loss,grad_norm,learning_rate,epoch,step
2.0842,0.39258643984794617,4.000000000000001e-06,0.09302325581395349,10
2.0411,0.31776493787765503,8.000000000000001e-06,0.18604651162790697,20
1.9467,0.2665407359600067,1.2e-05,0.27906976744186046,30
1.9131,0.27922648191452026,1.6000000000000003e-05,0.37209302325581395,40
1.7353,0.38724032044410706,2e-05,0.46511627906976744,50
1.5224,0.6602446436882019,2.4e-05,0.5581395348837209,60
1.0967,0.6640816926956177,2.8000000000000003e-05,0.6511627906976745,70
0.6702,0.26523134112358093,3.2000000000000005e-05,0.7441860465116279,80
0.5406,0.17431379854679108,3.6e-05,0.8372093023255814,90
0.5325,0.16474466025829315,4e-05,0.9302325581395349,100