Upload folder using huggingface_hub

scaling_mha_granularity_2/hd_128_seed_1339/dataloader_04374.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1e9422c936c4d62660261ba5451b4b72bd5055f233420bd7b0797ede3ccedd1e
+oid sha256:6031fd3e2855a036f7a5531cc24555aabd1115f9dd6618b8b2ca6f55279ef0b2
 size 964

scaling_mha_granularity_2/hd_128_seed_1339/log2.txt

@@ -1,529 +1,529 @@
 max_steps: 4375
-0 val loss 11.3314
-0 val perplexity 83399.0547
-0 train 11.334114 (lr=4.1958e-06) (hash(x)=161979327)
-10 train 9.911278 (lr=4.6154e-05) (hash(x)=165941070)
-20 train 9.239570 (lr=8.8112e-05) (hash(x)=152918033)
-30 train 8.526983 (lr=1.3007e-04) (hash(x)=169931607)
-40 train 7.758261 (lr=1.7203e-04) (hash(x)=174342675)
-50 train 7.437022 (lr=2.1399e-04) (hash(x)=163872554)
-60 train 7.409611 (lr=2.5594e-04) (hash(x)=161534908)
-70 train 7.016286 (lr=2.9790e-04) (hash(x)=156776712)
-80 train 7.018011 (lr=3.3986e-04) (hash(x)=164996929)
-90 train 6.854432 (lr=3.8182e-04) (hash(x)=152962169)
-100 val loss 6.7716
-100 val perplexity 872.7406
-100 train 6.629302 (lr=4.2378e-04) (hash(x)=147111701)
-110 train 6.636058 (lr=4.6573e-04) (hash(x)=167595270)
-120 train 6.513920 (lr=5.0769e-04) (hash(x)=152205426)
-130 train 6.414953 (lr=5.4965e-04) (hash(x)=158427335)
-140 train 6.414366 (lr=5.9161e-04) (hash(x)=174588384)
-150 train 6.518809 (lr=6.3357e-04) (hash(x)=157067270)
-160 train 6.484323 (lr=6.7552e-04) (hash(x)=161859595)
-170 train 6.389052 (lr=7.1748e-04) (hash(x)=162739537)
-180 train 6.337269 (lr=7.5944e-04) (hash(x)=170785304)
-190 train 6.278738 (lr=8.0140e-04) (hash(x)=168658088)
-200 val loss 6.2057
-200 val perplexity 495.5506
-200 train 6.240477 (lr=8.4336e-04) (hash(x)=150749086)
-210 train 6.136272 (lr=8.8531e-04) (hash(x)=148789576)
-220 train 6.083822 (lr=9.2727e-04) (hash(x)=153164408)
-230 train 6.216101 (lr=9.6923e-04) (hash(x)=184594366)
-240 train 5.941367 (lr=1.0112e-03) (hash(x)=157817319)
-250 train 6.019251 (lr=1.0531e-03) (hash(x)=149451562)
-260 train 5.897743 (lr=1.0951e-03) (hash(x)=165738161)
-270 train 5.864067 (lr=1.1371e-03) (hash(x)=165724764)
-280 train 5.790833 (lr=1.1790e-03) (hash(x)=155414136)
-290 train 5.734062 (lr=1.2210e-03) (hash(x)=165090903)
-300 val loss 5.8518
-300 val perplexity 347.8566
-300 train 5.618012 (lr=1.2629e-03) (hash(x)=154445545)
-310 train 5.567541 (lr=1.3049e-03) (hash(x)=158949603)
-320 train 5.644899 (lr=1.3469e-03) (hash(x)=161412973)
-330 train 5.773060 (lr=1.3888e-03) (hash(x)=164572249)
-340 train 5.783787 (lr=1.4308e-03) (hash(x)=162202191)
-350 train 5.753174 (lr=1.4727e-03) (hash(x)=158135392)
-360 train 5.697433 (lr=1.5147e-03) (hash(x)=163752391)
-370 train 5.725485 (lr=1.5566e-03) (hash(x)=172758799)
-380 train 5.527005 (lr=1.5986e-03) (hash(x)=144569488)
-390 train 5.502805 (lr=1.6406e-03) (hash(x)=150429229)
-400 val loss 5.5060
-400 val perplexity 246.1613
-400 train 5.485485 (lr=1.6825e-03) (hash(x)=154261530)
-410 train 5.433628 (lr=1.7245e-03) (hash(x)=159172506)
-420 train 5.463166 (lr=1.7664e-03) (hash(x)=154419867)
-430 train 5.486227 (lr=1.8084e-03) (hash(x)=167337099)
-440 train 5.366755 (lr=1.8503e-03) (hash(x)=154535255)
-450 train 5.364452 (lr=1.8923e-03) (hash(x)=153077071)
-460 train 5.264833 (lr=1.9343e-03) (hash(x)=150853070)
-470 train 5.081760 (lr=1.9762e-03) (hash(x)=160701113)
-480 train 5.236236 (lr=2.0182e-03) (hash(x)=154690870)
-490 train 5.033422 (lr=2.0601e-03) (hash(x)=163834286)
-500 val loss 5.3196
-500 val perplexity 204.2981
-500 train 5.108479 (lr=2.1021e-03) (hash(x)=162651717)
-510 train 5.232162 (lr=2.1441e-03) (hash(x)=160448311)
-520 train 5.279454 (lr=2.1860e-03) (hash(x)=165741643)
-530 train 5.392269 (lr=2.2280e-03) (hash(x)=153201264)
-540 train 5.299987 (lr=2.2699e-03) (hash(x)=159027160)
-550 train 5.262941 (lr=2.3119e-03) (hash(x)=170769095)
-560 train 5.283454 (lr=2.3538e-03) (hash(x)=153606190)
-570 train 5.259928 (lr=2.3958e-03) (hash(x)=159708689)
-580 train 5.182477 (lr=2.4378e-03) (hash(x)=179108359)
-590 train 5.232567 (lr=2.4797e-03) (hash(x)=157134241)
-600 val loss 5.1822
-600 val perplexity 178.0667
-600 train 5.210699 (lr=2.5217e-03) (hash(x)=162178356)
-610 train 5.128594 (lr=2.5636e-03) (hash(x)=167646928)
-620 train 5.097169 (lr=2.6056e-03) (hash(x)=160979601)
-630 train 5.002016 (lr=2.6476e-03) (hash(x)=158349468)
-640 train 5.003212 (lr=2.6895e-03) (hash(x)=150370348)
-650 train 4.951624 (lr=2.7315e-03) (hash(x)=152810075)
-660 train 4.946865 (lr=2.7734e-03) (hash(x)=145786262)
-670 train 4.898516 (lr=2.8154e-03) (hash(x)=160779604)
-680 train 4.794373 (lr=2.8573e-03) (hash(x)=152343688)
-690 train 4.806675 (lr=2.8993e-03) (hash(x)=166507600)
-700 val loss 5.0489
-700 val perplexity 155.8544
-700 train 4.790212 (lr=2.9413e-03) (hash(x)=162623700)
-710 train 5.097353 (lr=2.9832e-03) (hash(x)=162293403)
-720 train 5.049531 (lr=3.0000e-03) (hash(x)=159564657)
-730 train 5.073081 (lr=2.9999e-03) (hash(x)=154209107)
-740 train 5.002309 (lr=2.9997e-03) (hash(x)=156485709)
-750 train 4.968740 (lr=2.9994e-03) (hash(x)=160834702)
-760 train 4.981289 (lr=2.9990e-03) (hash(x)=152240275)
-770 train 4.954337 (lr=2.9985e-03) (hash(x)=168903075)
-780 train 5.047203 (lr=2.9979e-03) (hash(x)=166843738)
-790 train 5.076912 (lr=2.9972e-03) (hash(x)=166358104)
-800 val loss 4.8992
-800 val perplexity 134.1890
-800 train 4.869042 (lr=2.9964e-03) (hash(x)=153477747)
-810 train 4.808437 (lr=2.9955e-03) (hash(x)=157954385)
-820 train 4.812778 (lr=2.9945e-03) (hash(x)=174383505)
-830 train 4.808867 (lr=2.9934e-03) (hash(x)=145226181)
-840 train 4.824004 (lr=2.9922e-03) (hash(x)=157665743)
-850 train 4.807346 (lr=2.9909e-03) (hash(x)=158565836)
-860 train 4.792809 (lr=2.9896e-03) (hash(x)=171168105)
-870 train 5.209675 (lr=2.9881e-03) (hash(x)=159017777)
-880 train 4.957736 (lr=2.9865e-03) (hash(x)=168156768)
-890 train 4.873993 (lr=2.9848e-03) (hash(x)=145645245)
-900 val loss 4.7605
-900 val perplexity 116.8074
-900 train 4.708981 (lr=2.9830e-03) (hash(x)=151245521)
-910 train 4.718685 (lr=2.9811e-03) (hash(x)=182614672)
-920 train 4.701442 (lr=2.9792e-03) (hash(x)=167377432)
-930 train 4.634707 (lr=2.9771e-03) (hash(x)=171347228)
-940 train 4.652725 (lr=2.9749e-03) (hash(x)=171588755)
-950 train 4.748018 (lr=2.9726e-03) (hash(x)=159560168)
-960 train 4.641128 (lr=2.9703e-03) (hash(x)=171801357)
-970 train 4.635054 (lr=2.9678e-03) (hash(x)=172513884)
-980 train 4.664547 (lr=2.9652e-03) (hash(x)=170659685)
-990 train 4.691324 (lr=2.9626e-03) (hash(x)=165436358)
-1000 val loss 4.6496
-1000 val perplexity 104.5396
-1000 train 4.750644 (lr=2.9598e-03) (hash(x)=167684690)
-1010 train 4.726415 (lr=2.9570e-03) (hash(x)=127225492)
-1020 train 4.718803 (lr=2.9540e-03) (hash(x)=167575323)
-1030 train 4.641523 (lr=2.9510e-03) (hash(x)=161266302)
-1040 train 4.624722 (lr=2.9478e-03) (hash(x)=155607661)
-1050 train 4.662981 (lr=2.9446e-03) (hash(x)=165685359)
-1060 train 4.522145 (lr=2.9412e-03) (hash(x)=160417824)
-1070 train 4.649426 (lr=2.9378e-03) (hash(x)=173313882)
-1080 train 4.627924 (lr=2.9343e-03) (hash(x)=165485047)
-1090 train 4.481127 (lr=2.9307e-03) (hash(x)=159848305)
-1100 val loss 4.6183
-1100 val perplexity 101.3261
-1100 train 4.674006 (lr=2.9270e-03) (hash(x)=170890601)
-1110 train 4.570353 (lr=2.9231e-03) (hash(x)=172689983)
-1120 train 4.613821 (lr=2.9192e-03) (hash(x)=154762859)
-1130 train 4.539060 (lr=2.9152e-03) (hash(x)=168196871)
-1140 train 4.571562 (lr=2.9112e-03) (hash(x)=156715577)
-1150 train 4.690268 (lr=2.9070e-03) (hash(x)=166381788)
-1160 train 4.629351 (lr=2.9027e-03) (hash(x)=162095130)
-1170 train 4.588916 (lr=2.8983e-03) (hash(x)=166094750)
-1180 train 4.531403 (lr=2.8939e-03) (hash(x)=154965753)
-1190 train 4.516268 (lr=2.8893e-03) (hash(x)=172766447)
-1200 val loss 4.5065
-1200 val perplexity 90.6058
-1200 train 4.429872 (lr=2.8847e-03) (hash(x)=161185334)
-1210 train 4.598069 (lr=2.8800e-03) (hash(x)=167224144)
-1220 train 4.435797 (lr=2.8751e-03) (hash(x)=167252866)
-1230 train 4.447281 (lr=2.8702e-03) (hash(x)=155856924)
-1240 train 4.446458 (lr=2.8652e-03) (hash(x)=155956446)
-1250 train 4.360755 (lr=2.8601e-03) (hash(x)=140794796)
-1260 train 4.481310 (lr=2.8550e-03) (hash(x)=166863869)
-1270 train 4.539800 (lr=2.8497e-03) (hash(x)=166021892)
-1280 train 4.549823 (lr=2.8443e-03) (hash(x)=170503648)
-1290 train 4.454582 (lr=2.8389e-03) (hash(x)=164012719)
-1300 val loss 4.4658
-1300 val perplexity 86.9912
-1300 train 4.522130 (lr=2.8333e-03) (hash(x)=156519834)
-1310 train 4.478926 (lr=2.8277e-03) (hash(x)=165146652)
-1320 train 4.491203 (lr=2.8220e-03) (hash(x)=174391990)
-1330 train 4.497159 (lr=2.8162e-03) (hash(x)=161131201)
-1340 train 4.478579 (lr=2.8103e-03) (hash(x)=162240912)
-1350 train 4.391195 (lr=2.8044e-03) (hash(x)=152462931)
-1360 train 4.399606 (lr=2.7983e-03) (hash(x)=150443278)
-1370 train 4.345457 (lr=2.7922e-03) (hash(x)=153721711)
-1380 train 4.479351 (lr=2.7860e-03) (hash(x)=158070628)
-1390 train 4.506227 (lr=2.7797e-03) (hash(x)=163318827)
-1400 val loss 4.4511
-1400 val perplexity 85.7211
-1400 train 4.420422 (lr=2.7733e-03) (hash(x)=154210336)
-1410 train 4.467419 (lr=2.7668e-03) (hash(x)=162393933)
-1420 train 4.466781 (lr=2.7603e-03) (hash(x)=156659308)
-1430 train 4.430832 (lr=2.7536e-03) (hash(x)=171243439)
-1440 train 4.571663 (lr=2.7469e-03) (hash(x)=178459521)
-1450 train 4.427816 (lr=2.7401e-03) (hash(x)=159166350)
-1460 train 4.347125 (lr=2.7333e-03) (hash(x)=173554569)
-1470 train 4.445459 (lr=2.7263e-03) (hash(x)=158362521)
-1480 train 4.405853 (lr=2.7193e-03) (hash(x)=165899207)
-1490 train 4.432036 (lr=2.7121e-03) (hash(x)=156470939)
-1500 val loss 4.3974
-1500 val perplexity 81.2356
-1500 train 4.403383 (lr=2.7050e-03) (hash(x)=164197780)
-1510 train 4.299698 (lr=2.6977e-03) (hash(x)=146709896)
-1520 train 4.323417 (lr=2.6903e-03) (hash(x)=151304838)
-1530 train 4.348329 (lr=2.6829e-03) (hash(x)=163187333)
-1540 train 4.310416 (lr=2.6754e-03) (hash(x)=153630537)
-1550 train 4.323468 (lr=2.6678e-03) (hash(x)=157299103)
-1560 train 4.233965 (lr=2.6602e-03) (hash(x)=153786477)
-1570 train 4.562898 (lr=2.6525e-03) (hash(x)=155037209)
-1580 train 4.437846 (lr=2.6447e-03) (hash(x)=146157308)
-1590 train 4.480853 (lr=2.6368e-03) (hash(x)=147823317)
-1600 val loss 4.3716
-1600 val perplexity 79.1741
-1600 train 4.464280 (lr=2.6289e-03) (hash(x)=169140159)
-1610 train 4.243920 (lr=2.6208e-03) (hash(x)=157811493)
-1620 train 4.342987 (lr=2.6128e-03) (hash(x)=168940851)
-1630 train 4.322331 (lr=2.6046e-03) (hash(x)=151149213)
-1640 train 4.423075 (lr=2.5964e-03) (hash(x)=161822009)
-1650 train 4.832223 (lr=2.5881e-03) (hash(x)=151056313)
-1660 train 4.348910 (lr=2.5797e-03) (hash(x)=156920306)
-1670 train 4.260403 (lr=2.5713e-03) (hash(x)=159695815)
-1680 train 4.289220 (lr=2.5628e-03) (hash(x)=160232907)
-1690 train 4.273482 (lr=2.5542e-03) (hash(x)=158099357)
-1700 val loss 4.3752
-1700 val perplexity 79.4525
-1700 train 4.323026 (lr=2.5455e-03) (hash(x)=164127381)
-1710 train 4.502932 (lr=2.5368e-03) (hash(x)=156592040)
-1720 train 4.430570 (lr=2.5281e-03) (hash(x)=160220265)
-1730 train 4.362256 (lr=2.5192e-03) (hash(x)=156162204)
-1740 train 4.349756 (lr=2.5103e-03) (hash(x)=152767893)
-1750 train 4.341974 (lr=2.5014e-03) (hash(x)=162707306)
-1760 train 4.359845 (lr=2.4924e-03) (hash(x)=154472004)
-1770 train 4.214347 (lr=2.4833e-03) (hash(x)=160374234)
-1780 train 4.375378 (lr=2.4741e-03) (hash(x)=159709225)
-1790 train 4.410933 (lr=2.4649e-03) (hash(x)=152095944)
-1800 val loss 4.3278
-1800 val perplexity 75.7750
-1800 train 4.387454 (lr=2.4556e-03) (hash(x)=157920255)
-1810 train 4.124925 (lr=2.4463e-03) (hash(x)=167237680)
-1820 train 4.235739 (lr=2.4369e-03) (hash(x)=161255774)
-1830 train 4.280618 (lr=2.4275e-03) (hash(x)=154275481)
-1840 train 4.214444 (lr=2.4180e-03) (hash(x)=152398178)
-1850 train 4.330526 (lr=2.4084e-03) (hash(x)=147646022)
-1860 train 4.337341 (lr=2.3988e-03) (hash(x)=158895559)
-1870 train 4.278633 (lr=2.3891e-03) (hash(x)=162742598)
-1880 train 4.282332 (lr=2.3794e-03) (hash(x)=169911029)
-1890 train 4.317436 (lr=2.3696e-03) (hash(x)=172068270)
-1900 val loss 4.3337
-1900 val perplexity 76.2231
-1900 train 4.417892 (lr=2.3598e-03) (hash(x)=159975457)
-1910 train 4.396720 (lr=2.3499e-03) (hash(x)=178775667)
-1920 train 4.264014 (lr=2.3400e-03) (hash(x)=151180831)
-1930 train 4.279847 (lr=2.3300e-03) (hash(x)=154239412)
-1940 train 4.291859 (lr=2.3200e-03) (hash(x)=139144465)
-1950 train 4.248331 (lr=2.3099e-03) (hash(x)=194482413)
-1960 train 4.258940 (lr=2.2998e-03) (hash(x)=155812344)
-1970 train 4.270830 (lr=2.2896e-03) (hash(x)=157577031)
-1980 train 4.304917 (lr=2.2793e-03) (hash(x)=152045027)
-1990 train 4.201397 (lr=2.2691e-03) (hash(x)=141742221)
-2000 val loss 4.2779
-2000 val perplexity 72.0891
-2000 train 4.328719 (lr=2.2588e-03) (hash(x)=160555812)
-2010 train 4.258727 (lr=2.2484e-03) (hash(x)=160371335)
-2020 train 4.259352 (lr=2.2380e-03) (hash(x)=167890767)
-2030 train 4.331351 (lr=2.2275e-03) (hash(x)=165127715)
-2040 train 4.276251 (lr=2.2170e-03) (hash(x)=151587302)
-2050 train 4.299056 (lr=2.2065e-03) (hash(x)=172347989)
-2060 train 4.368150 (lr=2.1959e-03) (hash(x)=161301919)
-2070 train 4.305970 (lr=2.1853e-03) (hash(x)=161472934)
-2080 train 4.392349 (lr=2.1746e-03) (hash(x)=159777322)
-2090 train 4.298571 (lr=2.1639e-03) (hash(x)=151149770)
-2100 val loss 4.2569
-2100 val perplexity 70.5941
-2100 train 4.126604 (lr=2.1532e-03) (hash(x)=160283520)
-2110 train 4.302319 (lr=2.1424e-03) (hash(x)=154222025)
-2120 train 4.243830 (lr=2.1316e-03) (hash(x)=155638178)
-2130 train 4.160988 (lr=2.1208e-03) (hash(x)=149125384)
-2140 train 4.213496 (lr=2.1099e-03) (hash(x)=159496074)
-2150 train 4.303203 (lr=2.0990e-03) (hash(x)=162019154)
-2160 train 4.457883 (lr=2.0881e-03) (hash(x)=155466928)
-2170 train 4.334429 (lr=2.0771e-03) (hash(x)=173469994)
-2180 train 4.436232 (lr=2.0661e-03) (hash(x)=163448753)
-2190 train 4.273968 (lr=2.0550e-03) (hash(x)=166595123)
-2200 val loss 4.2361
-2200 val perplexity 69.1383
-2200 train 4.156641 (lr=2.0440e-03) (hash(x)=165958875)
-2210 train 4.292385 (lr=2.0329e-03) (hash(x)=165019039)
-2220 train 4.278936 (lr=2.0217e-03) (hash(x)=167325225)
-2230 train 4.106654 (lr=2.0106e-03) (hash(x)=154287439)
-2240 train 4.174267 (lr=1.9994e-03) (hash(x)=158565817)
-2250 train 4.131730 (lr=1.9882e-03) (hash(x)=160592835)
-2260 train 4.176505 (lr=1.9770e-03) (hash(x)=144591648)
-2270 train 4.059541 (lr=1.9657e-03) (hash(x)=150486876)
-2280 train 4.256087 (lr=1.9544e-03) (hash(x)=159194788)
-2290 train 4.332350 (lr=1.9431e-03) (hash(x)=164315345)
-2300 val loss 4.2120
-2300 val perplexity 67.4941
-2300 train 4.256531 (lr=1.9318e-03) (hash(x)=170470665)
-2310 train 4.215390 (lr=1.9205e-03) (hash(x)=176539068)
-2320 train 4.196702 (lr=1.9091e-03) (hash(x)=161008474)
-2330 train 4.274555 (lr=1.8977e-03) (hash(x)=168474380)
-2340 train 4.268914 (lr=1.8863e-03) (hash(x)=156072131)
-2350 train 4.346761 (lr=1.8749e-03) (hash(x)=165127087)
-2360 train 4.348551 (lr=1.8635e-03) (hash(x)=159462801)
-2370 train 4.187483 (lr=1.8520e-03) (hash(x)=161698394)
-2380 train 4.111777 (lr=1.8406e-03) (hash(x)=165588078)
-2390 train 4.118217 (lr=1.8291e-03) (hash(x)=171673672)
-2400 val loss 4.2061
-2400 val perplexity 67.0974
-2400 train 4.185753 (lr=1.8176e-03) (hash(x)=157381516)
-2410 train 4.191660 (lr=1.8061e-03) (hash(x)=163238532)
-2420 train 4.146212 (lr=1.7946e-03) (hash(x)=156178237)
-2430 train 4.033537 (lr=1.7830e-03) (hash(x)=155080617)
-2440 train 4.134524 (lr=1.7715e-03) (hash(x)=167490538)
-2450 train 4.110298 (lr=1.7600e-03) (hash(x)=163757352)
-2460 train 4.072205 (lr=1.7484e-03) (hash(x)=156734811)
-2470 train 4.064728 (lr=1.7368e-03) (hash(x)=161522341)
-2480 train 4.323094 (lr=1.7253e-03) (hash(x)=174032060)
-2490 train 4.279665 (lr=1.7137e-03) (hash(x)=163065818)
-2500 val loss 4.1815
-2500 val perplexity 65.4669
-2500 train 4.336280 (lr=1.7021e-03) (hash(x)=126959780)
-2510 train 4.268579 (lr=1.6906e-03) (hash(x)=153844939)
-2520 train 4.242063 (lr=1.6790e-03) (hash(x)=174018174)
-2530 train 4.223239 (lr=1.6674e-03) (hash(x)=150568427)
-2540 train 4.215667 (lr=1.6558e-03) (hash(x)=161725206)
-2550 train 4.180617 (lr=1.6442e-03) (hash(x)=181346417)
-2560 train 4.176907 (lr=1.6326e-03) (hash(x)=168730011)
-2570 train 4.065266 (lr=1.6210e-03) (hash(x)=155900214)
-2580 train 4.120285 (lr=1.6094e-03) (hash(x)=160014621)
-2590 train 4.096421 (lr=1.5979e-03) (hash(x)=148996689)
-2600 val loss 4.1624
-2600 val perplexity 64.2223
-2600 train 4.167160 (lr=1.5863e-03) (hash(x)=145183968)
-2610 train 3.916850 (lr=1.5747e-03) (hash(x)=159420240)
-2620 train 3.892455 (lr=1.5632e-03) (hash(x)=151169903)
-2630 train 3.906480 (lr=1.5516e-03) (hash(x)=149454811)
-2640 train 3.974618 (lr=1.5400e-03) (hash(x)=155603914)
-2650 train 3.969555 (lr=1.5285e-03) (hash(x)=169545195)
-2660 train 4.324354 (lr=1.5170e-03) (hash(x)=168454079)
-2670 train 4.209318 (lr=1.5054e-03) (hash(x)=163110417)
-2680 train 4.168423 (lr=1.4939e-03) (hash(x)=166075603)
-2690 train 4.197409 (lr=1.4824e-03) (hash(x)=174132258)
-2700 val loss 4.1564
-2700 val perplexity 63.8412
-2700 train 4.205038 (lr=1.4709e-03) (hash(x)=169846940)
-2710 train 4.097352 (lr=1.4594e-03) (hash(x)=151553731)
-2720 train 4.201156 (lr=1.4480e-03) (hash(x)=164024255)
-2730 train 4.218869 (lr=1.4365e-03) (hash(x)=156458097)
-2740 train 4.153926 (lr=1.4251e-03) (hash(x)=154663748)
-2750 train 4.014450 (lr=1.4137e-03) (hash(x)=157548181)
-2760 train 4.127584 (lr=1.4023e-03) (hash(x)=151584783)
-2770 train 4.149693 (lr=1.3909e-03) (hash(x)=152692594)
-2780 train 4.122634 (lr=1.3795e-03) (hash(x)=165952919)
-2790 train 4.017684 (lr=1.3682e-03) (hash(x)=157484183)
-2800 val loss 4.1386
-2800 val perplexity 62.7165
-2800 train 4.087901 (lr=1.3569e-03) (hash(x)=154790034)
-2810 train 3.960568 (lr=1.3456e-03) (hash(x)=157856097)
-2820 train 3.927202 (lr=1.3343e-03) (hash(x)=168903939)
-2830 train 4.053163 (lr=1.3230e-03) (hash(x)=167001475)
-2840 train 3.884875 (lr=1.3118e-03) (hash(x)=158987150)
-2850 train 4.119647 (lr=1.3006e-03) (hash(x)=168695407)
-2860 train 4.252012 (lr=1.2894e-03) (hash(x)=164999034)
-2870 train 4.099381 (lr=1.2783e-03) (hash(x)=191197316)
-2880 train 4.056838 (lr=1.2671e-03) (hash(x)=159625200)
-2890 train 4.213985 (lr=1.2560e-03) (hash(x)=157756126)
-2900 val loss 4.1125
-2900 val perplexity 61.1010
-2900 train 4.070239 (lr=1.2450e-03) (hash(x)=164721377)
-2910 train 4.121223 (lr=1.2339e-03) (hash(x)=155873181)
-2920 train 4.064056 (lr=1.2229e-03) (hash(x)=156477082)
-2930 train 4.073844 (lr=1.2119e-03) (hash(x)=151261633)
-2940 train 4.013025 (lr=1.2010e-03) (hash(x)=160090385)
-2950 train 4.150830 (lr=1.1901e-03) (hash(x)=157300618)
-2960 train 4.070534 (lr=1.1792e-03) (hash(x)=163095555)
-2970 train 4.101968 (lr=1.1684e-03) (hash(x)=149649212)
-2980 train 3.907556 (lr=1.1576e-03) (hash(x)=174517675)
-2990 train 3.881487 (lr=1.1468e-03) (hash(x)=161344491)
-3000 val loss 4.1129
-3000 val perplexity 61.1211
-3000 train 3.908909 (lr=1.1361e-03) (hash(x)=163816905)
-3010 train 3.817577 (lr=1.1254e-03) (hash(x)=157420680)
-3020 train 3.929490 (lr=1.1147e-03) (hash(x)=153843035)
-3030 train 4.120437 (lr=1.1041e-03) (hash(x)=149770373)
-3040 train 4.188701 (lr=1.0935e-03) (hash(x)=165320741)
-3050 train 4.155907 (lr=1.0830e-03) (hash(x)=170181151)
-3060 train 4.080693 (lr=1.0725e-03) (hash(x)=163837230)
-3070 train 4.075505 (lr=1.0620e-03) (hash(x)=156812797)
-3080 train 4.046247 (lr=1.0516e-03) (hash(x)=143663627)
-3090 train 4.073302 (lr=1.0412e-03) (hash(x)=165052478)
-3100 val loss 4.0906
-3100 val perplexity 59.7749
-3100 train 4.005958 (lr=1.0309e-03) (hash(x)=151458705)
-3110 train 4.055736 (lr=1.0207e-03) (hash(x)=158228317)
-3120 train 4.233213 (lr=1.0104e-03) (hash(x)=149856801)
-3130 train 3.893944 (lr=1.0002e-03) (hash(x)=146297847)
-3140 train 3.997954 (lr=9.9011e-04) (hash(x)=172795167)
-3150 train 4.058806 (lr=9.8002e-04) (hash(x)=145936074)
-3160 train 4.139550 (lr=9.6999e-04) (hash(x)=157564303)
-3170 train 4.235463 (lr=9.6000e-04) (hash(x)=157451362)
-3180 train 4.155211 (lr=9.5007e-04) (hash(x)=156366043)
-3190 train 4.121437 (lr=9.4019e-04) (hash(x)=162344697)
-3200 val loss 4.0748
-3200 val perplexity 58.8384
-3200 train 4.147734 (lr=9.3036e-04) (hash(x)=164206298)
-3210 train 4.210531 (lr=9.2058e-04) (hash(x)=149468251)
-3220 train 4.142988 (lr=9.1085e-04) (hash(x)=157808606)
-3230 train 4.289199 (lr=9.0118e-04) (hash(x)=157078690)
-3240 train 4.083640 (lr=8.9157e-04) (hash(x)=165699696)
-3250 train 4.031158 (lr=8.8201e-04) (hash(x)=181352842)
-3260 train 4.129960 (lr=8.7251e-04) (hash(x)=142534383)
-3270 train 4.085957 (lr=8.6307e-04) (hash(x)=165297359)
-3280 train 4.129749 (lr=8.5368e-04) (hash(x)=160326815)
-3290 train 4.051638 (lr=8.4435e-04) (hash(x)=150043063)
-3300 val loss 4.0650
-3300 val perplexity 58.2644
-3300 train 4.011019 (lr=8.3508e-04) (hash(x)=164133091)
-3310 train 3.947780 (lr=8.2588e-04) (hash(x)=145803863)
-3320 train 3.967892 (lr=8.1673e-04) (hash(x)=157282009)
-3330 train 3.912521 (lr=8.0764e-04) (hash(x)=159126415)
-3340 train 4.122174 (lr=7.9862e-04) (hash(x)=164426675)
-3350 train 4.147809 (lr=7.8966e-04) (hash(x)=170424924)
-3360 train 4.089585 (lr=7.8076e-04) (hash(x)=163161621)
-3370 train 4.149801 (lr=7.7192e-04) (hash(x)=160786731)
-3380 train 4.113340 (lr=7.6315e-04) (hash(x)=165954688)
-3390 train 4.160811 (lr=7.5445e-04) (hash(x)=147371107)
-3400 val loss 4.0439
-3400 val perplexity 57.0467
-3400 train 4.100399 (lr=7.4581e-04) (hash(x)=153067604)
-3410 train 4.098449 (lr=7.3724e-04) (hash(x)=161748196)
-3420 train 4.108459 (lr=7.2874e-04) (hash(x)=168297963)
-3430 train 4.102887 (lr=7.2030e-04) (hash(x)=145086892)
-3440 train 4.006596 (lr=7.1193e-04) (hash(x)=173125711)
-3450 train 4.106333 (lr=7.0363e-04) (hash(x)=158080216)
-3460 train 3.928622 (lr=6.9541e-04) (hash(x)=158055750)
-3470 train 4.027759 (lr=6.8725e-04) (hash(x)=151381400)
-3480 train 3.966299 (lr=6.7916e-04) (hash(x)=147262346)
-3490 train 3.940119 (lr=6.7114e-04) (hash(x)=154740086)
-3500 val loss 4.0414
-3500 val perplexity 56.9081
-3500 train 3.868088 (lr=6.6320e-04) (hash(x)=145730968)
-3510 train 4.013971 (lr=6.5533e-04) (hash(x)=151355929)
-3520 train 4.007189 (lr=6.4753e-04) (hash(x)=165136835)
-3530 train 4.075416 (lr=6.3981e-04) (hash(x)=163869056)
-3540 train 4.020398 (lr=6.3216e-04) (hash(x)=161580134)
-3550 train 4.035475 (lr=6.2458e-04) (hash(x)=168595690)
-3560 train 4.073230 (lr=6.1708e-04) (hash(x)=158576466)
-3570 train 4.034396 (lr=6.0966e-04) (hash(x)=157298619)
-3580 train 4.116951 (lr=6.0231e-04) (hash(x)=160892091)
-3590 train 3.982661 (lr=5.9504e-04) (hash(x)=165168008)
-3600 val loss 4.0216
-3600 val perplexity 55.7906
-3600 train 4.083446 (lr=5.8785e-04) (hash(x)=144911638)
-3610 train 4.045273 (lr=5.8074e-04) (hash(x)=150614513)
-3620 train 4.033285 (lr=5.7370e-04) (hash(x)=167695016)
-3630 train 3.949727 (lr=5.6675e-04) (hash(x)=182708233)
-3640 train 4.056235 (lr=5.5987e-04) (hash(x)=154146274)
-3650 train 4.027171 (lr=5.5308e-04) (hash(x)=159129968)
-3660 train 3.936507 (lr=5.4636e-04) (hash(x)=146209497)
-3670 train 3.958195 (lr=5.3973e-04) (hash(x)=145578104)
-3680 train 3.948595 (lr=5.3318e-04) (hash(x)=150243521)
-3690 train 3.906227 (lr=5.2671e-04) (hash(x)=154328313)
-3700 val loss 4.0240
-3700 val perplexity 55.9251
-3700 train 3.878004 (lr=5.2033e-04) (hash(x)=159904582)
-3710 train 4.217566 (lr=5.1402e-04) (hash(x)=175986417)
-3720 train 4.132571 (lr=5.0780e-04) (hash(x)=158509479)
-3730 train 4.153467 (lr=5.0167e-04) (hash(x)=156087760)
-3740 train 4.150502 (lr=4.9562e-04) (hash(x)=177667170)
-3750 train 4.119051 (lr=4.8965e-04) (hash(x)=145256842)
-3760 train 4.082419 (lr=4.8377e-04) (hash(x)=149855635)
-3770 train 4.114024 (lr=4.7798e-04) (hash(x)=168798886)
-3780 train 4.033966 (lr=4.7227e-04) (hash(x)=156277179)
-3790 train 4.004301 (lr=4.6665e-04) (hash(x)=149698945)
-3800 val loss 4.0062
-3800 val perplexity 54.9362
-3800 train 3.958048 (lr=4.6112e-04) (hash(x)=168039654)
-3810 train 4.072495 (lr=4.5567e-04) (hash(x)=154595574)
-3820 train 4.128203 (lr=4.5031e-04) (hash(x)=158303284)
-3830 train 4.135707 (lr=4.4504e-04) (hash(x)=164128644)
-3840 train 4.095457 (lr=4.3986e-04) (hash(x)=166842997)
-3850 train 4.004663 (lr=4.3477e-04) (hash(x)=177681163)
-3860 train 3.979341 (lr=4.2977e-04) (hash(x)=177062575)
-3870 train 4.020718 (lr=4.2486e-04) (hash(x)=158957159)
-3880 train 3.952075 (lr=4.2004e-04) (hash(x)=151567740)
-3890 train 4.036334 (lr=4.1530e-04) (hash(x)=153261675)
-3900 val loss 3.9958
-3900 val perplexity 54.3706
-3900 train 4.295433 (lr=4.1066e-04) (hash(x)=160268683)
-3910 train 4.024287 (lr=4.0611e-04) (hash(x)=153745240)
-3920 train 4.080532 (lr=4.0166e-04) (hash(x)=164537824)
-3930 train 4.023722 (lr=3.9729e-04) (hash(x)=162062116)
-3940 train 3.987105 (lr=3.9302e-04) (hash(x)=168626742)
-3950 train 4.044534 (lr=3.8884e-04) (hash(x)=148753561)
-3960 train 3.976094 (lr=3.8475e-04) (hash(x)=163526242)
-3970 train 4.187771 (lr=3.8076e-04) (hash(x)=159329323)
-3980 train 4.114512 (lr=3.7685e-04) (hash(x)=146502339)
-3990 train 3.810850 (lr=3.7305e-04) (hash(x)=158409708)
-4000 val loss 3.9911
-4000 val perplexity 54.1126
-4000 train 3.987913 (lr=3.6933e-04) (hash(x)=154645137)
-4010 train 3.992298 (lr=3.6572e-04) (hash(x)=160044026)
-4020 train 3.972251 (lr=3.6219e-04) (hash(x)=155544686)
-4030 train 3.958335 (lr=3.5876e-04) (hash(x)=147589421)
-4040 train 4.072530 (lr=3.5543e-04) (hash(x)=158607990)
-4050 train 3.984080 (lr=3.5219e-04) (hash(x)=154108872)
-4060 train 4.059793 (lr=3.4905e-04) (hash(x)=167101465)
-4070 train 4.021134 (lr=3.4600e-04) (hash(x)=160323363)
-4080 train 4.011046 (lr=3.4305e-04) (hash(x)=150423692)
-4090 train 3.988738 (lr=3.4019e-04) (hash(x)=167461885)
-4100 val loss 3.9801
-4100 val perplexity 53.5204
-4100 train 4.017893 (lr=3.3744e-04) (hash(x)=151783377)
-4110 train 4.020905 (lr=3.3477e-04) (hash(x)=165679649)
-4120 train 3.912448 (lr=3.3221e-04) (hash(x)=153643039)
-4130 train 3.964171 (lr=3.2974e-04) (hash(x)=149648749)
-4140 train 3.997967 (lr=3.2737e-04) (hash(x)=162751418)
-4150 train 3.993538 (lr=3.2510e-04) (hash(x)=157047478)
-4160 train 3.932080 (lr=3.2292e-04) (hash(x)=156391812)
-4170 train 4.005792 (lr=3.2085e-04) (hash(x)=180926985)
-4180 train 4.044862 (lr=3.1887e-04) (hash(x)=168699059)
-4190 train 3.971218 (lr=3.1699e-04) (hash(x)=155410372)
-4200 val loss 3.9797
-4200 val perplexity 53.4996
-4200 train 3.979416 (lr=3.1520e-04) (hash(x)=152615252)
-4210 train 3.948052 (lr=3.1352e-04) (hash(x)=155679453)
-4220 train 4.058191 (lr=3.1193e-04) (hash(x)=152954399)
-4230 train 4.061883 (lr=3.1044e-04) (hash(x)=164608084)
-4240 train 4.088196 (lr=3.0905e-04) (hash(x)=162463867)
-4250 train 4.000721 (lr=3.0776e-04) (hash(x)=168832088)
-4260 train 3.972380 (lr=3.0657e-04) (hash(x)=162698882)
-4270 train 4.005717 (lr=3.0548e-04) (hash(x)=180861492)
-4280 train 3.890420 (lr=3.0449e-04) (hash(x)=160824300)
-4290 train 3.954951 (lr=3.0359e-04) (hash(x)=157750938)
-4300 val loss 3.9767
-4300 val perplexity 53.3412
-4300 train 3.907433 (lr=3.0280e-04) (hash(x)=157216833)
-4310 train 3.906287 (lr=3.0210e-04) (hash(x)=176184661)
-4320 train 4.071271 (lr=3.0150e-04) (hash(x)=161537117)
-4330 train 4.023269 (lr=3.0101e-04) (hash(x)=160700216)
-4340 train 3.975991 (lr=3.0061e-04) (hash(x)=175519903)
-4350 train 3.980127 (lr=3.0031e-04) (hash(x)=167988701)
-4360 train 3.933335 (lr=3.0011e-04) (hash(x)=162619084)
-4370 train 4.315520 (lr=3.0001e-04) (hash(x)=166292684)
-4374 val loss 3.9697
-4374 val perplexity 52.9682
+0 val loss 11.3296
+0 val perplexity 83249.8203
+0 train 11.334114 (lr=1.2000e-05) (hash(x)=165975646)
+10 train 9.589714 (lr=1.3200e-04) (hash(x)=148013305)
+20 train 8.207848 (lr=2.5200e-04) (hash(x)=148456869)
+30 train 7.620950 (lr=3.7200e-04) (hash(x)=161872433)
+40 train 7.358727 (lr=4.9200e-04) (hash(x)=149957686)
+50 train 7.172750 (lr=6.1200e-04) (hash(x)=172543864)
+60 train 7.168416 (lr=7.3200e-04) (hash(x)=179765771)
+70 train 6.798749 (lr=8.5200e-04) (hash(x)=157826813)
+80 train 6.798883 (lr=9.7200e-04) (hash(x)=164392066)
+90 train 6.661848 (lr=1.0920e-03) (hash(x)=155935551)
+100 val loss 6.6047
+100 val perplexity 738.5858
+100 train 6.491619 (lr=1.2120e-03) (hash(x)=163224052)
+110 train 6.518770 (lr=1.3320e-03) (hash(x)=152737029)
+120 train 6.384511 (lr=1.4520e-03) (hash(x)=160672632)
+130 train 6.262604 (lr=1.5720e-03) (hash(x)=163241107)
+140 train 6.281211 (lr=1.6920e-03) (hash(x)=158622777)
+150 train 6.363368 (lr=1.8120e-03) (hash(x)=155912425)
+160 train 6.344102 (lr=1.9320e-03) (hash(x)=171017827)
+170 train 6.243967 (lr=2.0520e-03) (hash(x)=166742757)
+180 train 6.180585 (lr=2.1720e-03) (hash(x)=159924472)
+190 train 6.157166 (lr=2.2920e-03) (hash(x)=153494648)
+200 val loss 6.0887
+200 val perplexity 440.8489
+200 train 6.142984 (lr=2.4120e-03) (hash(x)=161358058)
+210 train 6.066473 (lr=2.5320e-03) (hash(x)=145035701)
+220 train 6.001383 (lr=2.6520e-03) (hash(x)=158648875)
+230 train 6.121311 (lr=2.7720e-03) (hash(x)=164635025)
+240 train 5.850390 (lr=2.8920e-03) (hash(x)=149742497)
+250 train 5.899929 (lr=3.0000e-03) (hash(x)=167005069)
+260 train 5.771887 (lr=3.0000e-03) (hash(x)=163726440)
+270 train 5.749442 (lr=2.9998e-03) (hash(x)=151610423)
+280 train 5.691358 (lr=2.9996e-03) (hash(x)=161421866)
+290 train 5.607024 (lr=2.9994e-03) (hash(x)=138340911)
+300 val loss 5.6924
+300 val perplexity 296.6138
+300 train 5.454905 (lr=2.9990e-03) (hash(x)=146894081)
+310 train 5.407044 (lr=2.9986e-03) (hash(x)=158876217)
+320 train 5.483162 (lr=2.9981e-03) (hash(x)=169472178)
+330 train 5.598046 (lr=2.9975e-03) (hash(x)=157958346)
+340 train 5.616897 (lr=2.9968e-03) (hash(x)=175684344)
+350 train 5.601511 (lr=2.9961e-03) (hash(x)=159051114)
+360 train 5.533087 (lr=2.9953e-03) (hash(x)=152215772)
+370 train 5.540503 (lr=2.9944e-03) (hash(x)=161579675)
+380 train 5.355555 (lr=2.9934e-03) (hash(x)=157026983)
+390 train 5.330606 (lr=2.9923e-03) (hash(x)=156370364)
+400 val loss 5.3227
+400 val perplexity 204.9415
+400 train 5.322221 (lr=2.9912e-03) (hash(x)=154702603)
+410 train 5.279710 (lr=2.9900e-03) (hash(x)=165261055)
+420 train 5.277102 (lr=2.9887e-03) (hash(x)=147555294)
+430 train 5.314711 (lr=2.9873e-03) (hash(x)=166093328)
+440 train 5.215892 (lr=2.9859e-03) (hash(x)=165254485)
+450 train 5.216350 (lr=2.9844e-03) (hash(x)=140386917)
+460 train 5.118093 (lr=2.9828e-03) (hash(x)=166690055)
+470 train 4.917212 (lr=2.9811e-03) (hash(x)=151425273)
+480 train 5.047673 (lr=2.9793e-03) (hash(x)=180686459)
+490 train 4.870615 (lr=2.9775e-03) (hash(x)=160473697)
+500 val loss 5.1584
+500 val perplexity 173.8936
+500 train 4.964884 (lr=2.9756e-03) (hash(x)=154665548)
+510 train 5.078107 (lr=2.9736e-03) (hash(x)=165066653)
+520 train 5.130062 (lr=2.9716e-03) (hash(x)=155375501)
+530 train 5.246745 (lr=2.9694e-03) (hash(x)=157291678)
+540 train 5.157645 (lr=2.9672e-03) (hash(x)=158751844)
+550 train 5.114317 (lr=2.9649e-03) (hash(x)=169683891)
+560 train 5.143666 (lr=2.9625e-03) (hash(x)=144456796)
+570 train 5.116349 (lr=2.9601e-03) (hash(x)=158365992)
+580 train 5.042166 (lr=2.9576e-03) (hash(x)=158188684)
+590 train 5.050703 (lr=2.9550e-03) (hash(x)=164060384)
+600 val loss 5.0012
+600 val perplexity 148.5954
+600 train 5.056788 (lr=2.9523e-03) (hash(x)=154544350)
+610 train 4.966457 (lr=2.9496e-03) (hash(x)=160151177)
+620 train 4.942435 (lr=2.9468e-03) (hash(x)=151113169)
+630 train 4.853900 (lr=2.9439e-03) (hash(x)=162956682)
+640 train 4.847920 (lr=2.9409e-03) (hash(x)=160228084)
+650 train 4.806967 (lr=2.9378e-03) (hash(x)=145530386)
+660 train 4.727909 (lr=2.9347e-03) (hash(x)=163254968)
+670 train 4.720093 (lr=2.9315e-03) (hash(x)=176416192)
+680 train 4.621030 (lr=2.9283e-03) (hash(x)=150118428)
+690 train 4.656602 (lr=2.9249e-03) (hash(x)=150414917)
+700 val loss 4.8629
+700 val perplexity 129.4005
+700 train 4.630126 (lr=2.9215e-03) (hash(x)=162596235)
+710 train 4.910139 (lr=2.9180e-03) (hash(x)=159363258)
+720 train 4.880068 (lr=2.9144e-03) (hash(x)=166036967)
+730 train 4.894136 (lr=2.9108e-03) (hash(x)=159938291)
+740 train 4.853999 (lr=2.9071e-03) (hash(x)=176297617)
+750 train 4.793022 (lr=2.9033e-03) (hash(x)=149845530)
+760 train 4.836494 (lr=2.8994e-03) (hash(x)=152777309)
+770 train 4.780898 (lr=2.8955e-03) (hash(x)=168903075)
+780 train 4.805902 (lr=2.8915e-03) (hash(x)=166843738)
+790 train 4.776241 (lr=2.8874e-03) (hash(x)=166358104)
+800 val loss 4.7443
+800 val perplexity 114.9245
+800 train 4.724244 (lr=2.8833e-03) (hash(x)=153477747)
+810 train 4.633126 (lr=2.8791e-03) (hash(x)=157954385)
+820 train 4.654481 (lr=2.8748e-03) (hash(x)=174383505)
+830 train 4.731977 (lr=2.8704e-03) (hash(x)=145226181)
+840 train 4.676006 (lr=2.8660e-03) (hash(x)=157665743)
+850 train 4.641937 (lr=2.8615e-03) (hash(x)=158565836)
+860 train 4.749852 (lr=2.8569e-03) (hash(x)=171168105)
+870 train 4.747385 (lr=2.8523e-03) (hash(x)=159017777)
+880 train 4.716692 (lr=2.8476e-03) (hash(x)=168156768)
+890 train 4.707652 (lr=2.8428e-03) (hash(x)=145645245)
+900 val loss 4.6663
+900 val perplexity 106.3069
+900 train 4.655340 (lr=2.8379e-03) (hash(x)=151245521)
+910 train 4.657638 (lr=2.8330e-03) (hash(x)=182614672)
+920 train 4.693944 (lr=2.8280e-03) (hash(x)=167377432)
+930 train 4.588878 (lr=2.8230e-03) (hash(x)=171347228)
+940 train 4.601975 (lr=2.8178e-03) (hash(x)=171588755)
+950 train 4.598028 (lr=2.8127e-03) (hash(x)=159560168)
+960 train 4.640528 (lr=2.8074e-03) (hash(x)=171801357)
+970 train 4.598306 (lr=2.8021e-03) (hash(x)=172513884)
+980 train 4.537321 (lr=2.7967e-03) (hash(x)=170659685)
+990 train 4.641826 (lr=2.7912e-03) (hash(x)=165436358)
+1000 val loss 4.5850
+1000 val perplexity 98.0058
+1000 train 4.612329 (lr=2.7857e-03) (hash(x)=167684690)
+1010 train 4.787028 (lr=2.7801e-03) (hash(x)=127225492)
+1020 train 4.682377 (lr=2.7744e-03) (hash(x)=167575323)
+1030 train 4.647320 (lr=2.7687e-03) (hash(x)=161266302)
+1040 train 4.552930 (lr=2.7629e-03) (hash(x)=155607661)
+1050 train 4.624825 (lr=2.7571e-03) (hash(x)=165685359)
+1060 train 4.621967 (lr=2.7512e-03) (hash(x)=160417824)
+1070 train 4.628582 (lr=2.7452e-03) (hash(x)=173313882)
+1080 train 4.514922 (lr=2.7391e-03) (hash(x)=165485047)
+1090 train 4.548283 (lr=2.7330e-03) (hash(x)=159848305)
+1100 val loss 4.5289
+1100 val perplexity 92.6528
+1100 train 4.579960 (lr=2.7269e-03) (hash(x)=170890601)
+1110 train 4.505666 (lr=2.7206e-03) (hash(x)=172689983)
+1120 train 4.536810 (lr=2.7143e-03) (hash(x)=154762859)
+1130 train 4.566061 (lr=2.7080e-03) (hash(x)=168196871)
+1140 train 4.418510 (lr=2.7016e-03) (hash(x)=156715577)
+1150 train 4.649400 (lr=2.6951e-03) (hash(x)=166381788)
+1160 train 4.584499 (lr=2.6886e-03) (hash(x)=162095130)
+1170 train 4.562027 (lr=2.6820e-03) (hash(x)=166094750)
+1180 train 4.502573 (lr=2.6753e-03) (hash(x)=154965753)
+1190 train 4.534716 (lr=2.6686e-03) (hash(x)=172766447)
+1200 val loss 4.4599
+1200 val perplexity 86.4808
+1200 train 4.396911 (lr=2.6618e-03) (hash(x)=161185334)
+1210 train 4.476095 (lr=2.6550e-03) (hash(x)=167224144)
+1220 train 4.458037 (lr=2.6481e-03) (hash(x)=167252866)
+1230 train 4.384659 (lr=2.6411e-03) (hash(x)=155856924)
+1240 train 4.454168 (lr=2.6341e-03) (hash(x)=155956446)
+1250 train 4.298338 (lr=2.6270e-03) (hash(x)=140794796)
+1260 train 4.454971 (lr=2.6199e-03) (hash(x)=166863869)
+1270 train 4.378241 (lr=2.6127e-03) (hash(x)=166021892)
+1280 train 4.449367 (lr=2.6055e-03) (hash(x)=170503648)
+1290 train 4.465008 (lr=2.5982e-03) (hash(x)=164012719)
+1300 val loss 4.4191
+1300 val perplexity 83.0204
+1300 train 4.431269 (lr=2.5909e-03) (hash(x)=156519834)
+1310 train 4.619505 (lr=2.5835e-03) (hash(x)=165146652)
+1320 train 4.369074 (lr=2.5760e-03) (hash(x)=174391990)
+1330 train 4.409378 (lr=2.5685e-03) (hash(x)=161131201)
+1340 train 4.515519 (lr=2.5609e-03) (hash(x)=162240912)
+1350 train 4.336414 (lr=2.5533e-03) (hash(x)=152462931)
+1360 train 4.450696 (lr=2.5457e-03) (hash(x)=150443278)
+1370 train 4.367658 (lr=2.5379e-03) (hash(x)=153721711)
+1380 train 4.363069 (lr=2.5302e-03) (hash(x)=158070628)
+1390 train 4.412385 (lr=2.5223e-03) (hash(x)=163318827)
+1400 val loss 4.4054
+1400 val perplexity 81.8917
+1400 train 4.370369 (lr=2.5145e-03) (hash(x)=154210336)
+1410 train 4.525505 (lr=2.5066e-03) (hash(x)=162393933)
+1420 train 4.593505 (lr=2.4986e-03) (hash(x)=156659308)
+1430 train 4.382072 (lr=2.4906e-03) (hash(x)=171243439)
+1440 train 4.513375 (lr=2.4825e-03) (hash(x)=178459521)
+1450 train 4.436228 (lr=2.4744e-03) (hash(x)=159166350)
+1460 train 4.349923 (lr=2.4662e-03) (hash(x)=173554569)
+1470 train 4.390570 (lr=2.4580e-03) (hash(x)=158362521)
+1480 train 4.436795 (lr=2.4497e-03) (hash(x)=165899207)
+1490 train 4.346364 (lr=2.4414e-03) (hash(x)=156470939)
+1500 val loss 4.3692
+1500 val perplexity 78.9842
+1500 train 4.405301 (lr=2.4331e-03) (hash(x)=164197780)
+1510 train 4.349495 (lr=2.4247e-03) (hash(x)=146709896)
+1520 train 4.380601 (lr=2.4162e-03) (hash(x)=151304838)
+1530 train 4.273098 (lr=2.4077e-03) (hash(x)=158179918)
+1540 train 4.303412 (lr=2.3992e-03) (hash(x)=161256064)
+1550 train 4.345532 (lr=2.3906e-03) (hash(x)=151591940)
+1560 train 4.280371 (lr=2.3820e-03) (hash(x)=147120266)
+1570 train 4.431362 (lr=2.3734e-03) (hash(x)=181404631)
+1580 train 4.337122 (lr=2.3647e-03) (hash(x)=157486637)
+1590 train 4.474000 (lr=2.3559e-03) (hash(x)=156440796)
+1600 val loss 4.3434
+1600 val perplexity 76.9667
+1600 train 4.376644 (lr=2.3471e-03) (hash(x)=172479098)
+1610 train 4.516253 (lr=2.3383e-03) (hash(x)=166059760)
+1620 train 4.388950 (lr=2.3294e-03) (hash(x)=165353726)
+1630 train 4.361284 (lr=2.3205e-03) (hash(x)=159323060)
+1640 train 4.458383 (lr=2.3116e-03) (hash(x)=168531548)
+1650 train 4.308321 (lr=2.3026e-03) (hash(x)=151279450)
+1660 train 4.362363 (lr=2.2936e-03) (hash(x)=149136290)
+1670 train 4.295530 (lr=2.2845e-03) (hash(x)=175415017)
+1680 train 4.186372 (lr=2.2754e-03) (hash(x)=160981851)
+1690 train 4.243690 (lr=2.2663e-03) (hash(x)=162044381)
+1700 val loss 4.3476
+1700 val perplexity 77.2918
+1700 train 4.423730 (lr=2.2572e-03) (hash(x)=154630837)
+1710 train 4.412186 (lr=2.2480e-03) (hash(x)=176778401)
+1720 train 4.412152 (lr=2.2387e-03) (hash(x)=154384524)
+1730 train 4.391443 (lr=2.2294e-03) (hash(x)=160342845)
+1740 train 4.389899 (lr=2.2201e-03) (hash(x)=167176158)
+1750 train 4.321714 (lr=2.2108e-03) (hash(x)=165200240)
+1760 train 4.393856 (lr=2.2014e-03) (hash(x)=152577632)
+1770 train 4.262097 (lr=2.1920e-03) (hash(x)=140044301)
+1780 train 4.346868 (lr=2.1826e-03) (hash(x)=175722780)
+1790 train 4.353307 (lr=2.1731e-03) (hash(x)=156350845)
+1800 val loss 4.3014
+1800 val perplexity 73.7999
+1800 train 4.398052 (lr=2.1637e-03) (hash(x)=155062509)
+1810 train 4.267949 (lr=2.1541e-03) (hash(x)=163860874)
+1820 train 4.141739 (lr=2.1446e-03) (hash(x)=158079613)
+1830 train 4.261207 (lr=2.1350e-03) (hash(x)=155660923)
+1840 train 4.233153 (lr=2.1254e-03) (hash(x)=158022360)
+1850 train 4.222778 (lr=2.1158e-03) (hash(x)=168432793)
+1860 train 4.384053 (lr=2.1061e-03) (hash(x)=157042286)
+1870 train 4.374121 (lr=2.0964e-03) (hash(x)=156443358)
+1880 train 4.264701 (lr=2.0867e-03) (hash(x)=155942931)
+1890 train 4.370747 (lr=2.0769e-03) (hash(x)=157175884)
+1900 val loss 4.2760
+1900 val perplexity 71.9552
+1900 train 4.279358 (lr=2.0672e-03) (hash(x)=162546009)
+1910 train 4.254579 (lr=2.0574e-03) (hash(x)=151139265)
+1920 train 4.289567 (lr=2.0476e-03) (hash(x)=154498634)
+1930 train 4.163076 (lr=2.0377e-03) (hash(x)=157286832)
+1940 train 4.254378 (lr=2.0279e-03) (hash(x)=138309582)
+1950 train 4.239661 (lr=2.0180e-03) (hash(x)=158819213)
+1960 train 4.150318 (lr=2.0081e-03) (hash(x)=151883403)
+1970 train 4.340694 (lr=1.9982e-03) (hash(x)=194409548)
+1980 train 4.238113 (lr=1.9882e-03) (hash(x)=160739690)
+1990 train 4.269607 (lr=1.9783e-03) (hash(x)=166445519)
+2000 val loss 4.2623
+2000 val perplexity 70.9697
+2000 train 4.287572 (lr=1.9683e-03) (hash(x)=159673845)
+2010 train 4.302210 (lr=1.9583e-03) (hash(x)=168806069)
+2020 train 4.319132 (lr=1.9483e-03) (hash(x)=160581536)
+2030 train 4.272835 (lr=1.9382e-03) (hash(x)=156141720)
+2040 train 4.389905 (lr=1.9282e-03) (hash(x)=170531572)
+2050 train 4.230220 (lr=1.9181e-03) (hash(x)=160037105)
+2060 train 4.460987 (lr=1.9080e-03) (hash(x)=150757653)
+2070 train 4.282402 (lr=1.8979e-03) (hash(x)=166026751)
+2080 train 4.297229 (lr=1.8878e-03) (hash(x)=157758002)
+2090 train 4.146669 (lr=1.8777e-03) (hash(x)=154520168)
+2100 val loss 4.2316
+2100 val perplexity 68.8305
+2100 train 4.180870 (lr=1.8675e-03) (hash(x)=164198408)
+2110 train 4.174623 (lr=1.8574e-03) (hash(x)=152821408)
+2120 train 4.186363 (lr=1.8472e-03) (hash(x)=158276857)
+2130 train 4.329699 (lr=1.8370e-03) (hash(x)=163521176)
+2140 train 4.364914 (lr=1.8268e-03) (hash(x)=168732683)
+2150 train 4.268414 (lr=1.8166e-03) (hash(x)=149299542)
+2160 train 4.298939 (lr=1.8064e-03) (hash(x)=157444871)
+2170 train 4.310306 (lr=1.7962e-03) (hash(x)=143188824)
+2180 train 4.280026 (lr=1.7860e-03) (hash(x)=158976812)
+2190 train 4.246160 (lr=1.7758e-03) (hash(x)=165191158)
+2200 val loss 4.2160
+2200 val perplexity 67.7592
+2200 train 4.219477 (lr=1.7655e-03) (hash(x)=152995704)
+2210 train 4.275416 (lr=1.7553e-03) (hash(x)=168187659)
+2220 train 4.175550 (lr=1.7450e-03) (hash(x)=155775581)
+2230 train 4.085768 (lr=1.7348e-03) (hash(x)=154680350)
+2240 train 4.176087 (lr=1.7245e-03) (hash(x)=147471278)
+2250 train 4.119020 (lr=1.7142e-03) (hash(x)=159695346)
+2260 train 4.145622 (lr=1.7040e-03) (hash(x)=151025882)
+2270 train 4.115958 (lr=1.6937e-03) (hash(x)=156521158)
+2280 train 4.436255 (lr=1.6834e-03) (hash(x)=156767674)
+2290 train 4.259054 (lr=1.6731e-03) (hash(x)=164457648)
+2300 val loss 4.1949
+2300 val perplexity 66.3475
+2300 train 4.198524 (lr=1.6629e-03) (hash(x)=169509455)
+2310 train 4.277474 (lr=1.6526e-03) (hash(x)=152810614)
+2320 train 4.364194 (lr=1.6423e-03) (hash(x)=158237154)
+2330 train 4.229253 (lr=1.6320e-03) (hash(x)=170356103)
+2340 train 4.317828 (lr=1.6217e-03) (hash(x)=158128269)
+2350 train 4.193096 (lr=1.6114e-03) (hash(x)=138332801)
+2360 train 4.257848 (lr=1.6012e-03) (hash(x)=170994284)
+2370 train 4.351678 (lr=1.5909e-03) (hash(x)=173415638)
+2380 train 4.225911 (lr=1.5806e-03) (hash(x)=154689617)
+2390 train 4.255297 (lr=1.5704e-03) (hash(x)=161793283)
+2400 val loss 4.1873
+2400 val perplexity 65.8436
+2400 train 4.166761 (lr=1.5601e-03) (hash(x)=152419100)
+2410 train 3.985188 (lr=1.5498e-03) (hash(x)=156948391)
+2420 train 4.107901 (lr=1.5396e-03) (hash(x)=158156048)
+2430 train 4.107360 (lr=1.5294e-03) (hash(x)=160653203)
+2440 train 4.079838 (lr=1.5191e-03) (hash(x)=161323851)
+2450 train 4.030266 (lr=1.5089e-03) (hash(x)=159253476)
+2460 train 3.930699 (lr=1.4987e-03) (hash(x)=144756346)
+2470 train 3.942721 (lr=1.4885e-03) (hash(x)=159239815)
+2480 train 4.222439 (lr=1.4783e-03) (hash(x)=155086855)
+2490 train 4.241613 (lr=1.4681e-03) (hash(x)=170243577)
+2500 val loss 4.1600
+2500 val perplexity 64.0719
+2500 train 4.197385 (lr=1.4579e-03) (hash(x)=171811405)
+2510 train 4.210459 (lr=1.4477e-03) (hash(x)=155717902)
+2520 train 4.196662 (lr=1.4375e-03) (hash(x)=160901663)
+2530 train 4.200081 (lr=1.4274e-03) (hash(x)=158088736)
+2540 train 4.235577 (lr=1.4173e-03) (hash(x)=157067679)
+2550 train 4.187673 (lr=1.4071e-03) (hash(x)=157320007)
+2560 train 4.105568 (lr=1.3970e-03) (hash(x)=153565165)
+2570 train 4.130184 (lr=1.3869e-03) (hash(x)=160010797)
+2580 train 4.058619 (lr=1.3769e-03) (hash(x)=168786235)
+2590 train 4.135674 (lr=1.3668e-03) (hash(x)=149540363)
+2600 val loss 4.1444
+2600 val perplexity 63.0822
+2600 train 4.109813 (lr=1.3568e-03) (hash(x)=152837043)
+2610 train 4.053783 (lr=1.3467e-03) (hash(x)=160015391)
+2620 train 3.999572 (lr=1.3367e-03) (hash(x)=170364902)
+2630 train 3.924249 (lr=1.3267e-03) (hash(x)=159798423)
+2640 train 3.956265 (lr=1.3168e-03) (hash(x)=139413884)
+2650 train 4.134111 (lr=1.3068e-03) (hash(x)=163952153)
+2660 train 4.154635 (lr=1.2969e-03) (hash(x)=163253609)
+2670 train 4.352928 (lr=1.2870e-03) (hash(x)=158449696)
+2680 train 4.244215 (lr=1.2771e-03) (hash(x)=160774597)
+2690 train 4.123965 (lr=1.2672e-03) (hash(x)=164911880)
+2700 val loss 4.1374
+2700 val perplexity 62.6398
+2700 train 4.250805 (lr=1.2573e-03) (hash(x)=157220460)
+2710 train 4.357726 (lr=1.2475e-03) (hash(x)=157423915)
+2720 train 4.095804 (lr=1.2377e-03) (hash(x)=151126931)
+2730 train 4.073377 (lr=1.2279e-03) (hash(x)=161890897)
+2740 train 4.050958 (lr=1.2182e-03) (hash(x)=156373676)
+2750 train 4.122328 (lr=1.2085e-03) (hash(x)=165480571)
+2760 train 4.103857 (lr=1.1988e-03) (hash(x)=130397806)
+2770 train 4.101247 (lr=1.1891e-03) (hash(x)=156871436)
+2780 train 4.047716 (lr=1.1794e-03) (hash(x)=168114417)
+2790 train 4.088237 (lr=1.1698e-03) (hash(x)=156585785)
+2800 val loss 4.1222
+2800 val perplexity 61.6965
+2800 train 3.941895 (lr=1.1602e-03) (hash(x)=163983863)
+2810 train 3.914798 (lr=1.1506e-03) (hash(x)=160405761)
+2820 train 3.935615 (lr=1.1411e-03) (hash(x)=163831817)
+2830 train 3.987344 (lr=1.1316e-03) (hash(x)=150112598)
+2840 train 3.956249 (lr=1.1221e-03) (hash(x)=162075903)
+2850 train 4.197121 (lr=1.1127e-03) (hash(x)=165111358)
+2860 train 4.179041 (lr=1.1033e-03) (hash(x)=153236786)
+2870 train 4.320436 (lr=1.0939e-03) (hash(x)=161816361)
+2880 train 4.167611 (lr=1.0845e-03) (hash(x)=151864522)
+2890 train 4.020428 (lr=1.0752e-03) (hash(x)=156002477)
+2900 val loss 4.0972
+2900 val perplexity 60.1736
+2900 train 4.075690 (lr=1.0659e-03) (hash(x)=154611558)
+2910 train 4.261486 (lr=1.0567e-03) (hash(x)=150322906)
+2920 train 4.144260 (lr=1.0474e-03) (hash(x)=136550018)
+2930 train 4.161936 (lr=1.0383e-03) (hash(x)=163368039)
+2940 train 4.048949 (lr=1.0291e-03) (hash(x)=140876015)
+2950 train 4.087054 (lr=1.0200e-03) (hash(x)=166845778)
+2960 train 4.049969 (lr=1.0109e-03) (hash(x)=163143219)
+2970 train 3.997180 (lr=1.0019e-03) (hash(x)=153818849)
+2980 train 4.066802 (lr=9.9289e-04) (hash(x)=156917850)
+2990 train 3.997714 (lr=9.8392e-04) (hash(x)=173189206)
+3000 val loss 4.0977
+3000 val perplexity 60.1994
+3000 train 3.945580 (lr=9.7500e-04) (hash(x)=149025011)
+3010 train 3.824667 (lr=9.6612e-04) (hash(x)=159783526)
+3020 train 3.944181 (lr=9.5727e-04) (hash(x)=153505500)
+3030 train 4.219179 (lr=9.4847e-04) (hash(x)=156235961)
+3040 train 4.060788 (lr=9.3970e-04) (hash(x)=167294230)
+3050 train 4.143058 (lr=9.3098e-04) (hash(x)=166095533)
+3060 train 4.120775 (lr=9.2230e-04) (hash(x)=118432481)
+3070 train 4.124558 (lr=9.1366e-04) (hash(x)=165105881)
+3080 train 4.202753 (lr=9.0506e-04) (hash(x)=165228401)
+3090 train 4.201431 (lr=8.9651e-04) (hash(x)=159692678)
+3100 val loss 4.0796
+3100 val perplexity 59.1226
+3100 train 4.107082 (lr=8.8800e-04) (hash(x)=158682056)
+3110 train 4.174942 (lr=8.7954e-04) (hash(x)=152985093)
+3120 train 3.867309 (lr=8.7112e-04) (hash(x)=175378301)
+3130 train 3.897981 (lr=8.6274e-04) (hash(x)=162798550)
+3140 train 3.975251 (lr=8.5441e-04) (hash(x)=155589161)
+3150 train 4.118763 (lr=8.4613e-04) (hash(x)=159716727)
+3160 train 4.170940 (lr=8.3789e-04) (hash(x)=166758382)
+3170 train 4.149882 (lr=8.2970e-04) (hash(x)=170684998)
+3180 train 4.166050 (lr=8.2156e-04) (hash(x)=145890459)
+3190 train 4.113280 (lr=8.1347e-04) (hash(x)=131736018)
+3200 val loss 4.0624
+3200 val perplexity 58.1162
+3200 train 4.141879 (lr=8.0542e-04) (hash(x)=163155085)
+3210 train 4.130290 (lr=7.9742e-04) (hash(x)=158434140)
+3220 train 4.135093 (lr=7.8948e-04) (hash(x)=153045472)
+3230 train 4.095717 (lr=7.8158e-04) (hash(x)=162298119)
+3240 train 4.156842 (lr=7.7373e-04) (hash(x)=154940868)
+3250 train 4.048419 (lr=7.6594e-04) (hash(x)=146164261)
+3260 train 4.064459 (lr=7.5819e-04) (hash(x)=180063993)
+3270 train 4.005987 (lr=7.5050e-04) (hash(x)=145912749)
+3280 train 4.038804 (lr=7.4286e-04) (hash(x)=160485262)
+3290 train 4.034842 (lr=7.3527e-04) (hash(x)=161184342)
+3300 val loss 4.0549
+3300 val perplexity 57.6801
+3300 train 3.941289 (lr=7.2774e-04) (hash(x)=152327761)
+3310 train 3.969263 (lr=7.2026e-04) (hash(x)=153447140)
+3320 train 4.061533 (lr=7.1283e-04) (hash(x)=147161965)
+3330 train 3.859341 (lr=7.0545e-04) (hash(x)=156739807)
+3340 train 4.121389 (lr=6.9814e-04) (hash(x)=159957642)
+3350 train 4.160736 (lr=6.9087e-04) (hash(x)=166328571)
+3360 train 4.151203 (lr=6.8367e-04) (hash(x)=151475934)
+3370 train 4.115513 (lr=6.7651e-04) (hash(x)=153931271)
+3380 train 4.112267 (lr=6.6942e-04) (hash(x)=171487011)
+3390 train 4.209892 (lr=6.6238e-04) (hash(x)=151804212)
+3400 val loss 4.0421
+3400 val perplexity 56.9445
+3400 train 4.113207 (lr=6.5540e-04) (hash(x)=153631674)
+3410 train 4.162016 (lr=6.4848e-04) (hash(x)=152678595)
+3420 train 4.137099 (lr=6.4161e-04) (hash(x)=153365144)
+3430 train 4.141204 (lr=6.3480e-04) (hash(x)=166893474)
+3440 train 4.094139 (lr=6.2806e-04) (hash(x)=165535548)
+3450 train 4.135862 (lr=6.2137e-04) (hash(x)=144370685)
+3460 train 4.009230 (lr=6.1474e-04) (hash(x)=154254791)
+3470 train 4.048392 (lr=6.0817e-04) (hash(x)=164497095)
+3480 train 4.061123 (lr=6.0166e-04) (hash(x)=168522376)
+3490 train 3.918683 (lr=5.9521e-04) (hash(x)=153896630)
+3500 val loss 4.0363
+3500 val perplexity 56.6141
+3500 train 3.796441 (lr=5.8883e-04) (hash(x)=154741045)
+3510 train 3.991416 (lr=5.8250e-04) (hash(x)=164266299)
+3520 train 3.917408 (lr=5.7624e-04) (hash(x)=158654950)
+3530 train 4.155674 (lr=5.7004e-04) (hash(x)=158230617)
+3540 train 4.089488 (lr=5.6390e-04) (hash(x)=161601342)
+3550 train 4.220429 (lr=5.5783e-04) (hash(x)=167990373)
+3560 train 4.058076 (lr=5.5182e-04) (hash(x)=169303878)
+3570 train 4.146748 (lr=5.4587e-04) (hash(x)=144079588)
+3580 train 3.988824 (lr=5.3998e-04) (hash(x)=166458297)
+3590 train 4.039760 (lr=5.3416e-04) (hash(x)=164303280)
+3600 val loss 4.0200
+3600 val perplexity 55.6997
+3600 train 4.077363 (lr=5.2841e-04) (hash(x)=159231031)
+3610 train 4.122448 (lr=5.2272e-04) (hash(x)=162424857)
+3620 train 4.015436 (lr=5.1710e-04) (hash(x)=154559262)
+3630 train 4.035576 (lr=5.1154e-04) (hash(x)=156608060)
+3640 train 4.046021 (lr=5.0604e-04) (hash(x)=157446611)
+3650 train 4.065583 (lr=5.0062e-04) (hash(x)=174610578)
+3660 train 3.975127 (lr=4.9526e-04) (hash(x)=158275381)
+3670 train 3.881399 (lr=4.8997e-04) (hash(x)=157437517)
+3680 train 3.826926 (lr=4.8474e-04) (hash(x)=165060839)
+3690 train 3.928437 (lr=4.7958e-04) (hash(x)=169543799)
+3700 val loss 4.0208
+3700 val perplexity 55.7448
+3700 train 4.048800 (lr=4.7449e-04) (hash(x)=140978922)
+3710 train 4.203963 (lr=4.6947e-04) (hash(x)=155829787)
+3720 train 4.081343 (lr=4.6452e-04) (hash(x)=171433940)
+3730 train 4.090452 (lr=4.5963e-04) (hash(x)=167899435)
+3740 train 4.379403 (lr=4.5482e-04) (hash(x)=164355528)
+3750 train 4.072184 (lr=4.5007e-04) (hash(x)=162327765)
+3760 train 4.113864 (lr=4.4540e-04) (hash(x)=156040613)
+3770 train 4.109349 (lr=4.4079e-04) (hash(x)=160180518)
+3780 train 4.026843 (lr=4.3625e-04) (hash(x)=150016441)
+3790 train 4.123217 (lr=4.3179e-04) (hash(x)=159325196)
+3800 val loss 4.0067
+3800 val perplexity 54.9655
+3800 train 4.002306 (lr=4.2739e-04) (hash(x)=151622199)
+3810 train 3.903182 (lr=4.2307e-04) (hash(x)=150340066)
+3820 train 4.060265 (lr=4.1881e-04) (hash(x)=185510998)
+3830 train 4.061746 (lr=4.1463e-04) (hash(x)=167230996)
+3840 train 3.986496 (lr=4.1052e-04) (hash(x)=165310104)
+3850 train 4.039975 (lr=4.0648e-04) (hash(x)=163449872)
+3860 train 4.020473 (lr=4.0252e-04) (hash(x)=159560710)
+3870 train 3.917598 (lr=3.9862e-04) (hash(x)=160392242)
+3880 train 4.009766 (lr=3.9480e-04) (hash(x)=145396414)
+3890 train 4.039589 (lr=3.9105e-04) (hash(x)=151853804)
+3900 val loss 3.9993
+3900 val perplexity 54.5626
+3900 train 3.988944 (lr=3.8738e-04) (hash(x)=171224252)
+3910 train 4.072381 (lr=3.8378e-04) (hash(x)=160454722)
+3920 train 4.077506 (lr=3.8025e-04) (hash(x)=164424859)
+3930 train 4.006749 (lr=3.7679e-04) (hash(x)=151089700)
+3940 train 4.055073 (lr=3.7341e-04) (hash(x)=160835936)
+3950 train 4.004356 (lr=3.7010e-04) (hash(x)=155067663)
+3960 train 4.001670 (lr=3.6687e-04) (hash(x)=155486171)
+3970 train 4.065631 (lr=3.6371e-04) (hash(x)=159851797)
+3980 train 4.033099 (lr=3.6063e-04) (hash(x)=156210123)
+3990 train 3.897400 (lr=3.5762e-04) (hash(x)=155100451)
+4000 val loss 3.9945
+4000 val perplexity 54.2968
+4000 train 3.983969 (lr=3.5468e-04) (hash(x)=154881551)
+4010 train 4.053800 (lr=3.5183e-04) (hash(x)=163349043)
+4020 train 3.983883 (lr=3.4904e-04) (hash(x)=155886322)
+4030 train 4.110051 (lr=3.4633e-04) (hash(x)=159537959)
+4040 train 4.043459 (lr=3.4370e-04) (hash(x)=152602379)
+4050 train 4.034940 (lr=3.4114e-04) (hash(x)=161610706)
+4060 train 4.008632 (lr=3.3866e-04) (hash(x)=154804860)
+4070 train 3.980394 (lr=3.3626e-04) (hash(x)=162296585)
+4080 train 4.067740 (lr=3.3393e-04) (hash(x)=157756706)
+4090 train 4.096369 (lr=3.3168e-04) (hash(x)=160386646)
+4100 val loss 3.9821
+4100 val perplexity 53.6282
+4100 train 4.033006 (lr=3.2950e-04) (hash(x)=163637730)
+4110 train 4.041345 (lr=3.2740e-04) (hash(x)=161965728)
+4120 train 4.042684 (lr=3.2538e-04) (hash(x)=156423546)
+4130 train 3.982870 (lr=3.2343e-04) (hash(x)=163053830)
+4140 train 3.971513 (lr=3.2156e-04) (hash(x)=153198863)
+4150 train 3.973744 (lr=3.1977e-04) (hash(x)=159899969)
+4160 train 3.937757 (lr=3.1806e-04) (hash(x)=164336555)
+4170 train 3.920032 (lr=3.1642e-04) (hash(x)=159546648)
+4180 train 3.987313 (lr=3.1486e-04) (hash(x)=159128343)
+4190 train 3.961296 (lr=3.1338e-04) (hash(x)=190404156)
+4200 val loss 3.9902
+4200 val perplexity 54.0643
+4200 train 3.982952 (lr=3.1197e-04) (hash(x)=160936605)
+4210 train 4.004089 (lr=3.1065e-04) (hash(x)=155496712)
+4220 train 4.030804 (lr=3.0940e-04) (hash(x)=155907702)
+4230 train 4.024456 (lr=3.0822e-04) (hash(x)=159562930)
+4240 train 4.018976 (lr=3.0713e-04) (hash(x)=169512573)
+4250 train 4.057533 (lr=3.0611e-04) (hash(x)=155431435)
+4260 train 4.045427 (lr=3.0517e-04) (hash(x)=155467567)
+4270 train 3.915394 (lr=3.0431e-04) (hash(x)=160350168)
+4280 train 3.984566 (lr=3.0353e-04) (hash(x)=155154478)
+4290 train 3.882757 (lr=3.0283e-04) (hash(x)=161981515)
+4300 val loss 3.9778
+4300 val perplexity 53.4000
+4300 train 4.063294 (lr=3.0220e-04) (hash(x)=147992100)
+4310 train 3.965492 (lr=3.0165e-04) (hash(x)=162522212)
+4320 train 3.976317 (lr=3.0118e-04) (hash(x)=156565952)
+4330 train 4.106852 (lr=3.0079e-04) (hash(x)=158960384)
+4340 train 4.014628 (lr=3.0048e-04) (hash(x)=154314808)
+4350 train 4.034998 (lr=3.0024e-04) (hash(x)=174193276)
+4360 train 3.960939 (lr=3.0009e-04) (hash(x)=157693382)
+4370 train 4.153555 (lr=3.0001e-04) (hash(x)=149511108)
+4374 val loss 3.9727
+4374 val perplexity 53.1301

scaling_mha_granularity_2/hd_128_seed_1339/model_04374.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:08e046638ac211d914272c598bf74f5abf5112689e26ab29e28d6296929ff86b
+oid sha256:3220aa11fad715d7b996ee186b2067b27379b5f3c4bebf6148d445c5a9721318
 size 89692138

scaling_mha_granularity_2/hd_128_seed_1339/optimizer_04374.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2a6c88f51d8faec21b4867cb8288b30f96b14866a62855d016b6c8504d5d6b82
+oid sha256:6fb4f3e6a9d9bb5816a16d2a6a37d53b2053eb3d4d55ac99073674b0b87e053f
 size 179406342