Upload folder using huggingface_hub

n_heads12_lr10e-4_total_batch_size61440_seed1339/args.json

@@ -0,0 +1 @@
+{"hellaswag": true, "attention_kind": "selective", "log_dir": "wider_is_better_11/n_heads12_lr10e-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": 0.001, "decay_lr": true, "readout_zero_init": false, "query_zero_init": false, "l1_loss": false, "debugpy": false, "key": "n_heads12_10e-4_61440", "n_embd": 768}

n_heads12_lr10e-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_lr10e-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=2.0000e-06) (hash(x)=84111305)
+10 train 9.805919 (lr=2.2000e-05) (hash(x)=76061209)
+20 train 9.400036 (lr=4.2000e-05) (hash(x)=72815142)
+30 train 9.092428 (lr=6.2000e-05) (hash(x)=74281158)
+40 train 8.515334 (lr=8.2000e-05) (hash(x)=70759787)
+50 train 8.091492 (lr=1.0200e-04) (hash(x)=67498044)
+60 train 7.754110 (lr=1.2200e-04) (hash(x)=74288138)
+70 train 7.565042 (lr=1.4200e-04) (hash(x)=72712110)
+80 train 7.567488 (lr=1.6200e-04) (hash(x)=76893978)
+90 train 7.350292 (lr=1.8200e-04) (hash(x)=69479499)
+100 val loss 7.2845
+100 val perplexity 1457.4838
+100 train 7.251658 (lr=2.0200e-04) (hash(x)=78805238)
+110 train 7.136649 (lr=2.2200e-04) (hash(x)=72379905)
+120 train 7.046024 (lr=2.4200e-04) (hash(x)=84941914)
+130 train 6.839346 (lr=2.6200e-04) (hash(x)=74520609)
+140 train 6.884595 (lr=2.8200e-04) (hash(x)=78600114)
+150 train 6.565871 (lr=3.0200e-04) (hash(x)=71364632)
+160 train 6.555339 (lr=3.2200e-04) (hash(x)=88692902)
+170 train 6.612408 (lr=3.4200e-04) (hash(x)=76813970)
+180 train 6.464255 (lr=3.6200e-04) (hash(x)=78277546)
+190 train 6.393544 (lr=3.8200e-04) (hash(x)=69274625)
+200 val loss 6.4242
+200 val perplexity 616.6071
+200 train 6.346609 (lr=4.0200e-04) (hash(x)=74752080)
+210 train 6.241256 (lr=4.2200e-04) (hash(x)=73366585)
+220 train 6.225573 (lr=4.4200e-04) (hash(x)=82441512)
+230 train 5.997927 (lr=4.6200e-04) (hash(x)=70370492)
+240 train 6.042721 (lr=4.8200e-04) (hash(x)=76271528)
+250 train 5.975184 (lr=5.0200e-04) (hash(x)=70391215)
+260 train 5.895554 (lr=5.2200e-04) (hash(x)=70932307)
+270 train 5.837888 (lr=5.4200e-04) (hash(x)=73903274)
+280 train 5.877750 (lr=5.6200e-04) (hash(x)=72833455)
+290 train 5.920525 (lr=5.8200e-04) (hash(x)=86446336)
+300 val loss 5.9734
+300 val perplexity 392.8230
+300 train 5.708982 (lr=6.0200e-04) (hash(x)=72231879)
+310 train 5.837598 (lr=6.2200e-04) (hash(x)=83289406)
+320 train 5.907551 (lr=6.4200e-04) (hash(x)=79292084)
+330 train 6.068362 (lr=6.6200e-04) (hash(x)=81963141)
+340 train 5.940867 (lr=6.8200e-04) (hash(x)=76688193)
+350 train 5.745305 (lr=7.0200e-04) (hash(x)=67974208)
+360 train 5.920948 (lr=7.2200e-04) (hash(x)=71759924)
+370 train 5.829780 (lr=7.4200e-04) (hash(x)=82538335)
+380 train 5.688867 (lr=7.6200e-04) (hash(x)=73719341)
+390 train 5.840536 (lr=7.8200e-04) (hash(x)=81731118)
+400 val loss 5.6736
+400 val perplexity 291.0920
+400 train 5.625833 (lr=8.0200e-04) (hash(x)=65538392)
+410 train 5.602108 (lr=8.2200e-04) (hash(x)=73037722)
+420 train 5.507573 (lr=8.4200e-04) (hash(x)=71698939)
+430 train 5.731693 (lr=8.6200e-04) (hash(x)=89649867)
+440 train 5.827013 (lr=8.8200e-04) (hash(x)=59262658)
+450 train 5.457302 (lr=9.0200e-04) (hash(x)=79057900)
+460 train 5.476742 (lr=9.2200e-04) (hash(x)=68890702)
+470 train 5.486958 (lr=9.4200e-04) (hash(x)=69474321)
+480 train 5.423604 (lr=9.6200e-04) (hash(x)=74715332)
+490 train 5.542636 (lr=9.8200e-04) (hash(x)=81929495)
+500 val loss 5.4476
+500 val perplexity 232.2077
+500 train 5.598450 (lr=1.0000e-03) (hash(x)=81405581)
+510 train 5.197465 (lr=1.0000e-03) (hash(x)=66596183)
+520 train 5.512326 (lr=9.9999e-04) (hash(x)=75298362)
+530 train 5.317410 (lr=9.9997e-04) (hash(x)=78449774)
+540 train 5.295600 (lr=9.9995e-04) (hash(x)=75749875)
+550 train 5.244550 (lr=9.9992e-04) (hash(x)=75371538)
+560 train 5.180340 (lr=9.9988e-04) (hash(x)=74708150)
+570 train 5.201348 (lr=9.9984e-04) (hash(x)=75984401)
+580 train 5.170445 (lr=9.9979e-04) (hash(x)=78930428)
+590 train 5.445223 (lr=9.9974e-04) (hash(x)=70858901)
+600 val loss 5.2832
+600 val perplexity 196.9989
+600 train 5.174394 (lr=9.9967e-04) (hash(x)=86811215)
+610 train 5.040601 (lr=9.9961e-04) (hash(x)=74537543)
+620 train 5.022754 (lr=9.9953e-04) (hash(x)=64425159)
+630 train 4.906867 (lr=9.9945e-04) (hash(x)=72575441)
+640 train 4.915975 (lr=9.9936e-04) (hash(x)=77886921)
+650 train 5.029866 (lr=9.9927e-04) (hash(x)=73649304)
+660 train 4.797788 (lr=9.9917e-04) (hash(x)=64469482)
+670 train 5.179537 (lr=9.9906e-04) (hash(x)=76580403)
+680 train 4.828611 (lr=9.9894e-04) (hash(x)=60374603)
+690 train 4.978015 (lr=9.9882e-04) (hash(x)=69827901)
+700 val loss 5.1498
+700 val perplexity 172.4033
+700 train 5.172194 (lr=9.9870e-04) (hash(x)=77235911)
+710 train 5.141053 (lr=9.9856e-04) (hash(x)=76598106)
+720 train 5.205838 (lr=9.9842e-04) (hash(x)=75315013)
+730 train 5.049296 (lr=9.9828e-04) (hash(x)=76940932)
+740 train 5.056776 (lr=9.9812e-04) (hash(x)=75540179)
+750 train 5.034538 (lr=9.9796e-04) (hash(x)=73283995)
+760 train 5.066340 (lr=9.9780e-04) (hash(x)=84096065)
+770 train 4.845884 (lr=9.9762e-04) (hash(x)=75832016)
+780 train 5.015874 (lr=9.9744e-04) (hash(x)=65611269)
+790 train 5.068170 (lr=9.9726e-04) (hash(x)=72352412)
+800 val loss 4.9891
+800 val perplexity 146.8019
+800 train 4.976930 (lr=9.9707e-04) (hash(x)=78191559)
+810 train 4.909904 (lr=9.9687e-04) (hash(x)=73255448)
+820 train 5.004680 (lr=9.9666e-04) (hash(x)=75336936)
+830 train 4.806792 (lr=9.9645e-04) (hash(x)=71181581)
+840 train 5.452914 (lr=9.9623e-04) (hash(x)=76217477)
+850 train 4.890632 (lr=9.9601e-04) (hash(x)=75090324)
+860 train 4.833434 (lr=9.9578e-04) (hash(x)=71125083)
+870 train 4.843399 (lr=9.9554e-04) (hash(x)=74444188)
+880 train 4.875022 (lr=9.9530e-04) (hash(x)=72143748)
+890 train 4.736992 (lr=9.9505e-04) (hash(x)=75638384)
+900 val loss 4.8579
+900 val perplexity 128.7546
+900 train 4.889030 (lr=9.9479e-04) (hash(x)=76967190)
+910 train 4.831389 (lr=9.9453e-04) (hash(x)=84611316)
+920 train 4.618568 (lr=9.9426e-04) (hash(x)=61365145)
+930 train 4.599793 (lr=9.9398e-04) (hash(x)=54695617)
+940 train 4.798724 (lr=9.9370e-04) (hash(x)=76251415)
+950 train 4.576065 (lr=9.9341e-04) (hash(x)=71503060)
+960 train 4.776672 (lr=9.9311e-04) (hash(x)=70925251)
+970 train 4.712270 (lr=9.9281e-04) (hash(x)=69495443)
+980 train 4.673240 (lr=9.9250e-04) (hash(x)=74994053)
+990 train 4.513422 (lr=9.9219e-04) (hash(x)=81245808)
+1000 val loss 4.7651
+1000 val perplexity 117.3448
+1000 train 4.737851 (lr=9.9187e-04) (hash(x)=73801537)
+1010 train 4.671319 (lr=9.9154e-04) (hash(x)=78784929)
+1020 train 4.420188 (lr=9.9121e-04) (hash(x)=84649514)
+1030 train 4.553526 (lr=9.9087e-04) (hash(x)=71200051)
+1040 train 4.421772 (lr=9.9052e-04) (hash(x)=68279861)
+1050 train 4.477155 (lr=9.9017e-04) (hash(x)=86748930)
+1060 train 4.336933 (lr=9.8981e-04) (hash(x)=71319577)
+1070 train 4.471417 (lr=9.8944e-04) (hash(x)=72682588)
+1080 train 4.437065 (lr=9.8907e-04) (hash(x)=71408336)
+1090 train 4.631516 (lr=9.8869e-04) (hash(x)=78863558)
+1100 val loss 4.6624
+1100 val perplexity 105.8934
+1100 train 4.911688 (lr=9.8831e-04) (hash(x)=75285186)
+1110 train 4.705091 (lr=9.8791e-04) (hash(x)=78202715)
+1120 train 4.503892 (lr=9.8752e-04) (hash(x)=75187925)
+1130 train 4.696259 (lr=9.8711e-04) (hash(x)=70507475)
+1140 train 4.596849 (lr=9.8670e-04) (hash(x)=77911876)
+1150 train 4.524327 (lr=9.8629e-04) (hash(x)=77470391)
+1160 train 4.520265 (lr=9.8586e-04) (hash(x)=76266017)
+1170 train 4.806881 (lr=9.8543e-04) (hash(x)=77505346)
+1180 train 4.600049 (lr=9.8500e-04) (hash(x)=78978095)
+1190 train 4.656416 (lr=9.8456e-04) (hash(x)=78602083)
+1200 val loss 4.5877
+1200 val perplexity 98.2684
+1200 train 4.681677 (lr=9.8411e-04) (hash(x)=81700834)
+1210 train 4.565983 (lr=9.8365e-04) (hash(x)=75776079)
+1220 train 4.549867 (lr=9.8319e-04) (hash(x)=95908213)
+1230 train 4.471623 (lr=9.8272e-04) (hash(x)=84269607)
+1240 train 4.567751 (lr=9.8225e-04) (hash(x)=69460140)
+1250 train 4.573816 (lr=9.8177e-04) (hash(x)=71522595)
+1260 train 4.603982 (lr=9.8129e-04) (hash(x)=81155508)
+1270 train 4.737506 (lr=9.8079e-04) (hash(x)=72842244)
+1280 train 4.575050 (lr=9.8030e-04) (hash(x)=80034345)
+1290 train 4.428643 (lr=9.7979e-04) (hash(x)=70948511)
+1300 val loss 4.5141
+1300 val perplexity 91.2984
+1300 train 4.476729 (lr=9.7928e-04) (hash(x)=69667983)
+1310 train 4.517889 (lr=9.7876e-04) (hash(x)=84077906)
+1320 train 4.234534 (lr=9.7824e-04) (hash(x)=67481757)
+1330 train 4.403744 (lr=9.7771e-04) (hash(x)=83761654)
+1340 train 4.505623 (lr=9.7717e-04) (hash(x)=76630114)
+1350 train 4.533320 (lr=9.7663e-04) (hash(x)=81080698)
+1360 train 4.420274 (lr=9.7608e-04) (hash(x)=71019125)
+1370 train 4.410006 (lr=9.7553e-04) (hash(x)=78550770)
+1380 train 4.418458 (lr=9.7497e-04) (hash(x)=67017249)
+1390 train 4.331968 (lr=9.7440e-04) (hash(x)=75430624)
+1400 val loss 4.4904
+1400 val perplexity 89.1552
+1400 train 4.436020 (lr=9.7383e-04) (hash(x)=72720863)
+1410 train 4.178013 (lr=9.7325e-04) (hash(x)=73824272)
+1420 train 4.305560 (lr=9.7267e-04) (hash(x)=69424571)
+1430 train 4.268951 (lr=9.7207e-04) (hash(x)=71030131)
+1440 train 4.217274 (lr=9.7148e-04) (hash(x)=80088993)
+1450 train 4.168185 (lr=9.7087e-04) (hash(x)=74332002)
+1460 train 4.093929 (lr=9.7026e-04) (hash(x)=75132184)
+1470 train 4.149832 (lr=9.6965e-04) (hash(x)=72644858)
+1480 train 4.152047 (lr=9.6903e-04) (hash(x)=70677454)
+1490 train 4.173537 (lr=9.6840e-04) (hash(x)=74408093)
+1500 val loss 4.4340
+1500 val perplexity 84.2718
+1500 train 4.407459 (lr=9.6777e-04) (hash(x)=82777829)
+1510 train 4.494979 (lr=9.6713e-04) (hash(x)=82481742)
+1520 train 4.431778 (lr=9.6648e-04) (hash(x)=81580164)
+1530 train 4.399837 (lr=9.6583e-04) (hash(x)=75892741)
+1540 train 4.336584 (lr=9.6517e-04) (hash(x)=70574468)
+1550 train 4.377357 (lr=9.6451e-04) (hash(x)=77451919)
+1560 train 4.380383 (lr=9.6384e-04) (hash(x)=73148801)
+1570 train 4.290299 (lr=9.6316e-04) (hash(x)=74661859)
+1580 train 4.432652 (lr=9.6248e-04) (hash(x)=85484059)
+1590 train 4.553872 (lr=9.6179e-04) (hash(x)=72848788)
+1600 val loss 4.3532
+1600 val perplexity 77.7236
+1600 train 4.358281 (lr=9.6110e-04) (hash(x)=79468408)
+1610 train 4.365561 (lr=9.6040e-04) (hash(x)=78654805)
+1620 train 4.405781 (lr=9.5969e-04) (hash(x)=72348939)
+1630 train 4.464393 (lr=9.5898e-04) (hash(x)=74107545)
+1640 train 4.529867 (lr=9.5826e-04) (hash(x)=75470148)
+1650 train 4.414195 (lr=9.5754e-04) (hash(x)=74478093)
+1660 train 4.320990 (lr=9.5681e-04) (hash(x)=74077733)
+1670 train 4.218833 (lr=9.5607e-04) (hash(x)=60922881)
+1680 train 4.476503 (lr=9.5533e-04) (hash(x)=79401992)
+1690 train 4.411097 (lr=9.5458e-04) (hash(x)=75405453)
+1700 val loss 4.3546
+1700 val perplexity 77.8328
+1700 train 4.361378 (lr=9.5383e-04) (hash(x)=67378635)
+1710 train 4.272304 (lr=9.5307e-04) (hash(x)=74327780)
+1720 train 4.447429 (lr=9.5231e-04) (hash(x)=83288144)
+1730 train 4.192892 (lr=9.5153e-04) (hash(x)=68094391)
+1740 train 4.245912 (lr=9.5076e-04) (hash(x)=74943031)
+1750 train 4.289745 (lr=9.4998e-04) (hash(x)=77148592)
+1760 train 4.265447 (lr=9.4919e-04) (hash(x)=71352508)
+1770 train 4.281930 (lr=9.4839e-04) (hash(x)=73819276)
+1780 train 4.179723 (lr=9.4759e-04) (hash(x)=75338495)
+1790 train 4.386617 (lr=9.4679e-04) (hash(x)=100973274)
+1800 val loss 4.3184
+1800 val perplexity 75.0707
+1800 train 4.403097 (lr=9.4598e-04) (hash(x)=82399749)
+1810 train 4.419158 (lr=9.4516e-04) (hash(x)=80859117)
+1820 train 4.453051 (lr=9.4434e-04) (hash(x)=74215317)
+1830 train 4.312271 (lr=9.4351e-04) (hash(x)=73527133)
+1840 train 4.476528 (lr=9.4268e-04) (hash(x)=70049534)
+1850 train 4.195002 (lr=9.4184e-04) (hash(x)=71891328)
+1860 train 4.443893 (lr=9.4099e-04) (hash(x)=76433977)
+1870 train 4.414962 (lr=9.4014e-04) (hash(x)=69103302)
+1880 train 4.432602 (lr=9.3928e-04) (hash(x)=73328951)
+1890 train 4.188231 (lr=9.3842e-04) (hash(x)=77718295)
+1900 val loss 4.2715
+1900 val perplexity 71.6254
+1900 train 4.375031 (lr=9.3755e-04) (hash(x)=74708596)
+1910 train 4.330207 (lr=9.3668e-04) (hash(x)=80918764)
+1920 train 4.263230 (lr=9.3580e-04) (hash(x)=77248658)
+1930 train 4.364061 (lr=9.3491e-04) (hash(x)=73945895)
+1940 train 4.235381 (lr=9.3402e-04) (hash(x)=72252382)
+1950 train 4.291055 (lr=9.3313e-04) (hash(x)=87410072)
+1960 train 4.343136 (lr=9.3223e-04) (hash(x)=79408896)
+1970 train 4.129812 (lr=9.3132e-04) (hash(x)=78848496)
+1980 train 4.207821 (lr=9.3041e-04) (hash(x)=77601406)
+1990 train 4.165915 (lr=9.2949e-04) (hash(x)=71775143)
+2000 val loss 4.2275
+2000 val perplexity 68.5449
+2000 train 4.252492 (lr=9.2856e-04) (hash(x)=73721824)
+2010 train 4.213508 (lr=9.2763e-04) (hash(x)=74757670)
+2020 train 4.272916 (lr=9.2670e-04) (hash(x)=86234926)
+2030 train 4.457973 (lr=9.2576e-04) (hash(x)=69882050)
+2040 train 4.181327 (lr=9.2481e-04) (hash(x)=81004021)
+2050 train 4.073867 (lr=9.2386e-04) (hash(x)=75592124)
+2060 train 4.002054 (lr=9.2291e-04) (hash(x)=74811474)
+2070 train 4.194228 (lr=9.2195e-04) (hash(x)=76209571)
+2080 train 4.145471 (lr=9.2098e-04) (hash(x)=74973104)
+2090 train 4.198992 (lr=9.2001e-04) (hash(x)=74977086)
+2100 val loss 4.2534
+2100 val perplexity 70.3429
+2100 train 4.148748 (lr=9.1903e-04) (hash(x)=79579234)
+2110 train 4.098315 (lr=9.1804e-04) (hash(x)=73442407)
+2120 train 4.274135 (lr=9.1706e-04) (hash(x)=80281222)
+2130 train 4.217751 (lr=9.1606e-04) (hash(x)=83139613)
+2140 train 4.273276 (lr=9.1506e-04) (hash(x)=69936205)
+2150 train 4.316131 (lr=9.1406e-04) (hash(x)=74768225)
+2160 train 4.338952 (lr=9.1305e-04) (hash(x)=67473891)
+2170 train 4.290520 (lr=9.1203e-04) (hash(x)=80351641)
+2180 train 4.299978 (lr=9.1101e-04) (hash(x)=76303033)
+2190 train 4.231803 (lr=9.0999e-04) (hash(x)=73585511)
+2200 val loss 4.1864
+2200 val perplexity 65.7852
+2200 train 4.158675 (lr=9.0896e-04) (hash(x)=73195726)
+2210 train 4.311562 (lr=9.0792e-04) (hash(x)=80614640)
+2220 train 4.077424 (lr=9.0688e-04) (hash(x)=73107823)
+2230 train 4.153723 (lr=9.0583e-04) (hash(x)=78921188)
+2240 train 4.173172 (lr=9.0478e-04) (hash(x)=82527812)
+2250 train 4.088119 (lr=9.0372e-04) (hash(x)=75909041)
+2260 train 4.266368 (lr=9.0266e-04) (hash(x)=75266897)
+2270 train 4.214224 (lr=9.0160e-04) (hash(x)=74444372)
+2280 train 4.073836 (lr=9.0052e-04) (hash(x)=71643352)
+2290 train 4.251841 (lr=8.9945e-04) (hash(x)=73081122)
+2300 val loss 4.1653
+2300 val perplexity 64.4138
+2300 train 4.039475 (lr=8.9836e-04) (hash(x)=73133506)
+2310 train 4.099180 (lr=8.9728e-04) (hash(x)=73796905)
+2320 train 4.075785 (lr=8.9618e-04) (hash(x)=79133958)
+2330 train 4.096443 (lr=8.9509e-04) (hash(x)=79177833)
+2340 train 4.130397 (lr=8.9398e-04) (hash(x)=72790081)
+2350 train 4.110607 (lr=8.9288e-04) (hash(x)=65070947)
+2360 train 4.264915 (lr=8.9177e-04) (hash(x)=73178485)
+2370 train 4.219382 (lr=8.9065e-04) (hash(x)=71904548)
+2380 train 4.043080 (lr=8.8953e-04) (hash(x)=73416656)
+2390 train 4.184124 (lr=8.8840e-04) (hash(x)=71563647)
+2400 val loss 4.1734
+2400 val perplexity 64.9370
+2400 train 4.261228 (lr=8.8727e-04) (hash(x)=82236758)
+2410 train 4.063186 (lr=8.8613e-04) (hash(x)=76229570)
+2420 train 4.272271 (lr=8.8499e-04) (hash(x)=71056939)
+2430 train 4.122241 (lr=8.8384e-04) (hash(x)=76280968)
+2440 train 3.875177 (lr=8.8269e-04) (hash(x)=84199966)
+2450 train 4.299001 (lr=8.8153e-04) (hash(x)=84985998)
+2460 train 4.048522 (lr=8.8037e-04) (hash(x)=94074119)
+2470 train 4.247514 (lr=8.7921e-04) (hash(x)=81246200)
+2480 train 4.191588 (lr=8.7804e-04) (hash(x)=76225235)
+2490 train 4.119070 (lr=8.7686e-04) (hash(x)=76541445)
+2500 val loss 4.1162
+2500 val perplexity 61.3264
+2500 train 4.196284 (lr=8.7568e-04) (hash(x)=77708978)
+2510 train 4.066960 (lr=8.7450e-04) (hash(x)=75319253)
+2520 train 4.092705 (lr=8.7331e-04) (hash(x)=67190186)
+2530 train 4.271063 (lr=8.7211e-04) (hash(x)=82496813)
+2540 train 4.037810 (lr=8.7091e-04) (hash(x)=78100449)
+2550 train 4.103054 (lr=8.6971e-04) (hash(x)=80190873)
+2560 train 4.018284 (lr=8.6850e-04) (hash(x)=77711923)
+2570 train 4.132126 (lr=8.6729e-04) (hash(x)=79825674)
+2580 train 3.981089 (lr=8.6607e-04) (hash(x)=77024199)
+2590 train 4.059968 (lr=8.6485e-04) (hash(x)=72139372)
+2600 val loss 4.1028
+2600 val perplexity 60.5065
+2600 train 4.015438 (lr=8.6362e-04) (hash(x)=71362290)
+2610 train 4.152131 (lr=8.6239e-04) (hash(x)=80679759)
+2620 train 3.976839 (lr=8.6116e-04) (hash(x)=75079422)
+2630 train 4.085567 (lr=8.5991e-04) (hash(x)=70323867)
+2640 train 3.985546 (lr=8.5867e-04) (hash(x)=81444590)
+2650 train 4.009567 (lr=8.5742e-04) (hash(x)=71086644)
+2660 train 3.953842 (lr=8.5617e-04) (hash(x)=73339459)
+2670 train 4.129621 (lr=8.5491e-04) (hash(x)=74666974)
+2680 train 4.159144 (lr=8.5365e-04) (hash(x)=79117538)
+2690 train 4.154873 (lr=8.5238e-04) (hash(x)=65928715)
+2700 val loss 4.0972
+2700 val perplexity 60.1725
+2700 train 4.159644 (lr=8.5111e-04) (hash(x)=79416673)
+2710 train 4.004745 (lr=8.4983e-04) (hash(x)=65110265)
+2720 train 4.219120 (lr=8.4855e-04) (hash(x)=60942352)
+2730 train 4.182342 (lr=8.4727e-04) (hash(x)=68786492)
+2740 train 4.071064 (lr=8.4598e-04) (hash(x)=71480596)
+2750 train 4.041572 (lr=8.4469e-04) (hash(x)=84054551)
+2760 train 4.004358 (lr=8.4339e-04) (hash(x)=94608646)
+2770 train 4.083294 (lr=8.4209e-04) (hash(x)=73708388)
+2780 train 4.005244 (lr=8.4078e-04) (hash(x)=88054328)
+2790 train 4.172704 (lr=8.3947e-04) (hash(x)=88920089)
+2800 val loss 4.0637
+2800 val perplexity 58.1873
+2800 train 4.058393 (lr=8.3816e-04) (hash(x)=80878641)
+2810 train 4.101612 (lr=8.3684e-04) (hash(x)=71310778)
+2820 train 4.096463 (lr=8.3552e-04) (hash(x)=69986577)
+2830 train 4.073813 (lr=8.3419e-04) (hash(x)=78866280)
+2840 train 4.161679 (lr=8.3286e-04) (hash(x)=84135925)
+2850 train 4.157485 (lr=8.3153e-04) (hash(x)=74153539)
+2860 train 4.047184 (lr=8.3019e-04) (hash(x)=74749949)
+2870 train 4.132497 (lr=8.2884e-04) (hash(x)=72237368)
+2880 train 3.981327 (lr=8.2750e-04) (hash(x)=77118077)
+2890 train 4.082878 (lr=8.2615e-04) (hash(x)=73010418)
+2900 val loss 4.0431
+2900 val perplexity 57.0048
+2900 train 3.968037 (lr=8.2479e-04) (hash(x)=70624604)
+2910 train 3.979253 (lr=8.2343e-04) (hash(x)=72862034)
+2920 train 3.907590 (lr=8.2207e-04) (hash(x)=74542687)
+2930 train 4.108903 (lr=8.2070e-04) (hash(x)=71424119)
+2940 train 4.007574 (lr=8.1933e-04) (hash(x)=68581990)
+2950 train 4.033174 (lr=8.1796e-04) (hash(x)=75184497)
+2960 train 4.161138 (lr=8.1658e-04) (hash(x)=79104818)
+2970 train 3.949968 (lr=8.1520e-04) (hash(x)=75872591)
+2980 train 4.025938 (lr=8.1381e-04) (hash(x)=78565892)
+2990 train 3.973404 (lr=8.1242e-04) (hash(x)=77347407)
+3000 val loss 4.0543
+3000 val perplexity 57.6469
+3000 train 3.968811 (lr=8.1103e-04) (hash(x)=73568709)
+3010 train 3.954822 (lr=8.0963e-04) (hash(x)=83954777)
+3020 train 3.950898 (lr=8.0823e-04) (hash(x)=91325052)
+3030 train 4.182109 (lr=8.0682e-04) (hash(x)=73073624)
+3040 train 4.013264 (lr=8.0541e-04) (hash(x)=78263491)
+3050 train 4.059776 (lr=8.0400e-04) (hash(x)=75006046)
+3060 train 3.809413 (lr=8.0258e-04) (hash(x)=77283473)
+3070 train 4.167386 (lr=8.0116e-04) (hash(x)=70498330)
+3080 train 4.015658 (lr=7.9974e-04) (hash(x)=74632023)
+3090 train 4.016079 (lr=7.9831e-04) (hash(x)=71582163)
+3100 val loss 4.0139
+3100 val perplexity 55.3628
+3100 train 4.008207 (lr=7.9688e-04) (hash(x)=76032144)
+3110 train 4.085329 (lr=7.9545e-04) (hash(x)=79935326)
+3120 train 4.104363 (lr=7.9401e-04) (hash(x)=81668940)
+3130 train 4.029022 (lr=7.9257e-04) (hash(x)=81530706)
+3140 train 4.021668 (lr=7.9112e-04) (hash(x)=75958994)
+3150 train 4.108933 (lr=7.8967e-04) (hash(x)=70375358)
+3160 train 4.026141 (lr=7.8822e-04) (hash(x)=81742382)
+3170 train 4.008487 (lr=7.8677e-04) (hash(x)=73892129)
+3180 train 4.024062 (lr=7.8531e-04) (hash(x)=69584996)
+3190 train 4.003625 (lr=7.8384e-04) (hash(x)=74711675)
+3200 val loss 4.0076
+3200 val perplexity 55.0135
+3200 train 4.113000 (lr=7.8238e-04) (hash(x)=72943579)
+3210 train 4.072075 (lr=7.8091e-04) (hash(x)=76110656)
+3220 train 3.854102 (lr=7.7944e-04) (hash(x)=70203094)
+3230 train 4.005814 (lr=7.7796e-04) (hash(x)=60287090)
+3240 train 3.941430 (lr=7.7648e-04) (hash(x)=74302555)
+3250 train 4.046152 (lr=7.7500e-04) (hash(x)=72366702)
+3260 train 3.946336 (lr=7.7351e-04) (hash(x)=73014004)
+3270 train 3.832401 (lr=7.7203e-04) (hash(x)=75491051)
+3280 train 4.064177 (lr=7.7053e-04) (hash(x)=82821096)
+3290 train 3.889500 (lr=7.6904e-04) (hash(x)=77913090)
+3300 val loss 4.0103
+3300 val perplexity 55.1629
+3300 train 4.036943 (lr=7.6754e-04) (hash(x)=72779327)
+3310 train 3.965623 (lr=7.6604e-04) (hash(x)=76032908)
+3320 train 3.836689 (lr=7.6453e-04) (hash(x)=66875476)
+3330 train 3.896916 (lr=7.6303e-04) (hash(x)=67498762)
+3340 train 4.034355 (lr=7.6151e-04) (hash(x)=77920169)
+3350 train 4.027437 (lr=7.6000e-04) (hash(x)=87527839)
+3360 train 3.859725 (lr=7.5848e-04) (hash(x)=78010101)
+3370 train 4.070470 (lr=7.5696e-04) (hash(x)=76688335)
+3380 train 4.059451 (lr=7.5544e-04) (hash(x)=80832602)
+3390 train 3.905339 (lr=7.5391e-04) (hash(x)=75944307)
+3400 val loss 3.9887
+3400 val perplexity 53.9863
+3400 train 4.184787 (lr=7.5238e-04) (hash(x)=67516962)
+3410 train 4.042036 (lr=7.5085e-04) (hash(x)=72960118)
+3420 train 4.062222 (lr=7.4932e-04) (hash(x)=77405757)
+3430 train 4.091372 (lr=7.4778e-04) (hash(x)=76094785)
+3440 train 3.901554 (lr=7.4624e-04) (hash(x)=68336851)
+3450 train 3.893520 (lr=7.4470e-04) (hash(x)=71910579)
+3460 train 3.936821 (lr=7.4315e-04) (hash(x)=74170641)
+3470 train 4.033819 (lr=7.4160e-04) (hash(x)=75423851)
+3480 train 3.986952 (lr=7.4005e-04) (hash(x)=73367119)
+3490 train 3.853799 (lr=7.3849e-04) (hash(x)=75558326)
+3500 val loss 3.9811
+3500 val perplexity 53.5768
+3500 train 3.876295 (lr=7.3694e-04) (hash(x)=69187503)
+3510 train 3.986142 (lr=7.3538e-04) (hash(x)=80052583)
+3520 train 3.978205 (lr=7.3381e-04) (hash(x)=73819848)
+3530 train 3.798000 (lr=7.3225e-04) (hash(x)=71358103)
+3540 train 3.952610 (lr=7.3068e-04) (hash(x)=95551988)
+3550 train 4.004719 (lr=7.2911e-04) (hash(x)=75088121)
+3560 train 4.011243 (lr=7.2754e-04) (hash(x)=80473866)
+3570 train 4.026552 (lr=7.2596e-04) (hash(x)=73228665)
+3580 train 3.925741 (lr=7.2438e-04) (hash(x)=76321895)
+3590 train 3.849825 (lr=7.2280e-04) (hash(x)=74530382)
+3600 val loss 3.9675
+3600 val perplexity 52.8528
+3600 train 3.929077 (lr=7.2122e-04) (hash(x)=69686188)
+3610 train 3.960974 (lr=7.1963e-04) (hash(x)=66842873)
+3620 train 3.952906 (lr=7.1804e-04) (hash(x)=74921650)
+3630 train 3.850386 (lr=7.1645e-04) (hash(x)=79834656)
+3640 train 3.970330 (lr=7.1486e-04) (hash(x)=75122572)
+3650 train 3.903623 (lr=7.1326e-04) (hash(x)=77134480)
+3660 train 4.063825 (lr=7.1167e-04) (hash(x)=73027661)
+3670 train 4.118287 (lr=7.1007e-04) (hash(x)=85366911)
+3680 train 4.000724 (lr=7.0846e-04) (hash(x)=75873712)
+3690 train 4.092496 (lr=7.0686e-04) (hash(x)=81324159)
+3700 val loss 3.9540
+3700 val perplexity 52.1413
+3700 train 3.987675 (lr=7.0525e-04) (hash(x)=75354374)
+3710 train 4.039375 (lr=7.0364e-04) (hash(x)=73820036)
+3720 train 4.044855 (lr=7.0203e-04) (hash(x)=69279349)
+3730 train 4.011619 (lr=7.0041e-04) (hash(x)=76885056)
+3740 train 3.887223 (lr=6.9880e-04) (hash(x)=66083153)
+3750 train 3.886369 (lr=6.9718e-04) (hash(x)=82585096)
+3760 train 3.855583 (lr=6.9556e-04) (hash(x)=74734456)
+3770 train 3.844720 (lr=6.9394e-04) (hash(x)=74077986)
+3780 train 3.848018 (lr=6.9231e-04) (hash(x)=73794394)
+3790 train 3.875733 (lr=6.9069e-04) (hash(x)=75877136)
+3800 val loss 3.9414
+3800 val perplexity 51.4908
+3800 train 3.972227 (lr=6.8906e-04) (hash(x)=72150028)
+3810 train 4.061382 (lr=6.8743e-04) (hash(x)=77316418)
+3820 train 4.255692 (lr=6.8579e-04) (hash(x)=79951144)
+3830 train 3.941373 (lr=6.8416e-04) (hash(x)=72971714)
+3840 train 3.965837 (lr=6.8252e-04) (hash(x)=78168722)
+3850 train 3.798503 (lr=6.8088e-04) (hash(x)=72103764)
+3860 train 3.848412 (lr=6.7924e-04) (hash(x)=70399281)
+3870 train 4.032396 (lr=6.7760e-04) (hash(x)=71911595)
+3880 train 3.836545 (lr=6.7596e-04) (hash(x)=72034433)
+3890 train 3.831523 (lr=6.7431e-04) (hash(x)=73270524)
+3900 val loss 3.9355
+3900 val perplexity 51.1879
+3900 train 3.823839 (lr=6.7266e-04) (hash(x)=71934546)
+3910 train 3.987900 (lr=6.7101e-04) (hash(x)=78906659)
+3920 train 3.882963 (lr=6.6936e-04) (hash(x)=67852098)
+3930 train 3.845071 (lr=6.6771e-04) (hash(x)=75387628)
+3940 train 3.956375 (lr=6.6605e-04) (hash(x)=72407767)
+3950 train 3.814538 (lr=6.6440e-04) (hash(x)=66515238)
+3960 train 3.997642 (lr=6.6274e-04) (hash(x)=73337537)
+3970 train 3.994295 (lr=6.6108e-04) (hash(x)=74639467)
+3980 train 3.989024 (lr=6.5942e-04) (hash(x)=76044746)
+3990 train 4.003694 (lr=6.5776e-04) (hash(x)=71769001)
+4000 val loss 3.9121
+4000 val perplexity 50.0062
+4000 train 3.877454 (lr=6.5609e-04) (hash(x)=70257299)
+4010 train 3.928067 (lr=6.5443e-04) (hash(x)=71314416)
+4020 train 4.046541 (lr=6.5276e-04) (hash(x)=82315713)
+4030 train 3.980618 (lr=6.5109e-04) (hash(x)=65813944)
+4040 train 3.869294 (lr=6.4942e-04) (hash(x)=78277446)
+4050 train 3.924748 (lr=6.4775e-04) (hash(x)=77850258)
+4060 train 4.026044 (lr=6.4607e-04) (hash(x)=69511868)
+4070 train 3.850436 (lr=6.4440e-04) (hash(x)=77262761)
+4080 train 3.899817 (lr=6.4272e-04) (hash(x)=79691185)
+4090 train 3.718890 (lr=6.4104e-04) (hash(x)=74848196)
+4100 val loss 3.9099
+4100 val perplexity 49.8921
+4100 train 3.877767 (lr=6.3937e-04) (hash(x)=72761342)
+4110 train 3.972394 (lr=6.3769e-04) (hash(x)=58710996)
+4120 train 3.854548 (lr=6.3600e-04) (hash(x)=75794649)
+4130 train 3.973106 (lr=6.3432e-04) (hash(x)=72213336)
+4140 train 3.886581 (lr=6.3264e-04) (hash(x)=75842443)
+4150 train 3.799202 (lr=6.3095e-04) (hash(x)=70329458)
+4160 train 3.786905 (lr=6.2927e-04) (hash(x)=88536084)
+4170 train 3.945808 (lr=6.2758e-04) (hash(x)=76887790)
+4180 train 3.830776 (lr=6.2589e-04) (hash(x)=76755599)
+4190 train 3.837867 (lr=6.2420e-04) (hash(x)=71379783)
+4200 val loss 3.9094
+4200 val perplexity 49.8693
+4200 train 3.822284 (lr=6.2251e-04) (hash(x)=112576746)
+4210 train 3.733948 (lr=6.2082e-04) (hash(x)=77101027)
+4220 train 3.912746 (lr=6.1913e-04) (hash(x)=71607880)
+4230 train 3.823435 (lr=6.1743e-04) (hash(x)=72892070)
+4240 train 3.765296 (lr=6.1574e-04) (hash(x)=70967675)
+4250 train 3.862766 (lr=6.1404e-04) (hash(x)=75050831)
+4260 train 3.921880 (lr=6.1235e-04) (hash(x)=75249845)
+4270 train 3.913503 (lr=6.1065e-04) (hash(x)=82744946)
+4280 train 3.895804 (lr=6.0895e-04) (hash(x)=68857061)
+4290 train 3.884595 (lr=6.0725e-04) (hash(x)=80434125)
+4300 val loss 3.8900
+4300 val perplexity 48.9130
+4300 train 3.851652 (lr=6.0555e-04) (hash(x)=84624166)
+4310 train 3.781058 (lr=6.0385e-04) (hash(x)=60925156)
+4320 train 3.923411 (lr=6.0215e-04) (hash(x)=75956316)
+4330 train 3.841059 (lr=6.0044e-04) (hash(x)=74068056)
+4340 train 4.084041 (lr=5.9874e-04) (hash(x)=94977597)
+4350 train 3.969670 (lr=5.9704e-04) (hash(x)=71781944)
+4360 train 4.021977 (lr=5.9533e-04) (hash(x)=74308923)
+4370 train 3.867145 (lr=5.9363e-04) (hash(x)=76645666)
+4380 train 3.847078 (lr=5.9192e-04) (hash(x)=73326958)
+4390 train 3.835787 (lr=5.9022e-04) (hash(x)=74743749)
+4400 val loss 3.8791
+4400 val perplexity 48.3830
+4400 train 3.962578 (lr=5.8851e-04) (hash(x)=73116611)
+4410 train 3.736036 (lr=5.8680e-04) (hash(x)=68578568)
+4420 train 3.968826 (lr=5.8509e-04) (hash(x)=78085026)
+4430 train 3.853173 (lr=5.8338e-04) (hash(x)=71504765)
+4440 train 4.046264 (lr=5.8168e-04) (hash(x)=73163972)
+4450 train 3.864668 (lr=5.7997e-04) (hash(x)=81967324)
+4460 train 3.835377 (lr=5.7826e-04) (hash(x)=67796275)
+4470 train 3.865762 (lr=5.7655e-04) (hash(x)=75136849)
+4480 train 3.824352 (lr=5.7483e-04) (hash(x)=80109861)
+4490 train 3.750341 (lr=5.7312e-04) (hash(x)=72094345)
+4500 val loss 3.8688
+4500 val perplexity 47.8852
+4500 train 3.763227 (lr=5.7141e-04) (hash(x)=77660356)
+4510 train 3.744982 (lr=5.6970e-04) (hash(x)=71303771)
+4520 train 3.958481 (lr=5.6799e-04) (hash(x)=67353213)
+4530 train 3.589906 (lr=5.6628e-04) (hash(x)=91984280)
+4540 train 3.740797 (lr=5.6456e-04) (hash(x)=74918518)
+4550 train 3.839872 (lr=5.6285e-04) (hash(x)=80129084)
+4560 train 3.768359 (lr=5.6114e-04) (hash(x)=74762343)
+4570 train 3.735575 (lr=5.5942e-04) (hash(x)=83264256)
+4580 train 3.844501 (lr=5.5771e-04) (hash(x)=77701601)
+4590 train 3.778143 (lr=5.5600e-04) (hash(x)=72730921)
+4600 val loss 3.8617
+4600 val perplexity 47.5478
+4600 train 3.828333 (lr=5.5428e-04) (hash(x)=77801190)
+4610 train 3.899195 (lr=5.5257e-04) (hash(x)=64563241)
+4620 train 3.833240 (lr=5.5086e-04) (hash(x)=75462951)
+4630 train 3.881160 (lr=5.4914e-04) (hash(x)=74708327)
+4640 train 3.785066 (lr=5.4743e-04) (hash(x)=83297632)
+4650 train 3.894285 (lr=5.4572e-04) (hash(x)=81914758)
+4660 train 3.915007 (lr=5.4400e-04) (hash(x)=71367268)
+4670 train 3.795990 (lr=5.4229e-04) (hash(x)=73587555)
+4680 train 3.910823 (lr=5.4058e-04) (hash(x)=74734602)
+4690 train 4.010897 (lr=5.3886e-04) (hash(x)=77508704)
+4700 val loss 3.8466
+4700 val perplexity 46.8340
+4700 train 3.811380 (lr=5.3715e-04) (hash(x)=73228682)
+4710 train 3.835506 (lr=5.3544e-04) (hash(x)=73851956)
+4720 train 3.883692 (lr=5.3372e-04) (hash(x)=71244190)
+4730 train 3.646925 (lr=5.3201e-04) (hash(x)=74416580)
+4740 train 3.879030 (lr=5.3030e-04) (hash(x)=69259353)
+4750 train 3.850823 (lr=5.2859e-04) (hash(x)=77662887)
+4760 train 3.792641 (lr=5.2688e-04) (hash(x)=73001257)
+4770 train 3.744576 (lr=5.2517e-04) (hash(x)=81466432)
+4780 train 3.687579 (lr=5.2345e-04) (hash(x)=76710780)
+4790 train 3.838884 (lr=5.2174e-04) (hash(x)=72440698)
+4800 val loss 3.8486
+4800 val perplexity 46.9251
+4800 train 3.720203 (lr=5.2003e-04) (hash(x)=75002116)
+4810 train 3.776699 (lr=5.1832e-04) (hash(x)=84873693)
+4820 train 3.623518 (lr=5.1662e-04) (hash(x)=68563264)
+4830 train 3.741002 (lr=5.1491e-04) (hash(x)=76811982)
+4840 train 3.655573 (lr=5.1320e-04) (hash(x)=73763746)
+4850 train 3.868413 (lr=5.1149e-04) (hash(x)=72834169)
+4860 train 3.808540 (lr=5.0978e-04) (hash(x)=71420794)
+4870 train 3.927446 (lr=5.0808e-04) (hash(x)=82083784)
+4880 train 3.902657 (lr=5.0637e-04) (hash(x)=71399278)
+4890 train 3.962420 (lr=5.0467e-04) (hash(x)=80791916)
+4900 val loss 3.8318
+4900 val perplexity 46.1453
+4900 train 3.848641 (lr=5.0296e-04) (hash(x)=76275522)
+4910 train 4.007618 (lr=5.0126e-04) (hash(x)=73846113)
+4920 train 3.935615 (lr=4.9956e-04) (hash(x)=65906497)
+4930 train 3.739518 (lr=4.9785e-04) (hash(x)=74564994)
+4940 train 4.004200 (lr=4.9615e-04) (hash(x)=78282849)
+4950 train 3.815697 (lr=4.9445e-04) (hash(x)=75618079)
+4960 train 3.959253 (lr=4.9275e-04) (hash(x)=78308363)
+4970 train 3.820936 (lr=4.9105e-04) (hash(x)=81122810)
+4980 train 3.962206 (lr=4.8935e-04) (hash(x)=72474525)
+4990 train 3.820616 (lr=4.8765e-04) (hash(x)=71682033)
+5000 val loss 3.8143
+5000 val perplexity 45.3471
+5000 train 3.771644 (lr=4.8596e-04) (hash(x)=74751266)
+5010 train 3.841506 (lr=4.8426e-04) (hash(x)=71876255)
+5020 train 4.012558 (lr=4.8257e-04) (hash(x)=67006293)
+5030 train 3.821250 (lr=4.8087e-04) (hash(x)=55446837)
+5040 train 3.931035 (lr=4.7918e-04) (hash(x)=72139304)
+5050 train 3.912959 (lr=4.7749e-04) (hash(x)=72848819)
+5060 train 3.840793 (lr=4.7580e-04) (hash(x)=74756891)
+5070 train 3.826256 (lr=4.7411e-04) (hash(x)=71425192)
+5080 train 3.704717 (lr=4.7242e-04) (hash(x)=91717336)
+5090 train 3.711437 (lr=4.7073e-04) (hash(x)=78489482)
+5100 val loss 3.8154
+5100 val perplexity 45.3969
+5100 train 3.703903 (lr=4.6905e-04) (hash(x)=102417092)
+5110 train 3.773282 (lr=4.6736e-04) (hash(x)=67631270)
+5120 train 3.762649 (lr=4.6568e-04) (hash(x)=70575439)
+5130 train 3.797820 (lr=4.6400e-04) (hash(x)=73933044)
+5140 train 3.664560 (lr=4.6231e-04) (hash(x)=75105935)
+5150 train 3.784578 (lr=4.6063e-04) (hash(x)=75662761)
+5160 train 3.784907 (lr=4.5896e-04) (hash(x)=77254705)
+5170 train 3.750488 (lr=4.5728e-04) (hash(x)=67051450)
+5180 train 3.575444 (lr=4.5560e-04) (hash(x)=69736129)
+5190 train 3.590010 (lr=4.5393e-04) (hash(x)=68256740)
+5200 val loss 3.8137
+5200 val perplexity 45.3177
+5200 train 3.551154 (lr=4.5225e-04) (hash(x)=69483022)
+5210 train 3.607293 (lr=4.5058e-04) (hash(x)=74805034)
+5220 train 3.584081 (lr=4.4891e-04) (hash(x)=75297212)
+5230 train 3.508024 (lr=4.4724e-04) (hash(x)=69772814)
+5240 train 3.622215 (lr=4.4557e-04) (hash(x)=72164549)
+5250 train 3.484519 (lr=4.4391e-04) (hash(x)=73896592)
+5260 train 3.629845 (lr=4.4224e-04) (hash(x)=75901314)
+5270 train 3.504444 (lr=4.4058e-04) (hash(x)=76054156)
+5280 train 3.843983 (lr=4.3892e-04) (hash(x)=85568643)
+5290 train 3.884977 (lr=4.3726e-04) (hash(x)=84053592)
+5300 val loss 3.7978
+5300 val perplexity 44.6012
+5300 train 3.767685 (lr=4.3560e-04) (hash(x)=73135718)
+5310 train 3.913254 (lr=4.3395e-04) (hash(x)=73149116)
+5320 train 3.891977 (lr=4.3229e-04) (hash(x)=77101540)
+5330 train 3.725110 (lr=4.3064e-04) (hash(x)=84274756)
+5340 train 3.748969 (lr=4.2899e-04) (hash(x)=74181619)
+5350 train 3.633959 (lr=4.2734e-04) (hash(x)=69664265)
+5360 train 3.860588 (lr=4.2569e-04) (hash(x)=68837592)
+5370 train 3.795830 (lr=4.2404e-04) (hash(x)=75148213)
+5380 train 3.886878 (lr=4.2240e-04) (hash(x)=68908917)
+5390 train 3.852165 (lr=4.2076e-04) (hash(x)=72753463)
+5400 val loss 3.7794
+5400 val perplexity 43.7889
+5400 train 3.749935 (lr=4.1912e-04) (hash(x)=69948122)
+5410 train 3.655521 (lr=4.1748e-04) (hash(x)=74530436)
+5420 train 3.821488 (lr=4.1584e-04) (hash(x)=67532641)
+5430 train 3.792599 (lr=4.1421e-04) (hash(x)=77416323)
+5440 train 3.771484 (lr=4.1257e-04) (hash(x)=77450277)
+5450 train 3.699586 (lr=4.1094e-04) (hash(x)=69637629)
+5460 train 3.750093 (lr=4.0931e-04) (hash(x)=77134731)
+5470 train 3.882771 (lr=4.0769e-04) (hash(x)=72637484)
+5480 train 3.819701 (lr=4.0606e-04) (hash(x)=76342282)
+5490 train 3.741092 (lr=4.0444e-04) (hash(x)=74794888)
+5500 val loss 3.7783
+5500 val perplexity 43.7397
+5500 train 3.739415 (lr=4.0282e-04) (hash(x)=70491787)
+5510 train 3.765331 (lr=4.0120e-04) (hash(x)=73673009)
+5520 train 3.698666 (lr=3.9959e-04) (hash(x)=73134241)
+5530 train 3.702623 (lr=3.9797e-04) (hash(x)=80260202)
+5540 train 3.737857 (lr=3.9636e-04) (hash(x)=75403578)
+5550 train 3.668324 (lr=3.9475e-04) (hash(x)=65589999)
+5560 train 3.749234 (lr=3.9314e-04) (hash(x)=71523917)
+5570 train 3.648072 (lr=3.9154e-04) (hash(x)=71871396)
+5580 train 3.463234 (lr=3.8993e-04) (hash(x)=77446225)
+5590 train 3.477172 (lr=3.8833e-04) (hash(x)=79892116)
+5600 val loss 3.7838
+5600 val perplexity 43.9822
+5600 train 3.696796 (lr=3.8674e-04) (hash(x)=72893174)
+5610 train 3.527065 (lr=3.8514e-04) (hash(x)=74362751)
+5620 train 3.661416 (lr=3.8355e-04) (hash(x)=75125741)
+5630 train 3.541162 (lr=3.8196e-04) (hash(x)=74303205)
+5640 train 3.547629 (lr=3.8037e-04) (hash(x)=65727805)
+5650 train 3.631219 (lr=3.7878e-04) (hash(x)=79694849)
+5660 train 3.649570 (lr=3.7720e-04) (hash(x)=74705862)
+5670 train 3.623400 (lr=3.7562e-04) (hash(x)=74962821)
+5680 train 3.995806 (lr=3.7404e-04) (hash(x)=76748843)
+5690 train 3.842822 (lr=3.7246e-04) (hash(x)=72562868)
+5700 val loss 3.7744
+5700 val perplexity 43.5693
+5700 train 3.787758 (lr=3.7089e-04) (hash(x)=83245544)
+5710 train 3.710258 (lr=3.6932e-04) (hash(x)=76485320)
+5720 train 3.824130 (lr=3.6775e-04) (hash(x)=75340650)
+5730 train 3.806679 (lr=3.6619e-04) (hash(x)=73202272)
+5740 train 3.782930 (lr=3.6462e-04) (hash(x)=75623558)
+5750 train 3.802449 (lr=3.6306e-04) (hash(x)=73893861)
+5760 train 3.771797 (lr=3.6151e-04) (hash(x)=78314196)
+5770 train 3.790672 (lr=3.5995e-04) (hash(x)=72732120)
+5780 train 3.768619 (lr=3.5840e-04) (hash(x)=79236217)
+5790 train 3.752296 (lr=3.5685e-04) (hash(x)=82223910)
+5800 val loss 3.7545
+5800 val perplexity 42.7123
+5800 train 3.766829 (lr=3.5530e-04) (hash(x)=71077079)
+5810 train 3.770199 (lr=3.5376e-04) (hash(x)=82793170)
+5820 train 3.768299 (lr=3.5222e-04) (hash(x)=100985291)
+5830 train 3.769893 (lr=3.5068e-04) (hash(x)=74375677)
+5840 train 3.903067 (lr=3.4915e-04) (hash(x)=71221972)
+5850 train 3.682836 (lr=3.4762e-04) (hash(x)=72588583)
+5860 train 3.768000 (lr=3.4609e-04) (hash(x)=88719929)
+5870 train 3.739300 (lr=3.4456e-04) (hash(x)=73369246)
+5880 train 3.800587 (lr=3.4304e-04) (hash(x)=73331315)
+5890 train 3.766808 (lr=3.4152e-04) (hash(x)=53449512)
+5900 val loss 3.7572
+5900 val perplexity 42.8272
+5900 train 3.655138 (lr=3.4000e-04) (hash(x)=75334660)
+5910 train 3.723900 (lr=3.3849e-04) (hash(x)=75037764)
+5920 train 3.792995 (lr=3.3697e-04) (hash(x)=72040912)
+5930 train 3.721459 (lr=3.3547e-04) (hash(x)=68296721)
+5940 train 3.685977 (lr=3.3396e-04) (hash(x)=72575829)
+5950 train 3.568504 (lr=3.3246e-04) (hash(x)=71274235)
+5960 train 3.753669 (lr=3.3096e-04) (hash(x)=70377631)
+5970 train 3.586278 (lr=3.2947e-04) (hash(x)=80434203)
+5980 train 3.455035 (lr=3.2797e-04) (hash(x)=71681516)
+5990 train 3.576262 (lr=3.2649e-04) (hash(x)=74917317)
+6000 val loss 3.7560
+6000 val perplexity 42.7777
+6000 train 3.502906 (lr=3.2500e-04) (hash(x)=69246326)
+6010 train 3.594422 (lr=3.2352e-04) (hash(x)=76209797)
+6020 train 3.700918 (lr=3.2204e-04) (hash(x)=77523561)
+6030 train 3.671709 (lr=3.2056e-04) (hash(x)=77066537)
+6040 train 3.640259 (lr=3.1909e-04) (hash(x)=81795462)
+6050 train 3.545828 (lr=3.1762e-04) (hash(x)=74328005)
+6060 train 3.560942 (lr=3.1616e-04) (hash(x)=76260120)
+6070 train 3.688489 (lr=3.1469e-04) (hash(x)=73864904)
+6080 train 3.660241 (lr=3.1323e-04) (hash(x)=83311739)
+6090 train 3.964492 (lr=3.1178e-04) (hash(x)=85943188)
+6100 val loss 3.7370
+6100 val perplexity 41.9723
+6100 train 3.626211 (lr=3.1033e-04) (hash(x)=70412657)
+6110 train 3.785763 (lr=3.0888e-04) (hash(x)=78569898)
+6120 train 3.839788 (lr=3.0743e-04) (hash(x)=53010374)
+6130 train 3.834085 (lr=3.0599e-04) (hash(x)=77387166)
+6140 train 3.817585 (lr=3.0455e-04) (hash(x)=68495380)
+6150 train 3.702678 (lr=3.0312e-04) (hash(x)=77674661)
+6160 train 3.782394 (lr=3.0169e-04) (hash(x)=73099224)
+6170 train 3.783549 (lr=3.0026e-04) (hash(x)=75550331)
+6180 train 3.852736 (lr=2.9884e-04) (hash(x)=70208558)
+6190 train 3.695030 (lr=2.9742e-04) (hash(x)=81474783)
+6200 val loss 3.7208
+6200 val perplexity 41.2980
+6200 train 3.688138 (lr=2.9600e-04) (hash(x)=72863420)
+6210 train 3.823990 (lr=2.9459e-04) (hash(x)=67338127)
+6220 train 3.749213 (lr=2.9318e-04) (hash(x)=72064676)
+6230 train 3.628724 (lr=2.9177e-04) (hash(x)=63929148)
+6240 train 3.661108 (lr=2.9037e-04) (hash(x)=75244700)
+6250 train 3.760086 (lr=2.8897e-04) (hash(x)=72131685)
+6260 train 3.820707 (lr=2.8758e-04) (hash(x)=77239186)
+6270 train 3.729449 (lr=2.8619e-04) (hash(x)=80827392)
+6280 train 3.692749 (lr=2.8480e-04) (hash(x)=68612370)
+6290 train 3.747125 (lr=2.8342e-04) (hash(x)=70664463)
+6300 val loss 3.7183
+6300 val perplexity 41.1954
+6300 train 3.686190 (lr=2.8204e-04) (hash(x)=73020224)
+6310 train 3.703270 (lr=2.8067e-04) (hash(x)=70614977)
+6320 train 3.693339 (lr=2.7930e-04) (hash(x)=76744877)
+6330 train 3.741690 (lr=2.7793e-04) (hash(x)=83384143)
+6340 train 3.599163 (lr=2.7657e-04) (hash(x)=72241681)
+6350 train 3.624433 (lr=2.7521e-04) (hash(x)=74037360)
+6360 train 3.471763 (lr=2.7385e-04) (hash(x)=70095163)
+6370 train 3.394009 (lr=2.7250e-04) (hash(x)=69510647)
+6380 train 3.567859 (lr=2.7116e-04) (hash(x)=74297517)
+6390 train 3.702486 (lr=2.6981e-04) (hash(x)=78058972)
+6400 val loss 3.7210
+6400 val perplexity 41.3055
+6400 train 3.615313 (lr=2.6847e-04) (hash(x)=82953514)
+6410 train 3.388636 (lr=2.6714e-04) (hash(x)=72289413)
+6420 train 3.525731 (lr=2.6581e-04) (hash(x)=66436985)
+6430 train 3.633868 (lr=2.6448e-04) (hash(x)=76842704)
+6440 train 3.581012 (lr=2.6316e-04) (hash(x)=77576509)
+6450 train 3.593360 (lr=2.6184e-04) (hash(x)=63757462)
+6460 train 3.353406 (lr=2.6053e-04) (hash(x)=74725544)
+6470 train 3.806491 (lr=2.5922e-04) (hash(x)=74976987)
+6480 train 3.939163 (lr=2.5791e-04) (hash(x)=68415455)
+6490 train 3.739840 (lr=2.5661e-04) (hash(x)=75902883)
+6500 val loss 3.7085
+6500 val perplexity 40.7919
+6500 train 3.678328 (lr=2.5531e-04) (hash(x)=76888442)
+6510 train 3.780519 (lr=2.5402e-04) (hash(x)=73767111)
+6520 train 3.782838 (lr=2.5273e-04) (hash(x)=76218061)
+6530 train 3.741245 (lr=2.5145e-04) (hash(x)=76573241)
+6540 train 3.724410 (lr=2.5017e-04) (hash(x)=71980504)
+6550 train 3.782207 (lr=2.4889e-04) (hash(x)=77715451)
+6560 train 3.660448 (lr=2.4762e-04) (hash(x)=73772644)
+6570 train 3.727784 (lr=2.4635e-04) (hash(x)=70907139)
+6580 train 3.736925 (lr=2.4509e-04) (hash(x)=68045832)
+6590 train 3.753581 (lr=2.4383e-04) (hash(x)=75212157)
+6600 val loss 3.7036
+6600 val perplexity 40.5946
+6600 train 3.692216 (lr=2.4258e-04) (hash(x)=69232309)
+6610 train 3.675817 (lr=2.4133e-04) (hash(x)=67915539)
+6620 train 3.538253 (lr=2.4009e-04) (hash(x)=73942199)
+6630 train 3.711311 (lr=2.3884e-04) (hash(x)=71108653)
+6640 train 3.702012 (lr=2.3761e-04) (hash(x)=70766590)
+6650 train 3.586818 (lr=2.3638e-04) (hash(x)=69205319)
+6660 train 3.693664 (lr=2.3515e-04) (hash(x)=81026148)
+6670 train 3.621096 (lr=2.3393e-04) (hash(x)=77657830)
+6680 train 3.530802 (lr=2.3271e-04) (hash(x)=76483678)
+6690 train 3.585719 (lr=2.3150e-04) (hash(x)=71151306)
+6700 val loss 3.7037
+6700 val perplexity 40.5980
+6700 train 3.682486 (lr=2.3029e-04) (hash(x)=74577530)
+6710 train 3.573588 (lr=2.2909e-04) (hash(x)=69171643)
+6720 train 3.615970 (lr=2.2789e-04) (hash(x)=67250925)
+6730 train 3.598993 (lr=2.2669e-04) (hash(x)=68971618)
+6740 train 3.557070 (lr=2.2550e-04) (hash(x)=75990195)
+6750 train 3.654644 (lr=2.2432e-04) (hash(x)=84198679)
+6760 train 3.735959 (lr=2.2314e-04) (hash(x)=100211939)
+6770 train 3.758752 (lr=2.2196e-04) (hash(x)=70028335)
+6780 train 3.821722 (lr=2.2079e-04) (hash(x)=76176743)
+6790 train 3.902557 (lr=2.1963e-04) (hash(x)=75642962)
+6800 val loss 3.6879
+6800 val perplexity 39.9615
+6800 train 3.631905 (lr=2.1847e-04) (hash(x)=73064800)
+6810 train 3.742045 (lr=2.1731e-04) (hash(x)=71307839)
+6820 train 3.662904 (lr=2.1616e-04) (hash(x)=80361526)
+6830 train 3.678871 (lr=2.1501e-04) (hash(x)=74155886)
+6840 train 3.762501 (lr=2.1387e-04) (hash(x)=83498789)
+6850 train 3.872828 (lr=2.1273e-04) (hash(x)=72686972)
+6860 train 3.821345 (lr=2.1160e-04) (hash(x)=74835805)
+6870 train 3.755506 (lr=2.1047e-04) (hash(x)=74809435)
+6880 train 3.814886 (lr=2.0935e-04) (hash(x)=71911947)
+6890 train 3.689610 (lr=2.0823e-04) (hash(x)=70201693)
+6900 val loss 3.6828
+6900 val perplexity 39.7581
+6900 train 3.712472 (lr=2.0712e-04) (hash(x)=77162446)
+6910 train 3.835165 (lr=2.0602e-04) (hash(x)=60395458)
+6920 train 3.740775 (lr=2.0491e-04) (hash(x)=78466934)
+6930 train 3.723298 (lr=2.0382e-04) (hash(x)=80996788)
+6940 train 3.615596 (lr=2.0272e-04) (hash(x)=74634433)
+6950 train 3.737562 (lr=2.0164e-04) (hash(x)=69618707)
+6960 train 3.601078 (lr=2.0055e-04) (hash(x)=77936797)
+6970 train 3.612643 (lr=1.9948e-04) (hash(x)=78698749)
+6980 train 3.609472 (lr=1.9840e-04) (hash(x)=71152904)
+6990 train 3.720703 (lr=1.9734e-04) (hash(x)=78876720)
+7000 val loss 3.6782
+7000 val perplexity 39.5765
+7000 train 3.698404 (lr=1.9628e-04) (hash(x)=70500963)
+7010 train 3.661447 (lr=1.9522e-04) (hash(x)=75732988)
+7020 train 3.834758 (lr=1.9417e-04) (hash(x)=72655885)
+7030 train 3.541782 (lr=1.9312e-04) (hash(x)=75132388)
+7040 train 3.611836 (lr=1.9208e-04) (hash(x)=88674279)
+7050 train 3.591294 (lr=1.9104e-04) (hash(x)=73050510)
+7060 train 3.629106 (lr=1.9001e-04) (hash(x)=74850157)
+7070 train 3.623950 (lr=1.8899e-04) (hash(x)=80733840)
+7080 train 3.614174 (lr=1.8797e-04) (hash(x)=68592323)
+7090 train 3.550707 (lr=1.8695e-04) (hash(x)=74142615)
+7100 val loss 3.6780
+7100 val perplexity 39.5670
+7100 train 3.566196 (lr=1.8594e-04) (hash(x)=78220018)
+7110 train 3.559711 (lr=1.8494e-04) (hash(x)=68388633)
+7120 train 3.380956 (lr=1.8394e-04) (hash(x)=71391877)
+7130 train 3.695735 (lr=1.8294e-04) (hash(x)=80144529)
+7140 train 3.789128 (lr=1.8196e-04) (hash(x)=79507532)
+7150 train 3.692015 (lr=1.8097e-04) (hash(x)=76982262)
+7160 train 3.855980 (lr=1.7999e-04) (hash(x)=82516157)
+7170 train 3.609463 (lr=1.7902e-04) (hash(x)=72371860)
+7180 train 3.862401 (lr=1.7805e-04) (hash(x)=72376921)
+7190 train 3.708340 (lr=1.7709e-04) (hash(x)=81008418)
+7200 val loss 3.6642
+7200 val perplexity 39.0257
+7200 train 3.771976 (lr=1.7614e-04) (hash(x)=73818590)
+7210 train 3.738631 (lr=1.7519e-04) (hash(x)=71683916)
+7220 train 3.691939 (lr=1.7424e-04) (hash(x)=77659620)
+7230 train 3.732988 (lr=1.7330e-04) (hash(x)=79997614)
+7240 train 3.748168 (lr=1.7237e-04) (hash(x)=74054734)
+7250 train 3.839936 (lr=1.7144e-04) (hash(x)=84630195)
+7260 train 3.742384 (lr=1.7051e-04) (hash(x)=83833114)
+7270 train 3.659254 (lr=1.6959e-04) (hash(x)=78625868)
+7280 train 3.868357 (lr=1.6868e-04) (hash(x)=76859769)
+7290 train 3.667498 (lr=1.6777e-04) (hash(x)=77079387)
+7300 val loss 3.6592
+7300 val perplexity 38.8299
+7300 train 3.661269 (lr=1.6687e-04) (hash(x)=74181120)
+7310 train 3.730875 (lr=1.6598e-04) (hash(x)=72220366)
+7320 train 3.669443 (lr=1.6509e-04) (hash(x)=77955423)
+7330 train 3.709283 (lr=1.6420e-04) (hash(x)=75397467)
+7340 train 3.703813 (lr=1.6332e-04) (hash(x)=73697940)
+7350 train 3.603831 (lr=1.6245e-04) (hash(x)=83531555)
+7360 train 3.717419 (lr=1.6158e-04) (hash(x)=72009097)
+7370 train 3.666575 (lr=1.6072e-04) (hash(x)=76354441)
+7380 train 3.706801 (lr=1.5986e-04) (hash(x)=77219348)
+7390 train 3.502160 (lr=1.5901e-04) (hash(x)=79816131)
+7400 val loss 3.6572
+7400 val perplexity 38.7531
+7400 train 3.686786 (lr=1.5816e-04) (hash(x)=67942598)
+7410 train 3.706089 (lr=1.5732e-04) (hash(x)=67638178)
+7420 train 3.642549 (lr=1.5649e-04) (hash(x)=71682874)
+7430 train 3.509881 (lr=1.5566e-04) (hash(x)=71297263)
+7440 train 3.518374 (lr=1.5484e-04) (hash(x)=73903999)
+7450 train 3.621816 (lr=1.5402e-04) (hash(x)=74071685)
+7460 train 3.416296 (lr=1.5321e-04) (hash(x)=71088160)
+7470 train 3.401508 (lr=1.5241e-04) (hash(x)=74178017)
+7480 train 3.435699 (lr=1.5161e-04) (hash(x)=84610585)
+7490 train 3.601688 (lr=1.5081e-04) (hash(x)=67695236)
+7500 val loss 3.6585
+7500 val perplexity 38.8026
+7500 train 3.542574 (lr=1.5002e-04) (hash(x)=72022604)
+7510 train 3.580838 (lr=1.4924e-04) (hash(x)=78266706)
+7520 train 3.673334 (lr=1.4847e-04) (hash(x)=72576417)
+7530 train 3.890779 (lr=1.4769e-04) (hash(x)=71494052)
+7540 train 3.816607 (lr=1.4693e-04) (hash(x)=76071936)
+7550 train 3.690997 (lr=1.4617e-04) (hash(x)=78379534)
+7560 train 3.653231 (lr=1.4542e-04) (hash(x)=75712040)
+7570 train 3.641760 (lr=1.4467e-04) (hash(x)=83109134)
+7580 train 3.700063 (lr=1.4393e-04) (hash(x)=78846949)
+7590 train 3.781287 (lr=1.4319e-04) (hash(x)=63919123)
+7600 val loss 3.6473
+7600 val perplexity 38.3707
+7600 train 3.658151 (lr=1.4246e-04) (hash(x)=73563619)
+7610 train 3.721436 (lr=1.4174e-04) (hash(x)=79745028)
+7620 train 3.701515 (lr=1.4102e-04) (hash(x)=68113846)
+7630 train 3.682285 (lr=1.4031e-04) (hash(x)=70369067)
+7640 train 3.685153 (lr=1.3960e-04) (hash(x)=79635086)
+7650 train 3.603169 (lr=1.3890e-04) (hash(x)=77348303)
+7660 train 3.607666 (lr=1.3821e-04) (hash(x)=74939810)
+7670 train 3.741526 (lr=1.3752e-04) (hash(x)=77151692)
+7680 train 3.555557 (lr=1.3684e-04) (hash(x)=69543853)
+7690 train 3.797997 (lr=1.3616e-04) (hash(x)=75604567)
+7700 val loss 3.6423
+7700 val perplexity 38.1779
+7700 train 3.772254 (lr=1.3549e-04) (hash(x)=84362557)
+7710 train 3.708710 (lr=1.3483e-04) (hash(x)=71030514)
+7720 train 3.663011 (lr=1.3417e-04) (hash(x)=76132027)
+7730 train 3.701341 (lr=1.3352e-04) (hash(x)=77799943)
+7740 train 3.681250 (lr=1.3287e-04) (hash(x)=74951280)
+7750 train 3.705631 (lr=1.3223e-04) (hash(x)=73706939)
+7760 train 3.619767 (lr=1.3160e-04) (hash(x)=78667790)
+7770 train 3.728979 (lr=1.3097e-04) (hash(x)=72949775)
+7780 train 3.659230 (lr=1.3035e-04) (hash(x)=74537611)
+7790 train 3.614482 (lr=1.2974e-04) (hash(x)=72876060)
+7800 val loss 3.6417
+7800 val perplexity 38.1550
+7800 train 3.715736 (lr=1.2913e-04) (hash(x)=73302610)
+7810 train 3.468505 (lr=1.2852e-04) (hash(x)=66480452)
+7820 train 3.486559 (lr=1.2793e-04) (hash(x)=74833008)
+7830 train 3.504539 (lr=1.2733e-04) (hash(x)=66729619)
+7840 train 3.607941 (lr=1.2675e-04) (hash(x)=87861820)
+7850 train 3.416695 (lr=1.2617e-04) (hash(x)=73130610)
+7860 train 3.629555 (lr=1.2560e-04) (hash(x)=67874016)
+7870 train 3.415697 (lr=1.2503e-04) (hash(x)=70600131)
+7880 train 3.554274 (lr=1.2447e-04) (hash(x)=68990980)
+7890 train 3.508148 (lr=1.2392e-04) (hash(x)=79666741)
+7900 val loss 3.6428
+7900 val perplexity 38.1974
+7900 train 3.526054 (lr=1.2337e-04) (hash(x)=74072340)
+7910 train 3.595364 (lr=1.2283e-04) (hash(x)=80452381)
+7920 train 3.728763 (lr=1.2229e-04) (hash(x)=83292891)
+7930 train 3.610550 (lr=1.2176e-04) (hash(x)=80375639)
+7940 train 3.523642 (lr=1.2124e-04) (hash(x)=82765856)
+7950 train 3.616413 (lr=1.2072e-04) (hash(x)=72962177)
+7960 train 3.688146 (lr=1.2021e-04) (hash(x)=74551129)
+7970 train 3.771068 (lr=1.1970e-04) (hash(x)=89185352)
+7980 train 3.750160 (lr=1.1921e-04) (hash(x)=81338443)
+7990 train 3.537557 (lr=1.1871e-04) (hash(x)=83820810)
+8000 val loss 3.6334
+8000 val perplexity 37.8429
+8000 train 3.709970 (lr=1.1823e-04) (hash(x)=72329617)
+8010 train 3.694041 (lr=1.1775e-04) (hash(x)=82188069)
+8020 train 3.716743 (lr=1.1728e-04) (hash(x)=78460596)
+8030 train 3.638503 (lr=1.1681e-04) (hash(x)=76416474)
+8040 train 3.805143 (lr=1.1635e-04) (hash(x)=70629903)
+8050 train 3.605256 (lr=1.1589e-04) (hash(x)=76561819)
+8060 train 3.640028 (lr=1.1544e-04) (hash(x)=71993036)
+8070 train 3.810941 (lr=1.1500e-04) (hash(x)=73270928)
+8080 train 3.708490 (lr=1.1457e-04) (hash(x)=82402459)
+8090 train 3.740438 (lr=1.1414e-04) (hash(x)=67208853)
+8100 val loss 3.6294
+8100 val perplexity 37.6890
+8100 train 3.817009 (lr=1.1371e-04) (hash(x)=85981056)
+8110 train 3.765708 (lr=1.1330e-04) (hash(x)=77469728)
+8120 train 3.669600 (lr=1.1289e-04) (hash(x)=84702034)
+8130 train 3.555752 (lr=1.1248e-04) (hash(x)=96254798)
+8140 train 3.626621 (lr=1.1209e-04) (hash(x)=71865027)
+8150 train 3.602555 (lr=1.1169e-04) (hash(x)=79902920)
+8160 train 3.749343 (lr=1.1131e-04) (hash(x)=70307892)
+8170 train 3.629095 (lr=1.1093e-04) (hash(x)=73853567)
+8180 train 3.631683 (lr=1.1056e-04) (hash(x)=72321642)
+8190 train 3.653842 (lr=1.1019e-04) (hash(x)=82101781)
+8200 val loss 3.6281
+8200 val perplexity 37.6394
+8200 train 3.666280 (lr=1.0983e-04) (hash(x)=66290405)
+8210 train 3.543933 (lr=1.0948e-04) (hash(x)=69697952)
+8220 train 3.793462 (lr=1.0913e-04) (hash(x)=72618212)
+8230 train 3.543591 (lr=1.0879e-04) (hash(x)=76210758)
+8240 train 3.550376 (lr=1.0846e-04) (hash(x)=70820601)
+8250 train 3.620776 (lr=1.0813e-04) (hash(x)=75548154)
+8260 train 3.554576 (lr=1.0781e-04) (hash(x)=73211139)
+8270 train 3.614822 (lr=1.0750e-04) (hash(x)=82252960)
+8280 train 3.567270 (lr=1.0719e-04) (hash(x)=69408116)
+8290 train 3.733830 (lr=1.0689e-04) (hash(x)=74462098)
+8300 val loss 3.6272
+8300 val perplexity 37.6068
+8300 train 3.739029 (lr=1.0659e-04) (hash(x)=79945329)
+8310 train 3.492974 (lr=1.0630e-04) (hash(x)=82459712)
+8320 train 3.675274 (lr=1.0602e-04) (hash(x)=81135573)
+8330 train 3.650322 (lr=1.0574e-04) (hash(x)=69939721)
+8340 train 3.594853 (lr=1.0547e-04) (hash(x)=84687517)
+8350 train 3.563389 (lr=1.0521e-04) (hash(x)=77515837)
+8360 train 3.567242 (lr=1.0495e-04) (hash(x)=72146536)
+8370 train 3.604628 (lr=1.0470e-04) (hash(x)=73198063)
+8380 train 3.607463 (lr=1.0446e-04) (hash(x)=68952542)
+8390 train 3.646948 (lr=1.0422e-04) (hash(x)=74385341)
+8400 val loss 3.6199
+8400 val perplexity 37.3320
+8400 train 3.467125 (lr=1.0399e-04) (hash(x)=65193622)
+8410 train 3.709282 (lr=1.0377e-04) (hash(x)=66981351)
+8420 train 3.604985 (lr=1.0355e-04) (hash(x)=97752378)
+8430 train 3.800304 (lr=1.0334e-04) (hash(x)=70936998)
+8440 train 3.715131 (lr=1.0313e-04) (hash(x)=71557836)
+8450 train 3.647805 (lr=1.0293e-04) (hash(x)=71340008)
+8460 train 3.696526 (lr=1.0274e-04) (hash(x)=69634745)
+8470 train 3.726096 (lr=1.0256e-04) (hash(x)=74224484)
+8480 train 3.669441 (lr=1.0238e-04) (hash(x)=75490845)
+8490 train 3.546716 (lr=1.0220e-04) (hash(x)=78619557)
+8500 val loss 3.6194
+8500 val perplexity 37.3161
+8500 train 3.561023 (lr=1.0204e-04) (hash(x)=70959421)
+8510 train 3.646967 (lr=1.0188e-04) (hash(x)=75975416)
+8520 train 3.360667 (lr=1.0172e-04) (hash(x)=76336317)
+8530 train 3.623477 (lr=1.0158e-04) (hash(x)=67836668)
+8540 train 3.584897 (lr=1.0144e-04) (hash(x)=71139919)
+8550 train 3.829602 (lr=1.0130e-04) (hash(x)=69476711)
+8560 train 3.578383 (lr=1.0118e-04) (hash(x)=75144101)
+8570 train 3.526005 (lr=1.0106e-04) (hash(x)=76246282)
+8580 train 3.624581 (lr=1.0094e-04) (hash(x)=69466257)
+8590 train 3.598452 (lr=1.0083e-04) (hash(x)=71601133)
+8600 val loss 3.6198
+8600 val perplexity 37.3288
+8600 train 3.568667 (lr=1.0073e-04) (hash(x)=72319507)
+8610 train 3.704520 (lr=1.0064e-04) (hash(x)=74618018)
+8620 train 3.557616 (lr=1.0055e-04) (hash(x)=76115665)
+8630 train 3.714321 (lr=1.0047e-04) (hash(x)=82063941)
+8640 train 3.677985 (lr=1.0039e-04) (hash(x)=79867543)
+8650 train 3.484867 (lr=1.0033e-04) (hash(x)=73802504)
+8660 train 3.593276 (lr=1.0026e-04) (hash(x)=74251531)
+8670 train 3.574250 (lr=1.0021e-04) (hash(x)=74637655)
+8680 train 3.671430 (lr=1.0016e-04) (hash(x)=82021614)
+8690 train 3.685838 (lr=1.0012e-04) (hash(x)=77938695)
+8700 val loss 3.6113
+8700 val perplexity 37.0127
+8700 train 3.720045 (lr=1.0008e-04) (hash(x)=72431581)
+8710 train 3.736783 (lr=1.0005e-04) (hash(x)=74875522)
+8720 train 3.672248 (lr=1.0003e-04) (hash(x)=69574551)
+8730 train 3.510911 (lr=1.0001e-04) (hash(x)=71568849)
+8740 train 3.752005 (lr=1.0000e-04) (hash(x)=77714444)
+8749 val loss 3.6109
+8749 val perplexity 36.9978

n_heads12_lr10e-4_total_batch_size61440_seed1339/model_08749.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cce7b2de6a2c120b4e1ea4805cddb5a8690e9b7bfd03cf50c6237639ca85854c
+size 498607490

n_heads12_lr10e-4_total_batch_size61440_seed1339/optimizer_08749.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1db9005365cfd4cb1062f1a9f51fe1bc4fa2e6a8507f6e78c4a32f9bd1a156d
+size 990934406