Exp B B1: r512k_gpt2b410 -> exp-b-b1/gpt2-denom-b410/metrics.csv
Browse files
exp-b-b1/gpt2-denom-b410/metrics.csv
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
step,tokens,timed_hours,gpu_hours,train_loss_ema,neutral_bpb,ownval_bpb,lr,wallclock_s
|
| 2 |
+
17,8912896,0.002119,0.016948,10.2781,3.384675,3.027201,7.29e-06,51.9
|
| 3 |
+
23,12058624,0.003938,0.031505,9.9731,3.358753,2.968319,9.86e-06,92.4
|
| 4 |
+
30,15728640,0.006060,0.048480,9.6662,3.294992,2.903269,1.29e-05,133.8
|
| 5 |
+
40,20971520,0.009089,0.072716,9.2847,3.230582,2.822304,1.71e-05,178.5
|
| 6 |
+
54,28311552,0.013333,0.106666,8.8197,3.124335,2.702263,2.31e-05,227.6
|
| 7 |
+
71,37224448,0.018485,0.147883,8.3361,2.995842,2.559287,3.04e-05,280.0
|
| 8 |
+
95,49807360,0.025759,0.206074,7.7181,2.833158,2.373873,4.07e-05,339.9
|
| 9 |
+
127,66584576,0.035453,0.283621,7.0526,2.708014,2.220260,5.44e-05,408.6
|
| 10 |
+
169,88604672,0.048183,0.385463,6.5932,2.605324,2.112979,7.24e-05,488.4
|
| 11 |
+
226,118489088,0.065461,0.523688,6.2582,2.494133,2.025813,9.69e-05,584.4
|
| 12 |
+
301,157810688,0.088195,0.705557,5.9999,2.396470,1.946403,1.29e-04,700.1
|
| 13 |
+
401,210239488,0.118496,0.947971,5.6963,2.304912,1.845770,1.72e-04,843.0
|
| 14 |
+
535,280494080,0.159098,1.272782,5.2585,2.165896,1.705865,2.29e-04,1023.0
|
| 15 |
+
713,373817344,0.213034,1.704271,4.7545,1.928700,1.544143,3.00e-04,1251.0
|
| 16 |
+
951,498597888,0.285141,2.281129,4.1643,1.626535,1.368991,3.00e-04,1544.4
|
| 17 |
+
1269,665321472,0.381480,3.051842,3.8269,1.518725,1.263685,2.99e-04,1925.0
|
| 18 |
+
1692,887095296,0.509774,4.078195,3.6100,1.441495,1.192874,2.98e-04,2420.8
|
| 19 |
+
2256,1182793728,0.680690,5.445517,3.4411,1.402528,1.140670,2.94e-04,3069.9
|
| 20 |
+
3009,1577582592,0.909091,7.272732,3.3152,1.364902,1.099913,2.87e-04,3926.0
|
| 21 |
+
4012,2103443456,1.213111,9.704888,3.2262,1.331337,1.065793,2.73e-04,5054.4
|
| 22 |
+
5350,2804940800,1.618703,12.949622,3.1201,1.313022,1.036432,2.49e-04,6548.5
|
| 23 |
+
7135,3740794880,2.159691,17.277527,3.0384,1.287835,1.010208,2.08e-04,8529.8
|
| 24 |
+
9514,4988076032,2.880694,23.045549,2.9642,1.259053,0.985247,1.47e-04,11159.4
|
| 25 |
+
12687,6651641856,3.842412,30.739299,2.9039,1.244802,0.962412,7.29e-05,14655.6
|
| 26 |
+
15227,7983333376,4.612213,36.897706,2.8631,1.242089,0.951682,3.72e-05,17460.8
|
| 27 |
+
15565,8160542720,4.714626,37.717007,2.8598,1.238105,0.950678,3.46e-05,17863.4
|
| 28 |
+
15904,8338276352,4.817360,38.538883,2.8656,1.235178,0.949846,3.26e-05,18267.0
|
| 29 |
+
16242,8515485696,4.919871,39.358965,2.8486,1.238443,0.949124,3.12e-05,18669.8
|
| 30 |
+
16581,8693219328,5.022700,40.181604,2.8538,1.241639,0.948510,3.03e-05,19073.8
|
| 31 |
+
16919,8870428672,5.125207,41.001660,2.8610,1.237696,0.947947,3.00e-05,19476.7
|