)=75749875)
+550 train 5.195624 (lr=1.3999e-03) (hash(x)=75371538)
+560 train 5.131622 (lr=1.3998e-03) (hash(x)=74708150)
+570 train 5.160957 (lr=1.3998e-03) (hash(x)=75984401)
+580 train 5.117284 (lr=1.3997e-03) (hash(x)=78930428)
+590 train 5.429266 (lr=1.3996e-03) (hash(x)=70858901)
+600 val loss 5.2380
+600 val perplexity 188.2875
+600 train 5.123903 (lr=1.3995e-03) (hash(x)=86811215)
+610 train 4.987842 (lr=1.3994e-03) (hash(x)=74537543)
+620 train 4.972369 (lr=1.3993e-03) (hash(x)=64425159)
+630 train 4.849003 (lr=1.3992e-03) (hash(x)=72575441)
+640 train 4.869482 (lr=1.3991e-03) (hash(x)=77886921)
+650 train 4.961205 (lr=1.3990e-03) (hash(x)=73649304)
+660 train 4.728675 (lr=1.3988e-03) (hash(x)=64469482)
+670 train 5.117930 (lr=1.3987e-03) (hash(x)=76580403)
+680 train 4.771859 (lr=1.3985e-03) (hash(x)=60374603)
+690 train 4.903076 (lr=1.3984e-03) (hash(x)=69827901)
+700 val loss 5.0885
+700 val perplexity 162.1519
+700 train 5.101073 (lr=1.3982e-03) (hash(x)=77235911)
+710 train 5.045017 (lr=1.3980e-03) (hash(x)=76598106)
+720 train 5.127827 (lr=1.3978e-03) (hash(x)=75315013)
+730 train 4.964180 (lr=1.3976e-03) (hash(x)=76940932)
+740 train 4.983181 (lr=1.3974e-03) (hash(x)=75540179)
+750 train 4.964045 (lr=1.3971e-03) (hash(x)=73283995)
+760 train 4.974654 (lr=1.3969e-03) (hash(x)=84096065)
+770 train 4.758175 (lr=1.3967e-03) (hash(x)=75832016)
+780 train 4.905328 (lr=1.3964e-03) (hash(x)=65611269)
+790 train 4.966614 (lr=1.3962e-03) (hash(x)=72352412)
+800 val loss 4.8796
+800 val perplexity 131.5824
+800 train 4.896714 (lr=1.3959e-03) (hash(x)=78191559)
+810 train 4.814936 (lr=1.3956e-03) (hash(x)=73255448)
+820 train 4.891327 (lr=1.3953e-03) (hash(x)=75336936)
+830 train 4.696508 (lr=1.3950e-03) (hash(x)=71181581)
+840 train 5.346991 (lr=1.3947e-03) (hash(x)=76217477)
+850 train 4.796478 (lr=1.3944e-03) (hash(x)=75090324)
+860 train 4.730086 (lr=1.3941e-03) (hash(x)=71125083)
+870 train 4.710831 (lr=1.3938e-03) (hash(x)=74444188)
+880 train 4.747270 (lr=1.3934e-03) (hash(x)=72143748)
+890 train 4.620478 (lr=1.3931e-03) (hash(x)=75638384)
+900 val loss 4.7292
+900 val perplexity 113.2096
+900 train 4.752657 (lr=1.3927e-03) (hash(x)=76967190)
+910 train 4.706180 (lr=1.3923e-03) (hash(x)=84611316)
+920 train 4.521179 (lr=1.3920e-03) (hash(x)=61365145)
+930 train 4.492158 (lr=1.3916e-03) (hash(x)=54695617)
+940 train 4.693443 (lr=1.3912e-03) (hash(x)=76251415)
+950 train 4.461804 (lr=1.3908e-03) (hash(x)=71503060)
+960 train 4.685982 (lr=1.3904e-03) (hash(x)=70925251)
+970 train 4.618879 (lr=1.3899e-03) (hash(x)=69495443)
+980 train 4.602659 (lr=1.3895e-03) (hash(x)=74994053)
+990 train 4.420630 (lr=1.3891e-03) (hash(x)=81245808)
+1000 val loss 4.6759
+1000 val perplexity 107.3307
+1000 train 4.652740 (lr=1.3886e-03) (hash(x)=73801537)
+1010 train 4.571343 (lr=1.3882e-03) (hash(x)=78784929)
+1020 train 4.339276 (lr=1.3877e-03) (hash(x)=84649514)
+1030 train 4.479644 (lr=1.3872e-03) (hash(x)=71200051)
+1040 train 4.348845 (lr=1.3867e-03) (hash(x)=68279861)
+1050 train 4.387001 (lr=1.3862e-03) (hash(x)=86748930)
+1060 train 4.276002 (lr=1.3857e-03) (hash(x)=71319577)
+1070 train 4.402288 (lr=1.3852e-03) (hash(x)=72682588)
+1080 train 4.379171 (lr=1.3847e-03) (hash(x)=71408336)
+1090 train 4.559654 (lr=1.3842e-03) (hash(x)=78863558)
+1100 val loss 4.5938
+1100 val perplexity 98.8741
+1100 train 4.828663 (lr=1.3836e-03) (hash(x)=75285186)
+1110 train 4.651353 (lr=1.3831e-03) (hash(x)=78202715)
+1120 train 4.439210 (lr=1.3825e-03) (hash(x)=75187925)
+1130 train 4.638708 (lr=1.3820e-03) (hash(x)=70507475)
+1140 train 4.535095 (lr=1.3814e-03) (hash(x)=77911876)
+1150 train 4.460696 (lr=1.3808e-03) (hash(x)=77470391)
+1160 train 4.473397 (lr=1.3802e-03) (hash(x)=76266017)
+1170 train 4.745687 (lr=1.3796e-03) (hash(x)=77505346)
+1180 train 4.540390 (lr=1.3790e-03) (hash(x)=78978095)
+1190 train 4.587722 (lr=1.3784e-03) (hash(x)=78602083)
+1200 val loss 4.5227
+1200 val perplexity 92.0864
+1200 train 4.625494 (lr=1.3778e-03) (hash(x)=81700834)
+1210 train 4.511156 (lr=1.3771e-03) (hash(x)=75776079)
+1220 train 4.497555 (lr=1.3765e-03) (hash(x)=95908213)
+1230 train 4.430666 (lr=1.3758e-03) (hash(x)=84269607)
+1240 train 4.526138 (lr=1.3752e-03) (hash(x)=69460140)
+1250 train 4.536676 (lr=1.3745e-03) (hash(x)=71522595)
+1260 train 4.556998 (lr=1.3738e-03) (hash(x)=81155508)
+1270 train 4.694812 (lr=1.3731e-03) (hash(x)=72842244)
+1280 train 4.537966 (lr=1.3724e-03) (hash(x)=80034345)
+1290 train 4.386315 (lr=1.3717e-03) (hash(x)=70948511)
+1300 val loss 4.4864
+1300 val perplexity 88.8050
+1300 train 4.439077 (lr=1.3710e-03) (hash(x)=69667983)
+1310 train 4.479475 (lr=1.3703e-03) (hash(x)=84077906)
+1320 train 4.206156 (lr=1.3695e-03) (hash(x)=67481757)
+1330 train 4.367138 (lr=1.3688e-03) (hash(x)=83761654)
+1340 train 4.473265 (lr=1.3680e-03) (hash(x)=76630114)
+1350 train 4.507352 (lr=1.3673e-03) (hash(x)=81080698)
+1360 train 4.389709 (lr=1.3665e-03) (hash(x)=71019125)
+1370 train 4.381732 (lr=1.3657e-03) (hash(x)=78550770)
+1380 train 4.396723 (lr=1.3650e-03) (hash(x)=67017249)
+1390 train 4.297318 (lr=1.3642e-03) (hash(x)=75430624)
+1400 val loss 4.4568
+1400 val perplexity 86.2089
+1400 train 4.392143 (lr=1.3634e-03) (hash(x)=72720863)
+1410 train 4.149557 (lr=1.3626e-03) (hash(x)=73824272)
+1420 train 4.276654 (lr=1.3617e-03) (hash(x)=69424571)
+1430 train 4.242927 (lr=1.3609e-03) (hash(x)=71030131)
+1440 train 4.190206 (lr=1.3601e-03) (hash(x)=80088993)
+1450 train 4.135458 (lr=1.3592e-03) (hash(x)=74332002)
+1460 train 4.056621 (lr=1.3584e-03) (hash(x)=75132184)
+1470 train 4.113497 (lr=1.3575e-03) (hash(x)=72644858)
+1480 train 4.125912 (lr=1.3566e-03) (hash(x)=70677454)
+1490 train 4.147674 (lr=1.3558e-03) (hash(x)=74408093)
+1500 val loss 4.4070
+1500 val perplexity 82.0231
+1500 train 4.387667 (lr=1.3549e-03) (hash(x)=82777829)
+1510 train 4.482785 (lr=1.3540e-03) (hash(x)=82481742)
+1520 train 4.400257 (lr=1.3531e-03) (hash(x)=81580164)
+1530 train 4.382079 (lr=1.3522e-03) (hash(x)=75892741)
+1540 train 4.307281 (lr=1.3512e-03) (hash(x)=70574468)
+1550 train 4.350969 (lr=1.3503e-03) (hash(x)=77451919)
+1560 train 4.369387 (lr=1.3494e-03) (hash(x)=73148801)
+1570 train 4.264351 (lr=1.3484e-03) (hash(x)=74661859)
+1580 train 4.416093 (lr=1.3475e-03) (hash(x)=85484059)
+1590 train 4.530247 (lr=1.3465e-03) (hash(x)=72848788)
+1600 val loss 4.3308
+1600 val perplexity 76.0084
+1600 train 4.339680 (lr=1.3455e-03) (hash(x)=79468408)
+1610 train 4.349786 (lr=1.3446e-03) (hash(x)=78654805)
+1620 train 4.389883 (lr=1.3436e-03) (hash(x)=72348939)
+1630 train 4.433557 (lr=1.3426e-03) (hash(x)=74107545)
+1640 train 4.505174 (lr=1.3416e-03) (hash(x)=75470148)
+1650 train 4.398847 (lr=1.3406e-03) (hash(x)=74478093)
+1660 train 4.294458 (lr=1.3395e-03) (hash(x)=74077733)
+1670 train 4.214479 (lr=1.3385e-03) (hash(x)=60922881)
+1680 train 4.457193 (lr=1.3375e-03) (hash(x)=79401992)
+1690 train 4.400383 (lr=1.3364e-03) (hash(x)=75405453)
+1700 val loss 4.3382
+1700 val perplexity 76.5683
+1700 train 4.341522 (lr=1.3354e-03) (hash(x)=67378635)
+1710 train 4.258157 (lr=1.3343e-03) (hash(x)=74327780)
+1720 train 4.438821 (lr=1.3332e-03) (hash(x)=83288144)
+1730 train 4.175040 (lr=1.3321e-03) (hash(x)=68094391)
+1740 train 4.234889 (lr=1.3311e-03) (hash(x)=74943031)
+1750 train 4.278042 (lr=1.3300e-03) (hash(x)=77148592)
+1760 train 4.256920 (lr=1.3289e-03) (hash(x)=71352508)
+1770 train 4.271409 (lr=1.3278e-03) (hash(x)=73819276)
+1780 train 4.167296 (lr=1.3266e-03) (hash(x)=75338495)
+1790 train 4.357256 (lr=1.3255e-03) (hash(x)=100973274)
+1800 val loss 4.3008
+1800 val perplexity 73.7598
+1800 train 4.385872 (lr=1.3244e-03) (hash(x)=82399749)
+1810 train 4.395938 (lr=1.3232e-03) (hash(x)=80859117)
+1820 train 4.436048 (lr=1.3221e-03) (hash(x)=74215317)
+1830 train 4.298370 (lr=1.3209e-03) (hash(x)=73527133)
+1840 train 4.457799 (lr=1.3197e-03) (hash(x)=70049534)
+1850 train 4.183886 (lr=1.3186e-03) (hash(x)=71891328)
+1860 train 4.436985 (lr=1.3174e-03) (hash(x)=76433977)
+1870 train 4.410738 (lr=1.3162e-03) (hash(x)=69103302)
+1880 train 4.422632 (lr=1.3150e-03) (hash(x)=73328951)
+1890 train 4.183946 (lr=1.3138e-03) (hash(x)=77718295)
+1900 val loss 4.2588
+1900 val perplexity 70.7280
+1900 train 4.359399 (lr=1.3126e-03) (hash(x)=74708596)
+1910 train 4.323352 (lr=1.3113e-03) (hash(x)=80918764)
+1920 train 4.254374 (lr=1.3101e-03) (hash(x)=77248658)
+1930 train 4.355618 (lr=1.3089e-03) (hash(x)=73945895)
+1940 train 4.231158 (lr=1.3076e-03) (hash(x)=72252382)
+1950 train 4.277661 (lr=1.3064e-03) (hash(x)=87410072)
+1960 train 4.329490 (lr=1.3051e-03) (hash(x)=79408896)
+1970 train 4.124446 (lr=1.3038e-03) (hash(x)=78848496)
+1980 train 4.195240 (lr=1.3026e-03) (hash(x)=77601406)
+1990 train 4.151528 (lr=1.3013e-03) (hash(x)=71775143)
+2000 val loss 4.2164
+2000 val perplexity 67.7889
+2000 train 4.237857 (lr=1.3000e-03) (hash(x)=73721824)
+2010 train 4.196913 (lr=1.2987e-03) (hash(x)=74757670)
+2020 train 4.242628 (lr=1.2974e-03) (hash(x)=86234926)
+2030 train 4.444455 (lr=1.2961e-03) (hash(x)=69882050)
+2040 train 4.167746 (lr=1.2947e-03) (hash(x)=81004021)
+2050 train 4.073251 (lr=1.2934e-03) (hash(x)=75592124)
+2060 train 3.989408 (lr=1.2921e-03) (hash(x)=74811474)
+2070 train 4.180940 (lr=1.2907e-03) (hash(x)=76209571)
+2080 train 4.127342 (lr=1.2894e-03) (hash(x)=74973104)
+2090 train 4.187265 (lr=1.2880e-03) (hash(x)=74977086)
+2100 val loss 4.2523
+2100 val perplexity 70.2638
+2100 train 4.148001 (lr=1.2866e-03) (hash(x)=79579234)
+2110 train 4.085246 (lr=1.2853e-03) (hash(x)=73442407)
+2120 train 4.268910 (lr=1.2839e-03) (hash(x)=80281222)
+2130 train 4.206561 (lr=1.2825e-03) (hash(x)=83139613)
+2140 train 4.273047 (lr=1.2811e-03) (hash(x)=69936205)
+2150 train 4.309901 (lr=1.2797e-03) (hash(x)=74768225)
+2160 train 4.326044 (lr=1.2783e-03) (hash(x)=67473891)
+2170 train 4.286726 (lr=1.2768e-03) (hash(x)=80351641)
+2180 train 4.286204 (lr=1.2754e-03) (hash(x)=76303033)
+2190 train 4.220602 (lr=1.2740e-03) (hash(x)=73585511)
+2200 val loss 4.1773
+2200 val perplexity 65.1920
+2200 train 4.153622 (lr=1.2725e-03) (hash(x)=73195726)
+2210 train 4.306350 (lr=1.2711e-03) (hash(x)=80614640)
+2220 train 4.064127 (lr=1.2696e-03) (hash(x)=73107823)
+2230 train 4.142232 (lr=1.2682e-03) (hash(x)=78921188)
+2240 train 4.171255 (lr=1.2667e-03) (hash(x)=82527812)
+2250 train 4.083915 (lr=1.2652e-03) (hash(x)=75909041)
+2260 train 4.259199 (lr=1.2637e-03) (hash(x)=75266897)
+2270 train 4.196569 (lr=1.2622e-03) (hash(x)=74444372)
+2280 train 4.062449 (lr=1.2607e-03) (hash(x)=71643352)
+2290 train 4.254333 (lr=1.2592e-03) (hash(x)=73081122)
+2300 val loss 4.1573
+2300 val perplexity 63.8963
+2300 train 4.021166 (lr=1.2577e-03) (hash(x)=73133506)
+2310 train 4.092760 (lr=1.2562e-03) (hash(x)=73796905)
+2320 train 4.074976 (lr=1.2547e-03) (hash(x)=79133958)
+2330 train 4.088642 (lr=1.2531e-03) (hash(x)=79177833)
+2340 train 4.129920 (lr=1.2516e-03) (hash(x)=72790081)
+2350 train 4.104806 (lr=1.2500e-03) (hash(x)=65070947)
+2360 train 4.260871 (lr=1.2485e-03) (hash(x)=73178485)
+2370 train 4.214497 (lr=1.2469e-03) (hash(x)=71904548)
+2380 train 4.039215 (lr=1.2453e-03) (hash(x)=73416656)
+2390 train 4.184001 (lr=1.2438e-03) (hash(x)=71563647)
+2400 val loss 4.1696
+2400 val perplexity 64.6918
+2400 train 4.256041 (lr=1.2422e-03) (hash(x)=82236758)
+2410 train 4.055654 (lr=1.2406e-03) (hash(x)=76229570)
+2420 train 4.273833 (lr=1.2390e-03) (hash(x)=71056939)
+2430 train 4.123327 (lr=1.2374e-03) (hash(x)=76280968)
+2440 train 3.874370 (lr=1.2358e-03) (hash(x)=84199966)
+2450 train 4.297363 (lr=1.2341e-03) (hash(x)=84985998)
+2460 train 4.018781 (lr=1.2325e-03) (hash(x)=94074119)
+2470 train 4.245076 (lr=1.2309e-03) (hash(x)=81246200)
+2480 train 4.190249 (lr=1.2293e-03) (hash(x)=76225235)
+2490 train 4.118330 (lr=1.2276e-03) (hash(x)=76541445)
+2500 val loss 4.1111
+2500 val perplexity 61.0137
+2500 train 4.189018 (lr=1.2260e-03) (hash(x)=77708978)
+2510 train 4.055840 (lr=1.2243e-03) (hash(x)=75319253)
+2520 train 4.090503 (lr=1.2226e-03) (hash(x)=67190186)
+2530 train 4.268934 (lr=1.2210e-03) (hash(x)=82496813)
+2540 train 4.035830 (lr=1.2193e-03) (hash(x)=78100449)
+2550 train 4.101653 (lr=1.2176e-03) (hash(x)=80190873)
+2560 train 4.011867 (lr=1.2159e-03) (hash(x)=77711923)
+2570 train 4.124818 (lr=1.2142e-03) (hash(x)=79825674)
+2580 train 3.981403 (lr=1.2125e-03) (hash(x)=77024199)
+2590 train 4.052171 (lr=1.2108e-03) (hash(x)=72139372)
+2600 val loss 4.1000
+2600 val perplexity 60.3430
+2600 train 4.012805 (lr=1.2091e-03) (hash(x)=71362290)
+2610 train 4.151912 (lr=1.2073e-03) (hash(x)=80679759)
+2620 train 3.979383 (lr=1.2056e-03) (hash(x)=75079422)
+2630 train 4.078722 (lr=1.2039e-03) (hash(x)=70323867)
+2640 train 3.983303 (lr=1.2021e-03) (hash(x)=81444590)
+2650 train 4.013614 (lr=1.2004e-03) (hash(x)=71086644)
+2660 train 3.951719 (lr=1.1986e-03) (hash(x)=73339459)
+2670 train 4.122108 (lr=1.1969e-03) (hash(x)=74666974)
+2680 train 4.158755 (lr=1.1951e-03) (hash(x)=79117538)
+2690 train 4.154376 (lr=1.1933e-03) (hash(x)=65928715)
+2700 val loss 4.0991
+2700 val perplexity 60.2854
+2700 train 4.155877 (lr=1.1916e-03) (hash(x)=79416673)
+2710 train 3.995750 (lr=1.1898e-03) (hash(x)=65110265)
+2720 train 4.216398 (lr=1.1880e-03) (hash(x)=60942352)
+2730 train 4.181613 (lr=1.1862e-03) (hash(x)=68786492)
+2740 train 4.064043 (lr=1.1844e-03) (hash(x)=71480596)
+2750 train 4.032422 (lr=1.1826e-03) (hash(x)=84054551)
+2760 train 3.992828 (lr=1.1807e-03) (hash(x)=94608646)
+2770 train 4.076508 (lr=1.1789e-03) (hash(x)=73708388)
+2780 train 3.999967 (lr=1.1771e-03) (hash(x)=88054328)
+2790 train 4.164315 (lr=1.1753e-03) (hash(x)=88920089)
+2800 val loss 4.0662
+2800 val perplexity 58.3360
+2800 train 4.063021 (lr=1.1734e-03) (hash(x)=80878641)
+2810 train 4.100664 (lr=1.1716e-03) (hash(x)=71310778)
+2820 train 4.092411 (lr=1.1697e-03) (hash(x)=69986577)
+2830 train 4.069861 (lr=1.1679e-03) (hash(x)=78866280)
+2840 train 4.156309 (lr=1.1660e-03) (hash(x)=84135925)
+2850 train 4.160354 (lr=1.1641e-03) (hash(x)=74153539)
+2860 train 4.053133 (lr=1.1623e-03) (hash(x)=74749949)
+2870 train 4.126884 (lr=1.1604e-03) (hash(x)=72237368)
+2880 train 3.974809 (lr=1.1585e-03) (hash(x)=77118077)
+2890 train 4.085336 (lr=1.1566e-03) (hash(x)=73010418)
+2900 val loss 4.0451
+2900 val perplexity 57.1181
+2900 train 3.974620 (lr=1.1547e-03) (hash(x)=70624604)
+2910 train 3.980934 (lr=1.1528e-03) (hash(x)=72862034)
+2920 train 3.907180 (lr=1.1509e-03) (hash(x)=74542687)
+2930 train 4.107670 (lr=1.1490e-03) (hash(x)=71424119)
+2940 train 4.006510 (lr=1.1471e-03) (hash(x)=68581990)
+2950 train 4.025000 (lr=1.1451e-03) (hash(x)=75184497)
+2960 train 4.158933 (lr=1.1432e-03) (hash(x)=79104818)
+2970 train 3.950874 (lr=1.1413e-03) (hash(x)=75872591)
+2980 train 4.023359 (lr=1.1393e-03) (hash(x)=78565892)
+2990 train 3.979622 (lr=1.1374e-03) (hash(x)=77347407)
+3000 val loss 4.0554
+3000 val perplexity 57.7093
+3000 train 3.970662 (lr=1.1354e-03) (hash(x)=73568709)
+3010 train 3.954880 (lr=1.1335e-03) (hash(x)=83954777)
+3020 train 3.950389 (lr=1.1315e-03) (hash(x)=91325052)
+3030 train 4.194590 (lr=1.1295e-03) (hash(x)=73073624)
+3040 train 4.012201 (lr=1.1276e-03) (hash(x)=78263491)
+3050 train 4.064857 (lr=1.1256e-03) (hash(x)=75006046)
+3060 train 3.803539 (lr=1.1236e-03) (hash(x)=77283473)
+3070 train 4.170357 (lr=1.1216e-03) (hash(x)=70498330)
+3080 train 4.018938 (lr=1.1196e-03) (hash(x)=74632023)
+3090 train 4.015738 (lr=1.1176e-03) (hash(x)=71582163)
+3100 val loss 4.0145
+3100 val perplexity 55.3959
+3100 train 4.004262 (lr=1.1156e-03) (hash(x)=76032144)
+3110 train 4.084311 (lr=1.1136e-03) (hash(x)=79935326)
+3120 train 4.106364 (lr=1.1116e-03) (hash(x)=81668940)
+3130 train 4.025532 (lr=1.1096e-03) (hash(x)=81530706)
+3140 train 4.021307 (lr=1.1076e-03) (hash(x)=75958994)
+3150 train 4.107799 (lr=1.1055e-03) (hash(x)=70375358)
+3160 train 4.019825 (lr=1.1035e-03) (hash(x)=81742382)
+3170 train 4.010878 (lr=1.1015e-03) (hash(x)=73892129)
+3180 train 4.021301 (lr=1.0994e-03) (hash(x)=69584996)
+3190 train 4.008098 (lr=1.0974e-03) (hash(x)=74711675)
+3200 val loss 4.0115
+3200 val perplexity 55.2301
+3200 train 4.115763 (lr=1.0953e-03) (hash(x)=72943579)
+3210 train 4.075553 (lr=1.0933e-03) (hash(x)=76110656)
+3220 train 3.852217 (lr=1.0912e-03) (hash(x)=70203094)
+3230 train 4.013988 (lr=1.0891e-03) (hash(x)=60287090)
+3240 train 3.944180 (lr=1.0871e-03) (hash(x)=74302555)
+3250 train 4.044456 (lr=1.0850e-03) (hash(x)=72366702)
+3260 train 3.940692 (lr=1.0829e-03) (hash(x)=73014004)
+3270 train 3.835338 (lr=1.0808e-03) (hash(x)=75491051)
+3280 train 4.059203 (lr=1.0787e-03) (hash(x)=82821096)
+3290 train 3.885968 (lr=1.0767e-03) (hash(x)=77913090)
+3300 val loss 4.0132
+3300 val perplexity 55.3231
+3300 train 4.033732 (lr=1.0746e-03) (hash(x)=72779327)
+3310 train 3.969360 (lr=1.0725e-03) (hash(x)=76032908)
+3320 train 3.841764 (lr=1.0703e-03) (hash(x)=66875476)
+3330 train 3.903355 (lr=1.0682e-03) (hash(x)=67498762)
+3340 train 4.037702 (lr=1.0661e-03) (hash(x)=77920169)
+3350 train 4.033425 (lr=1.0640e-03) (hash(x)=87527839)
+3360 train 3.854577 (lr=1.0619e-03) (hash(x)=78010101)
+3370 train 4.070465 (lr=1.0597e-03) (hash(x)=76688335)
+3380 train 4.061776 (lr=1.0576e-03) (hash(x)=80832602)
+3390 train 3.903405 (lr=1.0555e-03) (hash(x)=75944307)
+3400 val loss 3.9920
+3400 val perplexity 54.1609
+3400 train 4.181395 (lr=1.0533e-03) (hash(x)=67516962)
+3410 train 4.043400 (lr=1.0512e-03) (hash(x)=72960118)
+3420 train 4.063179 (lr=1.0490e-03) (hash(x)=77405757)
+3430 train 4.098376 (lr=1.0469e-03) (hash(x)=76094785)
+3440 train 3.905484 (lr=1.0447e-03) (hash(x)=68336851)
+3450 train 3.896772 (lr=1.0426e-03) (hash(x)=71910579)
+3460 train 3.932227 (lr=1.0404e-03) (hash(x)=74170641)
+3470 train 4.038262 (lr=1.0382e-03) (hash(x)=75423851)
+3480 train 3.989763 (lr=1.0361e-03) (hash(x)=73367119)
+3490 train 3.861373 (lr=1.0339e-03) (hash(x)=75558326)
+3500 val loss 3.9833
+3500 val perplexity 53.6950
+3500 train 3.875478 (lr=1.0317e-03) (hash(x)=69187503)
+3510 train 3.986794 (lr=1.0295e-03) (hash(x)=80052583)
+3520 train 3.971057 (lr=1.0273e-03) (hash(x)=73819848)
+3530 train 3.798421 (lr=1.0251e-03) (hash(x)=71358103)
+3540 train 3.957886 (lr=1.0230e-03) (hash(x)=95551988)
+3550 train 4.015412 (lr=1.0208e-03) (hash(x)=75088121)
+3560 train 4.014828 (lr=1.0186e-03) (hash(x)=80473866)
+3570 train 4.025586 (lr=1.0163e-03) (hash(x)=73228665)
+3580 train 3.921344 (lr=1.0141e-03) (hash(x)=76321895)
+3590 train 3.846807 (lr=1.0119e-03) (hash(x)=74530382)
+3600 val loss 3.9716
+3600 val perplexity 53.0709
+3600 train 3.929642 (lr=1.0097e-03) (hash(x)=69686188)
+3610 train 3.961621 (lr=1.0075e-03) (hash(x)=66842873)
+3620 train 3.950438 (lr=1.0053e-03) (hash(x)=74921650)
+3630 train 3.851181 (lr=1.0030e-03) (hash(x)=79834656)
+3640 train 3.978228 (lr=1.0008e-03) (hash(x)=75122572)
+3650 train 3.910312 (lr=9.9857e-04) (hash(x)=77134480)
+3660 train 4.063187 (lr=9.9633e-04) (hash(x)=73027661)
+3670 train 4.123003 (lr=9.9409e-04) (hash(x)=85366911)
+3680 train 4.003402 (lr=9.9185e-04) (hash(x)=75873712)
+3690 train 4.092031 (lr=9.8960e-04) (hash(x)=81324159)
+3700 val loss 3.9557
+3700 val perplexity 52.2348
+3700 train 3.992060 (lr=9.8735e-04) (hash(x)=75354374)
+3710 train 4.037969 (lr=9.8510e-04) (hash(x)=73820036)
+3720 train 4.048768 (lr=9.8284e-04) (hash(x)=69279349)
+3730 train 4.013989 (lr=9.8058e-04) (hash(x)=76885056)
+3740 train 3.888714 (lr=9.7832e-04) (hash(x)=66083153)
+3750 train 3.888757 (lr=9.7605e-04) (hash(x)=82585096)
+3760 train 3.854025 (lr=9.7378e-04) (hash(x)=74734456)
+3770 train 3.855695 (lr=9.7151e-04) (hash(x)=74077986)
+3780 train 3.847320 (lr=9.6924e-04) (hash(x)=73794394)
+3790 train 3.879459 (lr=9.6696e-04) (hash(x)=75877136)
+3800 val loss 3.9426
+3800 val perplexity 51.5529
+3800 train 3.967906 (lr=9.6468e-04) (hash(x)=72150028)
+3810 train 4.063464 (lr=9.6240e-04) (hash(x)=77316418)
+3820 train 4.255959 (lr=9.6011e-04) (hash(x)=79951144)
+3830 train 3.940696 (lr=9.5782e-04) (hash(x)=72971714)
+3840 train 3.969411 (lr=9.5553e-04) (hash(x)=78168722)
+3850 train 3.804637 (lr=9.5324e-04) (hash(x)=72103764)
+3860 train 3.850130 (lr=9.5094e-04) (hash(x)=70399281)
+3870 train 4.033202 (lr=9.4864e-04) (hash(x)=71911595)
+3880 train 3.841239 (lr=9.4634e-04) (hash(x)=72034433)
+3890 train 3.836830 (lr=9.4404e-04) (hash(x)=73270524)
+3900 val loss 3.9377
+3900 val perplexity 51.3013
+3900 train 3.830778 (lr=9.4173e-04) (hash(x)=71934546)
+3910 train 3.987936 (lr=9.3942e-04) (hash(x)=78906659)
+3920 train 3.892256 (lr=9.3711e-04) (hash(x)=67852098)
+3930 train 3.849916 (lr=9.3479e-04) (hash(x)=75387628)
+3940 train 3.963380 (lr=9.3248e-04) (hash(x)=72407767)
+3950 train 3.810200 (lr=9.3016e-04) (hash(x)=66515238)
+3960 train 4.003397 (lr=9.2784e-04) (hash(x)=73337537)
+3970 train 3.995071 (lr=9.2551e-04) (hash(x)=74639467)
+3980 train 3.992195 (lr=9.2319e-04) (hash(x)=76044746)
+3990 train 4.008162 (lr=9.2086e-04) (hash(x)=71769001)
+4000 val loss 3.9154
+4000 val perplexity 50.1673
+4000 train 3.885686 (lr=9.1853e-04) (hash(x)=70257299)
+4010 train 3.935741 (lr=9.1620e-04) (hash(x)=71314416)
+4020 train 4.042789 (lr=9.1386e-04) (hash(x)=82315713)
+4030 train 3.985622 (lr=9.1152e-04) (hash(x)=65813944)
+4040 train 3.876420 (lr=9.0919e-04) (hash(x)=78277446)
+4050 train 3.927946 (lr=9.0684e-04) (hash(x)=77850258)
+4060 train 4.024888 (lr=9.0450e-04) (hash(x)=69511868)
+4070 train 3.847049 (lr=9.0216e-04) (hash(x)=77262761)
+4080 train 3.908839 (lr=8.9981e-04) (hash(x)=79691185)
+4090 train 3.725063 (lr=8.9746e-04) (hash(x)=74848196)
+4100 val loss 3.9142
+4100 val perplexity 50.1092
+4100 train 3.877695 (lr=8.9511e-04) (hash(x)=72761342)
+4110 train 3.975305 (lr=8.9276e-04) (hash(x)=58710996)
+4120 train 3.860022 (lr=8.9041e-04) (hash(x)=75794649)
+4130 train 3.975359 (lr=8.8805e-04) (hash(x)=72213336)
+4140 train 3.886427 (lr=8.8569e-04) (hash(x)=75842443)
+4150 train 3.798686 (lr=8.8333e-04) (hash(x)=70329458)
+4160 train 3.788982 (lr=8.8097e-04) (hash(x)=88536084)
+4170 train 3.940527 (lr=8.7861e-04) (hash(x)=76887790)
+4180 train 3.834532 (lr=8.7625e-04) (hash(x)=76755599)
+4190 train 3.838089 (lr=8.7388e-04) (hash(x)=71379783)
+4200 val loss 3.9132
+4200 val perplexity 50.0610
+4200 train 3.822760 (lr=8.7151e-04) (hash(x)=112576746)
+4210 train 3.747171 (lr=8.6915e-04) (hash(x)=77101027)
+4220 train 3.919739 (lr=8.6678e-04) (hash(x)=71607880)
+4230 train 3.825302 (lr=8.6440e-04) (hash(x)=72892070)
+4240 train 3.767149 (lr=8.6203e-04) (hash(x)=70967675)
+4250 train 3.867950 (lr=8.5966e-04) (hash(x)=75050831)
+4260 train 3.925073 (lr=8.5728e-04) (hash(x)=75249845)
+4270 train 3.920346 (lr=8.5491e-04) (hash(x)=82744946)
+4280 train 3.903028 (lr=8.5253e-04) (hash(x)=68857061)
+4290 train 3.884757 (lr=8.5015e-04) (hash(x)=80434125)
+4300 val loss 3.8945
+4300 val perplexity 49.1306
+4300 train 3.863431 (lr=8.4777e-04) (hash(x)=84624166)
+4310 train 3.782129 (lr=8.4539e-04) (hash(x)=60925156)
+4320 train 3.925793 (lr=8.4301e-04) (hash(x)=75956316)
+4330 train 3.844311 (lr=8.4062e-04) (hash(x)=74068056)
+4340 train 4.077273 (lr=8.3824e-04) (hash(x)=94977597)
+4350 train 3.969797 (lr=8.3585e-04) (hash(x)=71781944)
+4360 train 4.028651 (lr=8.3347e-04) (hash(x)=74308923)
+4370 train 3.876070 (lr=8.3108e-04) (hash(x)=76645666)
+4380 train 3.857038 (lr=8.2869e-04) (hash(x)=73326958)
+4390 train 3.838163 (lr=8.2630e-04) (hash(x)=74743749)
+4400 val loss 3.8829
+4400 val perplexity 48.5656
+4400 train 3.967121 (lr=8.2391e-04) (hash(x)=73116611)
+4410 train 3.744528 (lr=8.2152e-04) (hash(x)=68578568)
+4420 train 3.971103 (lr=8.1913e-04) (hash(x)=78085026)
+4430 train 3.864100 (lr=8.1674e-04) (hash(x)=71504765)
+4440 train 4.051152 (lr=8.1435e-04) (hash(x)=73163972)
+4450 train 3.873592 (lr=8.1195e-04) (hash(x)=81967324)
+4460 train 3.841782 (lr=8.0956e-04) (hash(x)=67796275)
+4470 train 3.865763 (lr=8.0716e-04) (hash(x)=75136849)
+4480 train 3.822582 (lr=8.0477e-04) (hash(x)=80109861)
+4490 train 3.753750 (lr=8.0237e-04) (hash(x)=72094345)
+4500 val loss 3.8746
+4500 val perplexity 48.1630
+4500 train 3.764498 (lr=7.9998e-04) (hash(x)=77660356)
+4510 train 3.747440 (lr=7.9758e-04) (hash(x)=71303771)
+4520 train 3.965457 (lr=7.9518e-04) (hash(x)=67353213)
+4530 train 3.587553 (lr=7.9279e-04) (hash(x)=91984280)
+4540 train 3.742611 (lr=7.9039e-04) (hash(x)=74918518)
+4550 train 3.835837 (lr=7.8799e-04) (hash(x)=80129084)
+4560 train 3.777633 (lr=7.8559e-04) (hash(x)=74762343)
+4570 train 3.743225 (lr=7.8319e-04) (hash(x)=83264256)
+4580 train 3.846176 (lr=7.8080e-04) (hash(x)=77701601)
+4590 train 3.778602 (lr=7.7840e-04) (hash(x)=72730921)
+4600 val loss 3.8652
+4600 val perplexity 47.7121
+4600 train 3.828789 (lr=7.7600e-04) (hash(x)=77801190)
+4610 train 3.891238 (lr=7.7360e-04) (hash(x)=64563241)
+4620 train 3.842094 (lr=7.7120e-04) (hash(x)=75462951)
+4630 train 3.882444 (lr=7.6880e-04) (hash(x)=74708327)
+4640 train 3.783543 (lr=7.6640e-04) (hash(x)=83297632)
+4650 train 3.894357 (lr=7.6400e-04) (hash(x)=81914758)
+4660 train 3.914593 (lr=7.6160e-04) (hash(x)=71367268)
+4670 train 3.801260 (lr=7.5920e-04) (hash(x)=73587555)
+4680 train 3.918172 (lr=7.5681e-04) (hash(x)=74734602)
+4690 train 4.008985 (lr=7.5441e-04) (hash(x)=77508704)
+4700 val loss 3.8518
+4700 val perplexity 47.0773
+4700 train 3.815714 (lr=7.5201e-04) (hash(x)=73228682)
+4710 train 3.837982 (lr=7.4961e-04) (hash(x)=73851956)
+4720 train 3.887956 (lr=7.4721e-04) (hash(x)=71244190)
+4730 train 3.654194 (lr=7.4482e-04) (hash(x)=74416580)
+4740 train 3.886899 (lr=7.4242e-04) (hash(x)=69259353)
+4750 train 3.853163 (lr=7.4002e-04) (hash(x)=77662887)
+4760 train 3.802109 (lr=7.3763e-04) (hash(x)=73001257)
+4770 train 3.746603 (lr=7.3523e-04) (hash(x)=81466432)
+4780 train 3.692114 (lr=7.3284e-04) (hash(x)=76710780)
+4790 train 3.833987 (lr=7.3044e-04) (hash(x)=72440698)
+4800 val loss 3.8527
+4800 val perplexity 47.1212
+4800 train 3.723619 (lr=7.2805e-04) (hash(x)=75002116)
+4810 train 3.777115 (lr=7.2565e-04) (hash(x)=84873693)
+4820 train 3.621597 (lr=7.2326e-04) (hash(x)=68563264)
+4830 train 3.740219 (lr=7.2087e-04) (hash(x)=76811982)
+4840 train 3.657794 (lr=7.1848e-04) (hash(x)=73763746)
+4850 train 3.869951 (lr=7.1609e-04) (hash(x)=72834169)
+4860 train 3.808547 (lr=7.1370e-04) (hash(x)=71420794)
+4870 train 3.927580 (lr=7.1131e-04) (hash(x)=82083784)
+4880 train 3.903363 (lr=7.0892e-04) (hash(x)=71399278)
+4890 train 3.969394 (lr=7.0653e-04) (hash(x)=80791916)
+4900 val loss 3.8337
+4900 val perplexity 46.2311
+4900 train 3.854835 (lr=7.0415e-04) (hash(x)=76275522)
+4910 train 4.006440 (lr=7.0176e-04) (hash(x)=73846113)
+4920 train 3.932277 (lr=6.9938e-04) (hash(x)=65906497)
+4930 train 3.731254 (lr=6.9699e-04) (hash(x)=74564994)
+4940 train 3.998323 (lr=6.9461e-04) (hash(x)=78282849)
+4950 train 3.823729 (lr=6.9223e-04) (hash(x)=75618079)
+4960 train 3.954535 (lr=6.8985e-04) (hash(x)=78308363)
+4970 train 3.826117 (lr=6.8747e-04) (hash(x)=81122810)
+4980 train 3.968577 (lr=6.8509e-04) (hash(x)=72474525)
+4990 train 3.821536 (lr=6.8272e-04) (hash(x)=71682033)
+5000 val loss 3.8176
+5000 val perplexity 45.4964
+5000 train 3.778444 (lr=6.8034e-04) (hash(x)=74751266)
+5010 train 3.835905 (lr=6.7797e-04) (hash(x)=71876255)
+5020 train 4.015388 (lr=6.7560e-04) (hash(x)=67006293)
+5030 train 3.826584 (lr=6.7322e-04) (hash(x)=55446837)
+5040 train 3.937929 (lr=6.7085e-04) (hash(x)=72139304)
+5050 train 3.918955 (lr=6.6849e-04) (hash(x)=72848819)
+5060 train 3.836163 (lr=6.6612e-04) (hash(x)=74756891)
+5070 train 3.830682 (lr=6.6375e-04) (hash(x)=71425192)
+5080 train 3.701489 (lr=6.6139e-04) (hash(x)=91717336)
+5090 train 3.710880 (lr=6.5903e-04) (hash(x)=78489482)
+5100 val loss 3.8167
+5100 val perplexity 45.4531
+5100 train 3.705281 (lr=6.5667e-04) (hash(x)=102417092)
+5110 train 3.774787 (lr=6.5431e-04) (hash(x)=67631270)
+5120 train 3.766247 (lr=6.5195e-04) (hash(x)=70575439)
+5130 train 3.799153 (lr=6.4959e-04) (hash(x)=73933044)
+5140 train 3.667325 (lr=6.4724e-04) (hash(x)=75105935)
+5150 train 3.789569 (lr=6.4489e-04) (hash(x)=75662761)
+5160 train 3.781642 (lr=6.4254e-04) (hash(x)=77254705)
+5170 train 3.751958 (lr=6.4019e-04) (hash(x)=67051450)
+5180 train 3.584543 (lr=6.3784e-04) (hash(x)=69736129)
+5190 train 3.594507 (lr=6.3550e-04) (hash(x)=68256740)
+5200 val loss 3.8168
+5200 val perplexity 45.4606
+5200 train 3.559217 (lr=6.3316e-04) (hash(x)=69483022)
+5210 train 3.604229 (lr=6.3081e-04) (hash(x)=74805034)
+5220 train 3.591407 (lr=6.2848e-04) (hash(x)=75297212)
+5230 train 3.515170 (lr=6.2614e-04) (hash(x)=69772814)
+5240 train 3.630870 (lr=6.2380e-04) (hash(x)=72164549)
+5250 train 3.483792 (lr=6.2147e-04) (hash(x)=73896592)
+5260 train 3.632141 (lr=6.1914e-04) (hash(x)=75901314)
+5270 train 3.503164 (lr=6.1681e-04) (hash(x)=76054156)
+5280 train 3.849838 (lr=6.1449e-04) (hash(x)=85568643)
+5290 train 3.880060 (lr=6.1216e-04) (hash(x)=84053592)
+5300 val loss 3.8007
+5300 val perplexity 44.7305
+5300 train 3.774042 (lr=6.0984e-04) (hash(x)=73135718)
+5310 train 3.921485 (lr=6.0752e-04) (hash(x)=73149116)
+5320 train 3.890686 (lr=6.0521e-04) (hash(x)=77101540)
+5330 train 3.726930 (lr=6.0289e-04) (hash(x)=84274756)
+5340 train 3.754840 (lr=6.0058e-04) (hash(x)=74181619)
+5350 train 3.639080 (lr=5.9827e-04) (hash(x)=69664265)
+5360 train 3.873178 (lr=5.9596e-04) (hash(x)=68837592)
+5370 train 3.798954 (lr=5.9366e-04) (hash(x)=75148213)
+5380 train 3.895877 (lr=5.9136e-04) (hash(x)=68908917)
+5390 train 3.850991 (lr=5.8906e-04) (hash(x)=72753463)
+5400 val loss 3.7810
+5400 val perplexity 43.8586
+5400 train 3.749129 (lr=5.8676e-04) (hash(x)=69948122)
+5410 train 3.660287 (lr=5.8447e-04) (hash(x)=74530436)
+5420 train 3.824893 (lr=5.8218e-04) (hash(x)=67532641)
+5430 train 3.798412 (lr=5.7989e-04) (hash(x)=77416323)
+5440 train 3.777107 (lr=5.7760e-04) (hash(x)=77450277)
+5450 train 3.701837 (lr=5.7532e-04) (hash(x)=69637629)
+5460 train 3.748670 (lr=5.7304e-04) (hash(x)=77134731)
+5470 train 3.878119 (lr=5.7076e-04) (hash(x)=72637484)
+5480 train 3.823482 (lr=5.6849e-04) (hash(x)=76342282)
+5490 train 3.745006 (lr=5.6622e-04) (hash(x)=74794888)
+5500 val loss 3.7793
+5500 val perplexity 43.7864
+5500 train 3.738924 (lr=5.6395e-04) (hash(x)=70491787)
+5510 train 3.766233 (lr=5.6168e-04) (hash(x)=73673009)
+5520 train 3.704779 (lr=5.5942e-04) (hash(x)=73134241)
+5530 train 3.706710 (lr=5.5716e-04) (hash(x)=80260202)
+5540 train 3.742632 (lr=5.5490e-04) (hash(x)=75403578)
+5550 train 3.667290 (lr=5.5265e-04) (hash(x)=65589999)
+5560 train 3.751736 (lr=5.5040e-04) (hash(x)=71523917)
+5570 train 3.648241 (lr=5.4815e-04) (hash(x)=71871396)
+5580 train 3.461463 (lr=5.4591e-04) (hash(x)=77446225)
+5590 train 3.475391 (lr=5.4367e-04) (hash(x)=79892116)
+5600 val loss 3.7859
+5600 val perplexity 44.0733
+5600 train 3.700331 (lr=5.4143e-04) (hash(x)=72893174)
+5610 train 3.534003 (lr=5.3920e-04) (hash(x)=74362751)
+5620 train 3.666277 (lr=5.3697e-04) (hash(x)=75125741)
+5630 train 3.544599 (lr=5.3474e-04) (hash(x)=74303205)
+5640 train 3.549058 (lr=5.3252e-04) (hash(x)=65727805)
+5650 train 3.631223 (lr=5.3030e-04) (hash(x)=79694849)
+5660 train 3.649204 (lr=5.2808e-04) (hash(x)=74705862)
+5670 train 3.625037 (lr=5.2587e-04) (hash(x)=74962821)
+5680 train 4.000886 (lr=5.2366e-04) (hash(x)=76748843)
+5690 train 3.845444 (lr=5.2145e-04) (hash(x)=72562868)
+5700 val loss 3.7757
+5700 val perplexity 43.6300
+5700 train 3.788968 (lr=5.1925e-04) (hash(x)=83245544)
+5710 train 3.706428 (lr=5.1705e-04) (hash(x)=76485320)
+5720 train 3.824124 (lr=5.1485e-04) (hash(x)=75340650)
+5730 train 3.808181 (lr=5.1266e-04) (hash(x)=73202272)
+5740 train 3.787750 (lr=5.1047e-04) (hash(x)=75623558)
+5750 train 3.805197 (lr=5.0829e-04) (hash(x)=73893861)
+5760 train 3.777469 (lr=5.0611e-04) (hash(x)=78314196)
+5770 train 3.789562 (lr=5.0393e-04) (hash(x)=72732120)
+5780 train 3.775733 (lr=5.0176e-04) (hash(x)=79236217)
+5790 train 3.759450 (lr=4.9959e-04) (hash(x)=82223910)
+5800 val loss 3.7565
+5800 val perplexity 42.7965
+5800 train 3.774546 (lr=4.9743e-04) (hash(x)=71077079)
+5810 train 3.773899 (lr=4.9526e-04) (hash(x)=82793170)
+5820 train 3.765761 (lr=4.9311e-04) (hash(x)=100985291)
+5830 train 3.771655 (lr=4.9095e-04) (hash(x)=74375677)
+5840 train 3.900134 (lr=4.8881e-04) (hash(x)=71221972)
+5850 train 3.681459 (lr=4.8666e-04) (hash(x)=72588583)
+5860 train 3.767431 (lr=4.8452e-04) (hash(x)=88719929)
+5870 train 3.743673 (lr=4.8238e-04) (hash(x)=73369246)
+5880 train 3.798212 (lr=4.8025e-04) (hash(x)=73331315)
+5890 train 3.765932 (lr=4.7812e-04) (hash(x)=53449512)
+5900 val loss 3.7584
+5900 val perplexity 42.8793
+5900 train 3.654534 (lr=4.7600e-04) (hash(x)=75334660)
+5910 train 3.723772 (lr=4.7388e-04) (hash(x)=75037764)
+5920 train 3.796310 (lr=4.7176e-04) (hash(x)=72040912)
+5930 train 3.722481 (lr=4.6965e-04) (hash(x)=68296721)
+5940 train 3.682830 (lr=4.6755e-04) (hash(x)=72575829)
+5950 train 3.569261 (lr=4.6544e-04) (hash(x)=71274235)
+5960 train 3.759852 (lr=4.6335e-04) (hash(x)=70377631)
+5970 train 3.584262 (lr=4.6125e-04) (hash(x)=80434203)
+5980 train 3.455098 (lr=4.5916e-04) (hash(x)=71681516)
+5990 train 3.583311 (lr=4.5708e-04) (hash(x)=74917317)
+6000 val loss 3.7567
+6000 val perplexity 42.8078
+6000 train 3.504721 (lr=4.5500e-04) (hash(x)=69246326)
+6010 train 3.595223 (lr=4.5292e-04) (hash(x)=76209797)
+6020 train 3.697965 (lr=4.5085e-04) (hash(x)=77523561)
+6030 train 3.671243 (lr=4.4879e-04) (hash(x)=77066537)
+6040 train 3.641183 (lr=4.4673e-04) (hash(x)=81795462)
+6050 train 3.548830 (lr=4.4467e-04) (hash(x)=74328005)
+6060 train 3.557174 (lr=4.4262e-04) (hash(x)=76260120)
+6070 train 3.693014 (lr=4.4057e-04) (hash(x)=73864904)
+6080 train 3.657921 (lr=4.3853e-04) (hash(x)=83311739)
+6090 train 3.964921 (lr=4.3649e-04) (hash(x)=85943188)
+6100 val loss 3.7353
+6100 val perplexity 41.9024
+6100 train 3.630619 (lr=4.3446e-04) (hash(x)=70412657)
+6110 train 3.785141 (lr=4.3243e-04) (hash(x)=78569898)
+6120 train 3.837827 (lr=4.3041e-04) (hash(x)=53010374)
+6130 train 3.832522 (lr=4.2839e-04) (hash(x)=77387166)
+6140 train 3.818694 (lr=4.2637e-04) (hash(x)=68495380)
+6150 train 3.707014 (lr=4.2437e-04) (hash(x)=77674661)
+6160 train 3.782187 (lr=4.2236e-04) (hash(x)=73099224)
+6170 train 3.789567 (lr=4.2036e-04) (hash(x)=75550331)
+6180 train 3.854024 (lr=4.1837e-04) (hash(x)=70208558)
+6190 train 3.687845 (lr=4.1638e-04) (hash(x)=81474783)
+6200 val loss 3.7200
+6200 val perplexity 41.2630
+6200 train 3.686452 (lr=4.1440e-04) (hash(x)=72863420)
+6210 train 3.821372 (lr=4.1242e-04) (hash(x)=67338127)
+6220 train 3.746941 (lr=4.1045e-04) (hash(x)=72064676)
+6230 train 3.627480 (lr=4.0848e-04) (hash(x)=63929148)
+6240 train 3.660976 (lr=4.0652e-04) (hash(x)=75244700)
+6250 train 3.756036 (lr=4.0456e-04) (hash(x)=72131685)
+6260 train 3.828809 (lr=4.0261e-04) (hash(x)=77239186)
+6270 train 3.725954 (lr=4.0067e-04) (hash(x)=80827392)
+6280 train 3.690965 (lr=3.9873e-04) (hash(x)=68612370)
+6290 train 3.742420 (lr=3.9679e-04) (hash(x)=70664463)
+6300 val loss 3.7156
+6300 val perplexity 41.0852
+6300 train 3.686260 (lr=3.9486e-04) (hash(x)=73020224)
+6310 train 3.704697 (lr=3.9294e-04) (hash(x)=70614977)
+6320 train 3.697288 (lr=3.9102e-04) (hash(x)=76744877)
+6330 train 3.748580 (lr=3.8910e-04) (hash(x)=83384143)
+6340 train 3.601230 (lr=3.8719e-04) (hash(x)=72241681)
+6350 train 3.629720 (lr=3.8529e-04) (hash(x)=74037360)
+6360 train 3.468064 (lr=3.8339e-04) (hash(x)=70095163)
+6370 train 3.395054 (lr=3.8150e-04) (hash(x)=69510647)
+6380 train 3.564710 (lr=3.7962e-04) (hash(x)=74297517)
+6390 train 3.706028 (lr=3.7774e-04) (hash(x)=78058972)
+6400 val loss 3.7211
+6400 val perplexity 41.3094
+6400 train 3.612296 (lr=3.7586e-04) (hash(x)=82953514)
+6410 train 3.389762 (lr=3.7399e-04) (hash(x)=72289413)
+6420 train 3.527603 (lr=3.7213e-04) (hash(x)=66436985)
+6430 train 3.634292 (lr=3.7027e-04) (hash(x)=76842704)
+6440 train 3.582709 (lr=3.6842e-04) (hash(x)=77576509)
+6450 train 3.586757 (lr=3.6658e-04) (hash(x)=63757462)
+6460 train 3.354029 (lr=3.6474e-04) (hash(x)=74725544)
+6470 train 3.802495 (lr=3.6290e-04) (hash(x)=74976987)
+6480 train 3.938939 (lr=3.6108e-04) (hash(x)=68415455)
+6490 train 3.736585 (lr=3.5925e-04) (hash(x)=75902883)
+6500 val loss 3.7055
+6500 val perplexity 40.6710
+6500 train 3.680727 (lr=3.5744e-04) (hash(x)=76888442)
+6510 train 3.781392 (lr=3.5563e-04) (hash(x)=73767111)
+6520 train 3.774084 (lr=3.5382e-04) (hash(x)=76218061)
+6530 train 3.739107 (lr=3.5203e-04) (hash(x)=76573241)
+6540 train 3.718171 (lr=3.5023e-04) (hash(x)=71980504)
+6550 train 3.774028 (lr=3.4845e-04) (hash(x)=77715451)
+6560 train 3.657356 (lr=3.4667e-04) (hash(x)=73772644)
+6570 train 3.726677 (lr=3.4489e-04) (hash(x)=70907139)
+6580 train 3.734878 (lr=3.4313e-04) (hash(x)=68045832)
+6590 train 3.749266 (lr=3.4137e-04) (hash(x)=75212157)
+6600 val loss 3.7018
+6600 val perplexity 40.5204
+6600 train 3.694598 (lr=3.3961e-04) (hash(x)=69232309)
+6610 train 3.673433 (lr=3.3786e-04) (hash(x)=67915539)
+6620 train 3.536473 (lr=3.3612e-04) (hash(x)=73942199)
+6630 train 3.712419 (lr=3.3438e-04) (hash(x)=71108653)
+6640 train 3.704979 (lr=3.3265e-04) (hash(x)=70766590)
+6650 train 3.584315 (lr=3.3093e-04) (hash(x)=69205319)
+6660 train 3.695248 (lr=3.2921e-04) (hash(x)=81026148)
+6670 train 3.621076 (lr=3.2750e-04) (hash(x)=77657830)
+6680 train 3.524972 (lr=3.2580e-04) (hash(x)=76483678)
+6690 train 3.584049 (lr=3.2410e-04) (hash(x)=71151306)
+6700 val loss 3.7020
+6700 val perplexity 40.5278
+6700 train 3.678775 (lr=3.2241e-04) (hash(x)=74577530)
+6710 train 3.565653 (lr=3.2072e-04) (hash(x)=69171643)
+6720 train 3.617409 (lr=3.1904e-04) (hash(x)=67250925)
+6730 train 3.595423 (lr=3.1737e-04) (hash(x)=68971618)
+6740 train 3.551506 (lr=3.1571e-04) (hash(x)=75990195)
+6750 train 3.655202 (lr=3.1405e-04) (hash(x)=84198679)
+6760 train 3.732120 (lr=3.1240e-04) (hash(x)=100211939)
+6770 train 3.756465 (lr=3.1075e-04) (hash(x)=70028335)
+6780 train 3.819086 (lr=3.0911e-04) (hash(x)=76176743)
+6790 train 3.896934 (lr=3.0748e-04) (hash(x)=75642962)
+6800 val loss 3.6856
+6800 val perplexity 39.8702
+6800 train 3.630347 (lr=3.0585e-04) (hash(x)=73064800)
+6810 train 3.736848 (lr=3.0423e-04) (hash(x)=71307839)
+6820 train 3.654453 (lr=3.0262e-04) (hash(x)=80361526)
+6830 train 3.673291 (lr=3.0102e-04) (hash(x)=74155886)
+6840 train 3.759528 (lr=2.9942e-04) (hash(x)=83498789)
+6850 train 3.865271 (lr=2.9783e-04) (hash(x)=72686972)
+6860 train 3.822212 (lr=2.9624e-04) (hash(x)=74835805)
+6870 train 3.748828 (lr=2.9466e-04) (hash(x)=74809435)
+6880 train 3.816219 (lr=2.9309e-04) (hash(x)=71911947)
+6890 train 3.686652 (lr=2.9153e-04) (hash(x)=70201693)
+6900 val loss 3.6792
+6900 val perplexity 39.6160
+6900 train 3.713596 (lr=2.8997e-04) (hash(x)=77162446)
+6910 train 3.831896 (lr=2.8842e-04) (hash(x)=60395458)
+6920 train 3.736101 (lr=2.8688e-04) (hash(x)=78466934)
+6930 train 3.724213 (lr=2.8534e-04) (hash(x)=80996788)
+6940 train 3.614672 (lr=2.8381e-04) (hash(x)=74634433)
+6950 train 3.739762 (lr=2.8229e-04) (hash(x)=69618707)
+6960 train 3.597714 (lr=2.8078e-04) (hash(x)=77936797)
+6970 train 3.604413 (lr=2.7927e-04) (hash(x)=78698749)
+6980 train 3.606887 (lr=2.7777e-04) (hash(x)=71152904)
+6990 train 3.718963 (lr=2.7627e-04) (hash(x)=78876720)
+7000 val loss 3.6747
+7000 val perplexity 39.4386
+7000 train 3.691628 (lr=2.7479e-04) (hash(x)=70500963)
+7010 train 3.659212 (lr=2.7331e-04) (hash(x)=75732988)
+7020 train 3.832793 (lr=2.7183e-04) (hash(x)=72655885)
+7030 train 3.541362 (lr=2.7037e-04) (hash(x)=75132388)
+7040 train 3.611767 (lr=2.6891e-04) (hash(x)=88674279)
+7050 train 3.584797 (lr=2.6746e-04) (hash(x)=73050510)
+7060 train 3.628935 (lr=2.6602e-04) (hash(x)=74850157)
+7070 train 3.623876 (lr=2.6458e-04) (hash(x)=80733840)
+7080 train 3.607394 (lr=2.6315e-04) (hash(x)=68592323)
+7090 train 3.549088 (lr=2.6173e-04) (hash(x)=74142615)
+7100 val loss 3.6749
+7100 val perplexity 39.4452
+7100 train 3.557771 (lr=2.6032e-04) (hash(x)=78220018)
+7110 train 3.559073 (lr=2.5891e-04) (hash(x)=68388633)
+7120 train 3.377791 (lr=2.5751e-04) (hash(x)=71391877)
+7130 train 3.689851 (lr=2.5612e-04) (hash(x)=80144529)
+7140 train 3.788820 (lr=2.5474e-04) (hash(x)=79507532)
+7150 train 3.691935 (lr=2.5336e-04) (hash(x)=76982262)
+7160 train 3.845235 (lr=2.5199e-04) (hash(x)=82516157)
+7170 train 3.603332 (lr=2.5063e-04) (hash(x)=72371860)
+7180 train 3.848937 (lr=2.4928e-04) (hash(x)=72376921)
+7190 train 3.699514 (lr=2.4793e-04) (hash(x)=81008418)
+7200 val loss 3.6588
+7200 val perplexity 38.8150
+7200 train 3.768717 (lr=2.4659e-04) (hash(x)=73818590)
+7210 train 3.735803 (lr=2.4526e-04) (hash(x)=71683916)
+7220 train 3.682644 (lr=2.4394e-04) (hash(x)=77659620)
+7230 train 3.732181 (lr=2.4262e-04) (hash(x)=79997614)
+7240 train 3.744857 (lr=2.4131e-04) (hash(x)=74054734)
+7250 train 3.834733 (lr=2.4001e-04) (hash(x)=84630195)
+7260 train 3.738304 (lr=2.3872e-04) (hash(x)=83833114)
+7270 train 3.651428 (lr=2.3743e-04) (hash(x)=78625868)
+7280 train 3.867038 (lr=2.3615e-04) (hash(x)=76859769)
+7290 train 3.663935 (lr=2.3488e-04) (hash(x)=77079387)
+7300 val loss 3.6538
+7300 val perplexity 38.6223
+7300 train 3.658094 (lr=2.3362e-04) (hash(x)=74181120)
+7310 train 3.728942 (lr=2.3237e-04) (hash(x)=72220366)
+7320 train 3.662348 (lr=2.3112e-04) (hash(x)=77955423)
+7330 train 3.701967 (lr=2.2988e-04) (hash(x)=75397467)
+7340 train 3.697631 (lr=2.2865e-04) (hash(x)=73697940)
+7350 train 3.599217 (lr=2.2743e-04) (hash(x)=83531555)
+7360 train 3.709046 (lr=2.2621e-04) (hash(x)=72009097)
+7370 train 3.657749 (lr=2.2500e-04) (hash(x)=76354441)
+7380 train 3.700727 (lr=2.2381e-04) (hash(x)=77219348)
+7390 train 3.497204 (lr=2.2261e-04) (hash(x)=79816131)
+7400 val loss 3.6522
+7400 val perplexity 38.5579
+7400 train 3.681549 (lr=2.2143e-04) (hash(x)=67942598)
+7410 train 3.703395 (lr=2.2025e-04) (hash(x)=67638178)
+7420 train 3.636333 (lr=2.1909e-04) (hash(x)=71682874)
+7430 train 3.499865 (lr=2.1793e-04) (hash(x)=71297263)
+7440 train 3.508817 (lr=2.1678e-04) (hash(x)=73903999)
+7450 train 3.619799 (lr=2.1563e-04) (hash(x)=74071685)
+7460 train 3.405374 (lr=2.1450e-04) (hash(x)=71088160)
+7470 train 3.398161 (lr=2.1337e-04) (hash(x)=74178017)
+7480 train 3.424115 (lr=2.1225e-04) (hash(x)=84610585)
+7490 train 3.598638 (lr=2.1114e-04) (hash(x)=67695236)
+7500 val loss 3.6542
+7500 val perplexity 38.6372
+7500 train 3.536128 (lr=2.1003e-04) (hash(x)=72022604)
+7510 train 3.578765 (lr=2.0894e-04) (hash(x)=78266706)
+7520 train 3.672654 (lr=2.0785e-04) (hash(x)=72576417)
+7530 train 3.887101 (lr=2.0677e-04) (hash(x)=71494052)
+7540 train 3.816981 (lr=2.0570e-04) (hash(x)=76071936)
+7550 train 3.686994 (lr=2.0464e-04) (hash(x)=78379534)
+7560 train 3.642784 (lr=2.0358e-04) (hash(x)=75712040)
+7570 train 3.637986 (lr=2.0254e-04) (hash(x)=83109134)
+7580 train 3.697557 (lr=2.0150e-04) (hash(x)=78846949)
+7590 train 3.783383 (lr=2.0047e-04) (hash(x)=63919123)
+7600 val loss 3.6413
+7600 val perplexity 38.1414
+7600 train 3.658424 (lr=1.9945e-04) (hash(x)=73563619)
+7610 train 3.721145 (lr=1.9844e-04) (hash(x)=79745028)
+7620 train 3.699767 (lr=1.9743e-04) (hash(x)=68113846)
+7630 train 3.678173 (lr=1.9643e-04) (hash(x)=70369067)
+7640 train 3.676894 (lr=1.9545e-04) (hash(x)=79635086)
+7650 train 3.597596 (lr=1.9447e-04) (hash(x)=77348303)
+7660 train 3.600773 (lr=1.9349e-04) (hash(x)=74939810)
+7670 train 3.730974 (lr=1.9253e-04) (hash(x)=77151692)
+7680 train 3.550768 (lr=1.9158e-04) (hash(x)=69543853)
+7690 train 3.790066 (lr=1.9063e-04) (hash(x)=75604567)
+7700 val loss 3.6359
+7700 val perplexity 37.9364
+7700 train 3.757467 (lr=1.8969e-04) (hash(x)=84362557)
+7710 train 3.700763 (lr=1.8876e-04) (hash(x)=71030514)
+7720 train 3.657194 (lr=1.8784e-04) (hash(x)=76132027)
+7730 train 3.702238 (lr=1.8693e-04) (hash(x)=77799943)
+7740 train 3.673486 (lr=1.8602e-04) (hash(x)=74951280)
+7750 train 3.699217 (lr=1.8513e-04) (hash(x)=73706939)
+7760 train 3.614076 (lr=1.8424e-04) (hash(x)=78667790)
+7770 train 3.721556 (lr=1.8336e-04) (hash(x)=72949775)
+7780 train 3.655546 (lr=1.8249e-04) (hash(x)=74537611)
+7790 train 3.608297 (lr=1.8163e-04) (hash(x)=72876060)
+7800 val loss 3.6352
+7800 val perplexity 37.9110
+7800 train 3.711815 (lr=1.8078e-04) (hash(x)=73302610)
+7810 train 3.465721 (lr=1.7993e-04) (hash(x)=66480452)
+7820 train 3.479512 (lr=1.7910e-04) (hash(x)=74833008)
+7830 train 3.501681 (lr=1.7827e-04) (hash(x)=66729619)
+7840 train 3.601382 (lr=1.7745e-04) (hash(x)=87861820)
+7850 train 3.408096 (lr=1.7664e-04) (hash(x)=73130610)
+7860 train 3.624389 (lr=1.7584e-04) (hash(x)=67874016)
+7870 train 3.412145 (lr=1.7504e-04) (hash(x)=70600131)
+7880 train 3.544292 (lr=1.7426e-04) (hash(x)=68990980)
+7890 train 3.502559 (lr=1.7348e-04) (hash(x)=79666741)
+7900 val loss 3.6367
+7900 val perplexity 37.9649
+7900 train 3.524320 (lr=1.7271e-04) (hash(x)=74072340)
+7910 train 3.585784 (lr=1.7196e-04) (hash(x)=80452381)
+7920 train 3.721328 (lr=1.7121e-04) (hash(x)=83292891)
+7930 train 3.609617 (lr=1.7046e-04) (hash(x)=80375639)
+7940 train 3.513196 (lr=1.6973e-04) (hash(x)=82765856)
+7950 train 3.601940 (lr=1.6901e-04) (hash(x)=72962177)
+7960 train 3.681541 (lr=1.6829e-04) (hash(x)=74551129)
+7970 train 3.770118 (lr=1.6759e-04) (hash(x)=89185352)
+7980 train 3.744188 (lr=1.6689e-04) (hash(x)=81338443)
+7990 train 3.531917 (lr=1.6620e-04) (hash(x)=83820810)
+8000 val loss 3.6274
+8000 val perplexity 37.6167
+8000 train 3.702232 (lr=1.6552e-04) (hash(x)=72329617)
+8010 train 3.685851 (lr=1.6485e-04) (hash(x)=82188069)
+8020 train 3.715370 (lr=1.6419e-04) (hash(x)=78460596)
+8030 train 3.628606 (lr=1.6353e-04) (hash(x)=76416474)
+8040 train 3.806722 (lr=1.6289e-04) (hash(x)=70629903)
+8050 train 3.599672 (lr=1.6225e-04) (hash(x)=76561819)
+8060 train 3.633317 (lr=1.6162e-04) (hash(x)=71993036)
+8070 train 3.800936 (lr=1.6100e-04) (hash(x)=73270928)
+8080 train 3.702070 (lr=1.6039e-04) (hash(x)=82402459)
+8090 train 3.736435 (lr=1.5979e-04) (hash(x)=67208853)
+8100 val loss 3.6228
+8100 val perplexity 37.4414
+8100 train 3.804675 (lr=1.5920e-04) (hash(x)=85981056)
+8110 train 3.760064 (lr=1.5862e-04) (hash(x)=77469728)
+8120 train 3.665158 (lr=1.5804e-04) (hash(x)=84702034)
+8130 train 3.555000 (lr=1.5748e-04) (hash(x)=96254798)
+8140 train 3.616427 (lr=1.5692e-04) (hash(x)=71865027)
+8150 train 3.593187 (lr=1.5637e-04) (hash(x)=79902920)
+8160 train 3.739484 (lr=1.5583e-04) (hash(x)=70307892)
+8170 train 3.616855 (lr=1.5530e-04) (hash(x)=73853567)
+8180 train 3.624122 (lr=1.5478e-04) (hash(x)=72321642)
+8190 train 3.642168 (lr=1.5427e-04) (hash(x)=82101781)
+8200 val loss 3.6208
+8200 val perplexity 37.3686
+8200 train 3.660584 (lr=1.5377e-04) (hash(x)=66290405)
+8210 train 3.542098 (lr=1.5327e-04) (hash(x)=69697952)
+8220 train 3.784667 (lr=1.5279e-04) (hash(x)=72618212)
+8230 train 3.537820 (lr=1.5231e-04) (hash(x)=76210758)
+8240 train 3.538558 (lr=1.5184e-04) (hash(x)=70820601)
+8250 train 3.611426 (lr=1.5138e-04) (hash(x)=75548154)
+8260 train 3.551512 (lr=1.5094e-04) (hash(x)=73211139)
+8270 train 3.606651 (lr=1.5049e-04) (hash(x)=82252960)
+8280 train 3.558560 (lr=1.5006e-04) (hash(x)=69408116)
+8290 train 3.724331 (lr=1.4964e-04) (hash(x)=74462098)
+8300 val loss 3.6204
+8300 val perplexity 37.3523
+8300 train 3.731257 (lr=1.4923e-04) (hash(x)=79945329)
+8310 train 3.485023 (lr=1.4882e-04) (hash(x)=82459712)
+8320 train 3.665993 (lr=1.4843e-04) (hash(x)=81135573)
+8330 train 3.640266 (lr=1.4804e-04) (hash(x)=69939721)
+8340 train 3.584143 (lr=1.4766e-04) (hash(x)=84687517)
+8350 train 3.552184 (lr=1.4729e-04) (hash(x)=77515837)
+8360 train 3.555122 (lr=1.4693e-04) (hash(x)=72146536)
+8370 train 3.600583 (lr=1.4658e-04) (hash(x)=73198063)
+8380 train 3.601297 (lr=1.4624e-04) (hash(x)=68952542)
+8390 train 3.637936 (lr=1.4591e-04) (hash(x)=74385341)
+8400 val loss 3.6139
+8400 val perplexity 37.1088
+8400 train 3.456181 (lr=1.4559e-04) (hash(x)=65193622)
+8410 train 3.699759 (lr=1.4527e-04) (hash(x)=66981351)
+8420 train 3.591754 (lr=1.4497e-04) (hash(x)=97752378)
+8430 train 3.800407 (lr=1.4467e-04) (hash(x)=70936998)
+8440 train 3.710022 (lr=1.4438e-04) (hash(x)=71557836)
+8450 train 3.638443 (lr=1.4411e-04) (hash(x)=71340008)
+8460 train 3.689378 (lr=1.4384e-04) (hash(x)=69634745)
+8470 train 3.711257 (lr=1.4358e-04) (hash(x)=74224484)
+8480 train 3.663258 (lr=1.4333e-04) (hash(x)=75490845)
+8490 train 3.542517 (lr=1.4309e-04) (hash(x)=78619557)
+8500 val loss 3.6124
+8500 val perplexity 37.0539
+8500 train 3.557504 (lr=1.4285e-04) (hash(x)=70959421)
+8510 train 3.642726 (lr=1.4263e-04) (hash(x)=75975416)
+8520 train 3.353360 (lr=1.4241e-04) (hash(x)=76336317)
+8530 train 3.616197 (lr=1.4221e-04) (hash(x)=67836668)
+8540 train 3.580212 (lr=1.4201e-04) (hash(x)=71139919)
+8550 train 3.824033 (lr=1.4183e-04) (hash(x)=69476711)
+8560 train 3.572627 (lr=1.4165e-04) (hash(x)=75144101)
+8570 train 3.522046 (lr=1.4148e-04) (hash(x)=76246282)
+8580 train 3.621476 (lr=1.4132e-04) (hash(x)=69466257)
+8590 train 3.593442 (lr=1.4117e-04) (hash(x)=71601133)
+8600 val loss 3.6127
+8600 val perplexity 37.0656
+8600 train 3.560265 (lr=1.4103e-04) (hash(x)=72319507)
+8610 train 3.698308 (lr=1.4090e-04) (hash(x)=74618018)
+8620 train 3.548955 (lr=1.4077e-04) (hash(x)=76115665)
+8630 train 3.708197 (lr=1.4066e-04) (hash(x)=82063941)
+8640 train 3.675837 (lr=1.4055e-04) (hash(x)=79867543)
+8650 train 3.478401 (lr=1.4046e-04) (hash(x)=73802504)
+8660 train 3.584822 (lr=1.4037e-04) (hash(x)=74251531)
+8670 train 3.569309 (lr=1.4029e-04) (hash(x)=74637655)
+8680 train 3.660057 (lr=1.4022e-04) (hash(x)=82021614)
+8690 train 3.675864 (lr=1.4016e-04) (hash(x)=77938695)
+8700 val loss 3.6043
+8700 val perplexity 36.7569
+8700 train 3.712452 (lr=1.4011e-04) (hash(x)=72431581)
+8710 train 3.728469 (lr=1.4007e-04) (hash(x)=74875522)
+8720 train 3.667145 (lr=1.4004e-04) (hash(x)=69574551)
+8730 train 3.504078 (lr=1.4002e-04) (hash(x)=71568849)
+8740 train 3.744762 (lr=1.4000e-04) (hash(x)=77714444)
+8749 val loss 3.6038
+8749 val perplexity 36.7384

n_heads12_lr14e-4_total_batch_size61440_seed1339/model_08749.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15bf69800f3f4c1047f5c1e2cda5145a18e8015c11594292e19de7f972feea22
+size 498607490

n_heads12_lr14e-4_total_batch_size61440_seed1339/optimizer_08749.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:41094f24bc3ecf7e0f01276daf26c1a588650006452f05e1f5cbeb726cd6d216
+size 990934406