| PEER21_CONFIG {"ctrl_params": 598145536, "active_params_per_token": 70541312, "experts": 495616, "peer_layers": [4, 8, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 69663232 SPARSE_POOL 507510784 |
| step=100 tokens=1022000 ce=9.6171 ppl=15019.3 bal=1.006 mut=0.001 tok_s=18,463 avg_tok_s=18,463 |
| step=200 tokens=2044000 ce=8.5571 ppl=5203.4 bal=1.019 mut=0.015 tok_s=33,723 avg_tok_s=26,093 |
| step=300 tokens=3066000 ce=7.5502 ppl=1901.0 bal=1.003 mut=0.058 tok_s=33,777 avg_tok_s=28,655 |
| step=400 tokens=4088000 ce=7.0258 ppl=1125.3 bal=1.001 mut=0.160 tok_s=33,834 avg_tok_s=29,949 |
| step=500 tokens=5110000 ce=6.9713 ppl=1065.6 bal=1.001 mut=0.267 tok_s=33,836 avg_tok_s=30,727 |
| step=600 tokens=6132000 ce=6.6434 ppl=767.7 bal=1.002 mut=0.367 tok_s=34,081 avg_tok_s=31,286 |
| step=700 tokens=7154000 ce=6.5984 ppl=733.9 bal=1.002 mut=0.461 tok_s=34,257 avg_tok_s=31,710 |
| step=800 tokens=8176000 ce=6.3388 ppl=566.1 bal=1.002 mut=0.546 tok_s=34,082 avg_tok_s=32,007 |
| step=900 tokens=9198000 ce=6.3211 ppl=556.2 bal=1.003 mut=0.600 tok_s=34,372 avg_tok_s=32,269 |
| step=1000 tokens=10220000 ce=6.1051 ppl=448.1 bal=1.003 mut=0.643 tok_s=34,421 avg_tok_s=32,485 |
| step=1100 tokens=11242000 ce=6.1585 ppl=472.7 bal=1.006 mut=0.678 tok_s=34,546 avg_tok_s=32,672 |
| step=1200 tokens=12264000 ce=6.1807 ppl=483.3 bal=1.006 mut=0.716 tok_s=34,629 avg_tok_s=32,835 |
| step=1300 tokens=13286000 ce=6.0142 ppl=409.2 bal=1.006 mut=0.741 tok_s=34,729 avg_tok_s=32,981 |
| step=1400 tokens=14308000 ce=5.9990 ppl=403.0 bal=1.012 mut=0.756 tok_s=34,606 avg_tok_s=33,097 |
| step=1500 tokens=15330000 ce=5.6967 ppl=297.9 bal=1.015 mut=0.773 tok_s=34,657 avg_tok_s=33,201 |
| step=1600 tokens=16352000 ce=5.8050 ppl=332.0 bal=1.022 mut=0.794 tok_s=34,535 avg_tok_s=33,284 |
| step=1700 tokens=17374000 ce=5.8189 ppl=336.6 bal=1.043 mut=0.810 tok_s=34,768 avg_tok_s=33,372 |
| step=1800 tokens=18396000 ce=5.6231 ppl=276.8 bal=1.061 mut=0.825 tok_s=34,667 avg_tok_s=33,443 |
| step=1900 tokens=19418000 ce=5.7805 ppl=323.9 bal=1.101 mut=0.837 tok_s=34,426 avg_tok_s=33,495 |
| step=2000 tokens=20440000 ce=5.5480 ppl=256.7 bal=1.193 mut=0.856 tok_s=34,117 avg_tok_s=33,526 |
| CKPT step=2000 valid_ce=5.6409 valid_ppl=281.7 |
| step=2100 tokens=21462000 ce=5.8349 ppl=342.0 bal=1.321 mut=0.858 tok_s=23,507 avg_tok_s=33,049 |
| step=2200 tokens=22484000 ce=5.5210 ppl=249.9 bal=1.338 mut=0.854 tok_s=33,804 avg_tok_s=33,083 |
| step=2300 tokens=23506000 ce=5.6261 ppl=277.6 bal=1.446 mut=0.845 tok_s=34,123 avg_tok_s=33,129 |
| step=2400 tokens=24528000 ce=5.5507 ppl=257.4 bal=1.399 mut=0.836 tok_s=33,899 avg_tok_s=33,161 |
| step=2500 tokens=25550000 ce=5.4747 ppl=238.6 bal=1.554 mut=0.827 tok_s=33,876 avg_tok_s=33,189 |
| step=2600 tokens=26572000 ce=5.5843 ppl=266.2 bal=2.039 mut=0.817 tok_s=33,526 avg_tok_s=33,202 |
| step=2700 tokens=27594000 ce=5.4278 ppl=227.7 bal=1.551 mut=0.815 tok_s=33,683 avg_tok_s=33,220 |
| step=2800 tokens=28616000 ce=5.5181 ppl=249.2 bal=1.812 mut=0.799 tok_s=33,594 avg_tok_s=33,234 |
| step=2900 tokens=29638000 ce=5.4603 ppl=235.2 bal=1.731 mut=0.787 tok_s=33,748 avg_tok_s=33,251 |
| step=3000 tokens=30660000 ce=5.3871 ppl=218.6 bal=2.565 mut=0.770 tok_s=33,759 avg_tok_s=33,268 |
| step=3100 tokens=31682000 ce=5.4449 ppl=231.6 bal=1.962 mut=0.719 tok_s=33,846 avg_tok_s=33,287 |
| step=3200 tokens=32704000 ce=5.1785 ppl=177.4 bal=2.069 mut=0.639 tok_s=33,765 avg_tok_s=33,302 |
| step=3300 tokens=33726000 ce=5.2387 ppl=188.4 bal=2.143 mut=0.602 tok_s=33,947 avg_tok_s=33,321 |
| step=3400 tokens=34748000 ce=5.4613 ppl=235.4 bal=2.512 mut=0.602 tok_s=33,770 avg_tok_s=33,335 |
| step=3500 tokens=35770000 ce=5.1858 ppl=178.7 bal=4.191 mut=0.618 tok_s=33,992 avg_tok_s=33,353 |
| step=3600 tokens=36792000 ce=5.3097 ppl=202.3 bal=2.291 mut=0.656 tok_s=33,892 avg_tok_s=33,368 |
| step=3700 tokens=37814000 ce=5.3004 ppl=200.4 bal=2.296 mut=0.666 tok_s=33,994 avg_tok_s=33,385 |
| step=3800 tokens=38836000 ce=5.4552 ppl=234.0 bal=2.178 mut=0.609 tok_s=34,014 avg_tok_s=33,402 |
| step=3900 tokens=39858000 ce=5.2438 ppl=189.4 bal=2.286 mut=0.568 tok_s=33,718 avg_tok_s=33,410 |
| step=4000 tokens=40880000 ce=5.2567 ppl=191.9 bal=3.247 mut=0.543 tok_s=33,895 avg_tok_s=33,422 |
| CKPT step=4000 valid_ce=5.2336 valid_ppl=187.5 |
| step=4100 tokens=41902000 ce=5.1815 ppl=178.0 bal=2.538 mut=0.572 tok_s=29,881 avg_tok_s=33,336 |
| step=4200 tokens=42924000 ce=5.3228 ppl=205.0 bal=2.771 mut=0.599 tok_s=33,941 avg_tok_s=33,350 |
| step=4300 tokens=43946000 ce=5.1848 ppl=178.5 bal=4.126 mut=0.623 tok_s=33,678 avg_tok_s=33,358 |
| step=4400 tokens=44968000 ce=5.1099 ppl=165.7 bal=2.398 mut=0.622 tok_s=33,840 avg_tok_s=33,369 |
| step=4500 tokens=45990000 ce=5.3259 ppl=205.6 bal=2.584 mut=0.621 tok_s=33,906 avg_tok_s=33,381 |
| step=4600 tokens=47012000 ce=5.3504 ppl=210.7 bal=2.438 mut=0.611 tok_s=33,912 avg_tok_s=33,392 |
| step=4700 tokens=48034000 ce=4.9649 ppl=143.3 bal=2.337 mut=0.586 tok_s=33,687 avg_tok_s=33,398 |
| step=4800 tokens=49056000 ce=5.1726 ppl=176.4 bal=2.220 mut=0.586 tok_s=33,761 avg_tok_s=33,406 |
| step=4900 tokens=50078000 ce=5.0785 ppl=160.5 bal=2.316 mut=0.595 tok_s=33,918 avg_tok_s=33,416 |
| step=5000 tokens=51100000 ce=5.0890 ppl=162.2 bal=2.871 mut=0.601 tok_s=33,875 avg_tok_s=33,426 |
| step=5100 tokens=52122000 ce=4.9443 ppl=140.4 bal=2.315 mut=0.611 tok_s=33,854 avg_tok_s=33,434 |
| step=5200 tokens=53144000 ce=4.9766 ppl=145.0 bal=2.761 mut=0.609 tok_s=33,826 avg_tok_s=33,441 |
| step=5300 tokens=54166000 ce=5.1149 ppl=166.5 bal=2.294 mut=0.593 tok_s=33,820 avg_tok_s=33,449 |
| step=5400 tokens=55188000 ce=4.9950 ppl=147.7 bal=2.640 mut=0.569 tok_s=33,832 avg_tok_s=33,456 |
| step=5500 tokens=56210000 ce=4.9615 ppl=142.8 bal=2.426 mut=0.560 tok_s=33,485 avg_tok_s=33,456 |
| step=5600 tokens=57232000 ce=4.9050 ppl=135.0 bal=2.344 mut=0.560 tok_s=33,633 avg_tok_s=33,459 |
| step=5700 tokens=58254000 ce=4.9484 ppl=140.9 bal=2.408 mut=0.559 tok_s=33,679 avg_tok_s=33,463 |
| step=5800 tokens=59276000 ce=5.0004 ppl=148.5 bal=2.389 mut=0.565 tok_s=33,676 avg_tok_s=33,467 |
| step=5900 tokens=60298000 ce=4.9394 ppl=139.7 bal=2.374 mut=0.544 tok_s=33,624 avg_tok_s=33,470 |
| step=6000 tokens=61320000 ce=4.9389 ppl=139.6 bal=2.541 mut=0.545 tok_s=33,443 avg_tok_s=33,469 |
| CKPT step=6000 valid_ce=5.0876 valid_ppl=162.0 |
| step=6100 tokens=62342000 ce=4.9210 ppl=137.1 bal=2.385 mut=0.556 tok_s=29,236 avg_tok_s=33,400 |
| step=6200 tokens=63364000 ce=4.7055 ppl=110.6 bal=2.540 mut=0.546 tok_s=33,479 avg_tok_s=33,401 |
| step=6300 tokens=64386000 ce=5.0065 ppl=149.4 bal=2.731 mut=0.549 tok_s=33,486 avg_tok_s=33,402 |
| step=6400 tokens=65408000 ce=4.6567 ppl=105.3 bal=2.588 mut=0.578 tok_s=33,367 avg_tok_s=33,402 |
| step=6500 tokens=66430000 ce=4.8899 ppl=132.9 bal=2.382 mut=0.599 tok_s=33,409 avg_tok_s=33,402 |
| step=6600 tokens=67452000 ce=4.8022 ppl=121.8 bal=3.084 mut=0.600 tok_s=33,358 avg_tok_s=33,401 |
| step=6700 tokens=68474000 ce=4.9022 ppl=134.6 bal=2.546 mut=0.578 tok_s=33,334 avg_tok_s=33,400 |
| step=6800 tokens=69496000 ce=4.8258 ppl=124.7 bal=2.462 mut=0.586 tok_s=33,145 avg_tok_s=33,397 |
| step=6900 tokens=70518000 ce=4.8288 ppl=125.1 bal=2.482 mut=0.606 tok_s=32,974 avg_tok_s=33,390 |
| step=7000 tokens=71540000 ce=4.7357 ppl=113.9 bal=2.266 mut=0.618 tok_s=33,183 avg_tok_s=33,387 |
| step=7100 tokens=72562000 ce=4.9510 ppl=141.3 bal=2.559 mut=0.609 tok_s=33,302 avg_tok_s=33,386 |
| step=7200 tokens=73584000 ce=4.8792 ppl=131.5 bal=2.579 mut=0.603 tok_s=33,328 avg_tok_s=33,385 |
| step=7300 tokens=74606000 ce=4.8494 ppl=127.7 bal=2.413 mut=0.602 tok_s=33,596 avg_tok_s=33,388 |
| step=7400 tokens=75628000 ce=4.8537 ppl=128.2 bal=2.475 mut=0.619 tok_s=33,409 avg_tok_s=33,389 |
| step=7500 tokens=76650000 ce=4.6232 ppl=101.8 bal=2.694 mut=0.613 tok_s=33,649 avg_tok_s=33,392 |
| step=7600 tokens=77672000 ce=4.8696 ppl=130.3 bal=2.435 mut=0.617 tok_s=33,491 avg_tok_s=33,393 |
| step=7700 tokens=78694000 ce=4.6999 ppl=109.9 bal=2.337 mut=0.608 tok_s=33,580 avg_tok_s=33,396 |
| step=7800 tokens=79716000 ce=4.6800 ppl=107.8 bal=2.298 mut=0.611 tok_s=33,307 avg_tok_s=33,395 |
| step=7900 tokens=80738000 ce=5.0483 ppl=155.8 bal=2.868 mut=0.617 tok_s=33,506 avg_tok_s=33,396 |
| step=8000 tokens=81760000 ce=4.6872 ppl=108.5 bal=2.746 mut=0.632 tok_s=33,403 avg_tok_s=33,396 |
| CKPT step=8000 valid_ce=4.8020 valid_ppl=121.8 |
| step=8100 tokens=82782000 ce=4.7238 ppl=112.6 bal=2.205 mut=0.630 tok_s=29,320 avg_tok_s=33,346 |
| step=8200 tokens=83804000 ce=4.6063 ppl=100.1 bal=2.304 mut=0.637 tok_s=33,441 avg_tok_s=33,347 |
| step=8300 tokens=84826000 ce=4.7280 ppl=113.1 bal=2.461 mut=0.637 tok_s=33,194 avg_tok_s=33,345 |
| step=8400 tokens=85848000 ce=4.7639 ppl=117.2 bal=2.576 mut=0.645 tok_s=33,310 avg_tok_s=33,345 |
| step=8500 tokens=86870000 ce=4.7843 ppl=119.6 bal=2.413 mut=0.649 tok_s=33,337 avg_tok_s=33,345 |
| step=8600 tokens=87892000 ce=4.6088 ppl=100.4 bal=2.291 mut=0.642 tok_s=33,366 avg_tok_s=33,345 |
| step=8700 tokens=88914000 ce=4.6687 ppl=106.6 bal=2.177 mut=0.640 tok_s=33,315 avg_tok_s=33,345 |
| step=8800 tokens=89936000 ce=4.8576 ppl=128.7 bal=2.237 mut=0.649 tok_s=33,187 avg_tok_s=33,343 |
| step=8900 tokens=90958000 ce=4.6739 ppl=107.1 bal=2.313 mut=0.652 tok_s=33,287 avg_tok_s=33,342 |
| step=9000 tokens=91980000 ce=4.6667 ppl=106.4 bal=2.148 mut=0.665 tok_s=33,264 avg_tok_s=33,341 |
| step=9100 tokens=93002000 ce=4.6407 ppl=103.6 bal=2.344 mut=0.650 tok_s=33,063 avg_tok_s=33,338 |
| step=9200 tokens=94024000 ce=4.8073 ppl=122.4 bal=2.258 mut=0.658 tok_s=33,111 avg_tok_s=33,336 |
| step=9300 tokens=95046000 ce=4.7432 ppl=114.8 bal=2.300 mut=0.679 tok_s=33,139 avg_tok_s=33,334 |
| step=9400 tokens=96068000 ce=4.7647 ppl=117.3 bal=2.264 mut=0.670 tok_s=33,214 avg_tok_s=33,332 |
| step=9500 tokens=97090000 ce=4.8375 ppl=126.1 bal=2.328 mut=0.676 tok_s=33,178 avg_tok_s=33,331 |
| step=9600 tokens=98112000 ce=4.9331 ppl=138.8 bal=4.250 mut=0.669 tok_s=32,721 avg_tok_s=33,324 |
| step=9700 tokens=99134000 ce=4.5852 ppl=98.0 bal=2.140 mut=0.662 tok_s=33,031 avg_tok_s=33,321 |
| step=9800 tokens=100156000 ce=4.7557 ppl=116.2 bal=2.377 mut=0.663 tok_s=32,996 avg_tok_s=33,318 |
| step=9900 tokens=101178000 ce=4.8145 ppl=123.3 bal=2.296 mut=0.672 tok_s=32,907 avg_tok_s=33,314 |
| step=10000 tokens=102200000 ce=4.7192 ppl=112.1 bal=2.118 mut=0.677 tok_s=32,676 avg_tok_s=33,307 |
| CKPT step=10000 valid_ce=4.6544 valid_ppl=105.0 |
| step=10100 tokens=103222000 ce=4.5968 ppl=99.2 bal=2.286 mut=0.676 tok_s=29,004 avg_tok_s=33,265 |
| step=10200 tokens=104244000 ce=4.6356 ppl=103.1 bal=2.491 mut=0.680 tok_s=32,979 avg_tok_s=33,262 |
| step=10300 tokens=105266000 ce=4.6799 ppl=107.8 bal=2.890 mut=0.672 tok_s=33,007 avg_tok_s=33,260 |
| step=10400 tokens=106288000 ce=4.4171 ppl=82.9 bal=2.238 mut=0.674 tok_s=32,624 avg_tok_s=33,253 |
| step=10500 tokens=107310000 ce=4.8059 ppl=122.2 bal=3.518 mut=0.681 tok_s=32,984 avg_tok_s=33,251 |
| step=10600 tokens=108332000 ce=4.4743 ppl=87.7 bal=2.176 mut=0.668 tok_s=32,969 avg_tok_s=33,248 |
| step=10700 tokens=109354000 ce=4.4247 ppl=83.5 bal=2.134 mut=0.673 tok_s=32,888 avg_tok_s=33,245 |
| step=10800 tokens=110376000 ce=4.4697 ppl=87.3 bal=2.193 mut=0.671 tok_s=32,782 avg_tok_s=33,241 |
| step=10900 tokens=111398000 ce=4.3477 ppl=77.3 bal=2.114 mut=0.662 tok_s=32,821 avg_tok_s=33,237 |
| step=11000 tokens=112420000 ce=4.7481 ppl=115.4 bal=2.271 mut=0.673 tok_s=32,845 avg_tok_s=33,233 |
| step=11100 tokens=113442000 ce=4.7110 ppl=111.2 bal=2.307 mut=0.673 tok_s=32,801 avg_tok_s=33,229 |
| step=11200 tokens=114464000 ce=4.6077 ppl=100.3 bal=2.055 mut=0.672 tok_s=32,416 avg_tok_s=33,222 |
| step=11300 tokens=115486000 ce=4.4341 ppl=84.3 bal=2.088 mut=0.677 tok_s=32,716 avg_tok_s=33,218 |
| step=11400 tokens=116508000 ce=4.7856 ppl=119.8 bal=2.068 mut=0.668 tok_s=32,722 avg_tok_s=33,213 |
| step=11500 tokens=117530000 ce=4.7295 ppl=113.2 bal=2.172 mut=0.664 tok_s=32,652 avg_tok_s=33,208 |
| step=11600 tokens=118552000 ce=4.4490 ppl=85.5 bal=2.079 mut=0.677 tok_s=32,527 avg_tok_s=33,202 |
| step=11700 tokens=119574000 ce=4.4769 ppl=88.0 bal=2.050 mut=0.669 tok_s=32,381 avg_tok_s=33,195 |
| step=11800 tokens=120596000 ce=4.5109 ppl=91.0 bal=2.603 mut=0.666 tok_s=32,488 avg_tok_s=33,189 |
| step=11900 tokens=121618000 ce=4.5709 ppl=96.6 bal=2.392 mut=0.666 tok_s=32,434 avg_tok_s=33,183 |
| step=12000 tokens=122640000 ce=4.4658 ppl=87.0 bal=2.150 mut=0.669 tok_s=32,253 avg_tok_s=33,175 |
| CKPT step=12000 valid_ce=4.4680 valid_ppl=87.2 |
| step=12100 tokens=123662000 ce=4.4535 ppl=85.9 bal=2.309 mut=0.668 tok_s=28,311 avg_tok_s=33,135 |
| step=12200 tokens=124684000 ce=4.4123 ppl=82.5 bal=1.912 mut=0.663 tok_s=32,134 avg_tok_s=33,127 |
| step=12300 tokens=125706000 ce=4.4634 ppl=86.8 bal=2.090 mut=0.660 tok_s=32,054 avg_tok_s=33,118 |
| step=12400 tokens=126728000 ce=4.4676 ppl=87.1 bal=2.023 mut=0.661 tok_s=31,944 avg_tok_s=33,109 |
| step=12500 tokens=127750000 ce=4.6261 ppl=102.1 bal=2.281 mut=0.665 tok_s=31,756 avg_tok_s=33,098 |
| step=12600 tokens=128772000 ce=4.5410 ppl=93.8 bal=2.302 mut=0.663 tok_s=31,966 avg_tok_s=33,089 |
| step=12700 tokens=129794000 ce=4.5818 ppl=97.7 bal=2.251 mut=0.666 tok_s=31,985 avg_tok_s=33,080 |
| step=12800 tokens=130816000 ce=4.6595 ppl=105.6 bal=2.157 mut=0.664 tok_s=31,884 avg_tok_s=33,071 |
| step=12900 tokens=131838000 ce=4.4738 ppl=87.7 bal=2.020 mut=0.659 tok_s=32,007 avg_tok_s=33,063 |
| step=13000 tokens=132860000 ce=4.7628 ppl=117.1 bal=2.171 mut=0.656 tok_s=31,992 avg_tok_s=33,054 |
| step=13100 tokens=133882000 ce=4.5381 ppl=93.5 bal=2.138 mut=0.660 tok_s=32,046 avg_tok_s=33,047 |
| step=13200 tokens=134904000 ce=4.6693 ppl=106.6 bal=2.170 mut=0.659 tok_s=31,906 avg_tok_s=33,038 |
| step=13300 tokens=135926000 ce=4.5302 ppl=92.8 bal=2.011 mut=0.658 tok_s=31,791 avg_tok_s=33,029 |
| step=13400 tokens=136948000 ce=4.4537 ppl=85.9 bal=2.296 mut=0.655 tok_s=32,057 avg_tok_s=33,021 |
| step=13500 tokens=137970000 ce=4.3525 ppl=77.7 bal=2.007 mut=0.653 tok_s=32,074 avg_tok_s=33,014 |
| step=13600 tokens=138992000 ce=4.3803 ppl=79.9 bal=2.048 mut=0.650 tok_s=31,972 avg_tok_s=33,007 |
| step=13700 tokens=140014000 ce=4.6504 ppl=104.6 bal=2.173 mut=0.659 tok_s=32,075 avg_tok_s=33,000 |
| step=13800 tokens=141036000 ce=4.5861 ppl=98.1 bal=2.211 mut=0.658 tok_s=31,865 avg_tok_s=32,992 |
| step=13900 tokens=142058000 ce=4.7403 ppl=114.5 bal=2.076 mut=0.658 tok_s=32,088 avg_tok_s=32,985 |
| step=14000 tokens=143080000 ce=4.4038 ppl=81.8 bal=2.012 mut=0.659 tok_s=31,973 avg_tok_s=32,978 |
| CKPT step=14000 valid_ce=4.4876 valid_ppl=88.9 |
| step=14100 tokens=144102000 ce=4.4263 ppl=83.6 bal=2.351 mut=0.655 tok_s=28,384 avg_tok_s=32,945 |
| step=14200 tokens=145124000 ce=4.7503 ppl=115.6 bal=2.135 mut=0.660 tok_s=31,812 avg_tok_s=32,937 |
| step=14300 tokens=146146000 ce=4.7020 ppl=110.2 bal=2.185 mut=0.661 tok_s=31,985 avg_tok_s=32,931 |
| step=14400 tokens=147168000 ce=4.3480 ppl=77.3 bal=2.116 mut=0.661 tok_s=31,876 avg_tok_s=32,923 |
| step=14500 tokens=148190000 ce=4.8219 ppl=124.2 bal=2.253 mut=0.655 tok_s=31,982 avg_tok_s=32,917 |
| step=14600 tokens=149212000 ce=4.4184 ppl=83.0 bal=2.553 mut=0.662 tok_s=31,708 avg_tok_s=32,909 |
| step=14700 tokens=150234000 ce=4.5687 ppl=96.4 bal=2.218 mut=0.666 tok_s=31,958 avg_tok_s=32,902 |
| step=14800 tokens=151256000 ce=4.3264 ppl=75.7 bal=2.896 mut=0.667 tok_s=31,872 avg_tok_s=32,895 |
| step=14900 tokens=152278000 ce=4.4557 ppl=86.1 bal=2.100 mut=0.670 tok_s=32,001 avg_tok_s=32,889 |
| step=15000 tokens=153300000 ce=4.3610 ppl=78.3 bal=2.078 mut=0.677 tok_s=31,980 avg_tok_s=32,883 |
| step=15100 tokens=154322000 ce=4.4581 ppl=86.3 bal=2.096 mut=0.680 tok_s=31,987 avg_tok_s=32,877 |
| step=15200 tokens=155344000 ce=4.5006 ppl=90.1 bal=2.021 mut=0.682 tok_s=31,904 avg_tok_s=32,871 |
| step=15300 tokens=156366000 ce=4.5151 ppl=91.4 bal=2.206 mut=0.685 tok_s=31,988 avg_tok_s=32,865 |
| step=15400 tokens=157388000 ce=4.2949 ppl=73.3 bal=2.001 mut=0.683 tok_s=31,906 avg_tok_s=32,859 |
| step=15500 tokens=158410000 ce=4.6290 ppl=102.4 bal=2.250 mut=0.690 tok_s=31,712 avg_tok_s=32,851 |
| step=15600 tokens=159432000 ce=4.3108 ppl=74.5 bal=2.312 mut=0.689 tok_s=31,845 avg_tok_s=32,845 |
| step=15700 tokens=160454000 ce=4.4801 ppl=88.2 bal=2.015 mut=0.694 tok_s=31,947 avg_tok_s=32,839 |
| step=15800 tokens=161476000 ce=4.6816 ppl=107.9 bal=2.419 mut=0.695 tok_s=31,938 avg_tok_s=32,834 |
| step=15900 tokens=162498000 ce=4.4997 ppl=90.0 bal=2.004 mut=0.693 tok_s=31,911 avg_tok_s=32,828 |
| step=16000 tokens=163520000 ce=4.4825 ppl=88.5 bal=2.015 mut=0.694 tok_s=31,667 avg_tok_s=32,821 |
| CKPT step=16000 valid_ce=4.4235 valid_ppl=83.4 |
| step=16100 tokens=164542000 ce=4.3421 ppl=76.9 bal=2.104 mut=0.705 tok_s=28,318 avg_tok_s=32,793 |
| step=16200 tokens=165564000 ce=4.5445 ppl=94.1 bal=1.995 mut=0.697 tok_s=31,854 avg_tok_s=32,787 |
| step=16300 tokens=166586000 ce=4.5099 ppl=90.9 bal=2.140 mut=0.705 tok_s=31,846 avg_tok_s=32,781 |
| step=16400 tokens=167608000 ce=4.5148 ppl=91.4 bal=2.044 mut=0.701 tok_s=31,561 avg_tok_s=32,774 |
| step=16500 tokens=168630000 ce=4.4919 ppl=89.3 bal=2.233 mut=0.705 tok_s=31,851 avg_tok_s=32,768 |
| step=16600 tokens=169652000 ce=4.4794 ppl=88.2 bal=2.085 mut=0.710 tok_s=31,837 avg_tok_s=32,762 |
| step=16700 tokens=170674000 ce=4.5953 ppl=99.0 bal=2.097 mut=0.712 tok_s=31,850 avg_tok_s=32,757 |
| step=16800 tokens=171696000 ce=4.4448 ppl=85.2 bal=2.028 mut=0.711 tok_s=31,514 avg_tok_s=32,749 |
| step=16900 tokens=172718000 ce=4.3010 ppl=73.8 bal=2.224 mut=0.715 tok_s=31,801 avg_tok_s=32,744 |
| step=17000 tokens=173740000 ce=4.5496 ppl=94.6 bal=2.031 mut=0.727 tok_s=31,813 avg_tok_s=32,738 |
| step=17100 tokens=174762000 ce=4.5795 ppl=97.5 bal=2.025 mut=0.715 tok_s=31,819 avg_tok_s=32,733 |
| step=17200 tokens=175784000 ce=4.4149 ppl=82.7 bal=2.156 mut=0.722 tok_s=31,694 avg_tok_s=32,727 |
| step=17300 tokens=176806000 ce=4.4308 ppl=84.0 bal=2.088 mut=0.722 tok_s=31,807 avg_tok_s=32,722 |
| step=17400 tokens=177828000 ce=4.2871 ppl=72.8 bal=2.178 mut=0.721 tok_s=31,790 avg_tok_s=32,716 |
| step=17500 tokens=178850000 ce=4.2842 ppl=72.5 bal=2.119 mut=0.729 tok_s=31,804 avg_tok_s=32,711 |
| step=17600 tokens=179872000 ce=4.3283 ppl=75.8 bal=2.005 mut=0.733 tok_s=31,718 avg_tok_s=32,705 |
| step=17700 tokens=180894000 ce=4.3239 ppl=75.5 bal=2.071 mut=0.735 tok_s=31,616 avg_tok_s=32,699 |
| step=17800 tokens=181916000 ce=4.6011 ppl=99.6 bal=2.105 mut=0.724 tok_s=31,833 avg_tok_s=32,694 |
| step=17900 tokens=182938000 ce=4.4394 ppl=84.7 bal=2.156 mut=0.737 tok_s=31,814 avg_tok_s=32,690 |
| step=18000 tokens=183960000 ce=4.4358 ppl=84.4 bal=2.002 mut=0.738 tok_s=31,752 avg_tok_s=32,684 |
| CKPT step=18000 valid_ce=4.4930 valid_ppl=89.4 |
| step=18100 tokens=184982000 ce=4.4091 ppl=82.2 bal=2.204 mut=0.735 tok_s=28,234 avg_tok_s=32,660 |
| step=18200 tokens=186004000 ce=4.2531 ppl=70.3 bal=2.184 mut=0.731 tok_s=31,611 avg_tok_s=32,654 |
| step=18300 tokens=187026000 ce=4.5514 ppl=94.8 bal=2.114 mut=0.743 tok_s=31,803 avg_tok_s=32,649 |
| step=18400 tokens=188048000 ce=4.2024 ppl=66.8 bal=3.060 mut=0.741 tok_s=31,721 avg_tok_s=32,644 |
| step=18500 tokens=189070000 ce=4.3371 ppl=76.5 bal=2.024 mut=0.747 tok_s=31,783 avg_tok_s=32,640 |
| step=18600 tokens=190092000 ce=4.3770 ppl=79.6 bal=1.999 mut=0.739 tok_s=31,752 avg_tok_s=32,635 |
| step=18700 tokens=191114000 ce=4.2993 ppl=73.6 bal=2.072 mut=0.736 tok_s=31,575 avg_tok_s=32,629 |
| step=18800 tokens=192136000 ce=4.2962 ppl=73.4 bal=2.063 mut=0.734 tok_s=31,695 avg_tok_s=32,624 |
| step=18900 tokens=193158000 ce=4.4176 ppl=82.9 bal=2.190 mut=0.731 tok_s=31,773 avg_tok_s=32,620 |
| step=19000 tokens=194180000 ce=4.3451 ppl=77.1 bal=2.234 mut=0.746 tok_s=31,750 avg_tok_s=32,615 |
| step=19100 tokens=195202000 ce=4.5043 ppl=90.4 bal=1.972 mut=0.744 tok_s=31,519 avg_tok_s=32,609 |
| step=19200 tokens=196224000 ce=4.5551 ppl=95.1 bal=2.441 mut=0.740 tok_s=31,661 avg_tok_s=32,604 |
| step=19300 tokens=197246000 ce=4.4538 ppl=86.0 bal=2.187 mut=0.754 tok_s=31,716 avg_tok_s=32,600 |
| step=19400 tokens=198268000 ce=4.5115 ppl=91.1 bal=2.250 mut=0.734 tok_s=31,720 avg_tok_s=32,595 |
| step=19500 tokens=199290000 ce=4.1869 ppl=65.8 bal=1.993 mut=0.746 tok_s=31,716 avg_tok_s=32,591 |
| step=19600 tokens=200312000 ce=4.4149 ppl=82.7 bal=2.149 mut=0.746 tok_s=31,612 avg_tok_s=32,586 |
| step=19700 tokens=201334000 ce=4.5226 ppl=92.1 bal=2.143 mut=0.742 tok_s=31,696 avg_tok_s=32,581 |
| step=19800 tokens=202356000 ce=4.4682 ppl=87.2 bal=1.968 mut=0.749 tok_s=31,719 avg_tok_s=32,577 |
| step=19900 tokens=203378000 ce=4.4340 ppl=84.3 bal=2.007 mut=0.753 tok_s=31,474 avg_tok_s=32,571 |
| step=20000 tokens=204400000 ce=4.3799 ppl=79.8 bal=2.662 mut=0.744 tok_s=31,509 avg_tok_s=32,566 |
| CKPT step=20000 valid_ce=4.2913 valid_ppl=73.1 |
| step=20100 tokens=205422000 ce=4.2167 ppl=67.8 bal=2.001 mut=0.743 tok_s=11,834 avg_tok_s=32,463 |
| step=20200 tokens=206444000 ce=4.5649 ppl=96.1 bal=2.233 mut=0.749 tok_s=31,569 avg_tok_s=32,459 |
| step=20300 tokens=207466000 ce=4.3489 ppl=77.4 bal=2.109 mut=0.744 tok_s=31,547 avg_tok_s=32,454 |
| step=20400 tokens=208488000 ce=4.4098 ppl=82.3 bal=2.244 mut=0.747 tok_s=31,293 avg_tok_s=32,448 |
| step=20500 tokens=209510000 ce=4.2224 ppl=68.2 bal=2.253 mut=0.755 tok_s=31,581 avg_tok_s=32,444 |
| step=20600 tokens=210532000 ce=4.3954 ppl=81.1 bal=2.181 mut=0.761 tok_s=31,518 avg_tok_s=32,440 |
| step=20700 tokens=211554000 ce=4.5598 ppl=95.6 bal=2.403 mut=0.749 tok_s=31,509 avg_tok_s=32,435 |
| step=20800 tokens=212576000 ce=4.3643 ppl=78.6 bal=2.214 mut=0.750 tok_s=31,174 avg_tok_s=32,429 |
| step=20900 tokens=213598000 ce=4.5550 ppl=95.1 bal=2.350 mut=0.745 tok_s=31,515 avg_tok_s=32,425 |
| step=21000 tokens=214620000 ce=4.4218 ppl=83.2 bal=3.110 mut=0.739 tok_s=31,475 avg_tok_s=32,420 |
| step=21100 tokens=215642000 ce=4.3819 ppl=80.0 bal=2.766 mut=0.751 tok_s=31,450 avg_tok_s=32,416 |
| step=21200 tokens=216664000 ce=4.2682 ppl=71.4 bal=2.140 mut=0.748 tok_s=31,322 avg_tok_s=32,410 |
| step=21300 tokens=217686000 ce=4.3719 ppl=79.2 bal=2.283 mut=0.747 tok_s=31,432 avg_tok_s=32,406 |
| step=21400 tokens=218708000 ce=4.3376 ppl=76.5 bal=3.349 mut=0.742 tok_s=31,454 avg_tok_s=32,401 |
| step=21500 tokens=219730000 ce=4.4729 ppl=87.6 bal=2.083 mut=0.751 tok_s=31,469 avg_tok_s=32,397 |
| step=21600 tokens=220752000 ce=4.5233 ppl=92.1 bal=2.270 mut=0.747 tok_s=31,131 avg_tok_s=32,391 |
| step=21700 tokens=221774000 ce=4.1604 ppl=64.1 bal=2.119 mut=0.737 tok_s=31,398 avg_tok_s=32,387 |
| step=21800 tokens=222796000 ce=4.1957 ppl=66.4 bal=2.193 mut=0.747 tok_s=31,321 avg_tok_s=32,382 |
| step=21900 tokens=223818000 ce=4.3716 ppl=79.2 bal=2.017 mut=0.749 tok_s=31,310 avg_tok_s=32,377 |
| step=22000 tokens=224840000 ce=4.4996 ppl=90.0 bal=2.256 mut=0.751 tok_s=31,202 avg_tok_s=32,371 |
| CKPT step=22000 valid_ce=4.4185 valid_ppl=83.0 |
| step=22100 tokens=225862000 ce=4.2767 ppl=72.0 bal=2.301 mut=0.748 tok_s=27,813 avg_tok_s=32,351 |
| step=22200 tokens=226884000 ce=4.2952 ppl=73.3 bal=2.063 mut=0.752 tok_s=31,060 avg_tok_s=32,345 |
| step=22300 tokens=227906000 ce=4.1093 ppl=60.9 bal=2.298 mut=0.753 tok_s=31,304 avg_tok_s=32,340 |
| step=22400 tokens=228928000 ce=4.2816 ppl=72.4 bal=2.097 mut=0.749 tok_s=31,216 avg_tok_s=32,335 |
| step=22500 tokens=229950000 ce=4.2087 ppl=67.3 bal=2.653 mut=0.741 tok_s=31,335 avg_tok_s=32,331 |
| step=22600 tokens=230972000 ce=4.4534 ppl=85.9 bal=2.079 mut=0.739 tok_s=31,314 avg_tok_s=32,326 |
| step=22700 tokens=231994000 ce=4.5537 ppl=95.0 bal=2.818 mut=0.744 tok_s=31,187 avg_tok_s=32,321 |
| step=22800 tokens=233016000 ce=4.3431 ppl=76.9 bal=2.320 mut=0.754 tok_s=31,327 avg_tok_s=32,317 |
| step=22900 tokens=234038000 ce=4.4281 ppl=83.8 bal=2.067 mut=0.747 tok_s=31,431 avg_tok_s=32,313 |
| step=23000 tokens=235060000 ce=4.4255 ppl=83.6 bal=2.117 mut=0.754 tok_s=31,478 avg_tok_s=32,309 |
| step=23100 tokens=236082000 ce=4.2684 ppl=71.4 bal=2.124 mut=0.753 tok_s=31,257 avg_tok_s=32,305 |
| step=23200 tokens=237104000 ce=4.3995 ppl=81.4 bal=2.213 mut=0.749 tok_s=31,496 avg_tok_s=32,301 |
| step=23300 tokens=238126000 ce=4.3961 ppl=81.1 bal=2.217 mut=0.756 tok_s=31,564 avg_tok_s=32,298 |
| step=23400 tokens=239148000 ce=4.4501 ppl=85.6 bal=2.116 mut=0.756 tok_s=31,554 avg_tok_s=32,295 |
| step=23500 tokens=240170000 ce=4.2317 ppl=68.8 bal=2.014 mut=0.759 tok_s=31,598 avg_tok_s=32,292 |
| step=23600 tokens=241192000 ce=4.3525 ppl=77.7 bal=2.079 mut=0.755 tok_s=31,499 avg_tok_s=32,289 |
| step=23700 tokens=242214000 ce=4.3705 ppl=79.1 bal=2.029 mut=0.762 tok_s=31,751 avg_tok_s=32,287 |
| step=23800 tokens=243236000 ce=4.3727 ppl=79.3 bal=2.077 mut=0.752 tok_s=31,829 avg_tok_s=32,285 |
| step=23900 tokens=244258000 ce=4.3153 ppl=74.8 bal=2.204 mut=0.755 tok_s=31,941 avg_tok_s=32,283 |
| step=24000 tokens=245280000 ce=4.4217 ppl=83.2 bal=2.174 mut=0.762 tok_s=31,772 avg_tok_s=32,281 |
| CKPT step=24000 valid_ce=4.3858 valid_ppl=80.3 |
| step=24100 tokens=246302000 ce=4.4593 ppl=86.4 bal=2.057 mut=0.762 tok_s=28,510 avg_tok_s=32,265 |
| step=24200 tokens=247324000 ce=4.2580 ppl=70.7 bal=1.997 mut=0.759 tok_s=32,119 avg_tok_s=32,265 |
| step=24300 tokens=248346000 ce=4.2509 ppl=70.2 bal=2.131 mut=0.756 tok_s=32,235 avg_tok_s=32,265 |
| step=24400 tokens=249368000 ce=4.2324 ppl=68.9 bal=2.030 mut=0.758 tok_s=32,311 avg_tok_s=32,265 |
| step=24500 tokens=250390000 ce=4.2216 ppl=68.1 bal=2.163 mut=0.771 tok_s=32,342 avg_tok_s=32,265 |
| step=24600 tokens=251412000 ce=4.2904 ppl=73.0 bal=2.130 mut=0.775 tok_s=32,490 avg_tok_s=32,266 |
| step=24700 tokens=252434000 ce=3.9661 ppl=52.8 bal=2.200 mut=0.771 tok_s=32,656 avg_tok_s=32,268 |
| step=24800 tokens=253456000 ce=4.0501 ppl=57.4 bal=2.289 mut=0.769 tok_s=32,476 avg_tok_s=32,268 |
| step=24900 tokens=254478000 ce=4.2526 ppl=70.3 bal=2.041 mut=0.773 tok_s=32,373 avg_tok_s=32,269 |
| step=25000 tokens=255500000 ce=4.2544 ppl=70.4 bal=2.028 mut=0.765 tok_s=32,815 avg_tok_s=32,271 |
| step=25100 tokens=256522000 ce=4.2907 ppl=73.0 bal=2.102 mut=0.775 tok_s=32,656 avg_tok_s=32,273 |
| step=25200 tokens=257544000 ce=4.0903 ppl=59.8 bal=2.046 mut=0.777 tok_s=32,572 avg_tok_s=32,274 |
| step=25300 tokens=258566000 ce=4.4228 ppl=83.3 bal=2.243 mut=0.777 tok_s=32,622 avg_tok_s=32,275 |
| step=25400 tokens=259588000 ce=4.1977 ppl=66.5 bal=1.961 mut=0.769 tok_s=32,786 avg_tok_s=32,277 |
| step=25500 tokens=260610000 ce=3.9554 ppl=52.2 bal=2.148 mut=0.781 tok_s=32,804 avg_tok_s=32,279 |
| step=25600 tokens=261632000 ce=4.2434 ppl=69.6 bal=2.427 mut=0.772 tok_s=32,708 avg_tok_s=32,281 |
| step=25700 tokens=262654000 ce=4.2246 ppl=68.3 bal=2.049 mut=0.779 tok_s=32,531 avg_tok_s=32,282 |
| step=25800 tokens=263676000 ce=4.4723 ppl=87.6 bal=2.157 mut=0.791 tok_s=32,663 avg_tok_s=32,283 |
| step=25900 tokens=264698000 ce=4.3818 ppl=80.0 bal=2.151 mut=0.792 tok_s=32,730 avg_tok_s=32,285 |
| step=26000 tokens=265720000 ce=4.3965 ppl=81.2 bal=2.083 mut=0.773 tok_s=32,748 avg_tok_s=32,287 |
| CKPT step=26000 valid_ce=4.3764 valid_ppl=79.6 |
| step=26100 tokens=266742000 ce=4.1645 ppl=64.4 bal=1.959 mut=0.787 tok_s=28,843 avg_tok_s=32,274 |
| step=26200 tokens=267764000 ce=4.2944 ppl=73.3 bal=2.241 mut=0.786 tok_s=32,562 avg_tok_s=32,275 |
| step=26300 tokens=268786000 ce=4.3404 ppl=76.7 bal=2.115 mut=0.794 tok_s=32,280 avg_tok_s=32,275 |
| step=26400 tokens=269808000 ce=4.4159 ppl=82.8 bal=2.241 mut=0.786 tok_s=32,574 avg_tok_s=32,276 |
| step=26500 tokens=270830000 ce=4.4492 ppl=85.6 bal=2.119 mut=0.775 tok_s=32,688 avg_tok_s=32,278 |
| step=26600 tokens=271852000 ce=4.1895 ppl=66.0 bal=2.067 mut=0.782 tok_s=32,630 avg_tok_s=32,279 |
| step=26700 tokens=272874000 ce=4.3338 ppl=76.2 bal=2.142 mut=0.777 tok_s=32,809 avg_tok_s=32,281 |
| step=26800 tokens=273896000 ce=4.3987 ppl=81.3 bal=2.364 mut=0.786 tok_s=32,601 avg_tok_s=32,282 |
| step=26900 tokens=274918000 ce=4.2250 ppl=68.4 bal=2.188 mut=0.791 tok_s=32,716 avg_tok_s=32,284 |
| step=27000 tokens=275940000 ce=4.1923 ppl=66.2 bal=2.228 mut=0.786 tok_s=32,537 avg_tok_s=32,285 |
| step=27100 tokens=276962000 ce=4.2669 ppl=71.3 bal=2.077 mut=0.794 tok_s=32,194 avg_tok_s=32,284 |
| step=27200 tokens=277984000 ce=4.4345 ppl=84.3 bal=2.028 mut=0.799 tok_s=32,596 avg_tok_s=32,285 |
| step=27300 tokens=279006000 ce=4.3452 ppl=77.1 bal=2.044 mut=0.789 tok_s=32,551 avg_tok_s=32,286 |
| step=27400 tokens=280028000 ce=3.9898 ppl=54.0 bal=2.004 mut=0.791 tok_s=32,681 avg_tok_s=32,288 |
| step=27500 tokens=281050000 ce=4.2549 ppl=70.4 bal=2.060 mut=0.781 tok_s=32,639 avg_tok_s=32,289 |
| step=27600 tokens=282072000 ce=4.3141 ppl=74.7 bal=2.330 mut=0.792 tok_s=32,400 avg_tok_s=32,289 |
| step=27700 tokens=283094000 ce=4.3410 ppl=76.8 bal=2.205 mut=0.787 tok_s=32,357 avg_tok_s=32,290 |
| step=27800 tokens=284116000 ce=4.2482 ppl=70.0 bal=2.159 mut=0.796 tok_s=32,413 avg_tok_s=32,290 |
| step=27900 tokens=285138000 ce=4.1395 ppl=62.8 bal=2.095 mut=0.793 tok_s=32,588 avg_tok_s=32,291 |
| step=28000 tokens=286160000 ce=4.4905 ppl=89.2 bal=2.617 mut=0.795 tok_s=32,145 avg_tok_s=32,291 |
| CKPT step=28000 valid_ce=4.2317 valid_ppl=68.8 |
| step=28100 tokens=287182000 ce=4.2260 ppl=68.4 bal=2.206 mut=0.797 tok_s=28,894 avg_tok_s=32,279 |
| step=28200 tokens=288204000 ce=4.5399 ppl=93.7 bal=2.341 mut=0.796 tok_s=32,345 avg_tok_s=32,279 |
| step=28300 tokens=289226000 ce=4.0609 ppl=58.0 bal=1.985 mut=0.793 tok_s=32,646 avg_tok_s=32,280 |
| step=28400 tokens=290248000 ce=4.2586 ppl=70.7 bal=2.100 mut=0.784 tok_s=32,114 avg_tok_s=32,280 |
| step=28500 tokens=291270000 ce=4.2512 ppl=70.2 bal=1.970 mut=0.808 tok_s=32,487 avg_tok_s=32,280 |
| step=28600 tokens=292292000 ce=4.4139 ppl=82.6 bal=1.991 mut=0.794 tok_s=32,474 avg_tok_s=32,281 |
| step=28700 tokens=293314000 ce=4.3761 ppl=79.5 bal=2.184 mut=0.786 tok_s=32,550 avg_tok_s=32,282 |
| step=28800 tokens=294336000 ce=4.3136 ppl=74.7 bal=2.133 mut=0.787 tok_s=32,487 avg_tok_s=32,283 |
| step=28900 tokens=295358000 ce=4.2251 ppl=68.4 bal=2.065 mut=0.783 tok_s=32,623 avg_tok_s=32,284 |
| step=29000 tokens=296380000 ce=4.1031 ppl=60.5 bal=1.980 mut=0.802 tok_s=32,905 avg_tok_s=32,286 |
| step=29100 tokens=297402000 ce=4.1722 ppl=64.9 bal=1.963 mut=0.791 tok_s=32,671 avg_tok_s=32,287 |
| step=29200 tokens=298424000 ce=4.4467 ppl=85.3 bal=2.957 mut=0.801 tok_s=32,720 avg_tok_s=32,289 |
| step=29300 tokens=299446000 ce=4.1323 ppl=62.3 bal=2.069 mut=0.803 tok_s=32,307 avg_tok_s=32,289 |
| step=29400 tokens=300468000 ce=4.2694 ppl=71.5 bal=2.663 mut=0.795 tok_s=32,640 avg_tok_s=32,290 |
| step=29500 tokens=301490000 ce=4.1237 ppl=61.8 bal=2.237 mut=0.794 tok_s=32,357 avg_tok_s=32,290 |
| step=29600 tokens=302512000 ce=4.2699 ppl=71.5 bal=2.108 mut=0.801 tok_s=32,494 avg_tok_s=32,291 |
| step=29700 tokens=303534000 ce=4.3778 ppl=79.7 bal=2.220 mut=0.801 tok_s=32,677 avg_tok_s=32,292 |
| step=29800 tokens=304556000 ce=4.1894 ppl=66.0 bal=2.103 mut=0.801 tok_s=32,517 avg_tok_s=32,293 |
| step=29900 tokens=305578000 ce=4.0385 ppl=56.7 bal=2.062 mut=0.794 tok_s=32,736 avg_tok_s=32,294 |
| step=30000 tokens=306600000 ce=4.2229 ppl=68.2 bal=2.124 mut=0.789 tok_s=32,455 avg_tok_s=32,295 |
| CKPT step=30000 valid_ce=4.3050 valid_ppl=74.1 |
| step=30100 tokens=307622000 ce=4.2328 ppl=68.9 bal=1.990 mut=0.796 tok_s=28,934 avg_tok_s=32,284 |
| step=30200 tokens=308644000 ce=4.1355 ppl=62.5 bal=2.130 mut=0.787 tok_s=32,344 avg_tok_s=32,284 |
| step=30300 tokens=309666000 ce=4.1836 ppl=65.6 bal=2.291 mut=0.792 tok_s=32,473 avg_tok_s=32,285 |
| step=30400 tokens=310688000 ce=4.5620 ppl=95.8 bal=2.196 mut=0.801 tok_s=32,339 avg_tok_s=32,285 |
| step=30500 tokens=311710000 ce=4.0550 ppl=57.7 bal=1.982 mut=0.786 tok_s=32,597 avg_tok_s=32,286 |
| step=30600 tokens=312732000 ce=4.3065 ppl=74.2 bal=2.124 mut=0.811 tok_s=32,309 avg_tok_s=32,286 |
| step=30700 tokens=313754000 ce=4.3144 ppl=74.8 bal=1.916 mut=0.804 tok_s=32,601 avg_tok_s=32,287 |
| step=30800 tokens=314776000 ce=4.2014 ppl=66.8 bal=2.105 mut=0.791 tok_s=32,208 avg_tok_s=32,287 |
| step=30900 tokens=315798000 ce=4.2846 ppl=72.6 bal=2.150 mut=0.799 tok_s=32,585 avg_tok_s=32,288 |
| step=31000 tokens=316820000 ce=4.1812 ppl=65.4 bal=2.253 mut=0.790 tok_s=32,616 avg_tok_s=32,289 |
| step=31100 tokens=317842000 ce=4.1787 ppl=65.3 bal=2.056 mut=0.795 tok_s=32,631 avg_tok_s=32,290 |
| step=31200 tokens=318864000 ce=4.2489 ppl=70.0 bal=2.137 mut=0.793 tok_s=32,532 avg_tok_s=32,291 |
| step=31300 tokens=319886000 ce=4.0516 ppl=57.5 bal=3.024 mut=0.787 tok_s=32,557 avg_tok_s=32,291 |
| step=31400 tokens=320908000 ce=4.0925 ppl=59.9 bal=2.155 mut=0.777 tok_s=32,165 avg_tok_s=32,291 |
| step=31500 tokens=321930000 ce=4.1084 ppl=60.8 bal=2.000 mut=0.773 tok_s=32,449 avg_tok_s=32,292 |
| step=31600 tokens=322952000 ce=4.0928 ppl=59.9 bal=2.142 mut=0.790 tok_s=32,487 avg_tok_s=32,292 |
| step=31700 tokens=323974000 ce=4.3226 ppl=75.4 bal=1.934 mut=0.799 tok_s=32,431 avg_tok_s=32,293 |
| step=31800 tokens=324996000 ce=4.4395 ppl=84.7 bal=2.160 mut=0.806 tok_s=32,519 avg_tok_s=32,293 |
| step=31900 tokens=326018000 ce=4.2026 ppl=66.9 bal=2.220 mut=0.792 tok_s=32,403 avg_tok_s=32,294 |
| step=32000 tokens=327040000 ce=4.2731 ppl=71.7 bal=2.098 mut=0.786 tok_s=32,068 avg_tok_s=32,293 |
| CKPT step=32000 valid_ce=4.2093 valid_ppl=67.3 |
| step=32100 tokens=328062000 ce=4.3280 ppl=75.8 bal=2.170 mut=0.794 tok_s=28,591 avg_tok_s=32,281 |
| step=32200 tokens=329084000 ce=4.2465 ppl=69.9 bal=2.000 mut=0.803 tok_s=31,853 avg_tok_s=32,280 |
| step=32300 tokens=330106000 ce=4.2661 ppl=71.2 bal=2.363 mut=0.799 tok_s=31,274 avg_tok_s=32,277 |
| step=32400 tokens=331128000 ce=4.2261 ppl=68.4 bal=2.079 mut=0.801 tok_s=31,640 avg_tok_s=32,275 |
| step=32500 tokens=332150000 ce=4.0447 ppl=57.1 bal=2.215 mut=0.799 tok_s=31,891 avg_tok_s=32,274 |
| step=32600 tokens=333172000 ce=4.2131 ppl=67.6 bal=2.121 mut=0.807 tok_s=32,363 avg_tok_s=32,274 |
| step=32700 tokens=334194000 ce=4.3000 ppl=73.7 bal=2.364 mut=0.794 tok_s=31,997 avg_tok_s=32,273 |
| step=32800 tokens=335216000 ce=4.2453 ppl=69.8 bal=2.138 mut=0.803 tok_s=31,916 avg_tok_s=32,272 |
| step=32900 tokens=336238000 ce=4.2599 ppl=70.8 bal=2.183 mut=0.795 tok_s=32,292 avg_tok_s=32,272 |
| step=33000 tokens=337260000 ce=4.3797 ppl=79.8 bal=2.146 mut=0.808 tok_s=32,613 avg_tok_s=32,273 |
| step=33100 tokens=338282000 ce=4.2471 ppl=69.9 bal=2.057 mut=0.790 tok_s=32,407 avg_tok_s=32,274 |
| step=33200 tokens=339304000 ce=4.2399 ppl=69.4 bal=2.173 mut=0.795 tok_s=31,980 avg_tok_s=32,273 |
| step=33300 tokens=340326000 ce=4.1883 ppl=65.9 bal=2.084 mut=0.802 tok_s=32,220 avg_tok_s=32,273 |
| step=33400 tokens=341348000 ce=4.2188 ppl=67.9 bal=1.983 mut=0.812 tok_s=32,045 avg_tok_s=32,272 |
| step=33500 tokens=342370000 ce=4.1119 ppl=61.1 bal=2.067 mut=0.800 tok_s=32,044 avg_tok_s=32,271 |
| step=33600 tokens=343392000 ce=4.0807 ppl=59.2 bal=2.034 mut=0.793 tok_s=32,071 avg_tok_s=32,271 |
| step=33700 tokens=344414000 ce=3.9312 ppl=51.0 bal=2.148 mut=0.789 tok_s=31,784 avg_tok_s=32,269 |
| step=33800 tokens=345436000 ce=4.5588 ppl=95.5 bal=2.191 mut=0.802 tok_s=32,067 avg_tok_s=32,269 |
| step=33900 tokens=346458000 ce=4.3606 ppl=78.3 bal=2.096 mut=0.803 tok_s=32,465 avg_tok_s=32,269 |
| step=34000 tokens=347480000 ce=4.2469 ppl=69.9 bal=2.549 mut=0.797 tok_s=32,158 avg_tok_s=32,269 |
| CKPT step=34000 valid_ce=4.2781 valid_ppl=72.1 |
| step=34100 tokens=348502000 ce=4.4461 ppl=85.3 bal=2.214 mut=0.793 tok_s=28,135 avg_tok_s=32,257 |
| step=34200 tokens=349524000 ce=4.2518 ppl=70.2 bal=2.117 mut=0.789 tok_s=32,221 avg_tok_s=32,257 |
| step=34300 tokens=350546000 ce=4.3031 ppl=73.9 bal=2.293 mut=0.800 tok_s=32,219 avg_tok_s=32,257 |
| step=34400 tokens=351568000 ce=4.2985 ppl=73.6 bal=2.186 mut=0.800 tok_s=31,860 avg_tok_s=32,255 |
| step=34500 tokens=352590000 ce=4.1700 ppl=64.7 bal=2.055 mut=0.798 tok_s=32,315 avg_tok_s=32,256 |
| step=34600 tokens=353612000 ce=4.1462 ppl=63.2 bal=1.870 mut=0.796 tok_s=32,271 avg_tok_s=32,256 |
| step=34700 tokens=354634000 ce=4.2234 ppl=68.3 bal=2.065 mut=0.803 tok_s=32,262 avg_tok_s=32,256 |
| step=34800 tokens=355656000 ce=4.1370 ppl=62.6 bal=2.149 mut=0.813 tok_s=31,811 avg_tok_s=32,254 |
| step=34900 tokens=356678000 ce=4.3404 ppl=76.7 bal=1.942 mut=0.795 tok_s=31,718 avg_tok_s=32,253 |
| step=35000 tokens=357700000 ce=4.2873 ppl=72.8 bal=2.092 mut=0.795 tok_s=32,155 avg_tok_s=32,253 |
| step=35100 tokens=358722000 ce=4.4427 ppl=85.0 bal=1.993 mut=0.812 tok_s=32,405 avg_tok_s=32,253 |
| step=35200 tokens=359744000 ce=4.1636 ppl=64.3 bal=2.220 mut=0.790 tok_s=32,337 avg_tok_s=32,253 |
| step=35300 tokens=360766000 ce=4.0480 ppl=57.3 bal=2.017 mut=0.801 tok_s=31,906 avg_tok_s=32,252 |
| step=35400 tokens=361788000 ce=4.2406 ppl=69.5 bal=2.167 mut=0.793 tok_s=32,191 avg_tok_s=32,252 |
| step=35500 tokens=362810000 ce=4.3316 ppl=76.1 bal=2.090 mut=0.788 tok_s=31,966 avg_tok_s=32,251 |
| step=35600 tokens=363832000 ce=4.3014 ppl=73.8 bal=2.084 mut=0.792 tok_s=31,964 avg_tok_s=32,250 |
| step=35700 tokens=364854000 ce=4.2724 ppl=71.7 bal=2.119 mut=0.797 tok_s=31,996 avg_tok_s=32,250 |
| step=35800 tokens=365876000 ce=4.2769 ppl=72.0 bal=2.198 mut=0.799 tok_s=32,344 avg_tok_s=32,250 |
| step=35900 tokens=366898000 ce=4.1184 ppl=61.5 bal=1.931 mut=0.800 tok_s=32,129 avg_tok_s=32,250 |
| step=36000 tokens=367920000 ce=4.3035 ppl=74.0 bal=2.019 mut=0.783 tok_s=32,020 avg_tok_s=32,249 |
| CKPT step=36000 valid_ce=4.2039 valid_ppl=66.9 |
| step=36100 tokens=368942000 ce=4.2026 ppl=66.9 bal=1.984 mut=0.788 tok_s=27,720 avg_tok_s=32,236 |
| step=36200 tokens=369964000 ce=4.4054 ppl=81.9 bal=2.148 mut=0.797 tok_s=32,224 avg_tok_s=32,236 |
| step=36300 tokens=370986000 ce=4.1886 ppl=65.9 bal=2.283 mut=0.791 tok_s=32,374 avg_tok_s=32,237 |
| step=36400 tokens=372008000 ce=4.1554 ppl=63.8 bal=2.206 mut=0.790 tok_s=32,065 avg_tok_s=32,236 |
| step=36500 tokens=373030000 ce=4.1429 ppl=63.0 bal=2.087 mut=0.802 tok_s=32,044 avg_tok_s=32,236 |
| step=36600 tokens=374052000 ce=4.4436 ppl=85.1 bal=2.435 mut=0.795 tok_s=32,021 avg_tok_s=32,235 |
| step=36700 tokens=375074000 ce=4.1505 ppl=63.5 bal=2.013 mut=0.800 tok_s=32,306 avg_tok_s=32,235 |
| step=36800 tokens=376096000 ce=4.2641 ppl=71.1 bal=2.058 mut=0.794 tok_s=32,431 avg_tok_s=32,236 |
| step=36900 tokens=377118000 ce=4.2462 ppl=69.8 bal=2.241 mut=0.790 tok_s=32,470 avg_tok_s=32,237 |
| step=37000 tokens=378140000 ce=4.2356 ppl=69.1 bal=2.030 mut=0.792 tok_s=32,075 avg_tok_s=32,236 |
| step=37100 tokens=379162000 ce=4.3655 ppl=78.7 bal=2.170 mut=0.797 tok_s=32,421 avg_tok_s=32,237 |
| step=37200 tokens=380184000 ce=4.2462 ppl=69.8 bal=2.021 mut=0.787 tok_s=32,303 avg_tok_s=32,237 |
| step=37300 tokens=381206000 ce=4.1777 ppl=65.2 bal=2.125 mut=0.783 tok_s=32,232 avg_tok_s=32,237 |
| step=37400 tokens=382228000 ce=4.2857 ppl=72.7 bal=2.139 mut=0.792 tok_s=32,496 avg_tok_s=32,237 |
| step=37500 tokens=383250000 ce=3.9507 ppl=52.0 bal=2.177 mut=0.787 tok_s=31,957 avg_tok_s=32,237 |
| step=37600 tokens=384272000 ce=4.0746 ppl=58.8 bal=2.131 mut=0.792 tok_s=32,139 avg_tok_s=32,236 |
| step=37700 tokens=385294000 ce=4.1476 ppl=63.3 bal=2.073 mut=0.795 tok_s=32,207 avg_tok_s=32,236 |
| step=37800 tokens=386316000 ce=4.2075 ppl=67.2 bal=2.175 mut=0.784 tok_s=31,990 avg_tok_s=32,236 |
| step=37900 tokens=387338000 ce=4.2180 ppl=67.9 bal=2.146 mut=0.800 tok_s=32,157 avg_tok_s=32,236 |
| step=38000 tokens=388360000 ce=4.2674 ppl=71.3 bal=1.964 mut=0.795 tok_s=32,136 avg_tok_s=32,235 |
| CKPT step=38000 valid_ce=4.2137 valid_ppl=67.6 |
| step=38100 tokens=389382000 ce=4.3450 ppl=77.1 bal=2.139 mut=0.800 tok_s=28,531 avg_tok_s=32,226 |
| step=38200 tokens=390404000 ce=3.9227 ppl=50.5 bal=1.994 mut=0.783 tok_s=32,335 avg_tok_s=32,226 |
| step=38300 tokens=391426000 ce=4.1821 ppl=65.5 bal=2.219 mut=0.795 tok_s=32,276 avg_tok_s=32,226 |
| step=38400 tokens=392448000 ce=4.0778 ppl=59.0 bal=2.448 mut=0.783 tok_s=32,258 avg_tok_s=32,226 |
| step=38500 tokens=393470000 ce=4.3346 ppl=76.3 bal=2.140 mut=0.794 tok_s=32,477 avg_tok_s=32,227 |
| step=38600 tokens=394492000 ce=4.0749 ppl=58.8 bal=2.731 mut=0.795 tok_s=32,336 avg_tok_s=32,227 |
| step=38700 tokens=395514000 ce=4.3082 ppl=74.3 bal=2.139 mut=0.796 tok_s=32,407 avg_tok_s=32,227 |
| step=38800 tokens=396536000 ce=4.2138 ppl=67.6 bal=2.058 mut=0.793 tok_s=32,005 avg_tok_s=32,227 |
| step=38900 tokens=397558000 ce=4.0825 ppl=59.3 bal=2.025 mut=0.788 tok_s=32,477 avg_tok_s=32,228 |
| step=39000 tokens=398580000 ce=4.0756 ppl=58.9 bal=2.063 mut=0.781 tok_s=32,568 avg_tok_s=32,228 |
| step=39100 tokens=399602000 ce=4.3339 ppl=76.2 bal=2.169 mut=0.792 tok_s=32,519 avg_tok_s=32,229 |
| step=39200 tokens=400624000 ce=4.1215 ppl=61.7 bal=2.111 mut=0.800 tok_s=32,206 avg_tok_s=32,229 |
| step=39300 tokens=401646000 ce=4.1262 ppl=61.9 bal=2.093 mut=0.792 tok_s=32,386 avg_tok_s=32,229 |
| step=39400 tokens=402668000 ce=4.2513 ppl=70.2 bal=1.958 mut=0.792 tok_s=32,530 avg_tok_s=32,230 |
| step=39500 tokens=403690000 ce=4.4118 ppl=82.4 bal=2.460 mut=0.793 tok_s=32,524 avg_tok_s=32,231 |
| step=39600 tokens=404712000 ce=4.2174 ppl=67.9 bal=2.074 mut=0.790 tok_s=32,259 avg_tok_s=32,231 |
| step=39700 tokens=405734000 ce=4.4658 ppl=87.0 bal=2.394 mut=0.797 tok_s=32,264 avg_tok_s=32,231 |
| step=39800 tokens=406756000 ce=4.2208 ppl=68.1 bal=2.075 mut=0.797 tok_s=32,509 avg_tok_s=32,232 |
| step=39900 tokens=407778000 ce=4.1826 ppl=65.5 bal=2.071 mut=0.792 tok_s=32,521 avg_tok_s=32,233 |
| step=40000 tokens=408800000 ce=4.1427 ppl=63.0 bal=2.074 mut=0.794 tok_s=32,389 avg_tok_s=32,233 |
| CKPT step=40000 valid_ce=4.1485 valid_ppl=63.3 |
| step=40100 tokens=409822000 ce=4.4088 ppl=82.2 bal=2.192 mut=0.793 tok_s=9,917 avg_tok_s=32,177 |
| step=40200 tokens=410844000 ce=4.0076 ppl=55.0 bal=2.207 mut=0.796 tok_s=32,240 avg_tok_s=32,177 |
| step=40300 tokens=411866000 ce=4.0529 ppl=57.6 bal=2.043 mut=0.798 tok_s=32,439 avg_tok_s=32,178 |
| step=40400 tokens=412888000 ce=4.0632 ppl=58.2 bal=2.060 mut=0.788 tok_s=32,186 avg_tok_s=32,178 |
| step=40500 tokens=413910000 ce=4.1699 ppl=64.7 bal=2.080 mut=0.792 tok_s=32,433 avg_tok_s=32,179 |
| step=40600 tokens=414932000 ce=4.2897 ppl=72.9 bal=2.075 mut=0.800 tok_s=32,313 avg_tok_s=32,179 |
| step=40700 tokens=415954000 ce=4.0085 ppl=55.1 bal=1.880 mut=0.787 tok_s=32,164 avg_tok_s=32,179 |
| step=40800 tokens=416976000 ce=4.1204 ppl=61.6 bal=1.955 mut=0.784 tok_s=32,177 avg_tok_s=32,179 |
| step=40900 tokens=417998000 ce=4.0470 ppl=57.2 bal=2.042 mut=0.801 tok_s=32,512 avg_tok_s=32,180 |
| step=41000 tokens=419020000 ce=3.7568 ppl=42.8 bal=2.196 mut=0.795 tok_s=32,551 avg_tok_s=32,181 |
| step=41100 tokens=420042000 ce=3.9349 ppl=51.2 bal=1.988 mut=0.807 tok_s=32,263 avg_tok_s=32,181 |
| step=41200 tokens=421064000 ce=4.0771 ppl=59.0 bal=2.024 mut=0.795 tok_s=32,220 avg_tok_s=32,181 |
| step=41300 tokens=422086000 ce=4.2530 ppl=70.3 bal=1.903 mut=0.803 tok_s=32,363 avg_tok_s=32,182 |
| step=41400 tokens=423108000 ce=3.9665 ppl=52.8 bal=2.141 mut=0.797 tok_s=32,533 avg_tok_s=32,182 |
| step=41500 tokens=424130000 ce=4.4257 ppl=83.6 bal=2.036 mut=0.795 tok_s=32,517 avg_tok_s=32,183 |
| step=41600 tokens=425152000 ce=4.1735 ppl=64.9 bal=1.995 mut=0.805 tok_s=32,363 avg_tok_s=32,184 |
| step=41700 tokens=426174000 ce=4.1729 ppl=64.9 bal=1.983 mut=0.799 tok_s=32,166 avg_tok_s=32,184 |
| step=41800 tokens=427196000 ce=4.0976 ppl=60.2 bal=2.074 mut=0.792 tok_s=32,464 avg_tok_s=32,184 |
| step=41900 tokens=428218000 ce=4.2930 ppl=73.2 bal=2.084 mut=0.787 tok_s=32,284 avg_tok_s=32,184 |
| step=42000 tokens=429240000 ce=3.9987 ppl=54.5 bal=2.117 mut=0.793 tok_s=32,262 avg_tok_s=32,185 |
| CKPT step=42000 valid_ce=4.0637 valid_ppl=58.2 |
| step=42100 tokens=430262000 ce=4.3104 ppl=74.5 bal=2.048 mut=0.785 tok_s=28,583 avg_tok_s=32,176 |
| step=42200 tokens=431284000 ce=4.0113 ppl=55.2 bal=1.920 mut=0.795 tok_s=32,226 avg_tok_s=32,176 |
| step=42300 tokens=432306000 ce=4.1393 ppl=62.8 bal=1.941 mut=0.793 tok_s=32,332 avg_tok_s=32,177 |
| step=42400 tokens=433328000 ce=4.1853 ppl=65.7 bal=2.139 mut=0.793 tok_s=32,068 avg_tok_s=32,176 |
| step=42500 tokens=434350000 ce=4.1920 ppl=66.2 bal=2.070 mut=0.795 tok_s=32,406 avg_tok_s=32,177 |
| step=42600 tokens=435372000 ce=3.9795 ppl=53.5 bal=2.071 mut=0.794 tok_s=32,345 avg_tok_s=32,177 |
| step=42700 tokens=436394000 ce=4.0071 ppl=55.0 bal=2.063 mut=0.804 tok_s=32,092 avg_tok_s=32,177 |
| step=42800 tokens=437416000 ce=4.0596 ppl=57.9 bal=1.944 mut=0.807 tok_s=32,343 avg_tok_s=32,177 |
| step=42900 tokens=438438000 ce=4.2758 ppl=71.9 bal=1.954 mut=0.796 tok_s=32,483 avg_tok_s=32,178 |
| step=43000 tokens=439460000 ce=3.9806 ppl=53.5 bal=2.035 mut=0.802 tok_s=32,425 avg_tok_s=32,179 |
| step=43100 tokens=440482000 ce=4.1465 ppl=63.2 bal=2.115 mut=0.786 tok_s=32,472 avg_tok_s=32,179 |
| step=43200 tokens=441504000 ce=4.1595 ppl=64.0 bal=2.049 mut=0.803 tok_s=32,242 avg_tok_s=32,180 |
| step=43300 tokens=442526000 ce=4.2124 ppl=67.5 bal=2.127 mut=0.800 tok_s=32,126 avg_tok_s=32,179 |
| step=43400 tokens=443548000 ce=4.1648 ppl=64.4 bal=2.413 mut=0.796 tok_s=32,533 avg_tok_s=32,180 |
| step=43500 tokens=444570000 ce=4.3314 ppl=76.1 bal=2.191 mut=0.800 tok_s=32,345 avg_tok_s=32,181 |
| step=43600 tokens=445592000 ce=3.8652 ppl=47.7 bal=2.007 mut=0.802 tok_s=32,053 avg_tok_s=32,180 |
| step=43700 tokens=446614000 ce=4.2411 ppl=69.5 bal=2.015 mut=0.800 tok_s=32,361 avg_tok_s=32,181 |
| step=43800 tokens=447636000 ce=4.1758 ppl=65.1 bal=2.069 mut=0.790 tok_s=32,601 avg_tok_s=32,182 |
| step=43900 tokens=448658000 ce=4.0595 ppl=57.9 bal=2.488 mut=0.803 tok_s=32,495 avg_tok_s=32,182 |
| step=44000 tokens=449680000 ce=4.0750 ppl=58.8 bal=2.267 mut=0.785 tok_s=32,554 avg_tok_s=32,183 |
| CKPT step=44000 valid_ce=4.1829 valid_ppl=65.6 |
| step=44100 tokens=450702000 ce=4.0088 ppl=55.1 bal=1.945 mut=0.804 tok_s=28,802 avg_tok_s=32,176 |
| step=44200 tokens=451724000 ce=3.9063 ppl=49.7 bal=2.001 mut=0.793 tok_s=32,295 avg_tok_s=32,176 |
| step=44300 tokens=452746000 ce=4.3448 ppl=77.1 bal=1.788 mut=0.800 tok_s=32,556 avg_tok_s=32,177 |
| step=44400 tokens=453768000 ce=4.0670 ppl=58.4 bal=2.356 mut=0.800 tok_s=32,514 avg_tok_s=32,177 |
| step=44500 tokens=454790000 ce=4.0896 ppl=59.7 bal=2.081 mut=0.803 tok_s=32,362 avg_tok_s=32,178 |
| step=44600 tokens=455812000 ce=4.2580 ppl=70.7 bal=2.008 mut=0.805 tok_s=32,503 avg_tok_s=32,179 |
| step=44700 tokens=456834000 ce=4.2964 ppl=73.4 bal=2.189 mut=0.790 tok_s=32,609 avg_tok_s=32,180 |
| step=44800 tokens=457856000 ce=4.0986 ppl=60.3 bal=1.988 mut=0.803 tok_s=32,219 avg_tok_s=32,180 |
| step=44900 tokens=458878000 ce=4.3511 ppl=77.6 bal=2.152 mut=0.790 tok_s=32,504 avg_tok_s=32,180 |
| step=45000 tokens=459900000 ce=4.1199 ppl=61.6 bal=1.964 mut=0.791 tok_s=32,656 avg_tok_s=32,181 |
| step=45100 tokens=460922000 ce=4.0833 ppl=59.3 bal=1.990 mut=0.797 tok_s=32,284 avg_tok_s=32,182 |
| step=45200 tokens=461944000 ce=4.3400 ppl=76.7 bal=2.421 mut=0.787 tok_s=32,537 avg_tok_s=32,182 |
| step=45300 tokens=462966000 ce=4.0424 ppl=57.0 bal=2.050 mut=0.792 tok_s=32,650 avg_tok_s=32,184 |
| step=45400 tokens=463988000 ce=4.1313 ppl=62.3 bal=1.990 mut=0.804 tok_s=32,273 avg_tok_s=32,184 |
| step=45500 tokens=465010000 ce=4.1313 ppl=62.3 bal=1.916 mut=0.797 tok_s=32,344 avg_tok_s=32,184 |
| step=45600 tokens=466032000 ce=4.0543 ppl=57.6 bal=2.193 mut=0.799 tok_s=32,282 avg_tok_s=32,184 |
| step=45700 tokens=467054000 ce=4.4114 ppl=82.4 bal=2.022 mut=0.794 tok_s=32,422 avg_tok_s=32,185 |
| step=45800 tokens=468076000 ce=3.9523 ppl=52.1 bal=2.090 mut=0.804 tok_s=32,610 avg_tok_s=32,186 |
| step=45900 tokens=469098000 ce=4.2944 ppl=73.3 bal=1.983 mut=0.805 tok_s=32,520 avg_tok_s=32,186 |
| step=46000 tokens=470120000 ce=3.8515 ppl=47.1 bal=2.012 mut=0.797 tok_s=32,412 avg_tok_s=32,187 |
| CKPT step=46000 valid_ce=4.2345 valid_ppl=69.0 |
| step=46100 tokens=471142000 ce=4.0748 ppl=58.8 bal=1.989 mut=0.803 tok_s=28,977 avg_tok_s=32,180 |
| step=46200 tokens=472164000 ce=4.2426 ppl=69.6 bal=2.085 mut=0.798 tok_s=32,396 avg_tok_s=32,180 |
| step=46300 tokens=473186000 ce=3.9854 ppl=53.8 bal=2.035 mut=0.796 tok_s=32,613 avg_tok_s=32,181 |
| step=46400 tokens=474208000 ce=4.3394 ppl=76.7 bal=2.126 mut=0.800 tok_s=32,451 avg_tok_s=32,182 |
| step=46500 tokens=475230000 ce=4.4037 ppl=81.8 bal=2.402 mut=0.803 tok_s=32,566 avg_tok_s=32,183 |
| step=46600 tokens=476252000 ce=4.1381 ppl=62.7 bal=2.016 mut=0.804 tok_s=32,780 avg_tok_s=32,184 |
| step=46700 tokens=477274000 ce=4.0643 ppl=58.2 bal=1.920 mut=0.803 tok_s=32,685 avg_tok_s=32,185 |
| step=46800 tokens=478296000 ce=3.9698 ppl=53.0 bal=2.157 mut=0.801 tok_s=32,524 avg_tok_s=32,186 |
| step=46900 tokens=479318000 ce=4.2474 ppl=69.9 bal=2.145 mut=0.789 tok_s=32,695 avg_tok_s=32,187 |
| step=47000 tokens=480340000 ce=4.3003 ppl=73.7 bal=2.954 mut=0.798 tok_s=32,653 avg_tok_s=32,188 |
| step=47100 tokens=481362000 ce=4.1931 ppl=66.2 bal=1.966 mut=0.795 tok_s=32,546 avg_tok_s=32,189 |
| step=47200 tokens=482384000 ce=4.1294 ppl=62.1 bal=1.951 mut=0.802 tok_s=32,299 avg_tok_s=32,189 |
| step=47300 tokens=483406000 ce=4.1519 ppl=63.6 bal=2.115 mut=0.794 tok_s=32,461 avg_tok_s=32,190 |
| step=47400 tokens=484428000 ce=4.0118 ppl=55.2 bal=2.039 mut=0.812 tok_s=32,553 avg_tok_s=32,190 |
| step=47500 tokens=485450000 ce=4.1980 ppl=66.6 bal=2.006 mut=0.796 tok_s=32,585 avg_tok_s=32,191 |
| step=47600 tokens=486472000 ce=4.1112 ppl=61.0 bal=2.134 mut=0.803 tok_s=32,515 avg_tok_s=32,192 |
| step=47700 tokens=487494000 ce=4.1304 ppl=62.2 bal=1.946 mut=0.801 tok_s=32,610 avg_tok_s=32,193 |
| step=47800 tokens=488516000 ce=4.3355 ppl=76.4 bal=1.957 mut=0.804 tok_s=32,423 avg_tok_s=32,193 |
| step=47900 tokens=489538000 ce=4.1524 ppl=63.6 bal=1.982 mut=0.796 tok_s=32,669 avg_tok_s=32,194 |
| step=48000 tokens=490560000 ce=4.2666 ppl=71.3 bal=2.010 mut=0.801 tok_s=32,550 avg_tok_s=32,195 |
| CKPT step=48000 valid_ce=4.1724 valid_ppl=64.9 |
| step=48100 tokens=491582000 ce=4.0972 ppl=60.2 bal=2.086 mut=0.815 tok_s=28,871 avg_tok_s=32,188 |
| step=48200 tokens=492604000 ce=4.3257 ppl=75.6 bal=2.286 mut=0.809 tok_s=32,439 avg_tok_s=32,188 |
| step=48300 tokens=493626000 ce=4.1052 ppl=60.7 bal=2.167 mut=0.802 tok_s=32,716 avg_tok_s=32,190 |
| step=48400 tokens=494648000 ce=4.1738 ppl=65.0 bal=1.901 mut=0.805 tok_s=32,520 avg_tok_s=32,190 |
| step=48500 tokens=495670000 ce=4.1846 ppl=65.7 bal=2.103 mut=0.796 tok_s=32,400 avg_tok_s=32,191 |
| step=48600 tokens=496692000 ce=4.0735 ppl=58.8 bal=1.930 mut=0.799 tok_s=32,356 avg_tok_s=32,191 |
| step=48700 tokens=497714000 ce=4.1509 ppl=63.5 bal=1.947 mut=0.794 tok_s=32,634 avg_tok_s=32,192 |
| step=48800 tokens=498736000 ce=3.9102 ppl=49.9 bal=2.303 mut=0.801 tok_s=32,392 avg_tok_s=32,192 |
| step=48900 tokens=499758000 ce=4.3286 ppl=75.8 bal=2.108 mut=0.796 tok_s=32,548 avg_tok_s=32,193 |
| step=49000 tokens=500780000 ce=4.3095 ppl=74.4 bal=2.130 mut=0.793 tok_s=32,580 avg_tok_s=32,194 |
| step=49100 tokens=501802000 ce=4.2914 ppl=73.1 bal=2.119 mut=0.800 tok_s=32,495 avg_tok_s=32,194 |
| step=49200 tokens=502824000 ce=4.1434 ppl=63.0 bal=1.966 mut=0.792 tok_s=32,247 avg_tok_s=32,195 |
| step=49300 tokens=503846000 ce=4.2848 ppl=72.6 bal=2.076 mut=0.803 tok_s=32,410 avg_tok_s=32,195 |
| step=49400 tokens=504868000 ce=4.1789 ppl=65.3 bal=2.224 mut=0.792 tok_s=32,495 avg_tok_s=32,196 |
| step=49500 tokens=505890000 ce=4.0749 ppl=58.8 bal=1.998 mut=0.794 tok_s=32,271 avg_tok_s=32,196 |
| step=49600 tokens=506912000 ce=4.2679 ppl=71.4 bal=2.184 mut=0.805 tok_s=32,404 avg_tok_s=32,196 |
| step=49700 tokens=507934000 ce=4.0302 ppl=56.3 bal=2.030 mut=0.810 tok_s=32,501 avg_tok_s=32,197 |
| step=49800 tokens=508956000 ce=4.2162 ppl=67.8 bal=1.969 mut=0.795 tok_s=32,398 avg_tok_s=32,197 |
| step=49900 tokens=509978000 ce=4.1041 ppl=60.6 bal=2.024 mut=0.799 tok_s=32,429 avg_tok_s=32,198 |
| step=50000 tokens=511000000 ce=4.2117 ppl=67.5 bal=2.001 mut=0.800 tok_s=32,103 avg_tok_s=32,198 |
| CKPT step=50000 valid_ce=4.2247 valid_ppl=68.4 |
| step=50100 tokens=512022000 ce=4.1648 ppl=64.4 bal=2.204 mut=0.795 tok_s=28,750 avg_tok_s=32,191 |
| step=50200 tokens=513044000 ce=4.2191 ppl=68.0 bal=1.932 mut=0.804 tok_s=32,372 avg_tok_s=32,191 |
| step=50300 tokens=514066000 ce=4.4038 ppl=81.8 bal=1.985 mut=0.808 tok_s=32,291 avg_tok_s=32,191 |
| step=50400 tokens=515088000 ce=4.1334 ppl=62.4 bal=1.943 mut=0.811 tok_s=32,221 avg_tok_s=32,191 |
| step=50500 tokens=516110000 ce=4.0466 ppl=57.2 bal=2.145 mut=0.795 tok_s=32,130 avg_tok_s=32,191 |
| step=50600 tokens=517132000 ce=4.1867 ppl=65.8 bal=2.015 mut=0.813 tok_s=32,248 avg_tok_s=32,191 |
| step=50700 tokens=518154000 ce=4.0768 ppl=59.0 bal=2.254 mut=0.802 tok_s=32,259 avg_tok_s=32,191 |
| step=50800 tokens=519176000 ce=4.0422 ppl=57.0 bal=3.857 mut=0.792 tok_s=31,950 avg_tok_s=32,191 |
| step=50900 tokens=520198000 ce=3.9618 ppl=52.6 bal=2.025 mut=0.809 tok_s=32,313 avg_tok_s=32,191 |
| step=51000 tokens=521220000 ce=4.0557 ppl=57.7 bal=1.902 mut=0.805 tok_s=32,326 avg_tok_s=32,191 |
| step=51100 tokens=522242000 ce=4.3282 ppl=75.8 bal=2.007 mut=0.790 tok_s=32,285 avg_tok_s=32,192 |
| step=51200 tokens=523264000 ce=4.0434 ppl=57.0 bal=2.199 mut=0.796 tok_s=32,251 avg_tok_s=32,192 |
| step=51300 tokens=524286000 ce=4.1636 ppl=64.3 bal=1.943 mut=0.787 tok_s=32,171 avg_tok_s=32,192 |
| step=51400 tokens=525308000 ce=4.1212 ppl=61.6 bal=1.908 mut=0.796 tok_s=32,328 avg_tok_s=32,192 |
| step=51500 tokens=526330000 ce=3.9826 ppl=53.7 bal=2.021 mut=0.802 tok_s=32,326 avg_tok_s=32,192 |
| step=51600 tokens=527352000 ce=4.2248 ppl=68.4 bal=3.424 mut=0.804 tok_s=32,259 avg_tok_s=32,192 |
| step=51700 tokens=528374000 ce=4.0831 ppl=59.3 bal=1.947 mut=0.807 tok_s=32,191 avg_tok_s=32,192 |
| step=51800 tokens=529396000 ce=4.1307 ppl=62.2 bal=2.099 mut=0.798 tok_s=32,487 avg_tok_s=32,193 |
| step=51900 tokens=530418000 ce=4.0508 ppl=57.4 bal=2.021 mut=0.801 tok_s=32,433 avg_tok_s=32,193 |
| step=52000 tokens=531440000 ce=4.1402 ppl=62.8 bal=1.850 mut=0.797 tok_s=32,357 avg_tok_s=32,194 |
| CKPT step=52000 valid_ce=4.1258 valid_ppl=61.9 |
| step=52100 tokens=532462000 ce=3.7848 ppl=44.0 bal=2.108 mut=0.797 tok_s=28,815 avg_tok_s=32,187 |
| step=52200 tokens=533484000 ce=4.1298 ppl=62.2 bal=2.459 mut=0.800 tok_s=32,407 avg_tok_s=32,188 |
| step=52300 tokens=534506000 ce=4.1997 ppl=66.7 bal=2.159 mut=0.792 tok_s=32,068 avg_tok_s=32,187 |
| step=52400 tokens=535528000 ce=4.3673 ppl=78.8 bal=2.091 mut=0.811 tok_s=32,360 avg_tok_s=32,188 |
| step=52500 tokens=536550000 ce=4.1917 ppl=66.1 bal=2.175 mut=0.808 tok_s=32,481 avg_tok_s=32,188 |
| step=52600 tokens=537572000 ce=4.1084 ppl=60.8 bal=2.070 mut=0.822 tok_s=32,469 avg_tok_s=32,189 |
| step=52700 tokens=538594000 ce=4.1559 ppl=63.8 bal=2.096 mut=0.814 tok_s=32,286 avg_tok_s=32,189 |
| step=52800 tokens=539616000 ce=3.9803 ppl=53.5 bal=2.122 mut=0.811 tok_s=32,044 avg_tok_s=32,189 |
| step=52900 tokens=540638000 ce=4.2036 ppl=66.9 bal=2.076 mut=0.816 tok_s=31,987 avg_tok_s=32,188 |
| step=53000 tokens=541660000 ce=4.3863 ppl=80.3 bal=2.048 mut=0.822 tok_s=32,357 avg_tok_s=32,189 |
| step=53100 tokens=542682000 ce=3.8819 ppl=48.5 bal=2.022 mut=0.823 tok_s=32,233 avg_tok_s=32,189 |
| step=53200 tokens=543704000 ce=4.1616 ppl=64.2 bal=2.260 mut=0.812 tok_s=31,814 avg_tok_s=32,188 |
| step=53300 tokens=544726000 ce=4.0868 ppl=59.6 bal=2.339 mut=0.819 tok_s=32,275 avg_tok_s=32,188 |
| step=53400 tokens=545748000 ce=4.3070 ppl=74.2 bal=2.201 mut=0.824 tok_s=32,334 avg_tok_s=32,188 |
| step=53500 tokens=546770000 ce=4.2198 ppl=68.0 bal=1.987 mut=0.825 tok_s=32,263 avg_tok_s=32,189 |
| step=53600 tokens=547792000 ce=4.0876 ppl=59.6 bal=1.916 mut=0.810 tok_s=32,225 avg_tok_s=32,189 |
| step=53700 tokens=548814000 ce=4.0287 ppl=56.2 bal=2.347 mut=0.818 tok_s=32,413 avg_tok_s=32,189 |
| step=53800 tokens=549836000 ce=3.8710 ppl=48.0 bal=1.864 mut=0.822 tok_s=32,444 avg_tok_s=32,190 |
| step=53900 tokens=550858000 ce=4.2802 ppl=72.3 bal=2.323 mut=0.813 tok_s=32,368 avg_tok_s=32,190 |
| step=54000 tokens=551880000 ce=3.9009 ppl=49.4 bal=1.897 mut=0.825 tok_s=32,283 avg_tok_s=32,190 |
| CKPT step=54000 valid_ce=3.9937 valid_ppl=54.3 |
| step=54100 tokens=552902000 ce=4.1921 ppl=66.2 bal=1.938 mut=0.821 tok_s=28,548 avg_tok_s=32,183 |
| step=54200 tokens=553924000 ce=4.1540 ppl=63.7 bal=1.988 mut=0.814 tok_s=32,441 avg_tok_s=32,184 |
| step=54300 tokens=554946000 ce=4.1130 ppl=61.1 bal=1.893 mut=0.830 tok_s=32,403 avg_tok_s=32,184 |
| step=54400 tokens=555968000 ce=4.1347 ppl=62.5 bal=2.093 mut=0.826 tok_s=32,259 avg_tok_s=32,184 |
| step=54500 tokens=556990000 ce=4.0706 ppl=58.6 bal=2.033 mut=0.825 tok_s=32,441 avg_tok_s=32,185 |
| step=54600 tokens=558012000 ce=4.2314 ppl=68.8 bal=2.079 mut=0.819 tok_s=32,384 avg_tok_s=32,185 |
| step=54700 tokens=559034000 ce=4.2332 ppl=68.9 bal=2.056 mut=0.823 tok_s=32,324 avg_tok_s=32,185 |
| step=54800 tokens=560056000 ce=4.1638 ppl=64.3 bal=2.028 mut=0.830 tok_s=32,319 avg_tok_s=32,186 |
| step=54900 tokens=561078000 ce=3.9608 ppl=52.5 bal=1.928 mut=0.817 tok_s=32,283 avg_tok_s=32,186 |
| step=55000 tokens=562100000 ce=4.1914 ppl=66.1 bal=2.038 mut=0.825 tok_s=32,557 avg_tok_s=32,187 |
| step=55100 tokens=563122000 ce=4.0524 ppl=57.5 bal=2.020 mut=0.836 tok_s=32,584 avg_tok_s=32,187 |
| step=55200 tokens=564144000 ce=4.0208 ppl=55.7 bal=1.932 mut=0.820 tok_s=32,425 avg_tok_s=32,188 |
| step=55300 tokens=565166000 ce=4.1843 ppl=65.6 bal=2.041 mut=0.818 tok_s=32,519 avg_tok_s=32,188 |
| step=55400 tokens=566188000 ce=4.2898 ppl=73.0 bal=2.368 mut=0.821 tok_s=32,559 avg_tok_s=32,189 |
| step=55500 tokens=567210000 ce=4.0444 ppl=57.1 bal=2.125 mut=0.824 tok_s=32,511 avg_tok_s=32,190 |
| step=55600 tokens=568232000 ce=4.3203 ppl=75.2 bal=2.150 mut=0.819 tok_s=32,608 avg_tok_s=32,190 |
| step=55700 tokens=569254000 ce=4.2220 ppl=68.2 bal=2.721 mut=0.813 tok_s=32,792 avg_tok_s=32,191 |
| step=55800 tokens=570276000 ce=4.1069 ppl=60.8 bal=1.995 mut=0.831 tok_s=32,515 avg_tok_s=32,192 |
| step=55900 tokens=571298000 ce=4.1635 ppl=64.3 bal=2.124 mut=0.827 tok_s=32,772 avg_tok_s=32,193 |
| step=56000 tokens=572320000 ce=4.2987 ppl=73.6 bal=2.534 mut=0.826 tok_s=32,604 avg_tok_s=32,194 |
| CKPT step=56000 valid_ce=4.1474 valid_ppl=63.3 |
| step=56100 tokens=573342000 ce=4.2206 ppl=68.1 bal=1.969 mut=0.832 tok_s=28,933 avg_tok_s=32,188 |
| step=56200 tokens=574364000 ce=3.8779 ppl=48.3 bal=1.885 mut=0.835 tok_s=32,712 avg_tok_s=32,189 |
| step=56300 tokens=575386000 ce=4.2269 ppl=68.5 bal=2.152 mut=0.842 tok_s=32,474 avg_tok_s=32,189 |
| step=56400 tokens=576408000 ce=4.2219 ppl=68.2 bal=2.018 mut=0.832 tok_s=32,656 avg_tok_s=32,190 |
| step=56500 tokens=577430000 ce=4.0565 ppl=57.8 bal=1.894 mut=0.839 tok_s=32,898 avg_tok_s=32,191 |
| step=56600 tokens=578452000 ce=4.1096 ppl=60.9 bal=1.934 mut=0.826 tok_s=33,041 avg_tok_s=32,193 |
| step=56700 tokens=579474000 ce=4.2418 ppl=69.5 bal=1.887 mut=0.834 tok_s=32,995 avg_tok_s=32,194 |
| step=56800 tokens=580496000 ce=4.1076 ppl=60.8 bal=2.041 mut=0.837 tok_s=33,039 avg_tok_s=32,196 |
| step=56900 tokens=581518000 ce=4.0507 ppl=57.4 bal=2.025 mut=0.831 tok_s=33,081 avg_tok_s=32,197 |
| step=57000 tokens=582540000 ce=4.0815 ppl=59.2 bal=2.012 mut=0.839 tok_s=33,051 avg_tok_s=32,199 |
| step=57100 tokens=583562000 ce=4.0035 ppl=54.8 bal=1.922 mut=0.843 tok_s=33,064 avg_tok_s=32,200 |
| step=57200 tokens=584584000 ce=4.0051 ppl=54.9 bal=2.053 mut=0.840 tok_s=32,848 avg_tok_s=32,202 |
| step=57300 tokens=585606000 ce=4.1540 ppl=63.7 bal=2.151 mut=0.834 tok_s=33,274 avg_tok_s=32,203 |
| step=57400 tokens=586628000 ce=4.2110 ppl=67.4 bal=2.041 mut=0.839 tok_s=33,143 avg_tok_s=32,205 |
| step=57500 tokens=587650000 ce=4.2011 ppl=66.8 bal=2.147 mut=0.851 tok_s=33,208 avg_tok_s=32,207 |
| step=57600 tokens=588672000 ce=4.0422 ppl=57.0 bal=1.914 mut=0.847 tok_s=33,040 avg_tok_s=32,208 |
| step=57700 tokens=589694000 ce=4.1481 ppl=63.3 bal=1.857 mut=0.842 tok_s=32,898 avg_tok_s=32,209 |
| step=57800 tokens=590716000 ce=4.2700 ppl=71.5 bal=2.109 mut=0.840 tok_s=33,161 avg_tok_s=32,211 |
| step=57900 tokens=591738000 ce=4.1142 ppl=61.2 bal=1.932 mut=0.848 tok_s=33,076 avg_tok_s=32,213 |
| step=58000 tokens=592760000 ce=4.0972 ppl=60.2 bal=1.907 mut=0.837 tok_s=33,076 avg_tok_s=32,214 |
| CKPT step=58000 valid_ce=4.2227 valid_ppl=68.2 |
| step=58100 tokens=593782000 ce=4.0974 ppl=60.2 bal=2.172 mut=0.837 tok_s=29,111 avg_tok_s=32,209 |
| step=58200 tokens=594804000 ce=4.4223 ppl=83.3 bal=2.477 mut=0.847 tok_s=33,235 avg_tok_s=32,210 |
| step=58300 tokens=595826000 ce=4.1573 ppl=63.9 bal=2.851 mut=0.836 tok_s=33,159 avg_tok_s=32,212 |
| step=58400 tokens=596848000 ce=4.1679 ppl=64.6 bal=2.017 mut=0.835 tok_s=33,083 avg_tok_s=32,214 |
| step=58500 tokens=597870000 ce=4.1309 ppl=62.2 bal=2.119 mut=0.840 tok_s=32,951 avg_tok_s=32,215 |
| step=58600 tokens=598892000 ce=4.2399 ppl=69.4 bal=2.087 mut=0.832 tok_s=33,223 avg_tok_s=32,217 |
| step=58700 tokens=599914000 ce=4.0522 ppl=57.5 bal=2.036 mut=0.853 tok_s=33,226 avg_tok_s=32,218 |
| step=58800 tokens=600936000 ce=3.9954 ppl=54.3 bal=2.110 mut=0.844 tok_s=33,088 avg_tok_s=32,220 |
| step=58900 tokens=601958000 ce=4.0177 ppl=55.6 bal=2.033 mut=0.843 tok_s=33,168 avg_tok_s=32,221 |
| step=59000 tokens=602980000 ce=4.1094 ppl=60.9 bal=1.929 mut=0.855 tok_s=33,087 avg_tok_s=32,223 |
| step=59100 tokens=604002000 ce=4.0502 ppl=57.4 bal=1.991 mut=0.842 tok_s=33,166 avg_tok_s=32,224 |
| step=59200 tokens=605024000 ce=3.8444 ppl=46.7 bal=2.058 mut=0.852 tok_s=32,945 avg_tok_s=32,226 |
| step=59300 tokens=606046000 ce=4.2317 ppl=68.8 bal=2.004 mut=0.840 tok_s=32,995 avg_tok_s=32,227 |
| step=59400 tokens=607068000 ce=4.1113 ppl=61.0 bal=2.046 mut=0.841 tok_s=33,213 avg_tok_s=32,229 |
| step=59500 tokens=608090000 ce=3.9922 ppl=54.2 bal=1.929 mut=0.853 tok_s=33,207 avg_tok_s=32,230 |
| step=59600 tokens=609112000 ce=4.0496 ppl=57.4 bal=2.123 mut=0.853 tok_s=33,062 avg_tok_s=32,232 |
| step=59700 tokens=610134000 ce=3.8959 ppl=49.2 bal=2.118 mut=0.836 tok_s=33,210 avg_tok_s=32,233 |
| step=59800 tokens=611156000 ce=4.0762 ppl=58.9 bal=1.881 mut=0.847 tok_s=33,173 avg_tok_s=32,235 |
| step=59900 tokens=612178000 ce=4.2080 ppl=67.2 bal=1.942 mut=0.841 tok_s=32,943 avg_tok_s=32,236 |
| step=60000 tokens=613200000 ce=3.9714 ppl=53.1 bal=2.072 mut=0.854 tok_s=33,021 avg_tok_s=32,237 |
| CKPT step=60000 valid_ce=4.0102 valid_ppl=55.2 |
| step=60100 tokens=614222000 ce=4.0134 ppl=55.3 bal=2.110 mut=0.843 tok_s=13,239 avg_tok_s=32,206 |
| step=60200 tokens=615244000 ce=4.1979 ppl=66.5 bal=1.861 mut=0.847 tok_s=33,173 avg_tok_s=32,207 |
| step=60300 tokens=616266000 ce=4.2423 ppl=69.6 bal=1.947 mut=0.838 tok_s=33,157 avg_tok_s=32,209 |
| step=60400 tokens=617288000 ce=3.8038 ppl=44.9 bal=2.072 mut=0.844 tok_s=32,706 avg_tok_s=32,210 |
| step=60500 tokens=618310000 ce=4.0144 ppl=55.4 bal=1.846 mut=0.849 tok_s=33,175 avg_tok_s=32,211 |
| step=60600 tokens=619332000 ce=4.0725 ppl=58.7 bal=1.899 mut=0.850 tok_s=33,195 avg_tok_s=32,213 |
| step=60700 tokens=620354000 ce=4.2694 ppl=71.5 bal=2.073 mut=0.851 tok_s=33,203 avg_tok_s=32,215 |
| step=60800 tokens=621376000 ce=3.9333 ppl=51.1 bal=1.903 mut=0.849 tok_s=33,013 avg_tok_s=32,216 |
| step=60900 tokens=622398000 ce=3.8909 ppl=49.0 bal=1.926 mut=0.846 tok_s=33,163 avg_tok_s=32,217 |
| step=61000 tokens=623420000 ce=4.0484 ppl=57.3 bal=2.192 mut=0.845 tok_s=32,851 avg_tok_s=32,219 |
| step=61100 tokens=624442000 ce=4.0847 ppl=59.4 bal=2.001 mut=0.855 tok_s=33,181 avg_tok_s=32,220 |
| step=61200 tokens=625464000 ce=4.1790 ppl=65.3 bal=1.959 mut=0.847 tok_s=32,880 avg_tok_s=32,221 |
| step=61300 tokens=626486000 ce=4.1301 ppl=62.2 bal=1.878 mut=0.842 tok_s=32,830 avg_tok_s=32,222 |
| step=61400 tokens=627508000 ce=3.8864 ppl=48.7 bal=2.064 mut=0.851 tok_s=33,230 avg_tok_s=32,224 |
| step=61500 tokens=628530000 ce=3.9503 ppl=52.0 bal=1.975 mut=0.849 tok_s=33,179 avg_tok_s=32,225 |
| step=61600 tokens=629552000 ce=3.9513 ppl=52.0 bal=2.124 mut=0.843 tok_s=32,999 avg_tok_s=32,227 |
| step=61700 tokens=630574000 ce=4.1981 ppl=66.6 bal=1.927 mut=0.842 tok_s=33,132 avg_tok_s=32,228 |
| step=61800 tokens=631596000 ce=3.8572 ppl=47.3 bal=1.924 mut=0.844 tok_s=33,070 avg_tok_s=32,229 |
| step=61900 tokens=632618000 ce=4.1211 ppl=61.6 bal=1.960 mut=0.857 tok_s=33,034 avg_tok_s=32,231 |
| step=62000 tokens=633640000 ce=4.0235 ppl=55.9 bal=1.887 mut=0.843 tok_s=32,884 avg_tok_s=32,232 |
| CKPT step=62000 valid_ce=3.9794 valid_ppl=53.5 |
| step=62100 tokens=634662000 ce=4.1498 ppl=63.4 bal=1.997 mut=0.841 tok_s=28,981 avg_tok_s=32,227 |
| step=62200 tokens=635684000 ce=4.1268 ppl=62.0 bal=2.061 mut=0.846 tok_s=33,087 avg_tok_s=32,228 |
| step=62300 tokens=636706000 ce=4.0446 ppl=57.1 bal=1.913 mut=0.837 tok_s=32,875 avg_tok_s=32,229 |
| step=62400 tokens=637728000 ce=3.9314 ppl=51.0 bal=2.020 mut=0.844 tok_s=32,964 avg_tok_s=32,230 |
| step=62500 tokens=638750000 ce=4.1778 ppl=65.2 bal=1.874 mut=0.834 tok_s=33,242 avg_tok_s=32,232 |
| step=62600 tokens=639772000 ce=4.2430 ppl=69.6 bal=1.889 mut=0.843 tok_s=33,228 avg_tok_s=32,233 |
| step=62700 tokens=640794000 ce=4.1082 ppl=60.8 bal=1.938 mut=0.849 tok_s=32,932 avg_tok_s=32,234 |
| step=62800 tokens=641816000 ce=4.0985 ppl=60.3 bal=2.515 mut=0.850 tok_s=33,066 avg_tok_s=32,236 |
| step=62900 tokens=642838000 ce=4.2712 ppl=71.6 bal=2.014 mut=0.839 tok_s=32,880 avg_tok_s=32,237 |
| step=63000 tokens=643860000 ce=4.1706 ppl=64.8 bal=2.000 mut=0.858 tok_s=32,931 avg_tok_s=32,238 |
| step=63100 tokens=644882000 ce=4.0127 ppl=55.3 bal=1.986 mut=0.841 tok_s=33,164 avg_tok_s=32,239 |
| step=63200 tokens=645904000 ce=4.2281 ppl=68.6 bal=2.092 mut=0.843 tok_s=32,826 avg_tok_s=32,240 |
| step=63300 tokens=646926000 ce=4.1880 ppl=65.9 bal=2.197 mut=0.840 tok_s=33,128 avg_tok_s=32,242 |
| step=63400 tokens=647948000 ce=4.0663 ppl=58.3 bal=1.822 mut=0.832 tok_s=33,082 avg_tok_s=32,243 |
| step=63500 tokens=648970000 ce=4.1925 ppl=66.2 bal=1.907 mut=0.834 tok_s=32,949 avg_tok_s=32,244 |
| step=63600 tokens=649992000 ce=3.8921 ppl=49.0 bal=1.970 mut=0.849 tok_s=33,042 avg_tok_s=32,245 |
| step=63700 tokens=651014000 ce=4.3440 ppl=77.0 bal=2.088 mut=0.848 tok_s=33,265 avg_tok_s=32,247 |
| step=63800 tokens=652036000 ce=4.1726 ppl=64.9 bal=1.885 mut=0.845 tok_s=33,235 avg_tok_s=32,249 |
| step=63900 tokens=653058000 ce=4.0294 ppl=56.2 bal=1.927 mut=0.843 tok_s=32,843 avg_tok_s=32,249 |
| step=64000 tokens=654080000 ce=3.8904 ppl=48.9 bal=1.924 mut=0.836 tok_s=33,026 avg_tok_s=32,251 |
| CKPT step=64000 valid_ce=4.0709 valid_ppl=58.6 |
| step=64100 tokens=655102000 ce=4.0670 ppl=58.4 bal=1.867 mut=0.842 tok_s=29,224 avg_tok_s=32,246 |
| step=64200 tokens=656124000 ce=3.9643 ppl=52.7 bal=1.926 mut=0.840 tok_s=33,254 avg_tok_s=32,248 |
| step=64300 tokens=657146000 ce=3.9196 ppl=50.4 bal=1.963 mut=0.848 tok_s=32,944 avg_tok_s=32,249 |
| step=64400 tokens=658168000 ce=4.0870 ppl=59.6 bal=1.918 mut=0.849 tok_s=33,053 avg_tok_s=32,250 |
| step=64500 tokens=659190000 ce=4.1568 ppl=63.9 bal=1.927 mut=0.859 tok_s=33,042 avg_tok_s=32,251 |
| step=64600 tokens=660212000 ce=4.3196 ppl=75.2 bal=2.019 mut=0.842 tok_s=33,097 avg_tok_s=32,252 |
| step=64700 tokens=661234000 ce=3.9830 ppl=53.7 bal=1.860 mut=0.842 tok_s=33,128 avg_tok_s=32,254 |
| step=64800 tokens=662256000 ce=3.8898 ppl=48.9 bal=1.959 mut=0.845 tok_s=33,007 avg_tok_s=32,255 |
| step=64900 tokens=663278000 ce=4.1546 ppl=63.7 bal=2.095 mut=0.850 tok_s=33,076 avg_tok_s=32,256 |
| step=65000 tokens=664300000 ce=4.1438 ppl=63.0 bal=1.815 mut=0.845 tok_s=33,070 avg_tok_s=32,257 |
| step=65100 tokens=665322000 ce=3.9567 ppl=52.3 bal=2.015 mut=0.842 tok_s=33,098 avg_tok_s=32,259 |
| step=65200 tokens=666344000 ce=4.2868 ppl=72.7 bal=2.022 mut=0.835 tok_s=32,801 avg_tok_s=32,260 |
| step=65300 tokens=667366000 ce=3.9981 ppl=54.5 bal=1.907 mut=0.841 tok_s=33,248 avg_tok_s=32,261 |
| step=65400 tokens=668388000 ce=4.1529 ppl=63.6 bal=1.930 mut=0.853 tok_s=33,242 avg_tok_s=32,263 |
| step=65500 tokens=669410000 ce=4.0996 ppl=60.3 bal=1.897 mut=0.851 tok_s=33,189 avg_tok_s=32,264 |
| step=65600 tokens=670432000 ce=4.1304 ppl=62.2 bal=1.953 mut=0.855 tok_s=33,083 avg_tok_s=32,265 |
| step=65700 tokens=671454000 ce=3.8798 ppl=48.4 bal=2.035 mut=0.850 tok_s=32,940 avg_tok_s=32,266 |
| step=65800 tokens=672476000 ce=4.0268 ppl=56.1 bal=2.260 mut=0.852 tok_s=33,116 avg_tok_s=32,268 |
| step=65900 tokens=673498000 ce=4.0969 ppl=60.2 bal=1.938 mut=0.852 tok_s=33,184 avg_tok_s=32,269 |
| step=66000 tokens=674520000 ce=4.0761 ppl=58.9 bal=1.911 mut=0.841 tok_s=32,936 avg_tok_s=32,270 |
| CKPT step=66000 valid_ce=4.3348 valid_ppl=76.3 |
| step=66100 tokens=675542000 ce=4.0506 ppl=57.4 bal=1.946 mut=0.856 tok_s=29,144 avg_tok_s=32,265 |
| step=66200 tokens=676564000 ce=4.3399 ppl=76.7 bal=2.127 mut=0.848 tok_s=33,219 avg_tok_s=32,267 |
| step=66300 tokens=677586000 ce=4.2889 ppl=72.9 bal=2.072 mut=0.848 tok_s=33,172 avg_tok_s=32,268 |
| step=66400 tokens=678608000 ce=3.9349 ppl=51.2 bal=2.067 mut=0.847 tok_s=33,143 avg_tok_s=32,269 |
| step=66500 tokens=679630000 ce=4.2205 ppl=68.1 bal=1.988 mut=0.846 tok_s=32,774 avg_tok_s=32,270 |
| step=66600 tokens=680652000 ce=4.0830 ppl=59.3 bal=1.971 mut=0.854 tok_s=33,095 avg_tok_s=32,271 |
| step=66700 tokens=681674000 ce=4.1315 ppl=62.3 bal=1.868 mut=0.848 tok_s=33,122 avg_tok_s=32,273 |
| step=66800 tokens=682696000 ce=4.0615 ppl=58.1 bal=2.001 mut=0.828 tok_s=33,069 avg_tok_s=32,274 |
| step=66900 tokens=683718000 ce=4.2868 ppl=72.7 bal=2.071 mut=0.849 tok_s=33,104 avg_tok_s=32,275 |
| step=67000 tokens=684740000 ce=4.0494 ppl=57.4 bal=1.988 mut=0.844 tok_s=33,256 avg_tok_s=32,277 |
| step=67100 tokens=685762000 ce=4.2129 ppl=67.6 bal=2.122 mut=0.841 tok_s=33,172 avg_tok_s=32,278 |
| step=67200 tokens=686784000 ce=4.0153 ppl=55.4 bal=2.042 mut=0.850 tok_s=32,905 avg_tok_s=32,279 |
| step=67300 tokens=687806000 ce=3.9925 ppl=54.2 bal=2.057 mut=0.845 tok_s=33,192 avg_tok_s=32,280 |
| step=67400 tokens=688828000 ce=4.3521 ppl=77.6 bal=2.047 mut=0.856 tok_s=32,881 avg_tok_s=32,281 |
| step=67500 tokens=689850000 ce=3.8751 ppl=48.2 bal=2.068 mut=0.848 tok_s=33,192 avg_tok_s=32,282 |
| step=67600 tokens=690872000 ce=3.8203 ppl=45.6 bal=2.334 mut=0.851 tok_s=33,146 avg_tok_s=32,284 |
| step=67700 tokens=691894000 ce=4.0275 ppl=56.1 bal=1.995 mut=0.856 tok_s=33,232 avg_tok_s=32,285 |
| step=67800 tokens=692916000 ce=4.1427 ppl=63.0 bal=1.920 mut=0.847 tok_s=33,125 avg_tok_s=32,286 |
| step=67900 tokens=693938000 ce=4.0940 ppl=60.0 bal=1.888 mut=0.844 tok_s=33,034 avg_tok_s=32,287 |
| step=68000 tokens=694960000 ce=4.0955 ppl=60.1 bal=1.782 mut=0.848 tok_s=33,144 avg_tok_s=32,289 |
| CKPT step=68000 valid_ce=3.8693 valid_ppl=47.9 |
| step=68100 tokens=695982000 ce=4.0015 ppl=54.7 bal=1.969 mut=0.849 tok_s=29,079 avg_tok_s=32,284 |
| step=68200 tokens=697004000 ce=4.2160 ppl=67.8 bal=2.008 mut=0.849 tok_s=33,235 avg_tok_s=32,285 |
| step=68300 tokens=698026000 ce=4.0833 ppl=59.3 bal=2.063 mut=0.846 tok_s=33,169 avg_tok_s=32,287 |
| step=68400 tokens=699048000 ce=4.0886 ppl=59.7 bal=1.998 mut=0.839 tok_s=32,898 avg_tok_s=32,288 |
| step=68500 tokens=700070000 ce=4.0603 ppl=58.0 bal=2.021 mut=0.848 tok_s=33,340 avg_tok_s=32,289 |
| step=68600 tokens=701092000 ce=4.1011 ppl=60.4 bal=1.916 mut=0.848 tok_s=33,239 avg_tok_s=32,290 |
| step=68700 tokens=702114000 ce=4.1016 ppl=60.4 bal=1.879 mut=0.855 tok_s=32,933 avg_tok_s=32,291 |
| step=68800 tokens=703136000 ce=4.2154 ppl=67.7 bal=2.098 mut=0.848 tok_s=32,909 avg_tok_s=32,292 |
| step=68900 tokens=704158000 ce=3.9423 ppl=51.5 bal=1.951 mut=0.846 tok_s=32,853 avg_tok_s=32,293 |
| step=69000 tokens=705180000 ce=4.1669 ppl=64.5 bal=1.859 mut=0.847 tok_s=32,910 avg_tok_s=32,294 |
| step=69100 tokens=706202000 ce=4.0394 ppl=56.8 bal=2.087 mut=0.842 tok_s=33,037 avg_tok_s=32,295 |
| step=69200 tokens=707224000 ce=4.0899 ppl=59.7 bal=2.016 mut=0.850 tok_s=32,913 avg_tok_s=32,296 |
| step=69300 tokens=708246000 ce=4.0946 ppl=60.0 bal=2.224 mut=0.843 tok_s=33,095 avg_tok_s=32,297 |
| step=69400 tokens=709268000 ce=3.9184 ppl=50.3 bal=1.972 mut=0.853 tok_s=33,200 avg_tok_s=32,298 |
| step=69500 tokens=710290000 ce=3.9107 ppl=49.9 bal=1.925 mut=0.842 tok_s=32,989 avg_tok_s=32,299 |
| step=69600 tokens=711312000 ce=3.9907 ppl=54.1 bal=1.897 mut=0.843 tok_s=32,730 avg_tok_s=32,300 |
| step=69700 tokens=712334000 ce=4.0263 ppl=56.1 bal=2.090 mut=0.852 tok_s=33,117 avg_tok_s=32,301 |
| step=69800 tokens=713356000 ce=4.1018 ppl=60.4 bal=1.955 mut=0.855 tok_s=33,239 avg_tok_s=32,303 |
| step=69900 tokens=714378000 ce=3.8516 ppl=47.1 bal=1.968 mut=0.840 tok_s=32,976 avg_tok_s=32,304 |
| step=70000 tokens=715400000 ce=4.0243 ppl=55.9 bal=1.844 mut=0.846 tok_s=32,929 avg_tok_s=32,304 |
| CKPT step=70000 valid_ce=3.7608 valid_ppl=43.0 |
| step=70100 tokens=716422000 ce=3.9438 ppl=51.6 bal=1.903 mut=0.851 tok_s=29,239 avg_tok_s=32,300 |
| step=70200 tokens=717444000 ce=4.0795 ppl=59.1 bal=2.078 mut=0.856 tok_s=32,991 avg_tok_s=32,301 |
| step=70300 tokens=718466000 ce=4.0191 ppl=55.7 bal=1.839 mut=0.835 tok_s=33,050 avg_tok_s=32,302 |
| step=70400 tokens=719488000 ce=4.1423 ppl=62.9 bal=2.104 mut=0.857 tok_s=33,134 avg_tok_s=32,303 |
| step=70500 tokens=720510000 ce=3.9583 ppl=52.4 bal=1.912 mut=0.856 tok_s=33,273 avg_tok_s=32,305 |
| step=70600 tokens=721532000 ce=3.9876 ppl=53.9 bal=1.797 mut=0.848 tok_s=33,269 avg_tok_s=32,306 |
| step=70700 tokens=722554000 ce=3.7857 ppl=44.1 bal=2.290 mut=0.850 tok_s=32,869 avg_tok_s=32,307 |
| step=70800 tokens=723576000 ce=4.2131 ppl=67.6 bal=2.458 mut=0.849 tok_s=33,130 avg_tok_s=32,308 |
| step=70900 tokens=724598000 ce=4.0121 ppl=55.3 bal=1.872 mut=0.841 tok_s=33,199 avg_tok_s=32,309 |
| step=71000 tokens=725620000 ce=3.9124 ppl=50.0 bal=2.044 mut=0.841 tok_s=33,142 avg_tok_s=32,310 |
| step=71100 tokens=726642000 ce=3.9276 ppl=50.8 bal=2.154 mut=0.854 tok_s=32,980 avg_tok_s=32,311 |
| step=71200 tokens=727664000 ce=3.9363 ppl=51.2 bal=1.928 mut=0.854 tok_s=33,135 avg_tok_s=32,313 |
| step=71300 tokens=728686000 ce=4.0633 ppl=58.2 bal=1.869 mut=0.850 tok_s=33,157 avg_tok_s=32,314 |
| step=71400 tokens=729708000 ce=4.2277 ppl=68.6 bal=1.983 mut=0.845 tok_s=33,226 avg_tok_s=32,315 |
| step=71500 tokens=730730000 ce=3.9432 ppl=51.6 bal=2.097 mut=0.843 tok_s=33,192 avg_tok_s=32,316 |
| step=71600 tokens=731752000 ce=4.0334 ppl=56.5 bal=2.075 mut=0.852 tok_s=32,972 avg_tok_s=32,317 |
| step=71700 tokens=732774000 ce=3.9706 ppl=53.0 bal=2.043 mut=0.845 tok_s=33,193 avg_tok_s=32,318 |
| step=71800 tokens=733796000 ce=3.8582 ppl=47.4 bal=1.953 mut=0.861 tok_s=33,221 avg_tok_s=32,320 |
| step=71900 tokens=734818000 ce=4.0763 ppl=58.9 bal=2.031 mut=0.838 tok_s=32,816 avg_tok_s=32,320 |
| step=72000 tokens=735840000 ce=4.1454 ppl=63.1 bal=2.038 mut=0.854 tok_s=33,013 avg_tok_s=32,321 |
| CKPT step=72000 valid_ce=3.9559 valid_ppl=52.2 |
| step=72100 tokens=736862000 ce=4.2072 ppl=67.2 bal=2.129 mut=0.853 tok_s=29,206 avg_tok_s=32,317 |
| step=72200 tokens=737884000 ce=4.1768 ppl=65.2 bal=2.107 mut=0.851 tok_s=32,842 avg_tok_s=32,318 |
| step=72300 tokens=738906000 ce=3.9914 ppl=54.1 bal=1.882 mut=0.857 tok_s=33,158 avg_tok_s=32,319 |
| step=72400 tokens=739928000 ce=4.3111 ppl=74.5 bal=2.126 mut=0.857 tok_s=33,012 avg_tok_s=32,320 |
| step=72500 tokens=740950000 ce=4.0818 ppl=59.3 bal=2.054 mut=0.843 tok_s=32,703 avg_tok_s=32,320 |
| step=72600 tokens=741972000 ce=3.9222 ppl=50.5 bal=2.060 mut=0.852 tok_s=33,318 avg_tok_s=32,322 |
| step=72700 tokens=742994000 ce=3.9141 ppl=50.1 bal=1.910 mut=0.853 tok_s=32,974 avg_tok_s=32,323 |
| step=72800 tokens=744016000 ce=4.3281 ppl=75.8 bal=2.226 mut=0.850 tok_s=32,740 avg_tok_s=32,323 |
| step=72900 tokens=745038000 ce=4.0493 ppl=57.4 bal=1.925 mut=0.851 tok_s=33,042 avg_tok_s=32,324 |
| step=73000 tokens=746060000 ce=4.0445 ppl=57.1 bal=2.427 mut=0.845 tok_s=33,164 avg_tok_s=32,325 |
| step=73100 tokens=747082000 ce=4.0045 ppl=54.8 bal=1.958 mut=0.847 tok_s=32,786 avg_tok_s=32,326 |
| step=73200 tokens=748104000 ce=3.9010 ppl=49.5 bal=1.859 mut=0.848 tok_s=33,100 avg_tok_s=32,327 |
| step=73300 tokens=749126000 ce=4.2162 ppl=67.8 bal=2.198 mut=0.847 tok_s=32,800 avg_tok_s=32,328 |
| step=73400 tokens=750148000 ce=4.1422 ppl=62.9 bal=2.314 mut=0.850 tok_s=32,860 avg_tok_s=32,328 |
| step=73500 tokens=751170000 ce=3.9481 ppl=51.8 bal=2.517 mut=0.840 tok_s=33,029 avg_tok_s=32,329 |
| step=73600 tokens=752192000 ce=3.9950 ppl=54.3 bal=1.965 mut=0.846 tok_s=32,891 avg_tok_s=32,330 |
| step=73700 tokens=753214000 ce=4.0715 ppl=58.6 bal=1.972 mut=0.850 tok_s=33,246 avg_tok_s=32,331 |
| step=73800 tokens=754236000 ce=4.2792 ppl=72.2 bal=2.509 mut=0.857 tok_s=33,155 avg_tok_s=32,332 |
| step=73900 tokens=755258000 ce=3.9922 ppl=54.2 bal=2.010 mut=0.850 tok_s=33,028 avg_tok_s=32,333 |
| step=74000 tokens=756280000 ce=4.0450 ppl=57.1 bal=1.958 mut=0.844 tok_s=32,944 avg_tok_s=32,334 |
| CKPT step=74000 valid_ce=3.9024 valid_ppl=49.5 |
| step=74100 tokens=757302000 ce=4.2322 ppl=68.9 bal=2.157 mut=0.848 tok_s=29,126 avg_tok_s=32,330 |
| step=74200 tokens=758324000 ce=4.1929 ppl=66.2 bal=1.933 mut=0.850 tok_s=32,974 avg_tok_s=32,331 |
| step=74300 tokens=759346000 ce=3.8458 ppl=46.8 bal=2.025 mut=0.846 tok_s=33,031 avg_tok_s=32,332 |
| step=74400 tokens=760368000 ce=3.9855 ppl=53.8 bal=1.856 mut=0.852 tok_s=32,797 avg_tok_s=32,332 |
| step=74500 tokens=761390000 ce=4.0207 ppl=55.7 bal=2.069 mut=0.868 tok_s=33,224 avg_tok_s=32,333 |
| step=74600 tokens=762412000 ce=3.9860 ppl=53.8 bal=1.874 mut=0.845 tok_s=33,113 avg_tok_s=32,335 |
| step=74700 tokens=763434000 ce=4.1317 ppl=62.3 bal=2.068 mut=0.848 tok_s=33,183 avg_tok_s=32,336 |
| step=74800 tokens=764456000 ce=4.0238 ppl=55.9 bal=1.927 mut=0.845 tok_s=32,863 avg_tok_s=32,336 |
| step=74900 tokens=765478000 ce=3.7944 ppl=44.5 bal=2.114 mut=0.853 tok_s=33,165 avg_tok_s=32,337 |
| step=75000 tokens=766500000 ce=3.8678 ppl=47.8 bal=2.052 mut=0.851 tok_s=33,078 avg_tok_s=32,338 |
| step=75100 tokens=767522000 ce=4.1185 ppl=61.5 bal=1.988 mut=0.849 tok_s=32,844 avg_tok_s=32,339 |
| step=75200 tokens=768544000 ce=4.1579 ppl=63.9 bal=1.919 mut=0.851 tok_s=32,852 avg_tok_s=32,340 |
| step=75300 tokens=769566000 ce=3.9725 ppl=53.1 bal=1.910 mut=0.839 tok_s=33,239 avg_tok_s=32,341 |
| step=75400 tokens=770588000 ce=4.1026 ppl=60.5 bal=1.911 mut=0.858 tok_s=33,093 avg_tok_s=32,342 |
| step=75500 tokens=771610000 ce=4.0489 ppl=57.3 bal=1.902 mut=0.846 tok_s=33,043 avg_tok_s=32,343 |
| step=75600 tokens=772632000 ce=4.0623 ppl=58.1 bal=1.932 mut=0.843 tok_s=32,665 avg_tok_s=32,343 |
| step=75700 tokens=773654000 ce=4.0428 ppl=57.0 bal=2.339 mut=0.849 tok_s=33,156 avg_tok_s=32,344 |
| step=75800 tokens=774676000 ce=4.0987 ppl=60.3 bal=1.969 mut=0.854 tok_s=33,238 avg_tok_s=32,346 |
| step=75900 tokens=775698000 ce=3.9742 ppl=53.2 bal=1.943 mut=0.865 tok_s=33,229 avg_tok_s=32,347 |
| step=76000 tokens=776720000 ce=4.1530 ppl=63.6 bal=1.952 mut=0.858 tok_s=32,801 avg_tok_s=32,347 |
| CKPT step=76000 valid_ce=4.2758 valid_ppl=71.9 |
| step=76100 tokens=777742000 ce=3.8900 ppl=48.9 bal=1.974 mut=0.852 tok_s=28,937 avg_tok_s=32,343 |
| step=76200 tokens=778764000 ce=4.1485 ppl=63.3 bal=2.027 mut=0.849 tok_s=33,053 avg_tok_s=32,344 |
| step=76300 tokens=779786000 ce=4.1116 ppl=61.0 bal=1.844 mut=0.853 tok_s=32,897 avg_tok_s=32,345 |
| step=76400 tokens=780808000 ce=4.2452 ppl=69.8 bal=2.057 mut=0.854 tok_s=32,678 avg_tok_s=32,345 |
| step=76500 tokens=781830000 ce=4.0069 ppl=55.0 bal=1.972 mut=0.851 tok_s=33,168 avg_tok_s=32,346 |
| step=76600 tokens=782852000 ce=3.9394 ppl=51.4 bal=2.011 mut=0.852 tok_s=32,733 avg_tok_s=32,347 |
| step=76700 tokens=783874000 ce=4.2043 ppl=67.0 bal=1.944 mut=0.856 tok_s=33,166 avg_tok_s=32,348 |
| step=76800 tokens=784896000 ce=3.7871 ppl=44.1 bal=2.076 mut=0.855 tok_s=32,915 avg_tok_s=32,348 |
| step=76900 tokens=785918000 ce=4.2909 ppl=73.0 bal=2.197 mut=0.858 tok_s=33,157 avg_tok_s=32,349 |
| step=77000 tokens=786940000 ce=4.1753 ppl=65.1 bal=2.041 mut=0.854 tok_s=33,231 avg_tok_s=32,351 |
| step=77100 tokens=787962000 ce=3.9404 ppl=51.4 bal=2.045 mut=0.853 tok_s=33,204 avg_tok_s=32,352 |
| step=77200 tokens=788984000 ce=4.0228 ppl=55.9 bal=2.177 mut=0.852 tok_s=32,657 avg_tok_s=32,352 |
| step=77300 tokens=790006000 ce=3.9533 ppl=52.1 bal=3.766 mut=0.858 tok_s=32,939 avg_tok_s=32,353 |
| step=77400 tokens=791028000 ce=3.9939 ppl=54.3 bal=1.892 mut=0.854 tok_s=33,146 avg_tok_s=32,354 |
| step=77500 tokens=792050000 ce=4.1978 ppl=66.5 bal=1.958 mut=0.859 tok_s=33,330 avg_tok_s=32,355 |
| step=77600 tokens=793072000 ce=3.9591 ppl=52.4 bal=2.020 mut=0.859 tok_s=32,762 avg_tok_s=32,356 |
| step=77700 tokens=794094000 ce=4.2199 ppl=68.0 bal=1.940 mut=0.863 tok_s=33,328 avg_tok_s=32,357 |
| step=77800 tokens=795116000 ce=3.9229 ppl=50.5 bal=1.989 mut=0.867 tok_s=32,925 avg_tok_s=32,358 |
| step=77900 tokens=796138000 ce=3.9765 ppl=53.3 bal=2.122 mut=0.843 tok_s=33,129 avg_tok_s=32,359 |
| step=78000 tokens=797160000 ce=4.0672 ppl=58.4 bal=1.990 mut=0.852 tok_s=32,999 avg_tok_s=32,359 |
| CKPT step=78000 valid_ce=4.0184 valid_ppl=55.6 |
| step=78100 tokens=798182000 ce=4.1001 ppl=60.3 bal=2.587 mut=0.867 tok_s=28,937 avg_tok_s=32,355 |
| step=78200 tokens=799204000 ce=3.9219 ppl=50.5 bal=1.928 mut=0.860 tok_s=32,814 avg_tok_s=32,356 |
| step=78300 tokens=800226000 ce=3.7968 ppl=44.6 bal=2.035 mut=0.859 tok_s=33,034 avg_tok_s=32,357 |
| step=78400 tokens=801248000 ce=4.1041 ppl=60.6 bal=1.917 mut=0.867 tok_s=32,939 avg_tok_s=32,357 |
| step=78500 tokens=802270000 ce=4.3087 ppl=74.3 bal=2.135 mut=0.866 tok_s=33,096 avg_tok_s=32,358 |
| step=78600 tokens=803292000 ce=3.7246 ppl=41.5 bal=2.146 mut=0.869 tok_s=32,922 avg_tok_s=32,359 |
| step=78700 tokens=804314000 ce=4.3071 ppl=74.2 bal=2.347 mut=0.865 tok_s=32,933 avg_tok_s=32,360 |
| step=78800 tokens=805336000 ce=3.8922 ppl=49.0 bal=2.223 mut=0.864 tok_s=32,880 avg_tok_s=32,360 |
| step=78900 tokens=806358000 ce=3.9037 ppl=49.6 bal=1.978 mut=0.860 tok_s=33,091 avg_tok_s=32,361 |
| step=79000 tokens=807380000 ce=3.9756 ppl=53.3 bal=1.977 mut=0.868 tok_s=33,055 avg_tok_s=32,362 |
| step=79100 tokens=808402000 ce=4.1018 ppl=60.5 bal=2.020 mut=0.865 tok_s=33,301 avg_tok_s=32,363 |
| step=79200 tokens=809424000 ce=4.1009 ppl=60.4 bal=2.053 mut=0.856 tok_s=32,932 avg_tok_s=32,364 |
| step=79300 tokens=810446000 ce=3.9884 ppl=54.0 bal=1.905 mut=0.872 tok_s=33,261 avg_tok_s=32,365 |
| step=79400 tokens=811468000 ce=4.1271 ppl=62.0 bal=2.304 mut=0.863 tok_s=33,106 avg_tok_s=32,366 |
| step=79500 tokens=812490000 ce=3.9105 ppl=49.9 bal=2.052 mut=0.856 tok_s=32,917 avg_tok_s=32,367 |
| step=79600 tokens=813512000 ce=4.1563 ppl=63.8 bal=2.015 mut=0.864 tok_s=33,061 avg_tok_s=32,368 |
| step=79700 tokens=814534000 ce=3.8239 ppl=45.8 bal=1.938 mut=0.867 tok_s=33,239 avg_tok_s=32,369 |
| step=79800 tokens=815556000 ce=3.9972 ppl=54.4 bal=1.922 mut=0.870 tok_s=32,810 avg_tok_s=32,369 |
| step=79900 tokens=816578000 ce=3.8952 ppl=49.2 bal=2.025 mut=0.860 tok_s=33,230 avg_tok_s=32,370 |
| step=80000 tokens=817600000 ce=4.0549 ppl=57.7 bal=2.032 mut=0.864 tok_s=32,755 avg_tok_s=32,371 |
| CKPT step=80000 valid_ce=4.1229 valid_ppl=61.7 |
| step=80100 tokens=818622000 ce=3.9743 ppl=53.2 bal=1.957 mut=0.870 tok_s=13,802 avg_tok_s=32,348 |
| step=80200 tokens=819644000 ce=3.8745 ppl=48.2 bal=1.857 mut=0.868 tok_s=33,178 avg_tok_s=32,349 |
| step=80300 tokens=820666000 ce=3.9952 ppl=54.3 bal=1.985 mut=0.857 tok_s=33,189 avg_tok_s=32,350 |
| step=80400 tokens=821688000 ce=4.0830 ppl=59.3 bal=1.881 mut=0.861 tok_s=33,080 avg_tok_s=32,351 |
| step=80500 tokens=822710000 ce=4.1430 ppl=63.0 bal=1.924 mut=0.865 tok_s=32,953 avg_tok_s=32,351 |
| step=80600 tokens=823732000 ce=4.0977 ppl=60.2 bal=1.798 mut=0.868 tok_s=33,184 avg_tok_s=32,352 |
| step=80700 tokens=824754000 ce=4.1140 ppl=61.2 bal=2.049 mut=0.864 tok_s=33,169 avg_tok_s=32,353 |
| step=80800 tokens=825776000 ce=4.0799 ppl=59.1 bal=1.918 mut=0.868 tok_s=32,742 avg_tok_s=32,354 |
| step=80900 tokens=826798000 ce=3.9764 ppl=53.3 bal=2.049 mut=0.871 tok_s=33,148 avg_tok_s=32,355 |
| step=81000 tokens=827820000 ce=4.0138 ppl=55.4 bal=1.991 mut=0.873 tok_s=33,103 avg_tok_s=32,356 |
| step=81100 tokens=828842000 ce=3.8959 ppl=49.2 bal=1.931 mut=0.866 tok_s=33,273 avg_tok_s=32,357 |
| step=81200 tokens=829864000 ce=4.0036 ppl=54.8 bal=2.548 mut=0.866 tok_s=33,086 avg_tok_s=32,358 |
| step=81300 tokens=830886000 ce=4.1151 ppl=61.3 bal=2.005 mut=0.861 tok_s=32,920 avg_tok_s=32,359 |
| step=81400 tokens=831908000 ce=4.0296 ppl=56.2 bal=1.881 mut=0.860 tok_s=33,228 avg_tok_s=32,360 |
| step=81500 tokens=832930000 ce=4.0802 ppl=59.2 bal=1.840 mut=0.868 tok_s=33,063 avg_tok_s=32,360 |
| step=81600 tokens=833952000 ce=4.1239 ppl=61.8 bal=2.208 mut=0.866 tok_s=33,164 avg_tok_s=32,361 |
| step=81700 tokens=834974000 ce=4.1387 ppl=62.7 bal=2.057 mut=0.878 tok_s=32,627 avg_tok_s=32,362 |
| step=81800 tokens=835996000 ce=4.1023 ppl=60.5 bal=1.877 mut=0.855 tok_s=32,994 avg_tok_s=32,363 |
| step=81900 tokens=837018000 ce=4.3154 ppl=74.8 bal=2.214 mut=0.857 tok_s=33,227 avg_tok_s=32,364 |
| step=82000 tokens=838040000 ce=4.0697 ppl=58.5 bal=1.863 mut=0.864 tok_s=33,134 avg_tok_s=32,365 |
| CKPT step=82000 valid_ce=4.0729 valid_ppl=58.7 |
| step=82100 tokens=839062000 ce=4.1142 ppl=61.2 bal=1.868 mut=0.864 tok_s=29,040 avg_tok_s=32,361 |
| step=82200 tokens=840084000 ce=3.9750 ppl=53.3 bal=2.008 mut=0.868 tok_s=33,090 avg_tok_s=32,361 |
| step=82300 tokens=841106000 ce=4.0190 ppl=55.6 bal=1.961 mut=0.862 tok_s=32,947 avg_tok_s=32,362 |
| step=82400 tokens=842128000 ce=4.0823 ppl=59.3 bal=1.938 mut=0.865 tok_s=32,928 avg_tok_s=32,363 |
| step=82500 tokens=843150000 ce=3.9382 ppl=51.3 bal=1.846 mut=0.872 tok_s=33,222 avg_tok_s=32,364 |
| step=82600 tokens=844172000 ce=4.2597 ppl=70.8 bal=1.974 mut=0.862 tok_s=33,179 avg_tok_s=32,365 |
| step=82700 tokens=845194000 ce=3.7512 ppl=42.6 bal=1.922 mut=0.861 tok_s=33,042 avg_tok_s=32,366 |
| step=82800 tokens=846216000 ce=4.1278 ppl=62.0 bal=1.843 mut=0.864 tok_s=33,145 avg_tok_s=32,367 |
| step=82900 tokens=847238000 ce=4.1026 ppl=60.5 bal=2.060 mut=0.869 tok_s=33,195 avg_tok_s=32,368 |
| step=83000 tokens=848260000 ce=3.8682 ppl=47.9 bal=1.797 mut=0.872 tok_s=33,285 avg_tok_s=32,369 |
| step=83100 tokens=849282000 ce=3.9220 ppl=50.5 bal=2.349 mut=0.861 tok_s=32,930 avg_tok_s=32,369 |
| step=83200 tokens=850304000 ce=4.0606 ppl=58.0 bal=1.906 mut=0.862 tok_s=32,938 avg_tok_s=32,370 |
| step=83300 tokens=851326000 ce=3.9595 ppl=52.4 bal=1.853 mut=0.858 tok_s=33,057 avg_tok_s=32,371 |
| step=83400 tokens=852348000 ce=3.9557 ppl=52.2 bal=1.967 mut=0.858 tok_s=33,161 avg_tok_s=32,372 |
| step=83500 tokens=853370000 ce=4.2149 ppl=67.7 bal=2.048 mut=0.870 tok_s=33,024 avg_tok_s=32,373 |
| step=83600 tokens=854392000 ce=3.8876 ppl=48.8 bal=1.927 mut=0.861 tok_s=32,781 avg_tok_s=32,373 |
| step=83700 tokens=855414000 ce=3.9510 ppl=52.0 bal=1.972 mut=0.859 tok_s=33,159 avg_tok_s=32,374 |
| step=83800 tokens=856436000 ce=4.0741 ppl=58.8 bal=2.245 mut=0.853 tok_s=33,293 avg_tok_s=32,375 |
| step=83900 tokens=857458000 ce=3.9107 ppl=49.9 bal=2.016 mut=0.866 tok_s=33,248 avg_tok_s=32,376 |
| step=84000 tokens=858480000 ce=3.9459 ppl=51.7 bal=1.914 mut=0.867 tok_s=32,875 avg_tok_s=32,377 |
| CKPT step=84000 valid_ce=3.9896 valid_ppl=54.0 |
| step=84100 tokens=859502000 ce=4.0683 ppl=58.5 bal=1.951 mut=0.867 tok_s=29,316 avg_tok_s=32,373 |
| step=84200 tokens=860524000 ce=3.9901 ppl=54.1 bal=1.868 mut=0.874 tok_s=33,193 avg_tok_s=32,374 |
| step=84300 tokens=861546000 ce=3.8284 ppl=46.0 bal=2.269 mut=0.870 tok_s=33,197 avg_tok_s=32,375 |
| step=84400 tokens=862568000 ce=3.7341 ppl=41.8 bal=1.984 mut=0.870 tok_s=32,528 avg_tok_s=32,375 |
| step=84500 tokens=863590000 ce=3.7725 ppl=43.5 bal=1.891 mut=0.862 tok_s=33,009 avg_tok_s=32,376 |
| step=84600 tokens=864612000 ce=3.9101 ppl=49.9 bal=1.889 mut=0.864 tok_s=33,042 avg_tok_s=32,377 |
| step=84700 tokens=865634000 ce=4.1007 ppl=60.4 bal=2.037 mut=0.868 tok_s=33,205 avg_tok_s=32,378 |
| step=84800 tokens=866656000 ce=3.7813 ppl=43.9 bal=1.950 mut=0.867 tok_s=33,018 avg_tok_s=32,379 |
| step=84900 tokens=867678000 ce=3.7791 ppl=43.8 bal=1.829 mut=0.869 tok_s=33,079 avg_tok_s=32,379 |
| step=85000 tokens=868700000 ce=4.0019 ppl=54.7 bal=1.944 mut=0.869 tok_s=33,177 avg_tok_s=32,380 |
| step=85100 tokens=869722000 ce=4.2371 ppl=69.2 bal=1.987 mut=0.869 tok_s=33,226 avg_tok_s=32,381 |
| step=85200 tokens=870744000 ce=3.9322 ppl=51.0 bal=1.830 mut=0.861 tok_s=32,875 avg_tok_s=32,382 |
| step=85300 tokens=871766000 ce=3.8292 ppl=46.0 bal=2.027 mut=0.865 tok_s=32,828 avg_tok_s=32,382 |
| step=85400 tokens=872788000 ce=3.9431 ppl=51.6 bal=1.895 mut=0.864 tok_s=33,105 avg_tok_s=32,383 |
| step=85500 tokens=873810000 ce=4.1992 ppl=66.6 bal=2.076 mut=0.872 tok_s=33,044 avg_tok_s=32,384 |
| step=85600 tokens=874832000 ce=3.9895 ppl=54.0 bal=1.798 mut=0.870 tok_s=32,958 avg_tok_s=32,385 |
| step=85700 tokens=875854000 ce=4.0099 ppl=55.1 bal=1.830 mut=0.858 tok_s=33,161 avg_tok_s=32,386 |
| step=85800 tokens=876876000 ce=3.9697 ppl=53.0 bal=2.083 mut=0.862 tok_s=32,965 avg_tok_s=32,386 |
| step=85900 tokens=877898000 ce=4.2270 ppl=68.5 bal=2.156 mut=0.867 tok_s=33,358 avg_tok_s=32,387 |
| step=86000 tokens=878920000 ce=3.7736 ppl=43.5 bal=1.904 mut=0.869 tok_s=32,964 avg_tok_s=32,388 |
| CKPT step=86000 valid_ce=4.0768 valid_ppl=59.0 |
| step=86100 tokens=879942000 ce=3.9091 ppl=49.9 bal=2.551 mut=0.854 tok_s=29,211 avg_tok_s=32,384 |
| step=86200 tokens=880964000 ce=4.0043 ppl=54.8 bal=1.885 mut=0.864 tok_s=32,961 avg_tok_s=32,385 |
| step=86300 tokens=881986000 ce=4.0468 ppl=57.2 bal=2.009 mut=0.864 tok_s=33,059 avg_tok_s=32,386 |
| step=86400 tokens=883008000 ce=3.9494 ppl=51.9 bal=1.963 mut=0.857 tok_s=32,925 avg_tok_s=32,386 |
| step=86500 tokens=884030000 ce=3.9928 ppl=54.2 bal=2.053 mut=0.862 tok_s=33,104 avg_tok_s=32,387 |
| step=86600 tokens=885052000 ce=4.1281 ppl=62.1 bal=1.997 mut=0.860 tok_s=33,028 avg_tok_s=32,388 |
| step=86700 tokens=886074000 ce=4.0926 ppl=59.9 bal=1.937 mut=0.874 tok_s=33,130 avg_tok_s=32,389 |
| step=86800 tokens=887096000 ce=4.1887 ppl=65.9 bal=2.227 mut=0.860 tok_s=32,873 avg_tok_s=32,389 |
| step=86900 tokens=888118000 ce=4.0558 ppl=57.7 bal=1.893 mut=0.867 tok_s=33,179 avg_tok_s=32,390 |
| step=87000 tokens=889140000 ce=3.8953 ppl=49.2 bal=1.855 mut=0.871 tok_s=33,138 avg_tok_s=32,391 |
| step=87100 tokens=890162000 ce=3.9295 ppl=50.9 bal=1.897 mut=0.861 tok_s=32,999 avg_tok_s=32,392 |
| step=87200 tokens=891184000 ce=4.0470 ppl=57.2 bal=1.952 mut=0.871 tok_s=32,984 avg_tok_s=32,393 |
| step=87300 tokens=892206000 ce=3.9508 ppl=52.0 bal=1.934 mut=0.864 tok_s=33,206 avg_tok_s=32,393 |
| step=87400 tokens=893228000 ce=4.1311 ppl=62.2 bal=2.107 mut=0.865 tok_s=33,245 avg_tok_s=32,394 |
| step=87500 tokens=894250000 ce=3.8858 ppl=48.7 bal=1.868 mut=0.873 tok_s=32,985 avg_tok_s=32,395 |
| step=87600 tokens=895272000 ce=4.0413 ppl=56.9 bal=2.090 mut=0.875 tok_s=32,883 avg_tok_s=32,396 |
| step=87700 tokens=896294000 ce=4.0144 ppl=55.4 bal=1.810 mut=0.873 tok_s=33,194 avg_tok_s=32,397 |
| step=87800 tokens=897316000 ce=4.0628 ppl=58.1 bal=1.845 mut=0.861 tok_s=33,175 avg_tok_s=32,397 |
| step=87900 tokens=898338000 ce=4.1323 ppl=62.3 bal=1.928 mut=0.869 tok_s=33,319 avg_tok_s=32,399 |
| step=88000 tokens=899360000 ce=4.1815 ppl=65.5 bal=1.844 mut=0.864 tok_s=32,791 avg_tok_s=32,399 |
| CKPT step=88000 valid_ce=3.8730 valid_ppl=48.1 |
| step=88100 tokens=900382000 ce=4.0529 ppl=57.6 bal=1.964 mut=0.862 tok_s=29,279 avg_tok_s=32,395 |
| step=88200 tokens=901404000 ce=4.0646 ppl=58.2 bal=2.038 mut=0.869 tok_s=33,047 avg_tok_s=32,396 |
| step=88300 tokens=902426000 ce=4.1662 ppl=64.5 bal=1.779 mut=0.862 tok_s=32,926 avg_tok_s=32,397 |
| step=88400 tokens=903448000 ce=3.9768 ppl=53.3 bal=1.932 mut=0.860 tok_s=32,919 avg_tok_s=32,397 |
| step=88500 tokens=904470000 ce=4.0623 ppl=58.1 bal=1.886 mut=0.860 tok_s=32,886 avg_tok_s=32,398 |
| step=88600 tokens=905492000 ce=4.0506 ppl=57.4 bal=1.944 mut=0.860 tok_s=33,237 avg_tok_s=32,399 |
| step=88700 tokens=906514000 ce=3.8012 ppl=44.8 bal=1.990 mut=0.870 tok_s=33,143 avg_tok_s=32,400 |
| step=88800 tokens=907536000 ce=3.9529 ppl=52.1 bal=2.033 mut=0.861 tok_s=33,024 avg_tok_s=32,400 |
| step=88900 tokens=908558000 ce=3.8955 ppl=49.2 bal=2.357 mut=0.869 tok_s=32,842 avg_tok_s=32,401 |
| step=89000 tokens=909580000 ce=4.3740 ppl=79.4 bal=2.130 mut=0.869 tok_s=33,146 avg_tok_s=32,402 |
| step=89100 tokens=910602000 ce=4.0368 ppl=56.6 bal=1.868 mut=0.866 tok_s=33,129 avg_tok_s=32,403 |
| step=89200 tokens=911624000 ce=4.1170 ppl=61.4 bal=1.929 mut=0.867 tok_s=33,178 avg_tok_s=32,403 |
| step=89300 tokens=912646000 ce=4.0747 ppl=58.8 bal=1.943 mut=0.866 tok_s=33,096 avg_tok_s=32,404 |
| step=89400 tokens=913668000 ce=3.8590 ppl=47.4 bal=2.785 mut=0.855 tok_s=33,320 avg_tok_s=32,405 |
| step=89500 tokens=914690000 ce=3.8343 ppl=46.3 bal=1.891 mut=0.865 tok_s=33,266 avg_tok_s=32,406 |
| step=89600 tokens=915712000 ce=3.8940 ppl=49.1 bal=1.821 mut=0.871 tok_s=33,147 avg_tok_s=32,407 |
| step=89700 tokens=916734000 ce=3.7279 ppl=41.6 bal=1.974 mut=0.869 tok_s=32,826 avg_tok_s=32,407 |
| step=89800 tokens=917756000 ce=4.0942 ppl=60.0 bal=2.077 mut=0.873 tok_s=33,016 avg_tok_s=32,408 |
| step=89900 tokens=918778000 ce=4.1089 ppl=60.9 bal=1.999 mut=0.864 tok_s=33,226 avg_tok_s=32,409 |
| step=90000 tokens=919800000 ce=3.7397 ppl=42.1 bal=1.862 mut=0.863 tok_s=33,059 avg_tok_s=32,410 |
| CKPT step=90000 valid_ce=3.9522 valid_ppl=52.0 |
| step=90100 tokens=920822000 ce=4.1422 ppl=62.9 bal=1.951 mut=0.869 tok_s=29,417 avg_tok_s=32,406 |
| step=90200 tokens=921844000 ce=4.0133 ppl=55.3 bal=1.786 mut=0.865 tok_s=33,274 avg_tok_s=32,407 |
| step=90300 tokens=922866000 ce=3.9173 ppl=50.3 bal=1.879 mut=0.874 tok_s=33,285 avg_tok_s=32,408 |
| step=90400 tokens=923888000 ce=3.7747 ppl=43.6 bal=1.910 mut=0.868 tok_s=32,809 avg_tok_s=32,409 |
| step=90500 tokens=924910000 ce=3.8372 ppl=46.4 bal=1.957 mut=0.873 tok_s=33,209 avg_tok_s=32,410 |
| step=90600 tokens=925932000 ce=4.0624 ppl=58.1 bal=1.946 mut=0.875 tok_s=33,156 avg_tok_s=32,411 |
| step=90700 tokens=926954000 ce=4.1316 ppl=62.3 bal=1.920 mut=0.858 tok_s=33,114 avg_tok_s=32,411 |
| step=90800 tokens=927976000 ce=4.1130 ppl=61.1 bal=1.939 mut=0.870 tok_s=33,177 avg_tok_s=32,412 |
| step=90900 tokens=928998000 ce=3.9531 ppl=52.1 bal=1.995 mut=0.865 tok_s=32,970 avg_tok_s=32,413 |
| step=91000 tokens=930020000 ce=3.9799 ppl=53.5 bal=1.918 mut=0.867 tok_s=33,196 avg_tok_s=32,414 |
| step=91100 tokens=931042000 ce=4.0303 ppl=56.3 bal=1.821 mut=0.869 tok_s=32,964 avg_tok_s=32,414 |
| step=91200 tokens=932064000 ce=4.2110 ppl=67.4 bal=2.037 mut=0.864 tok_s=33,014 avg_tok_s=32,415 |
| step=91300 tokens=933086000 ce=4.0301 ppl=56.3 bal=2.094 mut=0.878 tok_s=32,978 avg_tok_s=32,416 |
| step=91400 tokens=934108000 ce=3.9621 ppl=52.6 bal=1.994 mut=0.866 tok_s=33,152 avg_tok_s=32,416 |
| step=91500 tokens=935130000 ce=4.0768 ppl=59.0 bal=1.761 mut=0.864 tok_s=33,243 avg_tok_s=32,417 |
| step=91600 tokens=936152000 ce=3.9488 ppl=51.9 bal=2.031 mut=0.859 tok_s=33,231 avg_tok_s=32,418 |
| step=91700 tokens=937174000 ce=4.2673 ppl=71.3 bal=1.991 mut=0.872 tok_s=33,233 avg_tok_s=32,419 |
| step=91800 tokens=938196000 ce=3.9811 ppl=53.6 bal=1.796 mut=0.881 tok_s=33,233 avg_tok_s=32,420 |
| step=91900 tokens=939218000 ce=4.2675 ppl=71.3 bal=1.885 mut=0.862 tok_s=32,983 avg_tok_s=32,421 |
| step=92000 tokens=940240000 ce=4.0542 ppl=57.6 bal=1.925 mut=0.864 tok_s=32,847 avg_tok_s=32,421 |
| CKPT step=92000 valid_ce=3.8600 valid_ppl=47.5 |
| step=92100 tokens=941262000 ce=3.9687 ppl=52.9 bal=1.963 mut=0.868 tok_s=28,980 avg_tok_s=32,417 |
| step=92200 tokens=942284000 ce=3.8592 ppl=47.4 bal=1.837 mut=0.864 tok_s=33,084 avg_tok_s=32,418 |
| step=92300 tokens=943306000 ce=4.1690 ppl=64.7 bal=1.855 mut=0.868 tok_s=33,256 avg_tok_s=32,419 |
| step=92400 tokens=944328000 ce=4.1114 ppl=61.0 bal=1.799 mut=0.861 tok_s=32,979 avg_tok_s=32,419 |
| step=92500 tokens=945350000 ce=4.0831 ppl=59.3 bal=2.023 mut=0.877 tok_s=33,158 avg_tok_s=32,420 |
| step=92600 tokens=946372000 ce=3.9877 ppl=53.9 bal=1.914 mut=0.862 tok_s=33,210 avg_tok_s=32,421 |
| step=92700 tokens=947394000 ce=3.6196 ppl=37.3 bal=1.977 mut=0.874 tok_s=33,240 avg_tok_s=32,422 |
| step=92800 tokens=948416000 ce=3.9875 ppl=53.9 bal=2.007 mut=0.862 tok_s=33,170 avg_tok_s=32,423 |
| step=92900 tokens=949438000 ce=3.9993 ppl=54.6 bal=1.909 mut=0.870 tok_s=33,170 avg_tok_s=32,424 |
| step=93000 tokens=950460000 ce=3.9246 ppl=50.6 bal=1.928 mut=0.875 tok_s=32,918 avg_tok_s=32,424 |
| step=93100 tokens=951482000 ce=4.0555 ppl=57.7 bal=1.870 mut=0.874 tok_s=33,117 avg_tok_s=32,425 |
| step=93200 tokens=952504000 ce=3.8500 ppl=47.0 bal=1.788 mut=0.879 tok_s=32,963 avg_tok_s=32,425 |
| step=93300 tokens=953526000 ce=4.0657 ppl=58.3 bal=1.882 mut=0.860 tok_s=33,038 avg_tok_s=32,426 |
| step=93400 tokens=954548000 ce=4.1758 ppl=65.1 bal=1.929 mut=0.868 tok_s=33,172 avg_tok_s=32,427 |
| step=93500 tokens=955570000 ce=4.0867 ppl=59.5 bal=2.038 mut=0.865 tok_s=32,893 avg_tok_s=32,427 |
| step=93600 tokens=956592000 ce=4.1395 ppl=62.8 bal=1.747 mut=0.860 tok_s=32,963 avg_tok_s=32,428 |
| step=93700 tokens=957614000 ce=3.9522 ppl=52.1 bal=2.105 mut=0.875 tok_s=33,018 avg_tok_s=32,429 |
| step=93800 tokens=958636000 ce=3.8559 ppl=47.3 bal=1.775 mut=0.885 tok_s=32,359 avg_tok_s=32,429 |
| step=93900 tokens=959658000 ce=4.1884 ppl=65.9 bal=2.042 mut=0.870 tok_s=33,158 avg_tok_s=32,429 |
| step=94000 tokens=960680000 ce=3.9347 ppl=51.1 bal=1.873 mut=0.873 tok_s=33,109 avg_tok_s=32,430 |
| CKPT step=94000 valid_ce=4.1560 valid_ppl=63.8 |
| step=94100 tokens=961702000 ce=3.8829 ppl=48.6 bal=2.110 mut=0.870 tok_s=29,252 avg_tok_s=32,427 |
| step=94200 tokens=962724000 ce=4.1494 ppl=63.4 bal=2.065 mut=0.878 tok_s=33,234 avg_tok_s=32,428 |
| step=94300 tokens=963746000 ce=3.8899 ppl=48.9 bal=1.912 mut=0.870 tok_s=33,020 avg_tok_s=32,428 |
| step=94400 tokens=964768000 ce=4.3161 ppl=74.9 bal=2.186 mut=0.874 tok_s=32,596 avg_tok_s=32,428 |
| step=94500 tokens=965790000 ce=4.0824 ppl=59.3 bal=1.996 mut=0.862 tok_s=33,056 avg_tok_s=32,429 |
| step=94600 tokens=966812000 ce=3.9592 ppl=52.4 bal=2.071 mut=0.880 tok_s=33,129 avg_tok_s=32,430 |
| step=94700 tokens=967834000 ce=4.0383 ppl=56.7 bal=2.112 mut=0.877 tok_s=33,089 avg_tok_s=32,430 |
| step=94800 tokens=968856000 ce=4.0956 ppl=60.1 bal=1.794 mut=0.867 tok_s=33,141 avg_tok_s=32,431 |
| step=94900 tokens=969878000 ce=3.8945 ppl=49.1 bal=2.146 mut=0.877 tok_s=32,966 avg_tok_s=32,432 |
| step=95000 tokens=970900000 ce=3.9719 ppl=53.1 bal=1.970 mut=0.867 tok_s=33,372 avg_tok_s=32,433 |
| step=95100 tokens=971922000 ce=3.7718 ppl=43.5 bal=1.920 mut=0.875 tok_s=33,258 avg_tok_s=32,434 |
| step=95200 tokens=972944000 ce=4.0104 ppl=55.2 bal=2.068 mut=0.871 tok_s=32,696 avg_tok_s=32,434 |
| step=95300 tokens=973966000 ce=4.0785 ppl=59.1 bal=1.868 mut=0.858 tok_s=33,122 avg_tok_s=32,435 |
| step=95400 tokens=974988000 ce=4.0576 ppl=57.8 bal=1.923 mut=0.871 tok_s=33,128 avg_tok_s=32,435 |
| step=95500 tokens=976010000 ce=4.0705 ppl=58.6 bal=2.012 mut=0.876 tok_s=33,262 avg_tok_s=32,436 |
| step=95600 tokens=977032000 ce=3.8472 ppl=46.9 bal=1.955 mut=0.868 tok_s=33,085 avg_tok_s=32,437 |
| step=95700 tokens=978054000 ce=3.9457 ppl=51.7 bal=1.836 mut=0.884 tok_s=32,719 avg_tok_s=32,437 |
| step=95800 tokens=979076000 ce=3.9018 ppl=49.5 bal=1.896 mut=0.864 tok_s=32,981 avg_tok_s=32,438 |
| step=95900 tokens=980098000 ce=4.0326 ppl=56.4 bal=1.901 mut=0.868 tok_s=33,031 avg_tok_s=32,438 |
| step=96000 tokens=981120000 ce=4.1878 ppl=65.9 bal=1.999 mut=0.879 tok_s=33,057 avg_tok_s=32,439 |
| CKPT step=96000 valid_ce=3.8711 valid_ppl=48.0 |
| step=96100 tokens=982142000 ce=4.1223 ppl=61.7 bal=1.925 mut=0.877 tok_s=29,176 avg_tok_s=32,436 |
| step=96200 tokens=983164000 ce=4.0037 ppl=54.8 bal=1.946 mut=0.868 tok_s=33,116 avg_tok_s=32,436 |
| step=96300 tokens=984186000 ce=3.9213 ppl=50.5 bal=1.958 mut=0.876 tok_s=33,146 avg_tok_s=32,437 |
| step=96400 tokens=985208000 ce=3.9273 ppl=50.8 bal=2.044 mut=0.871 tok_s=32,891 avg_tok_s=32,438 |
| step=96500 tokens=986230000 ce=4.1970 ppl=66.5 bal=1.994 mut=0.863 tok_s=32,658 avg_tok_s=32,438 |
| step=96600 tokens=987252000 ce=3.8370 ppl=46.4 bal=1.937 mut=0.876 tok_s=33,089 avg_tok_s=32,438 |
| step=96700 tokens=988274000 ce=3.9643 ppl=52.7 bal=1.892 mut=0.876 tok_s=33,213 avg_tok_s=32,439 |
| step=96800 tokens=989296000 ce=4.1096 ppl=60.9 bal=1.916 mut=0.863 tok_s=33,209 avg_tok_s=32,440 |
| step=96900 tokens=990318000 ce=4.0336 ppl=56.5 bal=1.949 mut=0.880 tok_s=33,023 avg_tok_s=32,441 |
| step=97000 tokens=991340000 ce=4.2064 ppl=67.1 bal=2.245 mut=0.875 tok_s=33,184 avg_tok_s=32,441 |
| step=97100 tokens=992362000 ce=4.1727 ppl=64.9 bal=1.982 mut=0.875 tok_s=33,056 avg_tok_s=32,442 |
| step=97200 tokens=993384000 ce=4.1644 ppl=64.4 bal=1.907 mut=0.870 tok_s=33,046 avg_tok_s=32,443 |
| step=97300 tokens=994406000 ce=4.1114 ppl=61.0 bal=1.852 mut=0.872 tok_s=32,438 avg_tok_s=32,443 |
| step=97400 tokens=995428000 ce=4.0883 ppl=59.6 bal=1.958 mut=0.875 tok_s=32,948 avg_tok_s=32,443 |
| step=97500 tokens=996450000 ce=4.0506 ppl=57.4 bal=1.829 mut=0.878 tok_s=32,997 avg_tok_s=32,444 |
| step=97600 tokens=997472000 ce=3.9158 ppl=50.2 bal=1.912 mut=0.871 tok_s=32,906 avg_tok_s=32,444 |
| step=97700 tokens=998494000 ce=4.2264 ppl=68.5 bal=1.995 mut=0.872 tok_s=32,761 avg_tok_s=32,445 |
| step=97800 tokens=999516000 ce=4.0929 ppl=59.9 bal=1.994 mut=0.881 tok_s=33,121 avg_tok_s=32,445 |
| step=97900 tokens=1000538000 ce=3.9752 ppl=53.3 bal=2.055 mut=0.875 tok_s=32,725 avg_tok_s=32,446 |
| step=98000 tokens=1001560000 ce=3.9518 ppl=52.0 bal=1.830 mut=0.868 tok_s=32,901 avg_tok_s=32,446 |
| CKPT step=98000 valid_ce=3.7395 valid_ppl=42.1 |
| step=98100 tokens=1002582000 ce=4.1378 ppl=62.7 bal=1.981 mut=0.873 tok_s=29,183 avg_tok_s=32,443 |
| step=98200 tokens=1003604000 ce=4.0957 ppl=60.1 bal=2.024 mut=0.870 tok_s=33,118 avg_tok_s=32,443 |
| step=98300 tokens=1004626000 ce=4.0458 ppl=57.2 bal=1.967 mut=0.878 tok_s=32,801 avg_tok_s=32,444 |
| step=98400 tokens=1005648000 ce=3.7775 ppl=43.7 bal=1.861 mut=0.871 tok_s=32,899 avg_tok_s=32,444 |
| step=98500 tokens=1006670000 ce=4.1124 ppl=61.1 bal=2.324 mut=0.881 tok_s=32,926 avg_tok_s=32,445 |
| step=98600 tokens=1007692000 ce=4.0864 ppl=59.5 bal=1.788 mut=0.878 tok_s=33,147 avg_tok_s=32,445 |
| step=98700 tokens=1008714000 ce=3.8764 ppl=48.2 bal=2.135 mut=0.884 tok_s=33,138 avg_tok_s=32,446 |
| step=98800 tokens=1009736000 ce=4.2768 ppl=72.0 bal=2.104 mut=0.872 tok_s=32,858 avg_tok_s=32,446 |
| step=98900 tokens=1010758000 ce=3.8627 ppl=47.6 bal=1.789 mut=0.881 tok_s=32,914 avg_tok_s=32,447 |
| step=99000 tokens=1011780000 ce=4.2409 ppl=69.5 bal=1.899 mut=0.880 tok_s=33,237 avg_tok_s=32,448 |
| step=99100 tokens=1012802000 ce=4.0968 ppl=60.1 bal=1.826 mut=0.875 tok_s=32,875 avg_tok_s=32,448 |
| step=99200 tokens=1013824000 ce=4.1374 ppl=62.6 bal=1.834 mut=0.875 tok_s=32,893 avg_tok_s=32,449 |
| step=99300 tokens=1014846000 ce=4.1125 ppl=61.1 bal=2.077 mut=0.873 tok_s=32,493 avg_tok_s=32,449 |
| step=99400 tokens=1015868000 ce=3.9931 ppl=54.2 bal=1.917 mut=0.876 tok_s=33,150 avg_tok_s=32,449 |
| step=99500 tokens=1016890000 ce=3.9867 ppl=53.9 bal=1.922 mut=0.882 tok_s=32,682 avg_tok_s=32,450 |
| step=99600 tokens=1017912000 ce=3.7425 ppl=42.2 bal=1.931 mut=0.876 tok_s=33,090 avg_tok_s=32,450 |
| step=99700 tokens=1018934000 ce=3.9011 ppl=49.5 bal=1.923 mut=0.881 tok_s=33,240 avg_tok_s=32,451 |
| step=99800 tokens=1019956000 ce=3.9390 ppl=51.4 bal=1.886 mut=0.871 tok_s=33,030 avg_tok_s=32,452 |
| step=99900 tokens=1020978000 ce=4.0029 ppl=54.8 bal=2.011 mut=0.878 tok_s=33,151 avg_tok_s=32,452 |
| step=100000 tokens=1022000000 ce=4.0667 ppl=58.4 bal=2.085 mut=0.881 tok_s=32,942 avg_tok_s=32,453 |
| CKPT step=100000 valid_ce=4.0396 valid_ppl=56.8 |
| step=100100 tokens=1023022000 ce=3.8943 ppl=49.1 bal=1.954 mut=0.871 tok_s=13,103 avg_tok_s=32,434 |
| step=100200 tokens=1024044000 ce=3.9647 ppl=52.7 bal=1.891 mut=0.884 tok_s=33,244 avg_tok_s=32,434 |
| step=100300 tokens=1025066000 ce=3.9579 ppl=52.3 bal=1.920 mut=0.878 tok_s=33,124 avg_tok_s=32,435 |
| step=100400 tokens=1026088000 ce=4.1104 ppl=61.0 bal=2.004 mut=0.881 tok_s=33,073 avg_tok_s=32,436 |
| step=100500 tokens=1027110000 ce=4.1311 ppl=62.2 bal=1.963 mut=0.871 tok_s=32,957 avg_tok_s=32,436 |
| step=100600 tokens=1028132000 ce=4.0693 ppl=58.5 bal=1.866 mut=0.880 tok_s=33,112 avg_tok_s=32,437 |
| step=100700 tokens=1029154000 ce=3.8488 ppl=46.9 bal=1.866 mut=0.880 tok_s=33,253 avg_tok_s=32,438 |
| step=100800 tokens=1030176000 ce=4.2359 ppl=69.1 bal=1.927 mut=0.868 tok_s=33,000 avg_tok_s=32,438 |
| step=100900 tokens=1031198000 ce=3.8605 ppl=47.5 bal=1.836 mut=0.878 tok_s=32,577 avg_tok_s=32,438 |
| step=101000 tokens=1032220000 ce=3.9369 ppl=51.3 bal=1.886 mut=0.872 tok_s=33,085 avg_tok_s=32,439 |
| step=101100 tokens=1033242000 ce=4.0127 ppl=55.3 bal=2.223 mut=0.886 tok_s=32,785 avg_tok_s=32,439 |
| step=101200 tokens=1034264000 ce=3.9720 ppl=53.1 bal=1.869 mut=0.868 tok_s=32,995 avg_tok_s=32,440 |
| step=101300 tokens=1035286000 ce=3.9558 ppl=52.2 bal=1.823 mut=0.865 tok_s=33,216 avg_tok_s=32,441 |
| step=101400 tokens=1036308000 ce=4.0311 ppl=56.3 bal=1.812 mut=0.873 tok_s=32,935 avg_tok_s=32,441 |
| step=101500 tokens=1037330000 ce=3.9845 ppl=53.8 bal=1.911 mut=0.874 tok_s=32,767 avg_tok_s=32,441 |
| step=101600 tokens=1038352000 ce=3.9034 ppl=49.6 bal=1.794 mut=0.873 tok_s=32,888 avg_tok_s=32,442 |
| step=101700 tokens=1039374000 ce=3.8296 ppl=46.0 bal=1.914 mut=0.870 tok_s=33,037 avg_tok_s=32,442 |
| step=101800 tokens=1040396000 ce=3.9383 ppl=51.3 bal=1.845 mut=0.865 tok_s=32,680 avg_tok_s=32,443 |
| step=101900 tokens=1041418000 ce=4.0379 ppl=56.7 bal=1.909 mut=0.884 tok_s=32,982 avg_tok_s=32,443 |
| step=102000 tokens=1042440000 ce=4.1722 ppl=64.9 bal=1.919 mut=0.868 tok_s=33,015 avg_tok_s=32,444 |
| CKPT step=102000 valid_ce=4.1427 valid_ppl=63.0 |
| step=102100 tokens=1043462000 ce=3.9146 ppl=50.1 bal=1.880 mut=0.882 tok_s=29,353 avg_tok_s=32,441 |
| step=102200 tokens=1044484000 ce=3.9440 ppl=51.6 bal=2.382 mut=0.871 tok_s=33,246 avg_tok_s=32,442 |
| step=102300 tokens=1045506000 ce=3.9561 ppl=52.3 bal=2.010 mut=0.874 tok_s=32,909 avg_tok_s=32,442 |
| step=102400 tokens=1046528000 ce=3.8052 ppl=44.9 bal=1.934 mut=0.875 tok_s=32,981 avg_tok_s=32,443 |
| step=102500 tokens=1047550000 ce=4.0057 ppl=54.9 bal=1.967 mut=0.871 tok_s=32,991 avg_tok_s=32,443 |
| step=102600 tokens=1048572000 ce=4.1081 ppl=60.8 bal=1.837 mut=0.869 tok_s=32,903 avg_tok_s=32,444 |
| step=102700 tokens=1049594000 ce=3.9180 ppl=50.3 bal=1.881 mut=0.873 tok_s=33,048 avg_tok_s=32,444 |
| step=102800 tokens=1050616000 ce=3.8913 ppl=49.0 bal=2.010 mut=0.870 tok_s=33,045 avg_tok_s=32,445 |
| step=102900 tokens=1051638000 ce=4.0521 ppl=57.5 bal=2.136 mut=0.878 tok_s=33,004 avg_tok_s=32,445 |
| step=103000 tokens=1052660000 ce=4.0443 ppl=57.1 bal=1.982 mut=0.886 tok_s=33,317 avg_tok_s=32,446 |
| step=103100 tokens=1053682000 ce=4.3206 ppl=75.2 bal=2.446 mut=0.872 tok_s=33,211 avg_tok_s=32,447 |
| step=103200 tokens=1054704000 ce=3.9314 ppl=51.0 bal=1.946 mut=0.866 tok_s=32,506 avg_tok_s=32,447 |
| step=103300 tokens=1055726000 ce=4.0158 ppl=55.5 bal=1.993 mut=0.872 tok_s=32,836 avg_tok_s=32,447 |
| step=103400 tokens=1056748000 ce=3.9847 ppl=53.8 bal=1.867 mut=0.880 tok_s=32,867 avg_tok_s=32,448 |
| step=103500 tokens=1057770000 ce=3.8596 ppl=47.4 bal=2.103 mut=0.879 tok_s=33,226 avg_tok_s=32,448 |
| step=103600 tokens=1058792000 ce=4.2499 ppl=70.1 bal=1.952 mut=0.877 tok_s=33,113 avg_tok_s=32,449 |
| step=103700 tokens=1059814000 ce=3.8379 ppl=46.4 bal=2.084 mut=0.871 tok_s=32,829 avg_tok_s=32,449 |
| step=103800 tokens=1060836000 ce=4.0778 ppl=59.0 bal=2.067 mut=0.887 tok_s=33,280 avg_tok_s=32,450 |
| step=103900 tokens=1061858000 ce=3.9787 ppl=53.4 bal=1.909 mut=0.888 tok_s=33,206 avg_tok_s=32,451 |
| step=104000 tokens=1062880000 ce=4.2144 ppl=67.7 bal=2.109 mut=0.882 tok_s=33,044 avg_tok_s=32,452 |
| CKPT step=104000 valid_ce=4.0017 valid_ppl=54.7 |
| step=104100 tokens=1063902000 ce=3.9059 ppl=49.7 bal=2.025 mut=0.872 tok_s=29,257 avg_tok_s=32,448 |
| step=104200 tokens=1064924000 ce=4.2850 ppl=72.6 bal=2.214 mut=0.879 tok_s=33,046 avg_tok_s=32,449 |
| step=104300 tokens=1065946000 ce=3.9556 ppl=52.2 bal=2.014 mut=0.880 tok_s=32,905 avg_tok_s=32,449 |
| step=104400 tokens=1066968000 ce=4.0331 ppl=56.4 bal=1.859 mut=0.875 tok_s=32,858 avg_tok_s=32,450 |
| step=104500 tokens=1067990000 ce=3.9174 ppl=50.3 bal=1.863 mut=0.878 tok_s=32,762 avg_tok_s=32,450 |
| step=104600 tokens=1069012000 ce=3.9283 ppl=50.8 bal=1.905 mut=0.876 tok_s=32,801 avg_tok_s=32,450 |
| step=104700 tokens=1070034000 ce=4.0397 ppl=56.8 bal=2.107 mut=0.878 tok_s=33,049 avg_tok_s=32,451 |
| step=104800 tokens=1071056000 ce=3.9899 ppl=54.1 bal=1.895 mut=0.880 tok_s=32,964 avg_tok_s=32,452 |
| step=104900 tokens=1072078000 ce=4.2518 ppl=70.2 bal=2.118 mut=0.887 tok_s=33,102 avg_tok_s=32,452 |
| step=105000 tokens=1073100000 ce=3.8426 ppl=46.6 bal=1.851 mut=0.877 tok_s=33,254 avg_tok_s=32,453 |
| step=105100 tokens=1074122000 ce=3.9232 ppl=50.6 bal=1.932 mut=0.886 tok_s=33,307 avg_tok_s=32,454 |
| step=105200 tokens=1075144000 ce=3.9688 ppl=52.9 bal=1.862 mut=0.878 tok_s=32,941 avg_tok_s=32,454 |
| step=105300 tokens=1076166000 ce=3.7737 ppl=43.5 bal=1.883 mut=0.881 tok_s=33,083 avg_tok_s=32,455 |
| step=105400 tokens=1077188000 ce=3.7972 ppl=44.6 bal=2.214 mut=0.889 tok_s=32,556 avg_tok_s=32,455 |
| step=105500 tokens=1078210000 ce=3.9493 ppl=51.9 bal=1.865 mut=0.887 tok_s=33,144 avg_tok_s=32,456 |
| step=105600 tokens=1079232000 ce=3.9705 ppl=53.0 bal=2.146 mut=0.879 tok_s=32,638 avg_tok_s=32,456 |
| step=105700 tokens=1080254000 ce=4.1629 ppl=64.3 bal=2.056 mut=0.877 tok_s=32,838 avg_tok_s=32,456 |
| step=105800 tokens=1081276000 ce=4.1717 ppl=64.8 bal=1.783 mut=0.887 tok_s=32,375 avg_tok_s=32,456 |
| step=105900 tokens=1082298000 ce=3.9142 ppl=50.1 bal=1.913 mut=0.884 tok_s=32,934 avg_tok_s=32,456 |
| step=106000 tokens=1083320000 ce=3.7959 ppl=44.5 bal=1.980 mut=0.885 tok_s=33,074 avg_tok_s=32,457 |
| CKPT step=106000 valid_ce=4.0804 valid_ppl=59.2 |
| step=106100 tokens=1084342000 ce=3.7434 ppl=42.2 bal=2.107 mut=0.888 tok_s=29,014 avg_tok_s=32,454 |
| step=106200 tokens=1085364000 ce=4.0948 ppl=60.0 bal=1.953 mut=0.889 tok_s=33,010 avg_tok_s=32,454 |
| step=106300 tokens=1086386000 ce=4.0131 ppl=55.3 bal=1.866 mut=0.879 tok_s=32,586 avg_tok_s=32,454 |
| step=106400 tokens=1087408000 ce=3.8794 ppl=48.4 bal=2.001 mut=0.887 tok_s=33,162 avg_tok_s=32,455 |
| step=106500 tokens=1088430000 ce=4.0169 ppl=55.5 bal=2.161 mut=0.886 tok_s=33,159 avg_tok_s=32,456 |
| step=106600 tokens=1089452000 ce=3.9017 ppl=49.5 bal=1.963 mut=0.893 tok_s=33,162 avg_tok_s=32,456 |
| step=106700 tokens=1090474000 ce=3.9070 ppl=49.7 bal=1.858 mut=0.881 tok_s=32,779 avg_tok_s=32,457 |
| step=106800 tokens=1091496000 ce=3.6559 ppl=38.7 bal=1.994 mut=0.883 tok_s=32,944 avg_tok_s=32,457 |
| step=106900 tokens=1092518000 ce=3.8555 ppl=47.3 bal=1.836 mut=0.880 tok_s=33,138 avg_tok_s=32,458 |
| step=107000 tokens=1093540000 ce=3.8708 ppl=48.0 bal=1.934 mut=0.877 tok_s=33,027 avg_tok_s=32,458 |
| step=107100 tokens=1094562000 ce=4.2334 ppl=69.0 bal=2.362 mut=0.882 tok_s=32,948 avg_tok_s=32,459 |
| step=107200 tokens=1095584000 ce=3.8881 ppl=48.8 bal=1.873 mut=0.893 tok_s=32,960 avg_tok_s=32,459 |
| step=107300 tokens=1096606000 ce=3.9721 ppl=53.1 bal=1.823 mut=0.890 tok_s=32,884 avg_tok_s=32,460 |
| step=107400 tokens=1097628000 ce=4.2618 ppl=70.9 bal=1.888 mut=0.880 tok_s=32,964 avg_tok_s=32,460 |
| step=107500 tokens=1098650000 ce=3.7624 ppl=43.1 bal=1.824 mut=0.893 tok_s=32,931 avg_tok_s=32,461 |
| step=107600 tokens=1099672000 ce=3.9935 ppl=54.2 bal=1.962 mut=0.889 tok_s=32,861 avg_tok_s=32,461 |
| step=107700 tokens=1100694000 ce=3.9905 ppl=54.1 bal=2.667 mut=0.886 tok_s=32,996 avg_tok_s=32,461 |
| step=107800 tokens=1101716000 ce=3.8504 ppl=47.0 bal=1.988 mut=0.883 tok_s=32,992 avg_tok_s=32,462 |
| step=107900 tokens=1102738000 ce=4.3188 ppl=75.1 bal=1.932 mut=0.885 tok_s=32,893 avg_tok_s=32,462 |
| step=108000 tokens=1103760000 ce=4.0419 ppl=56.9 bal=1.877 mut=0.889 tok_s=32,984 avg_tok_s=32,463 |
| CKPT step=108000 valid_ce=3.9928 valid_ppl=54.2 |
| step=108100 tokens=1104782000 ce=3.8938 ppl=49.1 bal=1.926 mut=0.891 tok_s=29,152 avg_tok_s=32,460 |
| step=108200 tokens=1105804000 ce=4.1179 ppl=61.4 bal=1.860 mut=0.887 tok_s=33,064 avg_tok_s=32,460 |
| step=108300 tokens=1106826000 ce=4.0810 ppl=59.2 bal=1.880 mut=0.895 tok_s=32,880 avg_tok_s=32,461 |
| step=108400 tokens=1107848000 ce=4.0057 ppl=54.9 bal=1.882 mut=0.889 tok_s=32,893 avg_tok_s=32,461 |
| step=108500 tokens=1108870000 ce=3.9241 ppl=50.6 bal=1.907 mut=0.889 tok_s=32,957 avg_tok_s=32,462 |
| step=108600 tokens=1109892000 ce=3.9239 ppl=50.6 bal=1.876 mut=0.878 tok_s=33,109 avg_tok_s=32,462 |
| step=108700 tokens=1110914000 ce=3.8781 ppl=48.3 bal=1.817 mut=0.885 tok_s=32,858 avg_tok_s=32,463 |
| step=108800 tokens=1111936000 ce=4.0556 ppl=57.7 bal=1.870 mut=0.895 tok_s=33,171 avg_tok_s=32,463 |
| step=108900 tokens=1112958000 ce=4.0002 ppl=54.6 bal=1.982 mut=0.891 tok_s=33,273 avg_tok_s=32,464 |
| step=109000 tokens=1113980000 ce=3.9969 ppl=54.4 bal=1.883 mut=0.883 tok_s=33,115 avg_tok_s=32,465 |
| step=109100 tokens=1115002000 ce=3.7716 ppl=43.4 bal=1.889 mut=0.890 tok_s=32,964 avg_tok_s=32,465 |
| step=109200 tokens=1116024000 ce=4.0079 ppl=55.0 bal=2.022 mut=0.895 tok_s=32,725 avg_tok_s=32,465 |
| step=109300 tokens=1117046000 ce=3.8854 ppl=48.7 bal=1.909 mut=0.901 tok_s=32,957 avg_tok_s=32,466 |
| step=109400 tokens=1118068000 ce=3.9528 ppl=52.1 bal=1.870 mut=0.893 tok_s=33,030 avg_tok_s=32,466 |
| step=109500 tokens=1119090000 ce=4.0938 ppl=60.0 bal=1.922 mut=0.883 tok_s=33,102 avg_tok_s=32,467 |
| step=109600 tokens=1120112000 ce=3.8680 ppl=47.8 bal=1.858 mut=0.891 tok_s=32,900 avg_tok_s=32,467 |
| step=109700 tokens=1121134000 ce=4.2623 ppl=71.0 bal=2.194 mut=0.895 tok_s=33,062 avg_tok_s=32,468 |
| step=109800 tokens=1122156000 ce=3.9248 ppl=50.6 bal=1.824 mut=0.888 tok_s=33,299 avg_tok_s=32,468 |
| step=109900 tokens=1123178000 ce=4.0267 ppl=56.1 bal=1.974 mut=0.895 tok_s=32,855 avg_tok_s=32,469 |
| step=110000 tokens=1124200000 ce=3.9774 ppl=53.4 bal=1.834 mut=0.882 tok_s=33,161 avg_tok_s=32,469 |
| CKPT step=110000 valid_ce=4.0672 valid_ppl=58.4 |
| step=110100 tokens=1125222000 ce=3.9153 ppl=50.2 bal=1.886 mut=0.890 tok_s=29,277 avg_tok_s=32,467 |
| step=110200 tokens=1126244000 ce=4.0585 ppl=57.9 bal=1.883 mut=0.888 tok_s=32,858 avg_tok_s=32,467 |
| step=110300 tokens=1127266000 ce=3.7996 ppl=44.7 bal=2.047 mut=0.883 tok_s=33,061 avg_tok_s=32,467 |
| step=110400 tokens=1128288000 ce=3.8304 ppl=46.1 bal=1.909 mut=0.893 tok_s=32,874 avg_tok_s=32,468 |
| step=110500 tokens=1129310000 ce=3.9761 ppl=53.3 bal=1.927 mut=0.898 tok_s=33,108 avg_tok_s=32,468 |
| step=110600 tokens=1130332000 ce=3.7760 ppl=43.6 bal=2.356 mut=0.890 tok_s=33,161 avg_tok_s=32,469 |
| step=110700 tokens=1131354000 ce=3.9678 ppl=52.9 bal=1.866 mut=0.895 tok_s=32,972 avg_tok_s=32,469 |
| step=110800 tokens=1132376000 ce=3.8087 ppl=45.1 bal=1.860 mut=0.884 tok_s=32,844 avg_tok_s=32,470 |
| step=110900 tokens=1133398000 ce=4.0360 ppl=56.6 bal=1.849 mut=0.884 tok_s=33,205 avg_tok_s=32,470 |
| step=111000 tokens=1134420000 ce=4.1372 ppl=62.6 bal=1.876 mut=0.880 tok_s=33,198 avg_tok_s=32,471 |
| step=111100 tokens=1135442000 ce=3.8599 ppl=47.5 bal=1.826 mut=0.887 tok_s=33,309 avg_tok_s=32,472 |
| step=111200 tokens=1136464000 ce=3.9559 ppl=52.2 bal=1.836 mut=0.888 tok_s=32,624 avg_tok_s=32,472 |
| step=111300 tokens=1137486000 ce=3.8882 ppl=48.8 bal=1.922 mut=0.885 tok_s=33,150 avg_tok_s=32,473 |
| step=111400 tokens=1138508000 ce=3.7385 ppl=42.0 bal=1.937 mut=0.896 tok_s=33,211 avg_tok_s=32,473 |
| step=111500 tokens=1139530000 ce=3.8596 ppl=47.4 bal=1.879 mut=0.886 tok_s=33,138 avg_tok_s=32,474 |
| step=111600 tokens=1140552000 ce=4.0037 ppl=54.8 bal=2.001 mut=0.886 tok_s=32,832 avg_tok_s=32,474 |
| step=111700 tokens=1141574000 ce=4.0582 ppl=57.9 bal=1.856 mut=0.894 tok_s=33,154 avg_tok_s=32,475 |
| step=111800 tokens=1142596000 ce=3.9751 ppl=53.3 bal=1.934 mut=0.889 tok_s=32,993 avg_tok_s=32,475 |
| step=111900 tokens=1143618000 ce=3.8719 ppl=48.0 bal=1.826 mut=0.889 tok_s=32,841 avg_tok_s=32,476 |
| step=112000 tokens=1144640000 ce=4.1057 ppl=60.7 bal=2.138 mut=0.895 tok_s=32,547 avg_tok_s=32,476 |
| CKPT step=112000 valid_ce=4.0317 valid_ppl=56.4 |
| step=112100 tokens=1145662000 ce=3.9073 ppl=49.8 bal=1.908 mut=0.891 tok_s=29,114 avg_tok_s=32,473 |
| step=112200 tokens=1146684000 ce=3.7379 ppl=42.0 bal=1.969 mut=0.891 tok_s=33,041 avg_tok_s=32,473 |
| step=112300 tokens=1147706000 ce=3.9239 ppl=50.6 bal=1.844 mut=0.890 tok_s=33,032 avg_tok_s=32,474 |
| step=112400 tokens=1148728000 ce=3.8128 ppl=45.3 bal=1.860 mut=0.895 tok_s=33,110 avg_tok_s=32,474 |
| step=112500 tokens=1149750000 ce=4.0208 ppl=55.7 bal=1.907 mut=0.889 tok_s=33,091 avg_tok_s=32,475 |
| step=112600 tokens=1150772000 ce=4.3423 ppl=76.9 bal=2.308 mut=0.893 tok_s=32,940 avg_tok_s=32,475 |
| step=112700 tokens=1151794000 ce=3.9001 ppl=49.4 bal=1.896 mut=0.894 tok_s=32,990 avg_tok_s=32,476 |
| step=112800 tokens=1152816000 ce=3.8946 ppl=49.1 bal=2.002 mut=0.884 tok_s=32,840 avg_tok_s=32,476 |
| step=112900 tokens=1153838000 ce=3.9694 ppl=53.0 bal=2.077 mut=0.894 tok_s=33,145 avg_tok_s=32,477 |
| step=113000 tokens=1154860000 ce=4.0000 ppl=54.6 bal=1.851 mut=0.884 tok_s=33,266 avg_tok_s=32,477 |
| step=113100 tokens=1155882000 ce=4.1082 ppl=60.8 bal=1.988 mut=0.893 tok_s=32,874 avg_tok_s=32,478 |
| step=113200 tokens=1156904000 ce=3.8109 ppl=45.2 bal=1.836 mut=0.889 tok_s=33,138 avg_tok_s=32,478 |
| step=113300 tokens=1157926000 ce=4.1840 ppl=65.6 bal=2.096 mut=0.885 tok_s=32,924 avg_tok_s=32,479 |
| step=113400 tokens=1158948000 ce=3.9090 ppl=49.8 bal=2.008 mut=0.891 tok_s=32,990 avg_tok_s=32,479 |
| step=113500 tokens=1159970000 ce=4.0208 ppl=55.7 bal=1.839 mut=0.897 tok_s=33,086 avg_tok_s=32,480 |
| step=113600 tokens=1160992000 ce=3.9973 ppl=54.5 bal=1.871 mut=0.886 tok_s=32,467 avg_tok_s=32,480 |
| step=113700 tokens=1162014000 ce=3.9136 ppl=50.1 bal=1.930 mut=0.887 tok_s=33,120 avg_tok_s=32,480 |
| PEER21_BASE_DONE {"valid_ce": 3.7246193885803223, "steps": 113738, "avg_tok_s": 32480.136284458968, "tokens": 1162402360} |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
| PEER21_CONFIG {"ctrl_params": 600193024, "active_params_per_token": 72761344, "experts": 495616, "peer_layers": [4, 12], "h_times_k": 8, "nk": 704, "pool_params": 507510784, "pool_dtype": "torch.float32", "codon_table_bytes": 536870912, "batch": 20, "seq": 512, "variant": "v21-shared-pool"} |
| MUON 20971520 ADAM 71710720 SPARSE_POOL 507510784 |
|
|