diff --git "a/weights/best_model_metadata.json" "b/weights/best_model_metadata.json" --- "a/weights/best_model_metadata.json" +++ "b/weights/best_model_metadata.json" @@ -1,196 +1,211 @@ { - "epoch": 1, + "epoch": 2, "optimizer_state_dict": { "state": { "0": { - "step": "tensor(2504.)", - "exp_avg": "tensor([[-2.5660e-05, 8.6654e-06, 6.0734e-05, ..., -6.8806e-05,\n 1.7010e-05, -1.2918e-05],\n [ 3.4550e-05, -2.6272e-05, -7.1295e-05, ..., 3.2710e-05,\n 4.9773e-05, -4.5199e-06],\n [-5.4501e-24, -1.1627e-23, -1.2960e-24, ..., -3.1867e-24,\n -3.0199e-24, 7.0640e-24],\n ...,\n [-1.1757e-05, -3.3489e-05, -5.5402e-05, ..., 1.1567e-05,\n -1.0570e-05, -2.4651e-05],\n [-2.2741e-05, -4.6035e-06, 2.8693e-05, ..., -8.8789e-06,\n 9.3655e-06, 9.3527e-06],\n [ 9.8680e-06, 2.4155e-06, -1.8717e-05, ..., -1.3777e-07,\n -1.6496e-06, 5.8280e-06]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.3867e-08, 1.5095e-08, 6.9419e-09, ..., 8.9740e-09, 7.5087e-09,\n 5.2236e-09],\n [9.8261e-09, 1.0136e-08, 8.8865e-09, ..., 6.9002e-09, 5.5357e-09,\n 4.8151e-09],\n [1.7317e-11, 2.0972e-11, 2.0908e-11, ..., 3.0171e-11, 5.8204e-12,\n 1.3405e-11],\n ...,\n [1.3815e-08, 1.2816e-08, 9.8299e-09, ..., 9.5702e-09, 7.5078e-09,\n 6.0434e-09],\n [1.6530e-08, 1.3552e-08, 1.1877e-08, ..., 1.0545e-08, 9.1210e-09,\n 7.8570e-09],\n [2.7099e-09, 4.2244e-09, 2.9066e-09, ..., 2.3599e-09, 1.9022e-09,\n 1.9200e-09]], device='cuda:0')" + "step": "tensor(3756.)", + "exp_avg": "tensor([[ 2.1517e-05, -2.6941e-05, -7.6814e-06, ..., -7.1187e-06,\n -2.2356e-06, 1.9026e-05],\n [ 7.0848e-06, -2.2839e-05, -2.4405e-07, ..., -3.2418e-05,\n -1.1959e-05, -2.4559e-05],\n [ 2.1240e-40, 3.4596e-40, 1.1568e-40, ..., -4.2081e-41,\n 1.2767e-40, -1.5001e-40],\n ...,\n [-1.1554e-05, -1.6690e-05, 1.2038e-05, ..., 1.7157e-05,\n -1.2429e-05, -4.7271e-06],\n [-1.9076e-05, 1.8564e-05, 7.2819e-06, ..., 1.5901e-05,\n 1.4939e-05, 2.8084e-05],\n [ 3.8714e-06, 1.4334e-06, 2.7298e-05, ..., -6.0566e-06,\n -1.2570e-05, -1.5604e-07]], device='cuda:0')", + "exp_avg_sq": "tensor([[1.3189e-08, 1.5605e-08, 6.9154e-09, ..., 8.6941e-09, 7.6699e-09,\n 4.9935e-09],\n [1.0543e-08, 9.8523e-09, 1.0035e-08, ..., 7.7223e-09, 5.9935e-09,\n 5.0743e-09],\n [4.9501e-12, 5.9935e-12, 5.9748e-12, ..., 8.6235e-12, 1.6632e-12,\n 3.8329e-12],\n ...,\n [1.3503e-08, 1.1356e-08, 9.2611e-09, ..., 8.3496e-09, 7.0566e-09,\n 5.9478e-09],\n [1.4595e-08, 1.2197e-08, 9.5062e-09, ..., 9.2816e-09, 7.8575e-09,\n 6.2278e-09],\n [3.1345e-09, 4.8947e-09, 3.4502e-09, ..., 2.2795e-09, 2.4005e-09,\n 2.1024e-09]], device='cuda:0')" }, "1": { - "step": "tensor(2504.)", - "exp_avg": "tensor([-1.9185e-03, 3.0139e-03, 3.1894e-22, ..., 9.6413e-04,\n -7.8702e-04, 3.0819e-04], device='cuda:0')", - "exp_avg_sq": "tensor([1.6176e-05, 1.4221e-05, 3.6212e-08, ..., 1.8365e-05, 2.1177e-05,\n 4.3719e-06], device='cuda:0')" + "step": "tensor(3756.)", + "exp_avg": "tensor([-1.4859e-04, 2.3281e-04, -9.1672e-39, ..., -6.7725e-04,\n -1.2363e-03, -1.0333e-04], device='cuda:0')", + "exp_avg_sq": "tensor([1.6299e-05, 1.3848e-05, 1.0350e-08, ..., 1.6846e-05, 1.6282e-05,\n 5.2225e-06], device='cuda:0')" }, "2": { - "step": "tensor(2504.)", - "exp_avg": "tensor([[-1.7566e-06, -3.6239e-06, 3.7904e-41, ..., 2.9396e-06,\n -2.2079e-06, 2.0373e-06],\n [ 1.1574e-29, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n ...,\n [ 0.0000e+00, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 0.0000e+00],\n [-6.0374e-06, 4.9641e-06, -5.2032e-41, ..., 6.7283e-06,\n 3.1242e-06, 5.2013e-06],\n [-7.6419e-06, 2.0507e-06, -7.4114e-40, ..., 6.3524e-06,\n 1.0472e-05, -6.2212e-07]], device='cuda:0')", - "exp_avg_sq": "tensor([[5.4391e-09, 2.5623e-09, 9.7713e-12, ..., 6.4433e-09, 1.9342e-08,\n 8.6120e-10],\n [1.0450e-11, 1.9139e-10, 0.0000e+00, ..., 3.5906e-11, 1.3284e-14,\n 1.3224e-10],\n [2.6480e-10, 9.3711e-10, 0.0000e+00, ..., 5.1448e-10, 4.5330e-15,\n 1.0542e-09],\n ...,\n [0.0000e+00, 4.4806e-17, 0.0000e+00, ..., 1.3126e-18, 1.5677e-18,\n 0.0000e+00],\n [2.7058e-09, 5.1170e-09, 1.4081e-12, ..., 2.1932e-09, 2.4470e-10,\n 6.7471e-09],\n [5.3852e-10, 6.1069e-10, 7.5062e-12, ..., 1.2926e-09, 1.0978e-08,\n 2.9124e-10]], device='cuda:0')" + "step": "tensor(3756.)", + "exp_avg": "tensor([[-9.9799e-06, -4.5945e-07, 5.6052e-45, ..., -2.7008e-06,\n -2.7476e-07, -1.2597e-06],\n [ 5.6052e-45, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [-6.7492e-08, -2.6027e-07, 0.0000e+00, ..., 8.6992e-08,\n 6.6784e-07, 5.4310e-12],\n ...,\n [ 0.0000e+00, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 0.0000e+00],\n [ 6.7450e-06, 7.5090e-07, -5.6052e-45, ..., 2.9990e-05,\n 6.8130e-07, 1.2804e-05],\n [ 1.9819e-06, -3.3921e-06, -5.6052e-45, ..., -9.3940e-06,\n 4.9197e-06, -9.4716e-07]], device='cuda:0')", + "exp_avg_sq": "tensor([[3.7625e-09, 1.1538e-09, 2.7922e-12, ..., 4.0106e-09, 6.2717e-09,\n 5.1303e-10],\n [2.9863e-12, 5.4691e-11, 0.0000e+00, ..., 1.0261e-11, 3.7961e-15,\n 3.7790e-11],\n [7.5890e-11, 2.7005e-10, 0.0000e+00, ..., 1.5915e-10, 1.1158e-11,\n 3.0124e-10],\n ...,\n [0.0000e+00, 1.2804e-17, 0.0000e+00, ..., 3.7509e-19, 4.4798e-19,\n 0.0000e+00],\n [6.1476e-09, 2.2487e-09, 4.0238e-13, ..., 3.7948e-09, 5.9777e-10,\n 4.3273e-09],\n [5.5344e-10, 9.7504e-10, 2.1450e-12, ..., 1.1998e-09, 9.3081e-09,\n 6.3259e-10]], device='cuda:0')" }, "3": { - "step": "tensor(2504.)", - "exp_avg": "tensor([ 1.4322e-04, 3.4533e-29, 6.5650e-11, -3.6571e-05, -6.6848e-05,\n 3.6434e-44, -9.7217e-05, 3.9947e-05, -1.7099e-04, 1.7532e-04,\n -3.9754e-04, 4.6762e-40, 1.3526e-04, 1.1270e-04, -5.7818e-04,\n 2.2875e-04, -5.5307e-05, 5.6052e-45, 5.6052e-45, -2.4281e-04,\n 5.6052e-45, -4.6699e-04, 1.1730e-04, 7.8408e-05, -2.5211e-04,\n -2.0803e-04, 1.4951e-04, 5.6052e-45, -6.7371e-05, 1.4645e-04,\n -4.7437e-05, 1.0607e-04, 1.0702e-04, 1.1955e-04, -1.3234e-05,\n -2.3357e-04, 5.6052e-45, 2.8735e-04, 8.1930e-05, -2.9507e-05,\n -4.6797e-04, -6.8474e-05, 3.4104e-04, -1.1711e-04, -2.0275e-04,\n -1.0806e-04, 6.7718e-05, -2.1533e-04, 1.7281e-04, 8.3198e-05,\n 5.6052e-45, -1.6009e-04, -1.3909e-04, -1.1147e-05, -1.9390e-04,\n -6.7170e-04, 2.2166e-17, -5.4672e-04, -5.7652e-16, -2.5850e-04,\n -8.6837e-05, -1.8233e-04, 1.1021e-04, 3.0163e-04, 2.4248e-05,\n -2.3600e-04, -6.3410e-05, 1.1184e-04, 5.6052e-45, 1.5415e-14,\n 3.8166e-04, 5.6052e-45, 3.0127e-05, 2.8628e-04, 2.5693e-05,\n 1.8250e-04, 5.6052e-45, 1.2301e-04, -2.1564e-04, 6.1370e-05,\n -1.8293e-08, -8.4287e-05, 7.3126e-05, 5.6052e-45, 2.4430e-05,\n 3.0213e-04, 1.8963e-04, 7.7238e-05, 2.1611e-04, 1.9266e-05,\n -9.2679e-05, -4.5723e-05, 1.0678e-04, -5.4022e-05, -3.1432e-04,\n -8.5947e-05, -9.7346e-05, 7.6312e-06, -1.8071e-04, -8.0175e-05,\n 2.9171e-04, -5.8724e-04, -4.9106e-05, 2.8464e-04, -4.5895e-05,\n 3.9347e-04, 3.7891e-05, -4.3048e-06, 4.7548e-04, -7.3704e-04,\n 7.4217e-05, -2.9715e-04, 3.5441e-05, 1.7988e-39, -1.6515e-04,\n 8.3038e-05, 1.1400e-04, -5.6444e-05, -4.2922e-07, 5.6052e-45,\n 2.7274e-05, 7.6963e-05, 2.6680e-05, 4.9977e-23, 4.0178e-05,\n 4.3908e-04, 9.0774e-05, 5.6052e-45, 3.1026e-04, 5.6052e-45,\n 6.6408e-05, 1.3555e-04, 5.6052e-45, 1.8844e-04, -1.5136e-18,\n 3.4767e-04, 1.6727e-18, 8.4772e-05, -3.9698e-05, 5.6052e-45,\n 7.9293e-30, -3.2685e-09, -3.3293e-06, 2.2372e-19, 5.6052e-45,\n 1.1082e-33, -2.3096e-04, -3.1814e-05, 5.6052e-45, -4.2766e-05,\n -2.2378e-05, 6.8677e-05, 3.7686e-05, -1.4206e-04, 2.1133e-30,\n 9.5181e-05, 5.6052e-45, -1.1860e-04, 5.6052e-45, 3.6836e-05,\n 8.6598e-06, 3.9245e-16, -1.4324e-04, -2.8696e-05, -8.9833e-05,\n -1.8606e-04, 7.8131e-18, 5.6052e-45, 4.8642e-05, 8.9579e-06,\n 1.0195e-04, 1.7994e-04, -5.4282e-05, 2.7488e-04, 4.0205e-04,\n -2.2136e-04, 1.4525e-10, 9.1646e-05, 2.0917e-04, 4.4284e-04,\n 5.7151e-04, 3.5650e-04, 2.0491e-04, 4.8932e-39, 9.9381e-24,\n 5.6490e-17, 5.3077e-04, 3.8222e-04, -9.6247e-05, -8.1921e-09,\n -5.0868e-05, 1.3761e-04, -4.5532e-05, 1.0723e-05, 3.4258e-04,\n 7.1746e-05, -7.6874e-05, -4.1260e-04, -4.6133e-05, 5.6052e-45,\n 5.6052e-45, 1.1629e-04, -6.1120e-04, 2.4319e-04, -5.4965e-06,\n -1.5510e-04, -1.7364e-05, 1.2989e-04, 5.6052e-45, 1.5166e-17,\n 2.1531e-05, 5.6052e-45, 1.2791e-05, 3.7798e-04, 2.7118e-04,\n -3.4054e-04, 4.8963e-04, 5.0734e-05, 2.3378e-05, -9.1435e-05,\n -4.1026e-05, 1.3175e-04, -2.0030e-05, -1.5977e-04, -1.6675e-04,\n 5.6052e-45, 2.2647e-08, 2.2562e-04, 1.9427e-04, 3.5270e-22,\n 1.1228e-04, -2.0312e-04, 5.9812e-05, 1.3641e-04, 1.0867e-41,\n 6.6329e-06, 3.3919e-04, 5.6052e-45, 1.2423e-04, 2.5550e-04,\n 5.6052e-45, 5.6052e-45, 2.7216e-04, -5.7222e-05, -2.0794e-05,\n -1.6444e-04, 3.4311e-04, -2.6302e-04, 9.6363e-05, 8.2548e-31,\n 7.3710e-05, 8.6174e-05, 8.6137e-06, 5.5764e-05, -1.7415e-04,\n 6.9814e-05, 5.6052e-45, 1.7774e-04, 5.6052e-45, -9.2232e-05,\n 5.6052e-45, 2.4521e-04, 7.6949e-05, -2.1472e-04, -7.1995e-05,\n 1.7149e-05, 3.9930e-05, -5.6790e-05, 2.5929e-04, -5.6052e-45,\n 8.2794e-05, 1.9250e-04, -1.3704e-04, -2.6978e-04, -3.1217e-04,\n -2.2599e-04, -2.3416e-04, 5.6052e-45, -1.6286e-04, 3.4079e-05,\n 5.6052e-45, -1.4093e-04, 5.6052e-45, 1.1633e-04, -2.1137e-05,\n -1.7026e-04, -3.3621e-04, 1.4733e-04, 5.6052e-45, -1.2498e-04,\n -1.9543e-04, 9.5395e-05, 5.6052e-45, -8.3478e-06, 2.9504e-04,\n 1.1157e-04, -5.2103e-04, -1.0053e-04, -2.2356e-04, -2.6378e-05,\n 4.5542e-05, -2.9292e-04, 9.5021e-05, -2.0734e-04, 5.6052e-45,\n 5.6052e-45, 1.0791e-14, -2.8591e-04, -3.9580e-05, -1.1984e-04,\n 3.0108e-04, 8.5024e-22, 1.6614e-05, 6.2520e-05, -7.9132e-05,\n 5.6052e-45, 1.5350e-12, 4.1572e-05, 9.6424e-10, 1.6133e-04,\n 8.9847e-05, -2.1067e-04, 1.4022e-05, 1.3755e-04, -1.1779e-04,\n 5.6052e-45, 2.0621e-04, 1.0332e-04, 2.5141e-04, 2.3641e-04,\n 7.7428e-05, -1.0527e-04, -2.6925e-05, 3.9117e-05, 1.4416e-07,\n -3.8762e-05, 3.6420e-04, 1.0793e-16, 5.6052e-45, 5.6052e-45,\n 4.9071e-05, 1.6219e-04, 7.7055e-05, 1.2669e-05, -1.1096e-05,\n 5.8260e-06, -1.1388e-04, -4.7981e-05, -3.0762e-05, 1.3368e-04,\n -5.9877e-06, 2.0230e-05, -2.0799e-05, 5.6052e-45, 3.9164e-05,\n 3.1191e-05, 2.5170e-10, 1.7391e-09, -7.1702e-06, 2.4974e-05,\n 9.8189e-05, -4.7543e-05, 5.6052e-45, 7.8405e-04, -1.0121e-05,\n 3.5048e-33, 2.9689e-04, -1.0116e-03, 8.3434e-05, -2.7620e-05,\n 4.8570e-05, -1.9765e-12, -1.9992e-05, -5.6363e-04, -1.4458e-04,\n 1.4438e-08, 5.6052e-45, 1.1052e-28, 2.3418e-28, -1.2132e-04,\n -1.2714e-05, -9.2795e-05, -1.6649e-04, 2.9925e-04, 5.6052e-45,\n -3.3937e-05, 5.6052e-45, -1.0752e-04, -1.6724e-04, -1.3534e-04,\n 3.9026e-04, 1.4994e-04, 5.6052e-45, -9.9074e-05, 6.4861e-05,\n 3.2268e-04, -3.2140e-07, 1.8217e-44, 3.2350e-24, 8.9012e-11,\n 1.6683e-04, 1.2794e-05, 1.8565e-04, 7.1517e-04, -2.7104e-04,\n 2.8276e-04, 5.6052e-45, 1.3167e-04, 2.6313e-04, -3.9169e-04,\n -1.1415e-04, 5.6052e-45, 4.6734e-05, -6.3397e-05, -2.5383e-06,\n 3.8966e-04, 5.6052e-45, 3.0111e-04, -3.4834e-04, 4.8336e-04,\n -4.9244e-04, 3.7109e-04, -2.0977e-04, 5.6052e-45, 5.6052e-45,\n 3.0751e-04, 7.8714e-07, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 7.1192e-05, -3.3616e-04, 2.8604e-04, 1.4790e-05, -8.3563e-05,\n 5.6052e-45, 7.7951e-05, -1.5048e-04, -6.7123e-04, 5.4905e-11,\n 5.6052e-45, 2.8351e-04, 7.1779e-06, 5.6052e-45, 8.5511e-05,\n 4.1778e-04, 1.6799e-04, -3.6826e-04, 6.0020e-12, 5.7071e-05,\n 4.0068e-04, -3.5485e-04, -2.8680e-04, -3.4258e-04, 1.4456e-04,\n -7.9463e-05, 2.8951e-13, -9.9962e-04, 9.8091e-45, -1.4106e-07,\n 4.6222e-05, -9.1472e-05, 5.6052e-45, 7.0233e-04, 5.6052e-45,\n 7.3851e-05, -6.0297e-05, -7.8765e-05, -1.4578e-04, 3.8607e-04,\n 3.4501e-04, 5.6052e-45, 5.6052e-45, 3.2170e-04, 3.3393e-05,\n 5.6052e-45, -9.2121e-05, 2.9594e-05, -1.0630e-19, -4.8263e-04,\n -4.9817e-04, -2.1474e-04, 1.5538e-04, 4.7020e-04, -6.3547e-05,\n -6.8207e-05, -9.1073e-05, 1.3091e-04, 2.1159e-04, 2.4608e-04,\n -1.0880e-03, -4.9823e-05, -4.4494e-05, -5.1705e-05, -1.8738e-04,\n -5.4606e-06, 1.3675e-04, -1.5774e-04, -1.0909e-03, -7.5536e-06,\n 5.6052e-45, -9.9769e-05, 4.9528e-05, 8.5745e-05, 3.3818e-04,\n 2.3689e-32, -2.5089e-04, 1.3137e-05, -2.9222e-05, 3.9141e-04,\n -7.5030e-05, 1.7190e-04, 2.6388e-04, -1.4062e-04, 3.9513e-24,\n 2.7132e-09, -7.4528e-05, -1.2651e-05, 1.5064e-42, 1.5316e-04,\n 2.4685e-04, 1.7313e-05, -8.7868e-05, 1.6602e-04, -5.1641e-08,\n -9.4725e-04, 4.4381e-33, -1.7365e-04, -2.2353e-04, 1.8567e-04,\n -9.2060e-05, -2.1319e-05, -6.2851e-05, -9.9499e-05, 6.4591e-05,\n -3.3036e-20, -6.8205e-05, 5.6052e-45, 1.3440e-04, -6.7255e-05,\n -9.4208e-05, 6.2517e-05, 3.3733e-05, -2.4861e-04, -2.4551e-04,\n 5.6052e-45, -8.3394e-05, 4.9821e-04, 2.5486e-04, 4.9623e-07,\n 6.6697e-05, 1.0025e-04, -2.7777e-04, -4.3108e-04, 9.7762e-06,\n 5.6052e-45, -1.8240e-05, 2.6031e-04, -8.4522e-05, 4.4094e-04,\n 5.6052e-45, 1.9444e-04, 8.8964e-23, 3.0764e-04, 1.7529e-04,\n -8.1915e-05, 2.2594e-04, 5.6052e-45, 4.6705e-04, 5.6052e-45,\n -3.4266e-04, -3.3431e-05, 2.1835e-04, 5.6052e-45, 2.0551e-04,\n 7.6803e-05, -2.4565e-04, -1.0524e-03, -2.0234e-04, 2.5220e-04,\n -1.4348e-04, 5.7707e-19, 1.6536e-05, -8.4527e-05, 5.6052e-45,\n 1.5139e-04, -2.0999e-04, 1.1975e-04, -2.1019e-04, -2.5023e-05,\n 6.7035e-05, 5.6052e-45, 9.4806e-05, 1.4627e-04, -4.8571e-05,\n 1.6756e-05, 1.0664e-04, 1.0193e-04, -7.7285e-06, -1.4664e-05,\n 9.5082e-05, 2.7579e-04, -3.3822e-04, 3.7190e-04, 1.8629e-04,\n -1.0317e-04, -1.3995e-05, 1.4553e-24, 6.9384e-04, 1.9461e-04,\n -3.0829e-05, -2.7351e-04, -3.0769e-06, -1.7694e-20, -4.8194e-04,\n 5.6052e-45, 1.6847e-04, 1.5109e-41, -2.3231e-04, 2.8603e-04,\n -2.0266e-04, 6.6389e-05, -9.3772e-05, 2.7580e-04, -3.4778e-05,\n 1.4767e-04, 5.6052e-45, 1.5654e-04, -2.2580e-05, -2.6966e-09,\n 8.6333e-05, 5.6052e-45, 5.6052e-45, 5.6052e-45, 6.6889e-05,\n 4.2646e-05, 5.6052e-45, -7.7037e-05, 3.5572e-04, 1.0305e-04,\n -2.8835e-04, 5.6052e-45, 3.4140e-05, 5.6052e-45, -1.0390e-05,\n -2.1077e-04, 1.5674e-05, 7.9610e-05, 3.5305e-04, -1.2788e-04,\n -6.8740e-05, 3.4249e-05, -1.8854e-04, 1.3930e-05, 2.3807e-05,\n -2.7075e-04, 3.4273e-04, 4.0979e-04, -1.2544e-04, 1.4097e-11,\n 4.4107e-04, 5.6052e-45, -8.7699e-06, 5.6052e-45, -3.8756e-05,\n -2.9300e-04, -8.4199e-05, 1.9581e-04, -1.3556e-04, 5.6052e-45,\n 2.4229e-04, 5.6052e-45, -1.0273e-04, 1.8812e-04, 1.6293e-04,\n 5.9536e-06, -1.6237e-04, 1.5821e-17, 5.6052e-45, 5.6052e-45,\n -2.1835e-04, 2.1701e-04, -2.3411e-04, 3.7574e-24, -4.9357e-04,\n 5.6052e-45, 6.1089e-04, 8.7090e-05, -7.2234e-05, -1.7241e-04,\n 4.5766e-08, 1.6947e-04, 5.6052e-45, 5.6281e-05, -5.6052e-45,\n 1.4642e-04, -1.6603e-04, -8.2690e-05, 5.6052e-45, 9.6808e-05,\n -2.9264e-04, 2.2817e-05, 4.3639e-04, 4.4569e-04, 5.7497e-05,\n -7.4553e-05, 5.6052e-45, -2.3102e-04, 1.0971e-04, 5.6052e-45,\n 1.1701e-04, 4.3481e-05, -1.0202e-05, -1.7046e-04, -1.6170e-04,\n 4.5094e-05, 6.3836e-04, 1.3358e-04, -1.0849e-04, -1.4310e-04,\n 5.6052e-45, -2.9752e-04, -8.1707e-05, 1.8192e-04, 5.6052e-45,\n 1.4066e-04, -1.1180e-04, 5.3549e-05, -7.0059e-05, -1.9310e-04,\n -1.1519e-04, -5.1895e-05, 5.6052e-45, -2.0593e-04, 5.6052e-45,\n -1.9151e-05, -3.2400e-04, -1.4885e-05, 2.6205e-04, 6.2573e-06,\n -1.2601e-04, 1.2892e-04, -3.9843e-05, 2.5519e-05, -2.7967e-05,\n 6.1098e-06, -1.5548e-04, -1.1650e-14, 2.0624e-04, 5.6052e-45,\n -3.3741e-04, 5.6052e-45, -1.5479e-06, 1.1650e-04, 2.4428e-04,\n 1.3543e-04, 1.4263e-04, 6.5750e-05, -1.8114e-04, 1.4803e-04,\n -1.9102e-04, 1.1764e-04, 2.6771e-04, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 1.1430e-04, -6.6519e-05], device='cuda:0')", - "exp_avg_sq": "tensor([1.2342e-06, 2.3553e-07, 7.4301e-07, 4.2122e-07, 3.1955e-06, 2.1949e-06,\n 4.0330e-06, 7.5430e-06, 2.1155e-06, 3.0190e-06, 2.0372e-06, 1.2122e-06,\n 7.9250e-07, 2.6306e-06, 1.1175e-06, 7.4447e-07, 5.6495e-06, 1.3400e-06,\n 1.6636e-07, 4.6131e-06, 2.7506e-08, 5.6136e-07, 2.2238e-06, 1.4625e-06,\n 2.6214e-06, 9.6434e-07, 7.0657e-07, 4.9843e-09, 2.2368e-06, 2.2791e-06,\n 3.2826e-07, 7.7211e-07, 1.5296e-06, 5.0952e-06, 1.0137e-06, 1.7005e-07,\n 2.6768e-06, 3.0430e-06, 1.2018e-06, 2.0201e-06, 6.1835e-06, 6.2264e-07,\n 2.0084e-06, 1.4445e-06, 2.8741e-06, 2.3127e-06, 7.6381e-07, 6.9796e-07,\n 1.0215e-05, 1.1382e-06, 4.2752e-06, 2.8962e-05, 1.5605e-06, 1.2622e-06,\n 3.6828e-06, 2.6060e-06, 5.9429e-06, 2.5466e-07, 1.3238e-06, 1.4978e-06,\n 2.7682e-08, 1.2220e-06, 3.4751e-06, 1.8988e-06, 2.0474e-07, 9.2864e-07,\n 2.1289e-06, 4.6202e-06, 2.3219e-06, 5.5272e-07, 1.2912e-06, 1.9449e-06,\n 4.1527e-07, 1.9172e-06, 9.4251e-07, 7.9661e-07, 1.3499e-08, 1.3153e-06,\n 2.4820e-07, 3.6930e-07, 5.6352e-06, 7.8537e-07, 3.5317e-06, 3.7269e-08,\n 1.1346e-05, 8.8569e-07, 2.0363e-06, 3.2290e-06, 7.0268e-07, 4.1262e-06,\n 8.6875e-07, 1.2242e-06, 1.0988e-07, 1.5875e-06, 2.5384e-07, 4.5176e-06,\n 4.4649e-06, 1.6223e-05, 3.0349e-06, 2.0404e-06, 2.9235e-06, 2.3948e-06,\n 9.2525e-06, 1.1088e-05, 1.4217e-06, 2.1893e-06, 6.8731e-07, 1.0942e-06,\n 1.2270e-06, 1.7645e-07, 1.0958e-06, 5.9989e-06, 1.8585e-06, 1.8697e-06,\n 4.0439e-06, 3.1063e-06, 1.0261e-06, 2.3009e-07, 1.8130e-06, 1.0204e-06,\n 2.7011e-06, 4.7989e-06, 1.6207e-06, 2.5569e-06, 1.2296e-06, 1.6923e-06,\n 4.1428e-06, 8.1518e-08, 2.1705e-06, 1.1381e-05, 1.8709e-06, 2.1573e-06,\n 1.1169e-07, 1.2208e-06, 1.1968e-06, 5.7163e-07, 4.5208e-10, 3.3164e-06,\n 8.9651e-07, 1.8343e-06, 1.1261e-06, 6.1926e-07, 4.7246e-07, 1.7635e-10,\n 2.3992e-09, 3.1913e-06, 1.9731e-06, 2.1343e-06, 8.5741e-06, 1.2596e-07,\n 1.5757e-07, 8.3277e-07, 2.4885e-07, 5.1176e-07, 7.3506e-09, 1.9737e-06,\n 2.7607e-06, 2.8011e-06, 3.4650e-06, 4.3590e-06, 2.7378e-07, 1.2630e-06,\n 1.1960e-06, 1.4520e-06, 1.8688e-06, 1.3642e-06, 1.3558e-06, 4.5111e-12,\n 1.1397e-06, 4.9661e-07, 4.1134e-06, 3.6591e-06, 9.3414e-07, 1.1707e-06,\n 1.8983e-06, 8.3217e-07, 4.6129e-10, 1.3218e-06, 3.4356e-06, 8.1310e-06,\n 3.5847e-06, 3.1714e-06, 1.4368e-06, 1.0372e-10, 1.8939e-10, 1.0894e-16,\n 2.5394e-06, 8.3050e-07, 7.3964e-08, 4.4076e-09, 1.2834e-06, 3.5154e-07,\n 1.9335e-06, 4.6954e-06, 2.7665e-06, 1.0292e-06, 1.3930e-06, 2.6483e-06,\n 5.0426e-07, 2.8870e-17, 3.4264e-06, 8.8617e-07, 2.0287e-06, 2.8348e-06,\n 1.7927e-06, 4.1582e-07, 3.0447e-06, 2.2217e-06, 1.1918e-06, 2.1640e-08,\n 2.2302e-06, 3.0361e-08, 9.6797e-07, 3.2734e-06, 1.8114e-06, 8.6986e-07,\n 1.4335e-06, 5.3629e-07, 2.2100e-06, 9.7094e-07, 1.0494e-06, 2.5223e-06,\n 1.5104e-06, 6.4713e-07, 5.0279e-06, 4.2745e-06, 9.5186e-06, 2.0408e-06,\n 2.0141e-06, 8.5056e-08, 1.3523e-06, 1.7987e-06, 1.3589e-06, 1.0986e-06,\n 2.7048e-06, 8.3487e-07, 5.1013e-06, 5.1290e-10, 9.5415e-07, 1.4967e-06,\n 1.7645e-06, 1.4070e-06, 2.2859e-06, 1.0473e-06, 6.2103e-07, 9.7371e-07,\n 6.4394e-07, 3.6812e-06, 4.5378e-07, 1.7947e-06, 8.2648e-06, 2.0629e-06,\n 3.4719e-06, 4.1655e-06, 8.9826e-07, 2.1467e-06, 5.4853e-10, 1.0637e-06,\n 4.0699e-15, 7.3606e-06, 3.2311e-15, 2.0414e-06, 1.3360e-06, 1.7683e-06,\n 7.6960e-06, 3.4635e-06, 4.4732e-07, 1.2941e-06, 4.9409e-06, 6.3597e-06,\n 2.8984e-06, 7.0937e-07, 1.0557e-06, 3.1518e-06, 2.7082e-06, 1.0173e-06,\n 2.7312e-06, 2.7974e-09, 1.4046e-07, 5.1184e-06, 7.9771e-06, 4.5327e-06,\n 4.6466e-06, 1.9182e-06, 3.1384e-06, 3.4927e-07, 1.9813e-06, 1.3051e-06,\n 1.6461e-09, 1.0402e-05, 3.6370e-07, 3.9338e-06, 1.0007e-05, 5.4587e-08,\n 1.4431e-06, 9.8578e-06, 2.9241e-06, 8.1838e-06, 1.1636e-06, 3.1689e-08,\n 7.3654e-07, 9.3861e-06, 1.0931e-07, 1.0095e-08, 2.9935e-07, 6.4418e-09,\n 7.9273e-08, 4.8544e-07, 1.6774e-06, 3.7217e-06, 4.1799e-06, 4.1914e-10,\n 1.5523e-06, 3.9449e-07, 8.8361e-07, 9.1999e-10, 3.5598e-09, 1.5150e-05,\n 1.8247e-06, 5.4027e-07, 2.1855e-06, 2.3093e-07, 3.3966e-10, 1.7560e-06,\n 4.5349e-07, 3.3377e-06, 6.7455e-06, 1.2701e-06, 1.4549e-06, 1.0958e-06,\n 1.6948e-06, 4.3207e-07, 2.2138e-06, 5.5533e-07, 3.2284e-08, 7.2661e-06,\n 7.3374e-07, 5.3816e-07, 2.5069e-15, 4.2402e-09, 5.2093e-06, 1.8741e-06,\n 1.1191e-05, 6.1929e-07, 3.8158e-06, 1.1430e-06, 1.5517e-06, 2.5159e-05,\n 9.2608e-07, 3.3793e-06, 2.4856e-06, 1.9730e-06, 1.7131e-07, 6.8248e-07,\n 8.2455e-07, 1.9525e-06, 5.9412e-09, 7.1935e-07, 6.0823e-07, 1.4402e-05,\n 1.9038e-06, 1.4084e-06, 2.8303e-06, 4.6445e-06, 1.0350e-06, 9.2705e-07,\n 3.6169e-07, 2.5135e-06, 7.7473e-08, 6.5724e-08, 1.2564e-06, 1.4505e-06,\n 4.2685e-07, 7.3962e-07, 9.4232e-07, 6.2950e-06, 4.9274e-07, 1.7835e-09,\n 1.3438e-06, 6.0871e-07, 1.8644e-06, 2.2317e-06, 7.4971e-07, 1.4943e-06,\n 2.8083e-06, 8.2071e-07, 7.2095e-09, 1.0187e-06, 4.5966e-06, 8.7032e-07,\n 3.4520e-06, 1.5587e-06, 1.9376e-09, 1.5169e-07, 2.1828e-06, 3.2811e-06,\n 5.3995e-08, 1.1754e-06, 9.6584e-06, 1.2510e-06, 5.4765e-06, 7.7861e-07,\n 1.8883e-06, 3.6099e-06, 1.6932e-06, 1.0361e-06, 8.1549e-06, 3.4259e-06,\n 2.3535e-06, 2.3915e-06, 3.6304e-06, 1.7386e-11, 3.4499e-06, 1.5176e-06,\n 1.0764e-07, 9.0875e-07, 1.6489e-09, 1.9576e-06, 5.0778e-07, 6.8152e-07,\n 3.3342e-06, 2.4314e-06, 1.7620e-06, 1.1689e-09, 7.7133e-17, 1.4296e-07,\n 2.4895e-06, 6.4866e-07, 1.7833e-09, 3.7294e-09, 1.6310e-06, 1.0366e-06,\n 3.8879e-06, 1.2527e-06, 3.5171e-06, 4.1335e-07, 1.3027e-06, 1.6604e-07,\n 1.8134e-06, 9.7241e-09, 1.4822e-06, 1.2445e-06, 9.6486e-09, 1.1792e-08,\n 6.0053e-07, 1.9073e-06, 9.7377e-07, 8.9003e-07, 9.6927e-10, 2.2416e-07,\n 9.8008e-07, 1.1416e-06, 2.0961e-06, 1.4093e-06, 1.7170e-06, 3.7436e-06,\n 6.4417e-07, 2.8944e-06, 3.6545e-07, 2.3491e-07, 1.0743e-09, 4.5410e-06,\n 8.9035e-11, 1.2580e-06, 7.6673e-07, 1.2939e-06, 2.3281e-07, 6.0000e-07,\n 3.6705e-06, 2.3029e-07, 1.0003e-06, 3.6043e-08, 2.2441e-08, 1.3760e-06,\n 6.0573e-08, 1.6006e-09, 4.1849e-07, 3.5218e-07, 8.0797e-06, 2.4469e-06,\n 2.2216e-06, 2.8469e-06, 1.5181e-06, 3.7686e-06, 8.0410e-07, 1.6776e-06,\n 3.2533e-06, 3.7778e-06, 1.0898e-06, 1.2510e-06, 2.8922e-06, 1.2033e-07,\n 2.8669e-06, 3.2232e-06, 8.4868e-07, 3.9920e-06, 1.6553e-06, 3.9551e-06,\n 1.5488e-06, 1.6749e-06, 1.9294e-05, 2.3888e-06, 1.0952e-06, 5.5434e-08,\n 2.4100e-06, 2.1610e-06, 9.5865e-07, 1.6894e-06, 4.5556e-07, 3.0249e-06,\n 2.6194e-06, 1.3612e-06, 9.6548e-07, 4.0148e-06, 1.1920e-07, 1.0298e-06,\n 6.8147e-07, 5.6453e-07, 2.6725e-06, 1.9617e-06, 2.0231e-06, 1.5765e-07,\n 1.4835e-06, 1.2730e-06, 5.8101e-07, 5.8430e-07, 7.6578e-11, 1.9978e-06,\n 9.6193e-07, 8.0785e-07, 2.4777e-06, 3.9382e-06, 1.8740e-06, 2.9094e-06,\n 6.3963e-07, 3.0339e-07, 3.4361e-06, 8.9237e-06, 2.7063e-06, 9.1326e-07,\n 2.1668e-06, 8.9477e-07, 3.0268e-06, 3.1117e-06, 1.0998e-05, 3.2109e-08,\n 1.3875e-06, 1.2396e-06, 1.5395e-06, 1.4016e-05, 5.2351e-06, 1.1303e-06,\n 2.7549e-06, 2.0478e-06, 3.5359e-07, 2.0251e-05, 6.5283e-07, 2.3265e-06,\n 4.2137e-07, 2.4615e-06, 5.4222e-10, 6.8261e-06, 2.6821e-07, 9.9649e-07,\n 2.7931e-06, 7.6039e-07, 4.8441e-07, 1.2724e-08, 4.8742e-06, 6.2060e-07,\n 6.1465e-07, 2.1435e-07, 1.1846e-06, 5.3122e-07, 2.8448e-06, 6.1203e-06,\n 1.5231e-06, 1.8601e-07, 1.9391e-06, 1.2727e-06, 2.5766e-06, 1.6111e-09,\n 7.5429e-08, 1.6444e-06, 1.5968e-13, 2.6782e-06, 1.3129e-06, 2.0987e-06,\n 6.8217e-07, 2.8695e-06, 5.1171e-07, 2.4898e-10, 2.5211e-06, 3.1402e-06,\n 1.6797e-06, 1.1125e-06, 1.0601e-06, 5.2986e-07, 9.6843e-07, 3.1576e-06,\n 4.1137e-07, 2.7210e-06, 2.3700e-06, 4.9400e-07, 4.9038e-07, 3.2205e-06,\n 9.0888e-07, 6.0443e-09, 1.9733e-06, 7.5287e-07, 2.0740e-06, 3.8841e-07,\n 1.6077e-06, 4.4487e-06, 3.2210e-06, 1.1261e-09, 1.7682e-06, 1.3066e-06,\n 7.4057e-07, 2.7610e-06, 2.0111e-06, 1.6941e-06, 4.3709e-06, 3.0512e-06,\n 4.6203e-06, 1.7072e-06, 1.1172e-10, 2.1940e-06, 8.7076e-07, 3.8197e-07,\n 1.0159e-06, 4.2075e-07, 2.9333e-08, 9.1423e-10, 6.5351e-07, 7.7729e-06,\n 3.4609e-12, 1.5363e-06, 1.2570e-06, 1.2254e-06, 4.2506e-06, 3.7096e-10,\n 2.6076e-06, 9.6842e-06, 4.3476e-09, 1.4353e-06, 2.0991e-06, 7.5750e-07,\n 5.3920e-06, 6.4765e-07, 1.6487e-06, 2.2763e-06, 1.0748e-06, 1.8438e-06,\n 8.8718e-07, 8.7676e-07, 3.3009e-06, 5.6425e-07, 6.0389e-06, 8.4232e-08,\n 4.0337e-06, 8.4054e-06, 2.7262e-06, 1.0005e-05, 1.0007e-06, 6.8124e-06,\n 2.5241e-06, 1.3547e-07, 5.2569e-06, 1.3317e-07, 1.4245e-06, 3.2605e-11,\n 6.3829e-07, 5.4756e-07, 1.3543e-06, 5.3845e-07, 2.5300e-06, 1.6444e-07,\n 1.4661e-06, 3.5104e-09, 2.1672e-06, 2.3464e-06, 2.4530e-06, 8.3048e-06,\n 1.2423e-06, 1.5335e-06, 2.5051e-06, 2.6706e-07, 1.2984e-06, 4.9161e-06,\n 1.3224e-08, 7.3419e-08, 1.2588e-05, 5.6375e-08, 7.4361e-08, 9.8109e-07,\n 3.6097e-06, 2.0040e-06, 9.9640e-10, 3.3286e-06, 1.6219e-06, 8.1969e-07,\n 2.0379e-06, 2.4106e-06, 2.6993e-06, 1.6373e-06, 7.9324e-09, 2.9212e-06,\n 6.6845e-07, 6.2520e-11, 1.6803e-06, 5.5341e-07, 9.8433e-07, 3.8181e-06,\n 2.4972e-06, 1.0685e-07, 1.1707e-06, 7.6252e-07, 3.6902e-07, 3.3998e-06,\n 5.5794e-06, 4.1344e-06, 1.4437e-06, 1.4553e-06, 8.4862e-06, 1.1338e-07,\n 1.2819e-06, 3.1262e-06, 1.8125e-06, 4.5328e-06, 7.7319e-07, 8.8453e-06,\n 6.4052e-11, 1.1625e-06, 7.4907e-06, 3.9933e-07, 3.3975e-06, 6.0516e-07,\n 6.7741e-07, 2.8122e-07, 1.4127e-06, 4.7937e-06, 1.5535e-05, 1.7112e-06,\n 3.4087e-07, 3.5979e-07, 6.6651e-06, 4.9908e-06, 2.3517e-06, 1.4996e-05,\n 1.2469e-06, 6.6386e-09, 6.1552e-07, 9.3368e-07, 6.8865e-07, 2.5898e-06,\n 2.6155e-07, 1.2554e-06, 8.9826e-07, 1.2451e-06, 1.3213e-06, 4.1140e-07,\n 3.4265e-06, 6.7042e-06, 7.9404e-09, 1.2002e-15, 5.4385e-06, 3.3372e-07],\n device='cuda:0')" + "step": "tensor(3756.)", + "exp_avg": "tensor([-2.7197e-04, 5.6052e-45, 6.1862e-06, -2.9774e-04, 2.9000e-04,\n 5.6052e-45, 1.5417e-04, -5.0991e-05, -2.9012e-05, -4.8245e-04,\n 1.2664e-04, -8.3094e-23, 1.3998e-04, -2.4330e-05, 1.9183e-04,\n -8.5183e-05, -2.9388e-05, 3.1215e-11, 2.2819e-05, -1.6405e-04,\n 1.0840e-12, -2.2573e-05, 2.5321e-05, 4.0705e-05, 1.6067e-04,\n 8.7898e-05, 2.1385e-05, 5.6052e-45, 5.9112e-05, -1.4310e-05,\n -5.6351e-06, 7.4121e-05, 5.4636e-05, -7.1182e-05, -9.7100e-05,\n -2.2994e-04, 5.6052e-45, -7.5110e-08, 4.1827e-05, -1.8324e-06,\n 2.1942e-04, -9.5022e-05, -1.5455e-04, 7.6379e-05, -1.6164e-04,\n 5.5655e-06, 1.0588e-04, 1.2678e-04, 5.1292e-05, -1.2479e-04,\n -1.6342e-35, -1.7196e-04, -6.5582e-05, -7.1310e-05, 1.3368e-04,\n 8.7433e-05, -4.4700e-06, 9.4420e-05, -7.3208e-05, -3.2428e-05,\n 4.9684e-05, -2.5186e-05, 8.6725e-05, -5.5476e-05, 3.7623e-05,\n -1.0068e-04, 3.6693e-05, 2.1892e-04, 5.6052e-45, 1.0065e-28,\n 1.8233e-04, 5.6052e-45, 1.7627e-04, 9.6067e-05, 1.6458e-04,\n 7.3967e-06, 5.6052e-45, -2.4910e-04, 2.5265e-05, -1.9294e-05,\n -4.6444e-06, 1.6491e-04, 1.8365e-04, 2.4745e-13, -8.0199e-04,\n -1.1084e-04, -2.0079e-04, -8.7441e-05, 2.9475e-04, -1.3690e-04,\n 6.7087e-05, -1.5930e-04, -3.6879e-06, -2.5063e-04, 8.6976e-05,\n -3.1316e-05, 1.9400e-04, -1.4038e-08, 1.8542e-05, 1.0500e-04,\n 1.1307e-04, -1.2929e-05, -5.4928e-05, 1.1323e-04, 2.0373e-04,\n 3.8913e-06, 4.9983e-05, 1.3431e-04, -1.6120e-05, -1.2842e-05,\n -5.3360e-05, -1.0518e-04, 4.9170e-05, 5.6052e-45, -2.7246e-04,\n 8.3844e-05, 2.9317e-05, -1.0397e-04, -5.0215e-06, 5.6052e-45,\n 3.7516e-05, 7.3719e-05, -4.2309e-05, 5.6052e-45, 7.9699e-05,\n 3.0156e-05, -1.7834e-04, -1.9200e-23, 1.0121e-05, 2.0687e-17,\n 2.4977e-04, 3.7485e-05, 5.6052e-45, -2.7051e-04, 1.6674e-13,\n 1.8268e-04, 5.6052e-45, -1.4405e-04, -2.7886e-04, -2.5406e-09,\n 5.6052e-45, 3.3899e-28, 4.6283e-05, 2.8794e-11, 5.6052e-45,\n 5.6052e-45, 9.9085e-05, -1.4151e-04, 5.6052e-45, -2.7200e-05,\n -1.5855e-05, -8.8834e-05, 4.9665e-05, -1.1635e-04, 5.6052e-45,\n 1.8282e-04, -7.7508e-05, -6.1249e-05, 5.6052e-45, -2.0261e-04,\n 1.0973e-04, 2.8602e-05, 1.8772e-04, 4.8478e-05, 3.2414e-05,\n -1.6312e-05, -3.6639e-13, 5.6052e-45, -1.0602e-04, 1.4927e-04,\n -3.5654e-05, 6.7847e-05, -6.3458e-05, 7.0233e-06, 3.2258e-05,\n -2.9439e-05, -1.1037e-07, 2.4340e-04, 1.0282e-04, 3.8703e-05,\n -9.1587e-05, 1.6932e-04, -3.6300e-05, 5.6052e-45, 2.3114e-11,\n 5.6052e-45, 6.8321e-05, -1.3482e-05, 1.5178e-05, 1.5293e-05,\n 6.8559e-05, -7.1032e-06, 1.8499e-04, 1.3980e-04, -2.6252e-05,\n -5.1081e-06, 3.6192e-05, 2.4450e-05, -9.5711e-05, 5.6052e-45,\n -1.3945e-05, 8.0197e-06, 3.0071e-04, -1.4174e-04, 1.5901e-04,\n -5.3548e-05, -3.4124e-04, 5.1332e-05, 2.5218e-30, 5.6052e-45,\n 3.7048e-05, 5.6052e-45, 8.9936e-05, 9.7511e-05, 1.8544e-05,\n -2.1783e-05, -1.2728e-04, -9.7020e-05, 1.9047e-05, -2.7507e-04,\n -5.8108e-05, -2.4093e-04, -4.6631e-05, 2.5515e-05, 7.9329e-05,\n 5.6052e-45, -1.7058e-04, -2.3092e-05, 2.5453e-04, 5.6052e-45,\n 2.2804e-04, 1.9779e-04, -2.2965e-05, -5.1846e-05, 5.6052e-45,\n -3.8675e-05, -1.9136e-05, 5.6052e-45, 6.2991e-05, 8.9380e-05,\n -4.7784e-05, 5.6052e-45, 1.6830e-04, 8.9307e-05, 6.5347e-06,\n -6.1316e-05, 1.1459e-04, -3.0487e-04, 3.1000e-05, 5.1483e-15,\n 8.2123e-05, 8.2426e-05, -2.4996e-04, -7.0940e-05, 2.2304e-05,\n 2.1345e-04, 5.6052e-45, -1.4850e-05, 5.6052e-45, -4.7127e-05,\n 5.6052e-45, 8.6978e-06, 1.9905e-04, -3.2898e-05, -1.2458e-04,\n 1.1960e-05, -1.2238e-04, 9.1834e-06, 2.6696e-04, -5.6052e-45,\n 3.1820e-04, -4.6181e-05, -6.7187e-05, -1.7674e-04, 2.2346e-04,\n 1.6948e-04, -1.4031e-04, 5.6052e-45, 5.9813e-05, -1.0007e-05,\n 5.6052e-45, 1.2687e-04, -4.2957e-05, 1.3414e-04, 2.5318e-05,\n 3.0422e-05, -1.2204e-04, -9.9087e-05, 3.8138e-05, 2.7195e-05,\n 1.1470e-05, 2.5279e-05, -6.5930e-17, 3.1055e-05, 1.5536e-04,\n -8.9411e-05, 1.3547e-04, 2.1328e-05, -7.0201e-05, -6.2227e-05,\n 2.0935e-05, -1.8993e-05, -3.1796e-05, -6.9201e-05, -6.7764e-19,\n 5.6052e-45, -1.1405e-07, 7.5524e-05, 2.0957e-05, 4.1945e-04,\n 7.0866e-05, 5.6052e-45, -5.8064e-04, 3.2967e-05, 1.0483e-04,\n 5.6052e-45, 8.4335e-06, -7.4314e-05, -1.0994e-04, -8.6618e-05,\n -4.1244e-05, -3.6992e-05, 2.1660e-05, 1.3446e-04, 6.7636e-05,\n 5.6052e-45, -2.3004e-04, -9.1055e-05, 3.8643e-05, 6.9657e-05,\n 2.2690e-04, 1.6341e-04, -2.4598e-05, 5.6057e-05, -2.0569e-04,\n 4.9524e-05, -8.1786e-05, 1.8336e-12, 5.6052e-45, 2.7056e-07,\n -1.9340e-04, -8.7592e-05, 9.5886e-05, -3.6200e-05, -1.2018e-04,\n -5.6607e-06, 3.6047e-05, 9.7618e-06, -2.0907e-04, 1.5156e-04,\n 1.0126e-04, 1.8733e-05, -1.6791e-04, 5.6052e-45, -1.3523e-04,\n 3.1359e-04, 4.0186e-05, -8.7976e-10, 1.5100e-04, -1.6622e-05,\n 2.8341e-04, 2.0688e-05, 2.5014e-32, -8.8644e-05, 2.0949e-04,\n -7.4947e-07, -7.6747e-05, 5.0886e-05, 5.1182e-05, -7.4682e-05,\n 9.6326e-05, -3.0874e-12, -1.6551e-04, 1.9047e-05, 1.0544e-04,\n -2.1925e-05, -6.5996e-09, 7.2938e-07, 2.9194e-19, 1.0881e-05,\n 5.9383e-05, -1.1777e-04, -3.3133e-05, 4.7211e-05, 5.6052e-45,\n -2.3213e-05, 5.6052e-45, -3.1118e-04, 4.9845e-06, -5.2574e-04,\n 4.0634e-05, -4.1606e-04, -8.4469e-05, 9.4994e-05, -2.9859e-05,\n 7.0320e-05, -2.2820e-09, 3.5886e-36, 5.8348e-16, -3.7702e-26,\n -4.5789e-05, 1.4645e-04, -9.3109e-06, -1.0897e-05, 6.3990e-05,\n 1.9604e-04, 1.0568e-18, -1.8391e-04, 5.2837e-05, -1.5669e-04,\n 1.1816e-05, 5.6052e-45, 1.3155e-05, 3.2457e-05, 1.1492e-06,\n -2.3563e-04, 2.4960e-27, 9.9720e-05, 2.6914e-05, 2.8514e-05,\n 4.7296e-05, -1.5961e-05, -1.8248e-04, 4.6205e-28, 5.6052e-45,\n 4.9044e-05, 2.6359e-05, 5.6052e-45, 5.0447e-44, 5.6052e-45,\n -1.9330e-04, -7.7495e-05, 9.9997e-06, -5.7853e-05, 2.1513e-04,\n 2.4750e-32, -8.4590e-05, -4.7036e-05, 7.2286e-05, 1.6814e-37,\n 6.0902e-09, -1.4020e-04, 2.1939e-06, 5.6052e-45, 8.4410e-05,\n -5.4605e-05, -1.0199e-04, -4.8959e-05, -6.2225e-05, 7.0820e-05,\n -2.4949e-04, -1.8071e-04, 3.1417e-04, 9.3810e-05, -1.8540e-04,\n -3.6795e-04, 2.2137e-13, -6.1994e-05, 5.6052e-45, -1.5517e-05,\n 5.7795e-05, 1.3899e-05, 5.6052e-45, -1.7321e-05, 5.6052e-45,\n -1.5170e-04, 8.5985e-05, -1.3560e-04, 6.6246e-05, 2.2613e-04,\n -6.4542e-05, 5.6052e-45, -5.1237e-10, 9.9993e-05, 5.6599e-05,\n 5.6052e-45, -2.4587e-04, 1.1440e-04, 4.5608e-05, 5.9847e-05,\n 6.7499e-05, 3.2050e-04, 2.8625e-04, 1.0268e-04, -5.2267e-05,\n 2.7519e-05, 3.9442e-05, -2.6537e-04, -1.9359e-04, 1.2529e-04,\n 5.0109e-05, -7.1149e-05, 3.2045e-04, 1.1430e-04, 7.7393e-05,\n 7.8939e-05, 7.9574e-05, 7.8275e-05, -8.9304e-05, -9.2056e-05,\n 5.6052e-45, 1.1124e-04, 1.6354e-04, -6.4147e-05, -1.1524e-04,\n -2.8964e-05, -5.1288e-05, 9.9683e-05, 1.0288e-04, -6.2678e-06,\n 1.0625e-04, -5.2388e-06, -1.9756e-05, -6.2773e-05, 3.8623e-05,\n -3.2490e-27, 5.2549e-05, 2.2645e-05, 5.6052e-45, 1.0446e-05,\n -1.1806e-04, 7.7663e-05, -2.0286e-04, -1.4781e-04, -4.3916e-07,\n 1.2635e-04, 5.6052e-45, 7.5277e-06, 3.0858e-05, 2.1860e-05,\n -1.4400e-04, -1.1795e-04, -2.2052e-04, 1.1407e-04, -6.0323e-05,\n 2.9382e-07, 1.0523e-04, 5.6052e-45, -8.6857e-05, -2.4203e-04,\n 4.6841e-05, -3.1884e-05, 8.5388e-05, -4.2835e-05, -3.5724e-05,\n -4.1905e-17, 7.5036e-05, 2.5012e-04, 3.9005e-04, -3.1379e-07,\n -1.4315e-04, 9.2852e-05, -4.9741e-05, -1.7292e-04, -1.8868e-04,\n 5.6052e-45, 7.4991e-05, 7.6131e-05, -9.2764e-05, -1.1851e-04,\n 5.6052e-45, 1.5513e-04, 5.6052e-45, 9.4664e-05, -1.3530e-04,\n -1.8565e-04, 9.2021e-06, 5.6052e-45, 1.4131e-04, 5.6052e-45,\n -3.4181e-05, 9.9309e-05, 1.9535e-04, 5.6052e-45, -1.2597e-05,\n -2.5968e-04, 2.1725e-04, -6.8999e-06, 1.9287e-05, -1.0047e-04,\n 2.4287e-04, -2.0028e-33, -5.9378e-05, -3.7998e-05, 1.6493e-39,\n 1.9648e-04, -2.8246e-05, 6.6649e-05, 3.9815e-05, 9.2276e-05,\n 2.0202e-04, 5.6052e-45, 1.1482e-04, 7.5223e-05, 3.0317e-05,\n 6.0908e-06, -7.4249e-05, -1.1544e-04, 6.8567e-05, -2.1952e-05,\n 3.1647e-04, -6.4769e-05, 2.3637e-04, -7.6074e-05, -1.6926e-04,\n 2.1360e-04, -1.4481e-04, -2.3779e-06, 1.1521e-05, -1.8595e-05,\n -1.2157e-04, 8.2667e-05, -1.3583e-04, 3.9066e-30, -1.7687e-05,\n 5.6052e-45, 1.8040e-04, 5.5298e-11, 1.5759e-04, 3.5618e-04,\n 1.9267e-04, -9.7537e-05, -1.3507e-04, -3.1299e-04, -8.7066e-05,\n 2.0294e-04, 8.2441e-15, 9.8858e-05, 7.6483e-05, -1.2555e-25,\n -4.1530e-05, 5.5543e-15, 5.6052e-45, 5.6052e-45, 8.3818e-05,\n 1.1757e-04, 5.6052e-45, 6.0960e-05, 7.0454e-05, -1.1895e-04,\n 6.1255e-05, 5.6052e-45, -5.6566e-05, 5.6052e-45, -2.4682e-04,\n -9.4397e-05, -2.6014e-04, 7.9740e-05, -9.1210e-05, -1.5157e-04,\n -4.8062e-05, -1.0497e-04, -2.5522e-04, -3.1553e-05, -5.2444e-05,\n 4.9597e-04, -1.1990e-04, 1.2030e-04, 6.6100e-05, -4.7959e-10,\n 1.2865e-04, 4.5508e-08, 4.2180e-04, -5.6052e-45, -3.3299e-05,\n 4.7784e-05, -5.7541e-05, -1.6514e-05, -1.5006e-04, 5.6052e-45,\n -1.1211e-04, 5.6052e-45, 9.0198e-05, 1.5240e-04, 1.6124e-04,\n -1.3589e-04, 1.4160e-04, -9.2734e-41, 5.6052e-45, 5.6052e-45,\n -3.8678e-05, -2.0310e-06, -7.0891e-05, 5.6052e-45, -1.9308e-04,\n 5.6052e-45, -1.9534e-06, -7.3181e-05, -5.1901e-05, 1.4131e-05,\n 1.3803e-42, 1.2441e-04, 5.6052e-45, -3.3919e-05, 5.6052e-45,\n 2.5494e-05, -2.2080e-05, 1.1969e-04, 5.6052e-45, 4.7694e-05,\n 2.3294e-04, -1.5167e-04, -9.1955e-05, -1.8953e-04, -1.0304e-04,\n -2.8700e-05, 5.6052e-45, -4.4384e-04, -5.4701e-05, 5.6052e-45,\n 2.3059e-05, -3.5691e-04, -1.3450e-04, 8.7911e-06, 1.3863e-05,\n -3.9465e-05, -2.9224e-04, 1.7523e-05, -1.5286e-04, -1.0216e-04,\n 5.6052e-45, 1.9340e-04, -4.3325e-05, 4.6669e-05, -2.3518e-10,\n -2.1366e-04, -1.6569e-04, -1.3570e-04, 2.3838e-04, 2.2108e-04,\n 1.5096e-04, 1.2108e-04, -3.3061e-05, -3.7859e-05, 5.6052e-45,\n -1.8971e-05, 1.1077e-04, -4.1044e-05, -1.4532e-04, 1.2329e-04,\n -1.9154e-04, -7.4180e-05, -1.9347e-05, -6.0849e-04, 1.7598e-05,\n -3.3117e-05, -1.8057e-04, 5.6052e-45, -1.8914e-04, -5.6052e-45,\n -2.2562e-05, -2.7832e-06, -1.2531e-05, 5.4463e-05, 2.1362e-04,\n -2.7446e-04, -1.7087e-04, 7.0099e-05, 1.4282e-05, 3.2202e-05,\n -2.0840e-05, -2.4095e-05, 5.4923e-05, -1.3832e-05, 5.6052e-45,\n 5.6052e-45, 1.8718e-04, 1.4130e-05], device='cuda:0')", + "exp_avg_sq": "tensor([5.8486e-07, 6.7305e-08, 2.1271e-07, 3.0549e-07, 1.3273e-06, 6.2720e-07,\n 1.5887e-06, 2.3324e-06, 7.2390e-07, 1.2560e-06, 8.7363e-07, 3.4640e-07,\n 4.6742e-07, 8.0218e-07, 5.8896e-07, 4.1006e-07, 1.9655e-06, 3.8291e-07,\n 5.0782e-08, 1.5670e-06, 7.8601e-09, 2.6791e-07, 7.2331e-07, 7.2412e-07,\n 1.1452e-06, 6.0510e-07, 3.1855e-07, 1.4243e-09, 9.1714e-07, 9.5761e-07,\n 2.0724e-07, 4.4289e-07, 6.4884e-07, 1.6348e-06, 4.7535e-07, 2.2039e-07,\n 7.6491e-07, 1.1045e-06, 6.0452e-07, 8.1290e-07, 2.0725e-06, 3.8307e-07,\n 9.1507e-07, 7.9083e-07, 1.1137e-06, 7.7224e-07, 4.5089e-07, 4.0657e-07,\n 3.4039e-06, 5.8783e-07, 1.2217e-06, 8.8047e-06, 6.6752e-07, 5.7676e-07,\n 1.2308e-06, 8.2569e-07, 1.7555e-06, 1.7820e-07, 3.9248e-07, 7.1190e-07,\n 7.6397e-08, 5.7588e-07, 1.5579e-06, 8.1114e-07, 1.5803e-07, 3.8394e-07,\n 1.0286e-06, 1.4942e-06, 6.6350e-07, 1.5794e-07, 6.3893e-07, 5.5578e-07,\n 2.7309e-07, 1.2165e-06, 5.5716e-07, 5.4762e-07, 3.8576e-09, 7.3741e-07,\n 2.0749e-07, 2.9355e-07, 1.6103e-06, 4.9930e-07, 1.4187e-06, 1.0650e-08,\n 3.5276e-06, 6.3142e-07, 9.2966e-07, 1.3082e-06, 4.4710e-07, 1.5247e-06,\n 4.5413e-07, 5.0332e-07, 1.6961e-07, 7.6558e-07, 2.1276e-07, 1.7768e-06,\n 1.5160e-06, 4.6359e-06, 1.0665e-06, 8.9231e-07, 1.0957e-06, 1.0281e-06,\n 3.1427e-06, 3.3817e-06, 6.7145e-07, 9.1689e-07, 3.5746e-07, 5.3008e-07,\n 5.8143e-07, 1.6473e-07, 5.7906e-07, 2.0707e-06, 7.9360e-07, 5.3427e-07,\n 1.5583e-06, 1.1565e-06, 5.3022e-07, 2.0895e-07, 5.3916e-07, 2.9158e-07,\n 9.9888e-07, 1.8331e-06, 5.4202e-07, 7.3067e-07, 5.9543e-07, 8.1662e-07,\n 1.3319e-06, 2.3294e-08, 1.0184e-06, 3.2522e-06, 7.8704e-07, 7.8408e-07,\n 3.1917e-08, 6.9062e-07, 3.4200e-07, 4.0409e-07, 1.2934e-10, 1.1698e-06,\n 5.1154e-07, 5.2417e-07, 3.2179e-07, 1.7696e-07, 3.0267e-07, 5.0393e-11,\n 6.8559e-10, 9.1193e-07, 9.1115e-07, 9.5189e-07, 2.4501e-06, 1.6752e-07,\n 1.9600e-07, 5.6579e-07, 2.3485e-07, 3.4101e-07, 2.1005e-09, 8.4503e-07,\n 7.9377e-07, 1.0071e-06, 9.9014e-07, 1.3660e-06, 3.2417e-07, 3.8327e-07,\n 6.9487e-07, 8.5898e-07, 6.7081e-07, 6.1049e-07, 3.8819e-07, 1.2891e-12,\n 6.0883e-07, 3.4680e-07, 1.4388e-06, 1.3092e-06, 5.0973e-07, 6.5577e-07,\n 8.4424e-07, 4.9309e-07, 1.3300e-10, 6.6550e-07, 1.3847e-06, 2.6365e-06,\n 1.4983e-06, 1.1356e-06, 6.9869e-07, 2.9638e-11, 5.4145e-11, 3.1130e-17,\n 1.1904e-06, 5.0787e-07, 9.1074e-08, 2.2184e-08, 5.7910e-07, 2.8942e-07,\n 8.9419e-07, 1.6842e-06, 1.0801e-06, 4.3451e-07, 8.4985e-07, 1.1028e-06,\n 3.7253e-07, 8.2499e-18, 9.7984e-07, 4.2011e-07, 9.0166e-07, 1.1720e-06,\n 9.7640e-07, 3.1239e-07, 1.2087e-06, 8.5500e-07, 3.4058e-07, 6.1838e-09,\n 1.0757e-06, 8.6760e-09, 4.3835e-07, 1.3780e-06, 7.6326e-07, 5.3528e-07,\n 6.5963e-07, 2.6045e-07, 9.0671e-07, 5.2666e-07, 3.8014e-07, 1.1267e-06,\n 8.1268e-07, 4.0897e-07, 1.5186e-06, 1.2215e-06, 2.8119e-06, 1.0392e-06,\n 1.0312e-06, 2.4305e-08, 6.4014e-07, 7.7685e-07, 5.9689e-07, 7.4781e-07,\n 7.7290e-07, 6.2347e-07, 1.9523e-06, 1.4656e-10, 4.4775e-07, 7.3036e-07,\n 5.0637e-07, 4.0205e-07, 1.0413e-06, 5.3219e-07, 3.5436e-07, 5.5503e-07,\n 4.7760e-07, 1.4421e-06, 3.0863e-07, 5.1285e-07, 2.6418e-06, 1.1078e-06,\n 1.2863e-06, 1.6111e-06, 4.6936e-07, 1.0095e-06, 1.5675e-10, 4.2362e-07,\n 1.1630e-15, 2.5368e-06, 9.2332e-16, 9.2061e-07, 6.4367e-07, 6.4506e-07,\n 2.5284e-06, 1.0492e-06, 3.0020e-07, 5.5645e-07, 1.7123e-06, 1.8173e-06,\n 1.2249e-06, 4.4884e-07, 4.8044e-07, 1.2102e-06, 1.2730e-06, 5.2293e-07,\n 1.1242e-06, 7.9942e-10, 1.5004e-07, 1.9010e-06, 2.2795e-06, 1.7769e-06,\n 1.3396e-06, 8.0307e-07, 1.2985e-06, 2.7493e-07, 9.8741e-07, 6.9259e-07,\n 3.6900e-08, 3.3425e-06, 2.6978e-07, 1.3398e-06, 2.8595e-06, 1.0765e-07,\n 7.5429e-07, 2.9283e-06, 1.0467e-06, 2.6411e-06, 5.4300e-07, 9.6525e-08,\n 4.9003e-07, 2.9545e-06, 1.3446e-07, 7.7174e-08, 8.5541e-08, 1.8408e-09,\n 2.2653e-08, 2.8438e-07, 7.1372e-07, 1.3210e-06, 1.4321e-06, 1.1977e-10,\n 7.5630e-07, 3.0710e-07, 5.1653e-07, 2.6290e-10, 1.4990e-08, 4.6652e-06,\n 5.5571e-07, 2.9821e-07, 8.8349e-07, 1.4700e-07, 4.7275e-08, 7.9849e-07,\n 2.7607e-07, 9.5377e-07, 2.1626e-06, 6.6508e-07, 6.9812e-07, 6.0029e-07,\n 8.3848e-07, 2.7599e-07, 8.9292e-07, 3.4591e-07, 7.7833e-08, 2.4878e-06,\n 4.4520e-07, 1.5379e-07, 7.1636e-16, 3.4817e-09, 2.0073e-06, 7.1493e-07,\n 3.4610e-06, 2.7943e-07, 1.4511e-06, 5.5975e-07, 6.9824e-07, 7.2428e-06,\n 5.5577e-07, 1.2911e-06, 1.0543e-06, 8.6477e-07, 1.4342e-07, 1.9502e-07,\n 4.2301e-07, 9.9569e-07, 1.2475e-08, 2.0556e-07, 4.0302e-07, 4.1481e-06,\n 8.0554e-07, 6.9444e-07, 8.0879e-07, 1.7548e-06, 5.1805e-07, 2.6492e-07,\n 2.4805e-07, 1.0640e-06, 1.8839e-07, 1.1453e-07, 5.2543e-07, 4.1449e-07,\n 2.9529e-07, 3.8180e-07, 5.0466e-07, 1.7990e-06, 1.4080e-07, 5.1564e-10,\n 3.8400e-07, 3.4136e-07, 1.0148e-06, 9.8705e-07, 3.4755e-07, 7.3117e-07,\n 8.0249e-07, 6.0858e-07, 2.0602e-09, 6.1735e-07, 1.8794e-06, 4.4716e-07,\n 1.2138e-06, 8.5758e-07, 5.0984e-08, 1.4056e-07, 1.0171e-06, 1.3524e-06,\n 1.5431e-08, 3.3588e-07, 2.7600e-06, 3.5748e-07, 1.7232e-06, 5.8755e-07,\n 8.9006e-07, 1.5649e-06, 8.2405e-07, 6.5520e-07, 2.3303e-06, 1.3869e-06,\n 9.7881e-07, 1.0458e-06, 1.2780e-06, 4.9681e-12, 1.1335e-06, 7.0893e-07,\n 7.8894e-08, 4.9371e-07, 4.7118e-10, 9.9218e-07, 4.0002e-07, 3.9096e-07,\n 1.1457e-06, 1.0767e-06, 8.8331e-07, 3.3403e-10, 2.2041e-17, 1.7459e-07,\n 7.8920e-07, 1.8536e-07, 5.0958e-10, 1.0657e-09, 8.1103e-07, 5.4135e-07,\n 1.3195e-06, 4.7695e-07, 1.4170e-06, 1.1812e-07, 6.9507e-07, 2.5021e-07,\n 8.4511e-07, 2.7787e-09, 4.2356e-07, 5.7534e-07, 1.0271e-07, 3.3696e-09,\n 2.7431e-07, 1.0747e-06, 4.8789e-07, 4.7937e-07, 2.9958e-08, 1.5074e-07,\n 7.0133e-07, 4.4668e-07, 9.0904e-07, 7.0668e-07, 8.5795e-07, 1.3056e-06,\n 1.8408e-07, 1.3976e-06, 1.0443e-07, 6.8383e-08, 9.1052e-08, 1.3667e-06,\n 2.5443e-11, 6.3652e-07, 2.1910e-07, 6.3879e-07, 1.9526e-07, 3.1935e-07,\n 1.3301e-06, 1.8577e-07, 4.1887e-07, 1.0300e-08, 6.4127e-09, 8.0370e-07,\n 1.0992e-07, 4.5740e-10, 3.2147e-07, 3.1744e-07, 2.3280e-06, 1.1528e-06,\n 9.1396e-07, 1.2202e-06, 7.2606e-07, 1.5138e-06, 4.6831e-07, 7.6289e-07,\n 1.4355e-06, 1.3786e-06, 5.9561e-07, 6.2240e-07, 1.0685e-06, 1.5100e-07,\n 1.1965e-06, 1.0568e-06, 5.5998e-07, 1.1728e-06, 7.5300e-07, 1.3953e-06,\n 8.1084e-07, 5.7104e-07, 5.5135e-06, 1.2401e-06, 5.1247e-07, 1.4647e-07,\n 1.0463e-06, 6.4147e-07, 5.5631e-07, 8.4690e-07, 3.8226e-07, 1.1633e-06,\n 9.1130e-07, 7.1735e-07, 6.6029e-07, 1.4952e-06, 7.4427e-08, 2.9427e-07,\n 3.6868e-07, 3.1176e-07, 7.6368e-07, 8.9063e-07, 8.2386e-07, 1.3683e-07,\n 6.1614e-07, 6.2491e-07, 3.5184e-07, 4.1233e-07, 2.1883e-11, 1.0088e-06,\n 4.8563e-07, 4.7879e-07, 1.0296e-06, 1.5800e-06, 8.2253e-07, 1.1187e-06,\n 3.6826e-07, 8.7401e-08, 1.1891e-06, 2.5500e-06, 8.8205e-07, 4.8460e-07,\n 1.0479e-06, 5.4647e-07, 1.2814e-06, 1.5067e-06, 3.4917e-06, 9.1755e-09,\n 5.7110e-07, 6.9218e-07, 7.8733e-07, 4.0053e-06, 1.8350e-06, 7.4901e-07,\n 1.0182e-06, 7.9896e-07, 2.7494e-07, 5.7868e-06, 4.2023e-07, 9.8054e-07,\n 2.8858e-07, 9.4421e-07, 1.5494e-10, 2.2449e-06, 7.6642e-08, 4.5723e-07,\n 9.7069e-07, 4.8916e-07, 3.1357e-07, 3.6359e-09, 1.6837e-06, 1.7734e-07,\n 4.2306e-07, 2.0147e-07, 6.2035e-07, 1.5180e-07, 9.5219e-07, 1.9761e-06,\n 7.6016e-07, 1.3163e-07, 7.5185e-07, 5.8537e-07, 1.0418e-06, 4.6039e-10,\n 1.4087e-07, 5.3565e-07, 6.1220e-14, 1.1441e-06, 6.3994e-07, 8.6216e-07,\n 3.9157e-07, 9.2755e-07, 3.4498e-07, 7.1148e-11, 9.0453e-07, 1.1713e-06,\n 5.8192e-07, 5.0824e-07, 4.3069e-07, 3.6747e-07, 4.1851e-07, 9.5970e-07,\n 4.0166e-07, 9.1797e-07, 1.0541e-06, 3.4054e-07, 3.8524e-07, 1.2907e-06,\n 6.1173e-07, 1.7278e-09, 8.2294e-07, 4.4052e-07, 8.9499e-07, 3.7058e-07,\n 5.4024e-07, 1.2713e-06, 1.2534e-06, 3.2180e-10, 7.6525e-07, 3.7338e-07,\n 4.7344e-07, 1.2873e-06, 9.9905e-07, 8.9615e-07, 1.7492e-06, 1.3731e-06,\n 1.4028e-06, 7.9521e-07, 3.1927e-11, 9.4321e-07, 5.5022e-07, 1.0915e-07,\n 3.5490e-07, 1.2023e-07, 8.3820e-09, 2.6125e-10, 3.9734e-07, 2.5277e-06,\n 9.8898e-13, 6.0548e-07, 6.5310e-07, 7.1258e-07, 1.6666e-06, 1.0600e-10,\n 1.1263e-06, 2.7673e-06, 7.3475e-08, 6.4961e-07, 6.9310e-07, 4.9224e-07,\n 2.0200e-06, 3.2431e-07, 7.2796e-07, 7.5071e-07, 6.7080e-07, 8.7545e-07,\n 4.5792e-07, 4.9774e-07, 1.4234e-06, 2.8189e-07, 1.9259e-06, 2.4070e-08,\n 1.7216e-06, 2.4019e-06, 1.1586e-06, 2.8589e-06, 4.4713e-07, 2.4244e-06,\n 1.0345e-06, 1.5155e-07, 1.7144e-06, 3.8055e-08, 6.3984e-07, 9.3172e-12,\n 3.5636e-07, 3.3954e-07, 6.0256e-07, 4.1759e-07, 1.1727e-06, 4.6991e-08,\n 4.1896e-07, 1.0031e-09, 1.1015e-06, 9.4829e-07, 1.0061e-06, 2.3732e-06,\n 6.1455e-07, 4.3822e-07, 1.1105e-06, 2.5620e-07, 6.1177e-07, 1.7062e-06,\n 3.7789e-09, 1.2076e-07, 3.5970e-06, 1.0530e-07, 2.1249e-08, 4.4448e-07,\n 1.4379e-06, 7.9509e-07, 2.8473e-10, 1.0784e-06, 7.9449e-07, 4.9015e-07,\n 9.5823e-07, 1.0845e-06, 1.1661e-06, 6.6448e-07, 2.2667e-09, 1.6183e-06,\n 5.3888e-07, 1.7866e-11, 8.7242e-07, 3.2521e-07, 3.9884e-07, 1.3442e-06,\n 1.0079e-06, 1.5342e-07, 6.2635e-07, 5.0330e-07, 2.6914e-07, 1.3293e-06,\n 1.5943e-06, 1.6918e-06, 6.5854e-07, 7.2117e-07, 2.4250e-06, 1.3364e-07,\n 5.8485e-07, 1.2992e-06, 1.1971e-06, 1.7535e-06, 3.1576e-07, 2.7208e-06,\n 6.3838e-08, 7.8176e-07, 2.1405e-06, 1.9674e-07, 1.3169e-06, 4.3461e-07,\n 4.1396e-07, 2.4920e-07, 5.9988e-07, 1.8233e-06, 4.4708e-06, 7.4075e-07,\n 2.0702e-07, 3.1660e-07, 2.1280e-06, 1.4262e-06, 9.3477e-07, 4.2852e-06,\n 5.9127e-07, 4.3328e-08, 1.8275e-07, 4.8662e-07, 4.7664e-07, 1.1740e-06,\n 2.1481e-07, 5.0966e-07, 5.2168e-07, 6.5028e-07, 4.5627e-07, 2.8062e-07,\n 1.3837e-06, 1.9243e-06, 2.2690e-09, 3.4297e-16, 1.7694e-06, 3.1399e-07],\n device='cuda:0')" }, "4": { - "step": "tensor(2504.)", - "exp_avg": "tensor([[ 1.6488e-05, -1.4424e-31, -1.3237e-14, ..., 5.6052e-45,\n 4.5902e-06, 7.1103e-06],\n [ 2.0914e-05, -2.4680e-31, 1.7005e-14, ..., -5.6052e-45,\n 1.2571e-05, 4.9623e-06],\n [-5.2860e-06, -4.6505e-31, 4.9668e-14, ..., -5.6052e-45,\n -9.4278e-06, -2.1789e-06],\n ...,\n [ 1.2960e-05, -1.4827e-32, 1.4761e-13, ..., -5.6052e-45,\n -6.5906e-06, -5.4997e-06],\n [ 1.3405e-05, 2.9384e-31, -2.1457e-13, ..., -5.6052e-45,\n 1.1084e-05, -2.2056e-05],\n [ 1.1067e-06, 1.1508e-31, 6.9854e-14, ..., -5.6052e-45,\n -1.9899e-05, -1.3157e-05]], device='cuda:0')", - "exp_avg_sq": "tensor([[9.0222e-09, 2.9364e-12, 6.5881e-10, ..., 3.3641e-18, 6.2046e-09,\n 5.3830e-10],\n [1.3911e-08, 5.4147e-11, 1.6519e-10, ..., 1.7193e-16, 2.4183e-09,\n 2.2894e-09],\n [2.0818e-08, 3.2965e-13, 4.4880e-11, ..., 7.4982e-17, 2.0536e-09,\n 2.5456e-09],\n ...,\n [1.9563e-08, 5.3595e-11, 1.3424e-11, ..., 6.5601e-17, 1.6696e-09,\n 1.4470e-09],\n [2.5444e-08, 4.4919e-12, 6.2387e-10, ..., 9.7767e-19, 3.3787e-09,\n 4.0470e-09],\n [1.8861e-08, 6.8987e-12, 4.9939e-10, ..., 6.2656e-17, 4.7172e-09,\n 3.3056e-09]], device='cuda:0')" + "step": "tensor(3756.)", + "exp_avg": "tensor([[-6.9723e-06, -5.6052e-45, -1.9480e-07, ..., 5.6052e-45,\n -1.7607e-05, 7.2223e-06],\n [-1.4869e-05, -5.6052e-45, 1.6776e-07, ..., -5.6052e-45,\n 2.9163e-06, 5.0429e-06],\n [ 2.3633e-05, -5.6052e-45, -2.0383e-07, ..., -5.6052e-45,\n -1.4050e-05, 1.1134e-05],\n ...,\n [ 1.6885e-05, -5.6052e-45, 2.7860e-07, ..., -5.6052e-45,\n 7.9984e-06, 1.5818e-06],\n [-9.7993e-06, 5.6052e-45, 2.6818e-07, ..., -5.6052e-45,\n -1.4768e-05, 1.9807e-05],\n [-1.4280e-05, 5.6052e-45, -5.4336e-09, ..., -5.6052e-45,\n -5.1312e-07, 1.2072e-05]], device='cuda:0')", + "exp_avg_sq": "tensor([[4.5730e-09, 8.3910e-13, 1.9574e-10, ..., 9.6131e-19, 3.1881e-09,\n 5.4013e-10],\n [7.2927e-09, 1.5473e-11, 5.9144e-11, ..., 4.9130e-17, 3.4340e-09,\n 2.6113e-09],\n [1.0876e-08, 9.4201e-14, 3.1303e-11, ..., 2.1427e-17, 3.5554e-09,\n 2.7200e-09],\n ...,\n [9.9474e-09, 1.5315e-11, 2.1602e-11, ..., 1.8746e-17, 2.0396e-09,\n 1.5691e-09],\n [1.2948e-08, 1.2836e-12, 1.8681e-10, ..., 2.7938e-19, 3.5793e-09,\n 4.2711e-09],\n [1.0005e-08, 1.9714e-12, 1.5010e-10, ..., 1.7904e-17, 3.4572e-09,\n 3.1526e-09]], device='cuda:0')" }, "5": { - "step": "tensor(1252.)", - "exp_avg": "tensor([[ 5.6052e-45, 3.9702e-26, 0.0000e+00, ..., 2.8262e-25,\n 5.6052e-45, 5.6052e-45],\n [ 2.8750e-06, -3.4558e-06, -1.9541e-39, ..., -1.7206e-06,\n -2.4016e-06, 1.0311e-05],\n [ 1.4448e-07, 3.8939e-07, -1.1574e-39, ..., -3.8548e-07,\n -2.4723e-07, 1.2512e-06],\n ...,\n [ 1.9303e-06, -9.3097e-06, -6.4635e-40, ..., -1.6230e-06,\n 3.4935e-08, -1.6698e-05],\n [-2.5685e-07, -3.0748e-06, -1.5835e-42, ..., -4.1745e-06,\n 1.5720e-06, 1.5001e-08],\n [-7.5800e-07, 1.2750e-06, -1.6718e-39, ..., 7.4812e-06,\n 1.5872e-06, 1.1566e-06]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.5982e-11, 1.5239e-11, 0.0000e+00, ..., 2.8985e-12, 7.7933e-11,\n 2.0601e-12],\n [5.5887e-10, 3.4808e-10, 3.6153e-12, ..., 1.0094e-09, 1.0207e-09,\n 1.4900e-10],\n [7.1554e-11, 7.0524e-11, 2.3953e-13, ..., 1.4232e-10, 3.5385e-10,\n 1.4243e-10],\n ...,\n [7.5388e-10, 4.4527e-10, 5.2433e-12, ..., 1.5097e-10, 2.0495e-10,\n 3.8287e-09],\n [1.1055e-11, 1.3709e-10, 5.0357e-14, ..., 2.9572e-10, 5.7164e-10,\n 7.6276e-11],\n [4.1424e-10, 2.5341e-10, 9.0561e-13, ..., 8.3957e-10, 2.4020e-10,\n 8.6336e-10]], device='cuda:0')" + "step": "tensor(2504.)", + "exp_avg": "tensor([[-1.1049e-37, -7.5380e-12, 0.0000e+00, ..., -1.9722e-10,\n 3.4509e-15, 5.6052e-45],\n [ 1.0658e-07, 9.5533e-07, -5.6052e-45, ..., -2.6661e-07,\n 4.6272e-08, 3.7274e-07],\n [-8.4002e-09, 1.2374e-07, -5.6052e-45, ..., -4.0115e-07,\n 1.4932e-06, 3.1902e-07],\n ...,\n [-1.4340e-06, -1.9331e-06, 2.6553e-41, ..., -2.3349e-07,\n 2.4483e-07, 1.3270e-06],\n [ 1.1407e-09, 2.2466e-06, -5.6052e-45, ..., -3.0531e-06,\n 1.5926e-06, 3.9220e-08],\n [ 2.7531e-06, -1.5339e-06, -5.6052e-45, ..., 7.7509e-07,\n 1.2155e-06, 3.3995e-06]], device='cuda:0')", + "exp_avg_sq": "tensor([[4.5670e-12, 4.3546e-12, 0.0000e+00, ..., 8.2909e-13, 2.2270e-11,\n 5.8867e-13],\n [3.8068e-10, 1.8270e-10, 1.0331e-12, ..., 6.5363e-10, 5.0998e-10,\n 7.3235e-11],\n [2.8487e-11, 2.9304e-11, 6.8448e-14, ..., 5.0903e-11, 1.7708e-10,\n 4.6909e-11],\n ...,\n [4.8653e-10, 2.1801e-10, 1.4983e-12, ..., 7.8652e-11, 9.7122e-11,\n 2.4574e-09],\n [7.9861e-12, 1.2800e-10, 1.4390e-14, ..., 1.7001e-10, 3.1079e-10,\n 2.2594e-11],\n [2.3274e-10, 1.1993e-10, 2.5879e-13, ..., 4.9258e-10, 1.1941e-10,\n 7.4302e-10]], device='cuda:0')" }, "6": { - "step": "tensor(1252.)", - "exp_avg": "tensor([-2.7876e-14, 2.1512e-05, -1.4557e-05, ..., 1.4872e-05,\n -6.0187e-06, -1.4926e-07], device='cuda:0')", - "exp_avg_sq": "tensor([1.5950e-09, 1.0827e-07, 2.4486e-08, ..., 6.4185e-08, 2.2175e-08,\n 6.6246e-08], device='cuda:0')" + "step": "tensor(2504.)", + "exp_avg": "tensor([ 1.1606e-08, -3.6526e-05, -3.7104e-06, ..., 2.8214e-05,\n 2.0651e-05, 3.5293e-05], device='cuda:0')", + "exp_avg_sq": "tensor([4.5584e-10, 5.1557e-08, 1.2877e-08, ..., 3.5296e-08, 1.3297e-08,\n 3.0710e-08], device='cuda:0')" }, "7": { - "step": "tensor(1252.)", - "exp_avg": "tensor([[-4.9243e-17, -7.5301e-08, -1.5296e-06, ..., -8.2674e-07,\n -2.0414e-06, 2.7959e-07],\n [ 1.6864e-16, -5.2434e-06, -9.0352e-07, ..., -2.9519e-06,\n 1.4543e-06, 1.2484e-07],\n [ 3.5097e-17, 4.5890e-06, -9.4728e-07, ..., -1.1508e-06,\n 3.1703e-07, -4.4113e-06],\n ...,\n [ 4.6054e-16, 3.5197e-07, -2.7686e-06, ..., -1.0305e-06,\n -1.2226e-07, 1.3850e-06],\n [-5.3855e-16, -3.4545e-06, -2.1339e-07, ..., -1.0924e-06,\n 4.2353e-08, -5.3837e-07],\n [ 9.1176e-16, -2.5962e-06, -1.3834e-06, ..., -6.8402e-07,\n 1.6414e-06, 4.9500e-08]], device='cuda:0')", - "exp_avg_sq": "tensor([[6.0808e-11, 5.7454e-11, 9.0231e-11, ..., 4.5614e-11, 5.4293e-11,\n 7.9972e-11],\n [3.0057e-10, 1.7043e-10, 1.3969e-10, ..., 1.3365e-10, 6.3195e-11,\n 1.2874e-10],\n [2.7659e-11, 1.1931e-10, 3.9111e-10, ..., 8.9023e-11, 7.4569e-11,\n 1.4400e-10],\n ...,\n [7.0189e-11, 1.6190e-10, 5.9274e-11, ..., 1.0649e-10, 1.3345e-10,\n 2.3981e-10],\n [8.4020e-11, 1.7071e-10, 1.5698e-10, ..., 1.0911e-10, 1.4494e-10,\n 1.3377e-10],\n [4.6804e-11, 1.5156e-10, 2.8898e-10, ..., 1.1255e-10, 5.3664e-11,\n 1.1334e-10]], device='cuda:0')" + "step": "tensor(2504.)", + "exp_avg": "tensor([[ 6.6865e-10, 8.2075e-07, -1.7039e-06, ..., 8.5299e-07,\n -1.4859e-06, -2.3474e-06],\n [-3.7056e-09, 2.6293e-06, 1.6175e-06, ..., 2.6888e-06,\n 7.2217e-07, 1.7641e-06],\n [-8.9940e-10, -8.0847e-07, 1.1742e-06, ..., 2.2207e-06,\n -4.1395e-07, -2.1736e-07],\n ...,\n [-3.4476e-09, -2.1548e-06, -1.7662e-07, ..., -2.8316e-07,\n 1.3020e-06, -2.6602e-06],\n [ 1.7088e-09, 1.2234e-06, 1.0442e-06, ..., 2.1703e-06,\n 8.6852e-07, -3.0169e-06],\n [-5.4819e-09, 2.3014e-06, 1.5491e-06, ..., -1.0597e-06,\n 6.9973e-07, 1.8270e-06]], device='cuda:0')", + "exp_avg_sq": "tensor([[1.7376e-11, 3.0709e-11, 5.3679e-11, ..., 2.6120e-11, 2.6402e-11,\n 3.8409e-11],\n [8.5892e-11, 8.3604e-11, 6.3620e-11, ..., 5.9595e-11, 3.4082e-11,\n 5.6347e-11],\n [7.9039e-12, 6.2825e-11, 1.2633e-10, ..., 4.9444e-11, 3.3846e-11,\n 7.3381e-11],\n ...,\n [2.0057e-11, 7.6697e-11, 3.4492e-11, ..., 5.6300e-11, 5.2138e-11,\n 1.0624e-10],\n [2.4010e-11, 9.0693e-11, 9.8801e-11, ..., 6.1589e-11, 5.1280e-11,\n 7.0476e-11],\n [1.3375e-11, 7.0728e-11, 2.0362e-10, ..., 5.5899e-11, 3.1467e-11,\n 5.7287e-11]], device='cuda:0')" }, "14": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([7.5368e-06], device='cuda:0')" + "exp_avg_sq": "tensor([2.1537e-06], device='cuda:0')" }, "15": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([ 5.6052e-45, -5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([9.7230e-09, 5.5040e-07, 4.1382e-07], device='cuda:0')" + "exp_avg_sq": "tensor([2.7784e-09, 1.5728e-07, 1.1825e-07], device='cuda:0')" }, "16": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([5.8550e-04, 6.0329e-05, 6.5278e-05, 6.9730e-05], device='cuda:0')" + "exp_avg_sq": "tensor([1.6731e-04, 1.7239e-05, 1.8654e-05, 1.9926e-05], device='cuda:0')" }, "18": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([[-5.6052e-45, -5.6052e-45, 0.0000e+00, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [-5.6052e-45, -5.6052e-45, 0.0000e+00, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n ...,\n [ 5.6052e-45, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [-5.6052e-45, -5.6052e-45, 0.0000e+00, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45]], device='cuda:0')", - "exp_avg_sq": "tensor([[6.2423e-10, 3.8968e-10, 0.0000e+00, ..., 2.5713e-09, 9.2475e-10,\n 2.9682e-11],\n [1.3318e-10, 3.4236e-10, 0.0000e+00, ..., 3.1475e-10, 1.1068e-09,\n 3.1192e-10],\n [4.0026e-11, 1.4126e-10, 0.0000e+00, ..., 4.7206e-11, 3.0604e-10,\n 1.0023e-10],\n ...,\n [5.0717e-12, 2.7278e-11, 0.0000e+00, ..., 3.8900e-11, 5.3871e-10,\n 3.0349e-12],\n [9.7692e-10, 1.0270e-09, 0.0000e+00, ..., 1.2431e-09, 5.2102e-09,\n 3.1268e-10],\n [3.0515e-12, 2.3715e-12, 0.0000e+00, ..., 4.6937e-12, 4.6558e-11,\n 1.1027e-11]], device='cuda:0')" + "exp_avg_sq": "tensor([[1.7838e-10, 1.1135e-10, 0.0000e+00, ..., 7.3476e-10, 2.6426e-10,\n 8.4820e-12],\n [3.8057e-11, 9.7831e-11, 0.0000e+00, ..., 8.9943e-11, 3.1627e-10,\n 8.9133e-11],\n [1.1438e-11, 4.0365e-11, 0.0000e+00, ..., 1.3489e-11, 8.7453e-11,\n 2.8642e-11],\n ...,\n [1.4493e-12, 7.7948e-12, 0.0000e+00, ..., 1.1116e-11, 1.5394e-10,\n 8.6724e-13],\n [2.7916e-10, 2.9348e-10, 0.0000e+00, ..., 3.5522e-10, 1.4889e-09,\n 8.9350e-11],\n [8.7198e-13, 6.7768e-13, 0.0000e+00, ..., 1.3413e-12, 1.3304e-11,\n 3.1512e-12]], device='cuda:0')" }, "19": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([8.8308e-07, 4.0264e-07, 9.1999e-08, 1.6702e-07, 6.0763e-08, 5.8881e-10,\n 1.0454e-08, 1.6582e-08, 4.2378e-08, 2.3990e-09, 1.3440e-07, 6.8348e-07,\n 2.8164e-07, 3.4117e-07, 5.1507e-08, 3.2122e-07, 7.6560e-07, 6.3462e-08,\n 4.2448e-07, 6.9534e-07, 8.3002e-09, 2.9215e-07, 6.3157e-08, 6.2446e-07,\n 1.2634e-07, 2.5567e-07, 1.1395e-07, 7.0883e-08, 1.9302e-08, 1.1958e-07,\n 3.5772e-07, 1.7928e-07, 6.5142e-09, 1.6444e-06, 8.5263e-07, 4.1454e-07,\n 4.2983e-08, 4.7403e-09, 3.2581e-07, 3.1752e-07, 1.5578e-06, 3.8533e-08,\n 6.8763e-09, 2.1776e-08, 5.4974e-08, 3.9941e-07, 1.7173e-07, 1.1195e-07,\n 1.0893e-07, 1.2770e-07, 5.2238e-09, 1.8388e-07, 4.6704e-07, 2.8493e-07,\n 6.0421e-08, 3.1508e-08, 2.1472e-07, 2.7351e-09, 8.2593e-08, 2.9695e-08,\n 9.2496e-08, 2.0435e-07, 1.3179e-07, 7.6380e-08, 7.4859e-09, 4.8749e-08,\n 1.7727e-07, 1.0428e-08, 2.2344e-07, 2.1031e-07, 2.0114e-09, 6.7896e-08,\n 1.7961e-07, 8.1649e-09, 7.2057e-08, 6.1735e-10, 4.7775e-08, 1.2200e-06,\n 2.4256e-08, 2.2419e-09, 3.7398e-07, 5.0057e-07, 8.2036e-11, 2.7913e-09,\n 8.9173e-08, 3.0973e-07, 9.3193e-09, 1.5852e-07, 2.0321e-08, 3.5246e-07,\n 5.5969e-08, 5.1121e-07, 6.5161e-07, 4.6108e-08, 6.4555e-08, 2.5698e-07,\n 1.6941e-08, 1.0103e-06, 3.3699e-07, 1.2070e-06, 3.4449e-07, 2.3870e-08,\n 4.7633e-09, 1.4577e-07, 1.2115e-08, 1.2033e-06, 3.0206e-08, 8.5584e-08,\n 1.9284e-07, 1.4878e-09, 1.5525e-06, 2.8761e-08, 4.6174e-07, 1.9610e-07,\n 6.6521e-08, 9.1507e-07, 2.2799e-08, 1.1092e-07, 3.0627e-07, 1.4654e-07,\n 1.5985e-07, 2.9578e-07, 2.9372e-08, 4.9372e-09, 1.9254e-09, 1.0756e-07,\n 2.5650e-06, 2.0001e-08, 1.6836e-07, 3.3391e-08, 4.4933e-09, 8.2488e-07,\n 6.9326e-08, 1.1611e-07, 4.6991e-07, 3.3555e-07, 2.8359e-07, 8.6947e-07,\n 3.7491e-07, 8.2480e-11, 3.4199e-09, 5.5487e-09, 3.0194e-07, 1.2391e-06,\n 6.9398e-07, 4.5595e-08, 4.6634e-08, 6.0201e-08, 3.7138e-07, 8.4102e-10,\n 1.9246e-08, 3.8663e-07, 1.3347e-07, 1.4865e-07, 3.3254e-07, 1.1440e-08,\n 1.0446e-07, 2.7445e-07, 5.3318e-08, 8.8894e-07, 4.2149e-07, 6.9165e-08,\n 1.6406e-07, 4.4545e-07, 2.8159e-08, 2.0147e-07, 1.9833e-06, 3.0135e-07,\n 2.0822e-07, 9.6013e-08, 1.3887e-08, 3.3126e-07, 1.3542e-06, 3.4060e-07,\n 4.4760e-07, 3.2605e-09, 4.5131e-08, 1.5583e-07, 1.7759e-07, 1.9039e-07,\n 2.3365e-07, 1.0910e-07, 2.1906e-08, 1.3154e-07, 2.5038e-07, 6.9478e-10,\n 6.5579e-08, 2.6934e-07, 6.8317e-08, 1.0857e-06, 8.1157e-07, 8.4168e-08,\n 5.5717e-08, 2.4134e-08, 1.6107e-10, 3.8891e-06, 4.5577e-08, 7.8472e-08,\n 8.6639e-08, 1.8995e-06, 1.8925e-11, 1.7739e-06, 3.3367e-07, 5.0767e-08,\n 3.2411e-07, 3.9825e-06, 3.6616e-07, 1.1427e-07, 1.6584e-09, 5.1402e-07,\n 4.1501e-08, 7.5014e-07, 1.1680e-07, 2.3798e-07, 1.8218e-09, 3.0304e-07,\n 2.3303e-06, 1.3192e-09, 2.5685e-08, 7.5333e-08, 5.5460e-08, 1.5503e-06,\n 1.9524e-07, 1.1147e-06, 1.5454e-07, 1.8765e-10, 3.9814e-09, 2.2809e-08,\n 8.4537e-09, 3.2547e-07, 2.7826e-09, 2.9742e-09, 5.6495e-07, 2.2094e-07,\n 1.7231e-07, 3.8392e-08, 4.5956e-08, 3.4786e-07, 4.3552e-08, 3.8269e-08,\n 4.4395e-07, 1.8068e-09, 9.7741e-08, 5.9932e-07, 1.0447e-07, 2.9928e-09,\n 5.5931e-09, 2.5151e-07, 2.8396e-08, 9.0457e-07, 4.7159e-07, 4.7385e-08,\n 4.0736e-07, 4.8099e-08, 1.3605e-06, 2.4056e-08], device='cuda:0')" + "exp_avg_sq": "tensor([2.5235e-07, 1.1506e-07, 2.6289e-08, 4.7726e-08, 1.7363e-08, 1.6826e-10,\n 2.9874e-09, 4.7385e-09, 1.2110e-08, 6.8553e-10, 3.8405e-08, 1.9531e-07,\n 8.0481e-08, 9.7491e-08, 1.4719e-08, 9.1790e-08, 2.1878e-07, 1.8135e-08,\n 1.2130e-07, 1.9870e-07, 2.3718e-09, 8.3483e-08, 1.8048e-08, 1.7844e-07,\n 3.6101e-08, 7.3060e-08, 3.2562e-08, 2.0256e-08, 5.5157e-09, 3.4170e-08,\n 1.0222e-07, 5.1230e-08, 1.8615e-09, 4.6991e-07, 2.4365e-07, 1.1846e-07,\n 1.2283e-08, 1.3546e-09, 9.3102e-08, 9.0735e-08, 4.4516e-07, 1.1011e-08,\n 1.9650e-09, 6.2227e-09, 1.5709e-08, 1.1413e-07, 4.9074e-08, 3.1989e-08,\n 3.1128e-08, 3.6491e-08, 1.4927e-09, 5.2545e-08, 1.3346e-07, 8.1420e-08,\n 1.7266e-08, 9.0035e-09, 6.1358e-08, 7.8158e-10, 2.3602e-08, 8.4855e-09,\n 2.6432e-08, 5.8395e-08, 3.7659e-08, 2.1826e-08, 2.1391e-09, 1.3930e-08,\n 5.0655e-08, 2.9798e-09, 6.3849e-08, 6.0097e-08, 5.7479e-10, 1.9402e-08,\n 5.1324e-08, 2.3332e-09, 2.0591e-08, 1.7641e-10, 1.3652e-08, 3.4862e-07,\n 6.9314e-09, 6.4063e-10, 1.0687e-07, 1.4304e-07, 2.3442e-11, 7.9764e-10,\n 2.5482e-08, 8.8509e-08, 2.6630e-09, 4.5297e-08, 5.8069e-09, 1.0072e-07,\n 1.5994e-08, 1.4608e-07, 1.8620e-07, 1.3176e-08, 1.8447e-08, 7.3434e-08,\n 4.8411e-09, 2.8869e-07, 9.6297e-08, 3.4492e-07, 9.8441e-08, 6.8211e-09,\n 1.3611e-09, 4.1655e-08, 3.4619e-09, 3.4387e-07, 8.6317e-09, 2.4456e-08,\n 5.5106e-08, 4.2516e-10, 4.4364e-07, 8.2186e-09, 1.3195e-07, 5.6036e-08,\n 1.9009e-08, 2.6149e-07, 6.5149e-09, 3.1698e-08, 8.7519e-08, 4.1874e-08,\n 4.5678e-08, 8.4521e-08, 8.3933e-09, 1.4109e-09, 5.5018e-10, 3.0735e-08,\n 7.3298e-07, 5.7154e-09, 4.8111e-08, 9.5418e-09, 1.2840e-09, 2.3571e-07,\n 1.9810e-08, 3.3179e-08, 1.3428e-07, 9.5886e-08, 8.1039e-08, 2.4846e-07,\n 1.0713e-07, 2.3569e-11, 9.7726e-10, 1.5856e-09, 8.6281e-08, 3.5408e-07,\n 1.9831e-07, 1.3029e-08, 1.3326e-08, 1.7203e-08, 1.0613e-07, 2.4033e-10,\n 5.4997e-09, 1.1048e-07, 3.8141e-08, 4.2477e-08, 9.5027e-08, 3.2692e-09,\n 2.9849e-08, 7.8425e-08, 1.5236e-08, 2.5402e-07, 1.2044e-07, 1.9764e-08,\n 4.6882e-08, 1.2729e-07, 8.0467e-09, 5.7572e-08, 5.6675e-07, 8.6112e-08,\n 5.9501e-08, 2.7437e-08, 3.9684e-09, 9.4661e-08, 3.8696e-07, 9.7328e-08,\n 1.2791e-07, 9.3170e-10, 1.2897e-08, 4.4531e-08, 5.0747e-08, 5.4406e-08,\n 6.6768e-08, 3.1176e-08, 6.2598e-09, 3.7589e-08, 7.1548e-08, 1.9854e-10,\n 1.8740e-08, 7.6965e-08, 1.9522e-08, 3.1025e-07, 2.3191e-07, 2.4052e-08,\n 1.5922e-08, 6.8965e-09, 4.6026e-11, 1.1113e-06, 1.3024e-08, 2.2424e-08,\n 2.4758e-08, 5.4281e-07, 5.4079e-12, 5.0690e-07, 9.5349e-08, 1.4507e-08,\n 9.2618e-08, 1.1380e-06, 1.0463e-07, 3.2654e-08, 4.7391e-10, 1.4689e-07,\n 1.1859e-08, 2.1436e-07, 3.3377e-08, 6.8006e-08, 5.2058e-10, 8.6597e-08,\n 6.6591e-07, 3.7697e-10, 7.3397e-09, 2.1527e-08, 1.5848e-08, 4.4301e-07,\n 5.5792e-08, 3.1854e-07, 4.4160e-08, 5.3624e-11, 1.1377e-09, 6.5178e-09,\n 2.4157e-09, 9.3006e-08, 7.9516e-10, 8.4989e-10, 1.6144e-07, 6.3136e-08,\n 4.9240e-08, 1.0971e-08, 1.3132e-08, 9.9402e-08, 1.2445e-08, 1.0936e-08,\n 1.2686e-07, 5.1630e-10, 2.7930e-08, 1.7126e-07, 2.9854e-08, 8.5523e-10,\n 1.5983e-09, 7.1870e-08, 8.1143e-09, 2.5849e-07, 1.3476e-07, 1.3541e-08,\n 1.1641e-07, 1.3745e-08, 3.8876e-07, 6.8741e-09], device='cuda:0')" }, "20": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([2.1787e-09, 1.9923e-09, 1.3398e-10, 8.7434e-10, 1.0257e-10, 8.4596e-12,\n 1.8602e-11, 8.6380e-11, 5.1293e-11, 3.0132e-12, 2.4756e-10, 1.2397e-09,\n 7.1375e-10, 1.6056e-09, 1.6528e-10, 7.5660e-10, 2.4816e-09, 1.1007e-10,\n 1.3731e-09, 1.9783e-09, 4.0998e-12, 1.2288e-09, 1.3894e-10, 2.5486e-09,\n 4.9537e-10, 4.9623e-10, 2.8818e-10, 2.3552e-10, 7.0048e-11, 2.1165e-10,\n 1.0925e-09, 5.3189e-10, 4.1487e-11, 4.4273e-09, 3.3036e-09, 1.0338e-09,\n 8.5337e-11, 8.6364e-14, 6.7701e-10, 5.7400e-10, 6.5895e-09, 9.1923e-11,\n 2.9254e-11, 9.5536e-11, 1.0983e-10, 8.9846e-10, 3.7408e-10, 2.3616e-10,\n 1.6660e-10, 1.9730e-10, 4.0930e-12, 5.0242e-10, 9.6966e-10, 6.2112e-10,\n 8.2766e-11, 1.4858e-10, 6.9466e-10, 5.7227e-12, 2.7259e-10, 8.4108e-11,\n 1.6700e-10, 1.5244e-09, 3.6965e-10, 1.8345e-10, 1.1328e-13, 7.7880e-11,\n 3.8705e-10, 5.6615e-12, 4.1256e-10, 6.8090e-10, 4.7333e-11, 1.5600e-10,\n 1.9942e-10, 2.7413e-11, 1.3179e-10, 3.9898e-11, 2.7857e-10, 6.1944e-09,\n 2.1230e-11, 6.3353e-15, 1.5664e-09, 1.2978e-09, 1.1927e-11, 1.3789e-11,\n 1.7541e-10, 6.9202e-10, 3.7138e-11, 7.4208e-10, 5.9557e-11, 6.5722e-10,\n 1.6953e-10, 1.1681e-09, 2.3687e-09, 5.8462e-11, 1.7860e-10, 9.0574e-10,\n 3.6683e-11, 2.1518e-09, 1.3115e-09, 3.3940e-09, 1.0325e-09, 1.1112e-10,\n 5.8248e-11, 2.5837e-10, 5.0535e-12, 2.8278e-09, 3.1576e-11, 2.0177e-10,\n 4.0153e-10, 2.1027e-12, 3.9626e-09, 8.9652e-11, 1.3086e-09, 4.1175e-10,\n 1.2138e-10, 3.6934e-09, 8.9823e-11, 2.3651e-10, 9.2345e-10, 3.5547e-10,\n 2.9411e-10, 6.1601e-10, 2.4684e-11, 4.7641e-13, 1.9648e-11, 1.6552e-10,\n 1.1599e-08, 7.5850e-11, 5.3774e-10, 5.0928e-11, 1.6465e-12, 2.4600e-09,\n 1.6965e-10, 3.4545e-10, 9.0204e-10, 7.5777e-10, 7.9654e-10, 2.2441e-09,\n 1.0256e-09, 1.3915e-11, 8.7698e-11, 3.1795e-11, 6.6344e-10, 3.6420e-09,\n 4.3607e-09, 1.8049e-10, 1.8733e-10, 1.4092e-10, 7.6932e-10, 2.1330e-13,\n 1.9265e-10, 8.4779e-10, 4.3093e-10, 1.6923e-10, 7.8013e-10, 6.9309e-11,\n 1.5416e-10, 4.6758e-10, 8.7710e-11, 3.5159e-09, 8.2192e-10, 7.8275e-11,\n 3.8223e-10, 1.4849e-09, 8.1830e-11, 2.6183e-10, 6.8177e-09, 5.7153e-10,\n 6.3715e-10, 2.4468e-10, 7.1168e-12, 1.3213e-09, 3.2619e-09, 1.2712e-09,\n 8.4256e-10, 1.2543e-11, 7.0528e-11, 6.9705e-10, 5.4397e-10, 4.9724e-10,\n 6.0340e-10, 2.1113e-10, 2.6602e-11, 1.9750e-10, 6.2679e-10, 3.7468e-16,\n 1.0103e-10, 5.9918e-10, 6.0165e-11, 2.7186e-09, 2.1182e-09, 1.0601e-10,\n 1.6570e-10, 1.3204e-10, 1.5182e-11, 1.5076e-08, 5.8023e-11, 1.2565e-10,\n 2.6221e-10, 5.2206e-09, 4.4975e-12, 4.7662e-09, 6.0211e-10, 6.9347e-11,\n 1.0781e-09, 1.0150e-08, 7.9311e-10, 2.8836e-10, 5.4363e-13, 8.0324e-10,\n 5.6846e-11, 1.8699e-09, 7.9309e-10, 7.2626e-10, 1.0887e-12, 5.2627e-10,\n 1.1087e-08, 4.5231e-13, 9.6430e-11, 9.0240e-11, 7.0839e-11, 6.8193e-09,\n 1.0864e-09, 2.7127e-09, 4.5687e-10, 3.9926e-12, 1.9353e-13, 6.7630e-11,\n 8.3731e-12, 9.1515e-10, 2.2491e-11, 2.0946e-12, 3.3325e-09, 9.7562e-10,\n 5.2151e-10, 8.9432e-11, 5.8780e-11, 1.0118e-09, 1.1953e-10, 3.8210e-11,\n 1.0993e-09, 4.2089e-13, 2.9326e-10, 1.5436e-09, 1.5843e-10, 5.1308e-13,\n 1.3174e-12, 4.8110e-10, 4.4786e-11, 2.4412e-09, 9.5826e-10, 6.6602e-11,\n 9.1244e-10, 9.5188e-11, 4.3569e-09, 1.2668e-10], device='cuda:0')" + "exp_avg_sq": "tensor([6.2257e-10, 5.6932e-10, 3.8286e-11, 2.4985e-10, 2.9311e-11, 2.4174e-12,\n 5.3155e-12, 2.4684e-11, 1.4657e-11, 8.6104e-13, 7.0743e-11, 3.5426e-10,\n 2.0396e-10, 4.5882e-10, 4.7230e-11, 2.1621e-10, 7.0913e-10, 3.1453e-11,\n 3.9236e-10, 5.6533e-10, 1.1715e-12, 3.5115e-10, 3.9704e-11, 7.2827e-10,\n 1.4156e-10, 1.4180e-10, 8.2349e-11, 6.7301e-11, 2.0017e-11, 6.0479e-11,\n 3.1219e-10, 1.5199e-10, 1.1855e-11, 1.2651e-09, 9.4403e-10, 2.9542e-10,\n 2.4386e-11, 2.4679e-14, 1.9346e-10, 1.6403e-10, 1.8830e-09, 2.6268e-11,\n 8.3597e-12, 2.7300e-11, 3.1384e-11, 2.5674e-10, 1.0690e-10, 6.7485e-11,\n 4.7607e-11, 5.6381e-11, 1.1696e-12, 1.4357e-10, 2.7709e-10, 1.7749e-10,\n 2.3651e-11, 4.2459e-11, 1.9850e-10, 1.6353e-12, 7.7894e-11, 2.4035e-11,\n 4.7723e-11, 4.3562e-10, 1.0563e-10, 5.2423e-11, 3.2372e-14, 2.2255e-11,\n 1.1060e-10, 1.6178e-12, 1.1789e-10, 1.9457e-10, 1.3526e-11, 4.4578e-11,\n 5.6985e-11, 7.8336e-12, 3.7659e-11, 1.1401e-11, 7.9605e-11, 1.7701e-09,\n 6.0667e-12, 1.8104e-15, 4.4762e-10, 3.7085e-10, 3.4081e-12, 3.9404e-12,\n 5.0124e-11, 1.9775e-10, 1.0613e-11, 2.1205e-10, 1.7019e-11, 1.8780e-10,\n 4.8445e-11, 3.3381e-10, 6.7688e-10, 1.6706e-11, 5.1036e-11, 2.5882e-10,\n 1.0482e-11, 6.1490e-10, 3.7476e-10, 9.6987e-10, 2.9504e-10, 3.1755e-11,\n 1.6645e-11, 7.3831e-11, 1.4441e-12, 8.0806e-10, 9.0230e-12, 5.7656e-11,\n 1.1474e-10, 6.0086e-13, 1.1323e-09, 2.5619e-11, 3.7393e-10, 1.1766e-10,\n 3.4686e-11, 1.0554e-09, 2.5668e-11, 6.7585e-11, 2.6388e-10, 1.0158e-10,\n 8.4045e-11, 1.7603e-10, 7.0538e-12, 1.3614e-13, 5.6146e-12, 4.7298e-11,\n 3.3144e-09, 2.1675e-11, 1.5366e-10, 1.4553e-11, 4.7051e-13, 7.0296e-10,\n 4.8478e-11, 9.8714e-11, 2.5776e-10, 2.1654e-10, 2.2762e-10, 6.4126e-10,\n 2.9308e-10, 3.9763e-12, 2.5060e-11, 9.0858e-12, 1.8958e-10, 1.0407e-09,\n 1.2461e-09, 5.1576e-11, 5.3532e-11, 4.0269e-11, 2.1984e-10, 6.0952e-14,\n 5.5052e-11, 2.4226e-10, 1.2314e-10, 4.8358e-11, 2.2293e-10, 1.9806e-11,\n 4.4051e-11, 1.3361e-10, 2.5064e-11, 1.0047e-09, 2.3487e-10, 2.2368e-11,\n 1.0922e-10, 4.2432e-10, 2.3383e-11, 7.4820e-11, 1.9482e-09, 1.6332e-10,\n 1.8207e-10, 6.9920e-11, 2.0337e-12, 3.7757e-10, 9.3210e-10, 3.6325e-10,\n 2.4077e-10, 3.5841e-12, 2.0154e-11, 1.9919e-10, 1.5544e-10, 1.4209e-10,\n 1.7243e-10, 6.0332e-11, 7.6017e-12, 5.6437e-11, 1.7911e-10, 1.0707e-16,\n 2.8869e-11, 1.7122e-10, 1.7193e-11, 7.7686e-10, 6.0531e-10, 3.0294e-11,\n 4.7349e-11, 3.7732e-11, 4.3385e-12, 4.3081e-09, 1.6581e-11, 3.5907e-11,\n 7.4929e-11, 1.4918e-09, 1.2852e-12, 1.3620e-09, 1.7206e-10, 1.9817e-11,\n 3.0808e-10, 2.9004e-09, 2.2664e-10, 8.2401e-11, 1.5535e-13, 2.2953e-10,\n 1.6244e-11, 5.3433e-10, 2.2663e-10, 2.0753e-10, 3.1112e-13, 1.5039e-10,\n 3.1682e-09, 1.2925e-13, 2.7556e-11, 2.5787e-11, 2.0243e-11, 1.9487e-09,\n 3.1046e-10, 7.7518e-10, 1.3056e-10, 1.1409e-12, 5.5301e-14, 1.9326e-11,\n 2.3927e-12, 2.6151e-10, 6.4271e-12, 5.9855e-13, 9.5229e-10, 2.7879e-10,\n 1.4902e-10, 2.5556e-11, 1.6797e-11, 2.8912e-10, 3.4157e-11, 1.0919e-11,\n 3.1412e-10, 1.2027e-13, 8.3802e-11, 4.4110e-10, 4.5271e-11, 1.4662e-13,\n 3.7645e-13, 1.3748e-10, 1.2798e-11, 6.9759e-10, 2.7383e-10, 1.9032e-11,\n 2.6074e-10, 2.7201e-11, 1.2450e-09, 3.6199e-11], device='cuda:0')" }, "21": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([3.7367e-09, 1.9074e-09, 2.6260e-10, 8.5836e-10, 1.3498e-10, 5.7339e-12,\n 2.6006e-11, 1.4126e-10, 8.6217e-11, 1.6896e-12, 4.1036e-10, 2.3162e-09,\n 8.1685e-10, 1.5918e-09, 2.7817e-10, 1.4673e-09, 3.3184e-09, 1.7545e-10,\n 1.2195e-09, 2.8793e-09, 4.6351e-12, 1.3759e-09, 3.1846e-10, 2.6883e-09,\n 6.6182e-10, 7.6636e-10, 5.3067e-10, 3.7236e-10, 1.4424e-10, 3.0853e-10,\n 1.6042e-09, 8.3467e-10, 6.4873e-11, 5.6596e-09, 3.6464e-09, 1.7809e-09,\n 6.1668e-11, 1.7925e-14, 1.4380e-09, 1.0921e-09, 5.0678e-09, 2.1382e-10,\n 7.0559e-11, 1.6877e-10, 1.8830e-10, 1.2839e-09, 7.9304e-10, 5.5275e-10,\n 2.7372e-10, 4.1199e-10, 6.6889e-12, 8.1296e-10, 1.5224e-09, 8.9301e-10,\n 1.5066e-10, 2.0383e-10, 9.9243e-10, 5.1600e-12, 4.6022e-10, 1.7675e-10,\n 2.2134e-10, 1.0736e-09, 6.0291e-10, 3.4940e-10, 1.8613e-12, 1.2184e-10,\n 8.3808e-10, 1.5298e-11, 9.9340e-10, 1.0192e-09, 5.2034e-11, 3.4463e-10,\n 6.2430e-10, 5.9158e-11, 1.7134e-10, 3.6567e-11, 2.9669e-10, 5.3912e-09,\n 3.1994e-11, 5.0843e-15, 1.8066e-09, 2.2541e-09, 1.3139e-11, 2.9832e-11,\n 4.5457e-10, 1.4423e-09, 6.5584e-11, 8.3927e-10, 6.5899e-11, 1.5511e-09,\n 2.9298e-10, 2.2003e-09, 2.0984e-09, 1.2229e-10, 2.9051e-10, 6.9408e-10,\n 4.8112e-11, 3.4497e-09, 1.5801e-09, 5.0554e-09, 1.5186e-09, 1.6587e-10,\n 6.6428e-11, 3.9735e-10, 7.4649e-12, 3.9717e-09, 5.3795e-11, 4.4788e-10,\n 5.4267e-10, 1.3953e-12, 5.2136e-09, 1.7501e-10, 2.0770e-09, 7.2249e-10,\n 1.5978e-10, 2.8906e-09, 1.2348e-10, 2.9912e-10, 1.4009e-09, 7.0843e-10,\n 4.5295e-10, 8.7247e-10, 4.5413e-11, 1.3138e-12, 3.6589e-11, 3.0854e-10,\n 1.0312e-08, 1.1225e-10, 8.5006e-10, 9.4024e-11, 1.9009e-12, 3.5724e-09,\n 3.5388e-10, 5.5007e-10, 1.4813e-09, 1.0951e-09, 1.2562e-09, 2.9323e-09,\n 1.6691e-09, 1.6970e-11, 6.6313e-11, 5.2007e-11, 9.6638e-10, 4.2389e-09,\n 3.2251e-09, 2.7022e-10, 2.6748e-10, 3.1463e-10, 1.2177e-09, 4.7620e-13,\n 1.7666e-10, 1.1632e-09, 6.7657e-10, 4.2638e-10, 9.2331e-10, 1.0538e-10,\n 2.6145e-10, 9.9497e-10, 1.4140e-10, 3.7677e-09, 1.3841e-09, 1.9244e-10,\n 3.8592e-10, 2.0033e-09, 1.5132e-10, 5.7542e-10, 7.0817e-09, 9.6056e-10,\n 9.5451e-10, 4.6903e-10, 1.9560e-11, 8.6265e-10, 4.6449e-09, 1.6144e-09,\n 1.3460e-09, 2.5625e-11, 1.0492e-10, 8.3307e-10, 4.0630e-10, 9.1025e-10,\n 6.1712e-10, 2.7260e-10, 4.4362e-11, 3.5937e-10, 1.0497e-09, 1.0771e-13,\n 2.0127e-10, 7.6857e-10, 1.8364e-10, 4.4568e-09, 2.7086e-09, 2.2223e-10,\n 3.0319e-10, 1.5929e-10, 1.9942e-11, 1.5958e-08, 9.2200e-11, 1.6943e-10,\n 4.2067e-10, 6.6937e-09, 1.1249e-11, 7.3854e-09, 1.0737e-09, 1.1271e-10,\n 1.5202e-09, 1.3770e-08, 1.0973e-09, 2.6858e-10, 1.3732e-12, 1.7870e-09,\n 6.8768e-11, 3.2351e-09, 6.6576e-10, 1.1497e-09, 1.3016e-12, 1.0108e-09,\n 7.8253e-09, 3.0491e-13, 1.6230e-10, 1.6811e-10, 1.0301e-10, 6.5011e-09,\n 1.0128e-09, 4.7426e-09, 7.9409e-10, 7.3511e-12, 1.3266e-13, 1.5367e-10,\n 2.1361e-11, 1.3989e-09, 4.4995e-11, 4.5242e-12, 2.5267e-09, 1.0445e-09,\n 7.9497e-10, 1.9224e-10, 1.1901e-10, 1.5605e-09, 2.4798e-10, 9.2774e-11,\n 2.0036e-09, 3.9776e-13, 4.7573e-10, 1.8753e-09, 2.9735e-10, 9.9184e-13,\n 1.2169e-12, 8.0111e-10, 4.4725e-11, 3.8947e-09, 1.5048e-09, 1.1038e-10,\n 1.2230e-09, 1.6150e-10, 4.5114e-09, 1.7652e-10], device='cuda:0')" + "exp_avg_sq": "tensor([1.0678e-09, 5.4506e-10, 7.5041e-11, 2.4528e-10, 3.8571e-11, 1.6385e-12,\n 7.4314e-12, 4.0365e-11, 2.4637e-11, 4.8282e-13, 1.1726e-10, 6.6187e-10,\n 2.3342e-10, 4.5486e-10, 7.9489e-11, 4.1930e-10, 9.4825e-10, 5.0137e-11,\n 3.4848e-10, 8.2278e-10, 1.3245e-12, 3.9317e-10, 9.1003e-11, 7.6821e-10,\n 1.8912e-10, 2.1899e-10, 1.5164e-10, 1.0640e-10, 4.1219e-11, 8.8164e-11,\n 4.5842e-10, 2.3851e-10, 1.8538e-11, 1.6173e-09, 1.0420e-09, 5.0890e-10,\n 1.7622e-11, 5.1223e-15, 4.1091e-10, 3.1208e-10, 1.4482e-09, 6.1102e-11,\n 2.0163e-11, 4.8226e-11, 5.3808e-11, 3.6689e-10, 2.2662e-10, 1.5795e-10,\n 7.8218e-11, 1.1773e-10, 1.9114e-12, 2.3231e-10, 4.3504e-10, 2.5519e-10,\n 4.3054e-11, 5.8245e-11, 2.8359e-10, 1.4745e-12, 1.3151e-10, 5.0507e-11,\n 6.3250e-11, 3.0680e-10, 1.7229e-10, 9.9845e-11, 5.3187e-13, 3.4816e-11,\n 2.3949e-10, 4.3716e-12, 2.8387e-10, 2.9125e-10, 1.4869e-11, 9.8482e-11,\n 1.7840e-10, 1.6905e-11, 4.8963e-11, 1.0449e-11, 8.4781e-11, 1.5406e-09,\n 9.1426e-12, 1.4529e-15, 5.1624e-10, 6.4413e-10, 3.7546e-12, 8.5247e-12,\n 1.2990e-10, 4.1216e-10, 1.8741e-11, 2.3983e-10, 1.8831e-11, 4.4323e-10,\n 8.3723e-11, 6.2874e-10, 5.9965e-10, 3.4944e-11, 8.3014e-11, 1.9834e-10,\n 1.3748e-11, 9.8577e-10, 4.5153e-10, 1.4446e-09, 4.3395e-10, 4.7400e-11,\n 1.8982e-11, 1.1355e-10, 2.1332e-12, 1.1349e-09, 1.5372e-11, 1.2798e-10,\n 1.5507e-10, 3.9873e-13, 1.4898e-09, 5.0010e-11, 5.9352e-10, 2.0646e-10,\n 4.5658e-11, 8.2602e-10, 3.5286e-11, 8.5474e-11, 4.0032e-10, 2.0244e-10,\n 1.2943e-10, 2.4932e-10, 1.2977e-11, 3.7542e-13, 1.0455e-11, 8.8167e-11,\n 2.9466e-09, 3.2075e-11, 2.4291e-10, 2.6868e-11, 5.4319e-13, 1.0208e-09,\n 1.0112e-10, 1.5719e-10, 4.2328e-10, 3.1293e-10, 3.5898e-10, 8.3792e-10,\n 4.7695e-10, 4.8494e-12, 1.8949e-11, 1.4862e-11, 2.7615e-10, 1.2113e-09,\n 9.2158e-10, 7.7218e-11, 7.6436e-11, 8.9909e-11, 3.4796e-10, 1.3608e-13,\n 5.0483e-11, 3.3239e-10, 1.9334e-10, 1.2184e-10, 2.6384e-10, 3.0113e-11,\n 7.4712e-11, 2.8432e-10, 4.0406e-11, 1.0767e-09, 3.9551e-10, 5.4993e-11,\n 1.1028e-10, 5.7245e-10, 4.3240e-11, 1.6443e-10, 2.0237e-09, 2.7449e-10,\n 2.7276e-10, 1.3403e-10, 5.5895e-12, 2.4651e-10, 1.3273e-09, 4.6133e-10,\n 3.8463e-10, 7.3227e-12, 2.9980e-11, 2.3806e-10, 1.1610e-10, 2.6011e-10,\n 1.7635e-10, 7.7898e-11, 1.2677e-11, 1.0269e-10, 2.9996e-10, 3.0779e-14,\n 5.7516e-11, 2.1963e-10, 5.2477e-11, 1.2736e-09, 7.7399e-10, 6.3505e-11,\n 8.6638e-11, 4.5518e-11, 5.6985e-12, 4.5601e-09, 2.6347e-11, 4.8416e-11,\n 1.2021e-10, 1.9128e-09, 3.2144e-12, 2.1104e-09, 3.0681e-10, 3.2207e-11,\n 4.3441e-10, 3.9348e-09, 3.1358e-10, 7.6750e-11, 3.9239e-13, 5.1066e-10,\n 1.9651e-11, 9.2445e-10, 1.9025e-10, 3.2853e-10, 3.7195e-13, 2.8885e-10,\n 2.2361e-09, 8.7131e-14, 4.6379e-11, 4.8040e-11, 2.9437e-11, 1.8577e-09,\n 2.8941e-10, 1.3552e-09, 2.2692e-10, 2.1006e-12, 3.7908e-14, 4.3911e-11,\n 6.1040e-12, 3.9976e-10, 1.2858e-11, 1.2928e-12, 7.2203e-10, 2.9847e-10,\n 2.2717e-10, 5.4934e-11, 3.4007e-11, 4.4593e-10, 7.0862e-11, 2.6511e-11,\n 5.7254e-10, 1.1366e-13, 1.3594e-10, 5.3587e-10, 8.4971e-11, 2.8343e-13,\n 3.4773e-13, 2.2892e-10, 1.2780e-11, 1.1130e-09, 4.3000e-10, 3.1543e-11,\n 3.4949e-10, 4.6150e-11, 1.2892e-09, 5.0442e-11], device='cuda:0')" }, "22": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([[-5.6052e-45, -5.6052e-45, 0.0000e+00, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [-5.6052e-45, -5.6052e-45, 0.0000e+00, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n ...,\n [ 5.6052e-45, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [-5.6052e-45, -5.6052e-45, 0.0000e+00, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45]], device='cuda:0')", - "exp_avg_sq": "tensor([[2.6403e-10, 9.9180e-10, 0.0000e+00, ..., 1.4441e-09, 2.3615e-09,\n 3.6132e-10],\n [1.2111e-10, 1.3516e-10, 0.0000e+00, ..., 2.2305e-10, 6.8801e-10,\n 4.3733e-13],\n [1.6259e-10, 8.8325e-11, 0.0000e+00, ..., 1.4853e-10, 3.1735e-10,\n 2.5558e-11],\n ...,\n [2.7772e-10, 3.4241e-10, 0.0000e+00, ..., 4.2293e-10, 8.0028e-10,\n 2.5457e-10],\n [8.4557e-10, 2.6848e-10, 0.0000e+00, ..., 2.6689e-10, 1.6935e-09,\n 1.6122e-10],\n [4.1389e-12, 5.7963e-12, 0.0000e+00, ..., 6.2501e-11, 9.9615e-11,\n 2.0332e-11]], device='cuda:0')" + "exp_avg_sq": "tensor([[7.5448e-11, 2.8341e-10, 0.0000e+00, ..., 4.1268e-10, 6.7483e-10,\n 1.0325e-10],\n [3.4607e-11, 3.8623e-11, 0.0000e+00, ..., 6.3737e-11, 1.9660e-10,\n 1.2497e-13],\n [4.6463e-11, 2.5240e-11, 0.0000e+00, ..., 4.2445e-11, 9.0686e-11,\n 7.3035e-12],\n ...,\n [7.9361e-11, 9.7847e-11, 0.0000e+00, ..., 1.2085e-10, 2.2869e-10,\n 7.2745e-11],\n [2.4163e-10, 7.6720e-11, 0.0000e+00, ..., 7.6265e-11, 4.8393e-10,\n 4.6070e-11],\n [1.1827e-12, 1.6563e-12, 0.0000e+00, ..., 1.7860e-11, 2.8466e-11,\n 5.8100e-12]], device='cuda:0')" }, "23": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([1.0314e-06, 7.9389e-08, 1.7535e-07, 1.3196e-07, 4.4997e-08, 2.4137e-08,\n 1.0382e-07, 6.1771e-11, 4.1400e-08, 7.0221e-09, 7.7954e-07, 1.5436e-07,\n 2.0767e-07, 1.3474e-07, 3.9432e-08, 9.7314e-08, 4.0781e-07, 2.6676e-08,\n 7.7207e-08, 8.2238e-07, 3.2463e-09, 3.1222e-07, 1.3591e-07, 7.8519e-08,\n 1.0507e-08, 3.2748e-07, 1.0345e-07, 1.6447e-08, 1.1500e-08, 9.1599e-08,\n 4.5246e-07, 9.4210e-08, 1.1557e-08, 3.3104e-06, 3.2699e-07, 2.2913e-07,\n 3.9239e-08, 2.8437e-12, 3.0575e-07, 1.5420e-06, 1.9393e-07, 1.1838e-07,\n 2.0927e-08, 1.2039e-07, 7.0296e-07, 5.2158e-07, 2.9495e-07, 1.7258e-07,\n 4.3481e-08, 7.4903e-07, 8.3990e-09, 1.3539e-06, 9.3841e-07, 2.1401e-07,\n 2.1717e-08, 3.9236e-09, 5.1483e-07, 1.6369e-08, 2.3507e-07, 4.7398e-08,\n 1.1062e-07, 1.2685e-07, 6.7488e-07, 3.5822e-07, 1.6345e-08, 4.2290e-08,\n 2.5242e-07, 4.5016e-08, 3.5346e-07, 8.4981e-09, 1.2484e-11, 1.4502e-07,\n 2.6671e-07, 4.6853e-09, 1.1312e-07, 1.3824e-08, 3.9202e-08, 7.5237e-07,\n 1.1061e-08, 4.7283e-09, 6.0448e-08, 5.3012e-07, 1.5119e-11, 3.9288e-08,\n 1.1008e-08, 3.9934e-07, 1.6070e-07, 9.0046e-08, 1.5887e-07, 2.3271e-06,\n 3.1172e-08, 5.2469e-07, 6.6913e-07, 1.0120e-07, 1.2697e-06, 9.4870e-08,\n 1.1234e-07, 7.1323e-07, 2.3968e-07, 2.7949e-07, 6.2087e-07, 3.4955e-08,\n 6.7330e-09, 2.5084e-07, 9.2959e-09, 1.3389e-06, 7.5424e-08, 1.5267e-07,\n 4.6634e-08, 2.1565e-10, 1.0949e-06, 7.8345e-08, 1.0874e-06, 3.3195e-07,\n 3.6318e-08, 6.0410e-08, 7.1975e-08, 6.1165e-08, 4.6152e-08, 1.1217e-07,\n 3.3574e-07, 1.9635e-07, 3.3360e-08, 2.6611e-11, 1.9746e-10, 1.2082e-07,\n 6.7224e-07, 4.3588e-07, 3.5271e-08, 5.8564e-07, 4.5953e-10, 2.8942e-06,\n 1.3167e-07, 2.4476e-07, 4.1990e-07, 1.1274e-06, 3.5768e-07, 1.1365e-06,\n 3.1003e-07, 1.4143e-10, 4.0555e-09, 2.1175e-08, 1.3611e-07, 1.2255e-06,\n 2.1369e-07, 5.7867e-08, 5.1207e-08, 1.9847e-08, 1.2182e-06, 2.6159e-08,\n 2.1364e-08, 1.0177e-07, 1.1120e-07, 4.4258e-07, 1.5088e-07, 2.0207e-08,\n 1.0652e-07, 1.1313e-06, 3.1985e-08, 1.5576e-06, 3.9590e-07, 5.7553e-07,\n 6.0081e-08, 5.2372e-07, 4.0060e-07, 3.3008e-07, 3.1803e-06, 4.4740e-07,\n 4.9205e-07, 6.3635e-08, 2.0298e-08, 8.5590e-08, 5.8985e-07, 1.8146e-07,\n 3.3843e-07, 3.6247e-09, 2.9814e-08, 1.9887e-08, 4.0371e-08, 1.4213e-07,\n 2.0878e-08, 1.1517e-07, 7.6004e-08, 1.0187e-07, 7.8592e-07, 2.6712e-09,\n 7.9605e-08, 2.4373e-07, 4.6810e-08, 1.2635e-06, 1.3322e-07, 1.9034e-07,\n 2.0556e-07, 1.6243e-08, 6.9105e-11, 3.6148e-06, 4.2973e-08, 4.4549e-08,\n 2.6594e-07, 6.9245e-07, 2.8075e-11, 1.9658e-07, 7.4494e-07, 8.8139e-09,\n 4.4774e-07, 1.4902e-06, 2.3483e-07, 4.8235e-08, 1.8017e-10, 2.0367e-06,\n 4.1867e-08, 6.9847e-07, 2.1462e-08, 9.5967e-08, 1.6239e-08, 1.9371e-07,\n 2.5609e-07, 8.8062e-10, 1.9487e-08, 2.5225e-08, 8.4243e-08, 3.2731e-07,\n 6.0639e-08, 1.1604e-06, 2.0453e-07, 8.3556e-09, 2.5821e-09, 1.2472e-08,\n 4.6353e-08, 1.0622e-07, 3.2231e-09, 1.1319e-08, 1.8506e-07, 1.4986e-07,\n 6.0200e-07, 1.8861e-07, 4.4087e-08, 3.3620e-07, 5.0288e-08, 8.8206e-08,\n 3.2244e-07, 6.6979e-10, 4.2170e-07, 5.2322e-07, 1.8169e-07, 4.1324e-10,\n 1.4430e-11, 1.8029e-06, 8.7044e-09, 6.6254e-07, 2.2731e-07, 3.9090e-08,\n 2.9846e-07, 3.1226e-07, 7.2478e-07, 2.0439e-08], device='cuda:0')" + "exp_avg_sq": "tensor([2.9474e-07, 2.2686e-08, 5.0107e-08, 3.7709e-08, 1.2858e-08, 6.8972e-09,\n 2.9668e-08, 1.7652e-11, 1.1830e-08, 2.0066e-09, 2.2276e-07, 4.4108e-08,\n 5.9344e-08, 3.8504e-08, 1.1268e-08, 2.7808e-08, 1.1653e-07, 7.6230e-09,\n 2.2063e-08, 2.3500e-07, 9.2765e-10, 8.9220e-08, 3.8838e-08, 2.2437e-08,\n 3.0026e-09, 9.3580e-08, 2.9560e-08, 4.6999e-09, 3.2863e-09, 2.6175e-08,\n 1.2929e-07, 2.6921e-08, 3.3025e-09, 9.4598e-07, 9.3441e-08, 6.5475e-08,\n 1.1213e-08, 8.1260e-13, 8.7371e-08, 4.4065e-07, 5.5418e-08, 3.3828e-08,\n 5.9799e-09, 3.4402e-08, 2.0088e-07, 1.4905e-07, 8.4285e-08, 4.9317e-08,\n 1.2425e-08, 2.1404e-07, 2.4001e-09, 3.8688e-07, 2.6816e-07, 6.1156e-08,\n 6.2057e-09, 1.1212e-09, 1.4712e-07, 4.6775e-09, 6.7174e-08, 1.3544e-08,\n 3.1611e-08, 3.6250e-08, 1.9285e-07, 1.0236e-07, 4.6707e-09, 1.2085e-08,\n 7.2131e-08, 1.2864e-08, 1.0100e-07, 2.4284e-09, 3.5674e-12, 4.1442e-08,\n 7.6216e-08, 1.3389e-09, 3.2324e-08, 3.9503e-09, 1.1202e-08, 2.1500e-07,\n 3.1608e-09, 1.3512e-09, 1.7273e-08, 1.5148e-07, 4.3203e-12, 1.1227e-08,\n 3.1457e-09, 1.1411e-07, 4.5921e-08, 2.5731e-08, 4.5399e-08, 6.6497e-07,\n 8.9078e-09, 1.4993e-07, 1.9121e-07, 2.8918e-08, 3.6283e-07, 2.7110e-08,\n 3.2101e-08, 2.0381e-07, 6.8492e-08, 7.9866e-08, 1.7742e-07, 9.9885e-09,\n 1.9240e-09, 7.1680e-08, 2.6564e-09, 3.8260e-07, 2.1553e-08, 4.3626e-08,\n 1.3326e-08, 6.1623e-11, 3.1288e-07, 2.2388e-08, 3.1074e-07, 9.4857e-08,\n 1.0378e-08, 1.7263e-08, 2.0567e-08, 1.7479e-08, 1.3188e-08, 3.2052e-08,\n 9.5942e-08, 5.6108e-08, 9.5329e-09, 7.6042e-12, 5.6424e-11, 3.4526e-08,\n 1.9210e-07, 1.2456e-07, 1.0079e-08, 1.6735e-07, 1.3131e-10, 8.2704e-07,\n 3.7626e-08, 6.9943e-08, 1.1999e-07, 3.2216e-07, 1.0221e-07, 3.2478e-07,\n 8.8593e-08, 4.0415e-11, 1.1589e-09, 6.0510e-09, 3.8893e-08, 3.5021e-07,\n 6.1064e-08, 1.6536e-08, 1.4633e-08, 5.6715e-09, 3.4812e-07, 7.4751e-09,\n 6.1049e-09, 2.9080e-08, 3.1776e-08, 1.2647e-07, 4.3114e-08, 5.7744e-09,\n 3.0440e-08, 3.2329e-07, 9.1400e-09, 4.4511e-07, 1.1313e-07, 1.6446e-07,\n 1.7169e-08, 1.4966e-07, 1.1448e-07, 9.4322e-08, 9.0880e-07, 1.2785e-07,\n 1.4061e-07, 1.8184e-08, 5.8004e-09, 2.4458e-08, 1.6856e-07, 5.1853e-08,\n 9.6708e-08, 1.0358e-09, 8.5195e-09, 5.6828e-09, 1.1536e-08, 4.0613e-08,\n 5.9660e-09, 3.2910e-08, 2.1719e-08, 2.9110e-08, 2.2458e-07, 7.6331e-10,\n 2.2748e-08, 6.9649e-08, 1.3376e-08, 3.6105e-07, 3.8068e-08, 5.4390e-08,\n 5.8740e-08, 4.6417e-09, 1.9747e-11, 1.0330e-06, 1.2280e-08, 1.2730e-08,\n 7.5995e-08, 1.9787e-07, 8.0226e-12, 5.6174e-08, 2.1287e-07, 2.5187e-09,\n 1.2794e-07, 4.2583e-07, 6.7104e-08, 1.3783e-08, 5.1484e-11, 5.8199e-07,\n 1.1964e-08, 1.9959e-07, 6.1330e-09, 2.7423e-08, 4.6404e-09, 5.5355e-08,\n 7.3181e-08, 2.5165e-10, 5.5686e-09, 7.2082e-09, 2.4073e-08, 9.3531e-08,\n 1.7328e-08, 3.3159e-07, 5.8447e-08, 2.3877e-09, 7.3785e-10, 3.5639e-09,\n 1.3246e-08, 3.0352e-08, 9.2102e-10, 3.2344e-09, 5.2881e-08, 4.2824e-08,\n 1.7203e-07, 5.3898e-08, 1.2598e-08, 9.6070e-08, 1.4370e-08, 2.5205e-08,\n 9.2141e-08, 1.9140e-10, 1.2050e-07, 1.4951e-07, 5.1921e-08, 1.1809e-10,\n 4.1236e-12, 5.1520e-07, 2.4874e-09, 1.8933e-07, 6.4955e-08, 1.1170e-08,\n 8.5287e-08, 8.9230e-08, 2.0711e-07, 5.8407e-09], device='cuda:0')" }, "24": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([2.8910e-09, 2.4425e-10, 2.7698e-10, 5.9937e-10, 7.0104e-11, 9.4659e-11,\n 2.4339e-10, 8.1458e-13, 4.9805e-11, 4.7780e-12, 2.7800e-09, 3.7917e-10,\n 4.3195e-10, 3.2146e-10, 1.1046e-10, 2.6500e-10, 9.8219e-10, 5.4281e-11,\n 1.6941e-10, 1.8284e-09, 4.1289e-14, 1.0713e-09, 4.1615e-10, 2.1669e-10,\n 4.4286e-11, 6.0204e-10, 2.3088e-10, 5.7465e-11, 4.5118e-11, 2.2619e-10,\n 1.7318e-09, 1.7387e-10, 2.8579e-11, 1.3107e-08, 7.5306e-10, 4.8266e-10,\n 9.5395e-11, 8.3629e-13, 8.4191e-10, 3.4256e-09, 3.7601e-10, 3.9596e-10,\n 4.7780e-11, 9.9642e-10, 2.0982e-09, 1.0257e-09, 1.3423e-09, 4.4117e-10,\n 6.9384e-11, 2.8707e-09, 7.3828e-12, 4.0780e-09, 2.3353e-09, 3.6360e-10,\n 1.8551e-11, 1.3092e-11, 1.6899e-09, 1.0587e-11, 6.5581e-10, 1.1819e-10,\n 2.4822e-10, 3.6530e-10, 2.9090e-09, 9.8370e-10, 1.1563e-11, 5.9325e-11,\n 6.7320e-10, 4.9651e-11, 7.2811e-10, 4.2343e-11, 1.4796e-11, 2.9326e-10,\n 5.4388e-10, 2.7991e-11, 2.0282e-10, 9.4765e-11, 1.2540e-10, 1.9387e-09,\n 1.7731e-11, 5.0382e-12, 1.4371e-10, 1.3950e-09, 1.3431e-12, 3.1384e-10,\n 5.0799e-11, 9.7052e-10, 7.2219e-10, 2.0609e-10, 2.8792e-10, 8.2625e-09,\n 8.5437e-11, 1.2518e-09, 2.3414e-09, 1.4998e-10, 8.1789e-09, 1.8216e-10,\n 1.4495e-10, 1.4331e-09, 6.2346e-10, 7.1890e-10, 2.6302e-09, 2.6358e-10,\n 3.3767e-11, 6.5187e-10, 3.8892e-12, 2.4877e-09, 1.4744e-10, 3.8329e-10,\n 7.3102e-11, 9.2876e-12, 2.3828e-09, 2.9747e-10, 6.8002e-09, 6.7797e-10,\n 3.1525e-11, 1.0631e-10, 1.9955e-10, 9.2903e-11, 1.6642e-10, 3.7560e-10,\n 9.6420e-10, 4.9965e-10, 3.7209e-11, 1.1857e-13, 4.6514e-12, 2.0262e-10,\n 1.4792e-09, 2.5889e-09, 7.8714e-11, 1.9707e-09, 5.8558e-13, 9.3773e-09,\n 3.2858e-10, 5.7082e-10, 6.2289e-10, 4.0953e-09, 8.1183e-10, 3.2996e-09,\n 7.3840e-10, 1.0519e-12, 4.1634e-11, 7.7865e-11, 2.3094e-10, 3.8259e-09,\n 4.5069e-10, 1.9612e-10, 1.5908e-10, 5.6550e-11, 4.2458e-09, 3.7377e-11,\n 9.2349e-11, 1.7059e-10, 3.2409e-10, 1.5201e-09, 2.4941e-10, 9.8808e-11,\n 1.5801e-10, 3.3620e-09, 7.1773e-11, 4.7778e-09, 8.3173e-10, 2.2132e-09,\n 8.4358e-11, 1.7736e-09, 2.4360e-09, 7.3850e-10, 1.0903e-08, 8.4801e-10,\n 2.1291e-09, 2.3737e-10, 1.5058e-11, 1.2260e-10, 1.1116e-09, 5.5352e-10,\n 6.2434e-10, 1.1934e-11, 3.4284e-11, 4.7257e-11, 7.1957e-11, 2.4602e-10,\n 3.7922e-11, 2.6928e-10, 1.2869e-10, 1.7498e-10, 1.6833e-09, 1.7160e-14,\n 1.4418e-10, 5.9296e-10, 6.1769e-11, 3.4325e-09, 1.6650e-10, 4.9148e-10,\n 9.4292e-10, 5.8734e-11, 1.7081e-12, 1.4275e-08, 7.8633e-11, 6.8914e-11,\n 1.0583e-09, 9.9502e-10, 4.6915e-12, 3.3347e-10, 2.5448e-09, 1.0461e-11,\n 1.3722e-09, 2.7974e-09, 3.9131e-10, 7.5898e-11, 5.8679e-14, 6.4835e-09,\n 8.7427e-11, 1.5656e-09, 1.0186e-10, 2.1025e-10, 1.8898e-11, 3.7240e-10,\n 4.7976e-10, 3.5809e-13, 5.9874e-11, 3.0844e-11, 2.0509e-10, 9.4448e-10,\n 2.7659e-10, 3.7310e-09, 7.1032e-10, 3.2681e-11, 1.1749e-12, 1.3076e-10,\n 7.3235e-11, 3.0534e-10, 1.6743e-11, 1.8413e-11, 3.1716e-10, 6.0249e-10,\n 1.5179e-09, 6.1553e-10, 6.9084e-11, 7.3785e-10, 1.5477e-10, 1.3226e-10,\n 8.1898e-10, 2.8407e-13, 3.5653e-09, 1.2771e-09, 2.9748e-10, 2.8136e-13,\n 5.1243e-12, 8.5054e-09, 9.1946e-12, 2.0528e-09, 4.3371e-10, 5.8567e-11,\n 6.8353e-10, 6.4738e-10, 1.5805e-09, 7.6657e-11], device='cuda:0')" + "exp_avg_sq": "tensor([8.2612e-10, 6.9797e-11, 7.9150e-11, 1.7127e-10, 2.0033e-11, 2.7050e-11,\n 6.9550e-11, 2.3277e-13, 1.4232e-11, 1.3653e-12, 7.9440e-10, 1.0835e-10,\n 1.2343e-10, 9.1859e-11, 3.1564e-11, 7.5724e-11, 2.8067e-10, 1.5511e-11,\n 4.8410e-11, 5.2248e-10, 1.1799e-14, 3.0615e-10, 1.1892e-10, 6.1920e-11,\n 1.2655e-11, 1.7204e-10, 6.5976e-11, 1.6421e-11, 1.2893e-11, 6.4635e-11,\n 4.9488e-10, 4.9684e-11, 8.1668e-12, 3.7454e-09, 2.1519e-10, 1.3792e-10,\n 2.7260e-11, 2.3898e-13, 2.4058e-10, 9.7890e-10, 1.0745e-10, 1.1315e-10,\n 1.3654e-11, 2.8474e-10, 5.9959e-10, 2.9310e-10, 3.8357e-10, 1.2607e-10,\n 1.9827e-11, 8.2034e-10, 2.1097e-12, 1.1653e-09, 6.6732e-10, 1.0390e-10,\n 5.3012e-12, 3.7413e-12, 4.8290e-10, 3.0252e-12, 1.8740e-10, 3.3773e-11,\n 7.0932e-11, 1.0439e-10, 8.3128e-10, 2.8110e-10, 3.3044e-12, 1.6952e-11,\n 1.9237e-10, 1.4188e-11, 2.0806e-10, 1.2100e-11, 4.2280e-12, 8.3802e-11,\n 1.5542e-10, 7.9986e-12, 5.7957e-11, 2.7080e-11, 3.5834e-11, 5.5401e-10,\n 5.0666e-12, 1.4397e-12, 4.1067e-11, 3.9865e-10, 3.8381e-13, 8.9683e-11,\n 1.4516e-11, 2.7733e-10, 2.0637e-10, 5.8891e-11, 8.2275e-11, 2.3611e-09,\n 2.4414e-11, 3.5772e-10, 6.6908e-10, 4.2858e-11, 2.3372e-09, 5.2052e-11,\n 4.1421e-11, 4.0951e-10, 1.7816e-10, 2.0543e-10, 7.5159e-10, 7.5319e-11,\n 9.6491e-12, 1.8628e-10, 1.1114e-12, 7.1087e-10, 4.2132e-11, 1.0953e-10,\n 2.0889e-11, 2.6540e-12, 6.8090e-10, 8.5004e-11, 1.9432e-09, 1.9374e-10,\n 9.0086e-12, 3.0378e-11, 5.7022e-11, 2.6548e-11, 4.7555e-11, 1.0733e-10,\n 2.7553e-10, 1.4278e-10, 1.0633e-11, 3.3883e-14, 1.3292e-12, 5.7899e-11,\n 4.2269e-10, 7.3981e-10, 2.2493e-11, 5.6314e-10, 1.6733e-13, 2.6796e-09,\n 9.3895e-11, 1.6311e-10, 1.7799e-10, 1.1703e-09, 2.3199e-10, 9.4290e-10,\n 2.1100e-10, 3.0059e-13, 1.1897e-11, 2.2251e-11, 6.5992e-11, 1.0933e-09,\n 1.2879e-10, 5.6042e-11, 4.5457e-11, 1.6160e-11, 1.2133e-09, 1.0681e-11,\n 2.6389e-11, 4.8747e-11, 9.2612e-11, 4.3438e-10, 7.1270e-11, 2.8235e-11,\n 4.5153e-11, 9.6071e-10, 2.0510e-11, 1.3653e-09, 2.3767e-10, 6.3244e-10,\n 2.4106e-11, 5.0682e-10, 6.9612e-10, 2.1103e-10, 3.1157e-09, 2.4233e-10,\n 6.0842e-10, 6.7829e-11, 4.3028e-12, 3.5035e-11, 3.1766e-10, 1.5817e-10,\n 1.7841e-10, 3.4103e-12, 9.7969e-12, 1.3504e-11, 2.0562e-11, 7.0302e-11,\n 1.0836e-11, 7.6948e-11, 3.6773e-11, 5.0001e-11, 4.8102e-10, 4.9036e-15,\n 4.1201e-11, 1.6944e-10, 1.7651e-11, 9.8086e-10, 4.7578e-11, 1.4045e-10,\n 2.6945e-10, 1.6784e-11, 4.8811e-13, 4.0792e-09, 2.2470e-11, 1.9693e-11,\n 3.0242e-10, 2.8434e-10, 1.3406e-12, 9.5291e-11, 7.2720e-10, 2.9892e-12,\n 3.9213e-10, 7.9938e-10, 1.1182e-10, 2.1689e-11, 1.6768e-14, 1.8527e-09,\n 2.4983e-11, 4.4739e-10, 2.9107e-11, 6.0082e-11, 5.4002e-12, 1.0642e-10,\n 1.3709e-10, 1.0233e-13, 1.7110e-11, 8.8138e-12, 5.8607e-11, 2.6989e-10,\n 7.9037e-11, 1.0662e-09, 2.0298e-10, 9.3388e-12, 3.3574e-13, 3.7366e-11,\n 2.0928e-11, 8.7254e-11, 4.7843e-12, 5.2616e-12, 9.0630e-11, 1.7217e-10,\n 4.3375e-10, 1.7589e-10, 1.9741e-11, 2.1085e-10, 4.4226e-11, 3.7795e-11,\n 2.3403e-10, 8.1175e-14, 1.0188e-09, 3.6493e-10, 8.5008e-11, 8.0401e-14,\n 1.4643e-12, 2.4305e-09, 2.6274e-12, 5.8661e-10, 1.2394e-10, 1.6736e-11,\n 1.9532e-10, 1.8500e-10, 4.5163e-10, 2.1905e-11], device='cuda:0')" }, "25": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([4.1962e-09, 3.6998e-10, 4.9922e-10, 6.7708e-10, 1.0504e-10, 1.5038e-10,\n 3.0447e-10, 9.5619e-13, 9.0088e-11, 8.3277e-12, 2.4868e-09, 5.5937e-10,\n 6.2489e-10, 6.2432e-10, 1.9984e-10, 4.8667e-10, 1.7780e-09, 7.5574e-11,\n 2.4714e-10, 3.5545e-09, 6.4834e-13, 1.4232e-09, 6.7768e-10, 3.6731e-10,\n 5.5759e-11, 1.0692e-09, 4.9065e-10, 1.0094e-10, 8.0760e-11, 2.7122e-10,\n 2.0502e-09, 4.1929e-10, 7.1569e-11, 1.1670e-08, 1.4304e-09, 9.9725e-10,\n 7.4202e-11, 6.3238e-13, 1.3725e-09, 5.4211e-09, 6.7915e-10, 5.7206e-10,\n 1.1912e-10, 6.6244e-10, 2.2987e-09, 1.8235e-09, 1.3556e-09, 8.3325e-10,\n 1.1265e-10, 2.4135e-09, 7.4750e-12, 5.5284e-09, 3.0485e-09, 6.8096e-10,\n 3.9167e-11, 3.1913e-11, 2.2961e-09, 2.9293e-11, 1.0100e-09, 2.6417e-10,\n 2.8503e-10, 5.9511e-10, 2.9000e-09, 1.5560e-09, 2.0069e-11, 1.1256e-10,\n 1.1221e-09, 1.0472e-10, 1.5023e-09, 5.2300e-11, 1.2226e-11, 6.9193e-10,\n 8.5769e-10, 2.9556e-11, 3.2366e-10, 1.0688e-10, 2.3032e-10, 3.1187e-09,\n 2.9082e-11, 4.4656e-12, 3.2007e-10, 2.3751e-09, 3.5850e-12, 2.6573e-10,\n 6.6634e-11, 1.6986e-09, 8.0699e-10, 4.0713e-10, 4.9491e-10, 9.3517e-09,\n 1.5398e-10, 2.2737e-09, 2.1309e-09, 3.1462e-10, 5.4412e-09, 2.9246e-10,\n 3.5481e-10, 2.5209e-09, 1.0683e-09, 1.3067e-09, 2.6724e-09, 2.3490e-10,\n 6.4001e-11, 7.2928e-10, 1.1269e-11, 4.5741e-09, 1.8439e-10, 7.3902e-10,\n 1.3380e-10, 1.6573e-11, 3.7593e-09, 4.0259e-10, 4.6406e-09, 1.1435e-09,\n 9.5934e-11, 1.9630e-10, 3.7090e-10, 1.8520e-10, 2.2537e-10, 5.7999e-10,\n 9.9288e-10, 5.9818e-10, 7.4884e-11, 7.7641e-14, 4.9223e-12, 3.6930e-10,\n 2.8658e-09, 1.9833e-09, 1.7836e-10, 1.7836e-09, 7.7717e-13, 1.1736e-08,\n 5.9674e-10, 1.0514e-09, 1.3388e-09, 3.7768e-09, 1.5931e-09, 3.8947e-09,\n 1.3347e-09, 2.2892e-12, 5.2807e-11, 1.4194e-10, 4.2366e-10, 4.1241e-09,\n 9.0676e-10, 3.1261e-10, 2.5758e-10, 1.0395e-10, 4.1296e-09, 4.2263e-11,\n 1.5020e-10, 3.3708e-10, 5.3842e-10, 1.3107e-09, 4.3497e-10, 1.4387e-10,\n 2.8643e-10, 3.8327e-09, 9.5235e-11, 6.3504e-09, 1.3158e-09, 1.7334e-09,\n 1.4230e-10, 2.3399e-09, 1.8594e-09, 1.0370e-09, 1.1101e-08, 1.5191e-09,\n 2.0929e-09, 3.1110e-10, 3.8838e-11, 2.4298e-10, 2.0430e-09, 8.5169e-10,\n 1.0995e-09, 2.8016e-11, 6.6245e-11, 9.6710e-11, 9.3148e-11, 6.5191e-10,\n 5.5625e-11, 2.8587e-10, 2.0825e-10, 3.1532e-10, 3.2840e-09, 2.8905e-13,\n 2.5856e-10, 7.2244e-10, 1.2293e-10, 5.2505e-09, 4.4743e-10, 5.5343e-10,\n 9.7615e-10, 9.3321e-11, 4.4993e-12, 1.4502e-08, 1.0250e-10, 1.1806e-10,\n 1.2364e-09, 2.3758e-09, 1.1351e-11, 8.7310e-10, 2.3436e-09, 1.9772e-11,\n 1.8539e-09, 5.2318e-09, 7.7247e-10, 1.2438e-10, 1.0408e-13, 7.3417e-09,\n 8.1489e-11, 2.9334e-09, 1.5162e-10, 4.6777e-10, 2.5530e-11, 6.4956e-10,\n 8.5415e-10, 3.5156e-12, 1.1866e-10, 5.7183e-11, 1.7144e-10, 1.4314e-09,\n 3.2372e-10, 4.8605e-09, 9.6749e-10, 6.7156e-11, 9.3204e-13, 1.1538e-10,\n 9.7238e-11, 4.9800e-10, 3.5785e-11, 1.9833e-11, 8.3216e-10, 7.6357e-10,\n 2.5865e-09, 8.8907e-10, 1.0243e-10, 1.4218e-09, 2.9260e-10, 2.5076e-10,\n 1.4301e-09, 6.2332e-13, 1.9849e-09, 1.7402e-09, 5.3162e-10, 2.1169e-12,\n 1.0598e-11, 6.1348e-09, 1.4780e-11, 2.6888e-09, 6.9445e-10, 9.0168e-11,\n 9.2302e-10, 9.3279e-10, 2.3570e-09, 1.3844e-10], device='cuda:0')" + "exp_avg_sq": "tensor([1.1991e-09, 1.0572e-10, 1.4266e-10, 1.9348e-10, 3.0017e-11, 4.2972e-11,\n 8.7004e-11, 2.7324e-13, 2.5743e-11, 2.3797e-12, 7.1062e-10, 1.5985e-10,\n 1.7857e-10, 1.7840e-10, 5.7107e-11, 1.3907e-10, 5.0809e-10, 2.1596e-11,\n 7.0621e-11, 1.0157e-09, 1.8527e-13, 4.0668e-10, 1.9365e-10, 1.0496e-10,\n 1.5933e-11, 3.0554e-10, 1.4021e-10, 2.8844e-11, 2.3078e-11, 7.7503e-11,\n 5.8586e-10, 1.1981e-10, 2.0451e-11, 3.3347e-09, 4.0876e-10, 2.8497e-10,\n 2.1204e-11, 1.8071e-13, 3.9220e-10, 1.5491e-09, 1.9407e-10, 1.6347e-10,\n 3.4039e-11, 1.8930e-10, 6.5687e-10, 5.2109e-10, 3.8737e-10, 2.3811e-10,\n 3.2190e-11, 6.8966e-10, 2.1360e-12, 1.5798e-09, 8.7115e-10, 1.9459e-10,\n 1.1192e-11, 9.1193e-12, 6.5612e-10, 8.3707e-12, 2.8860e-10, 7.5488e-11,\n 8.1449e-11, 1.7006e-10, 8.2869e-10, 4.4465e-10, 5.7348e-12, 3.2166e-11,\n 3.2065e-10, 2.9924e-11, 4.2928e-10, 1.4945e-11, 3.4936e-12, 1.9772e-10,\n 2.4509e-10, 8.4458e-12, 9.2488e-11, 3.0541e-11, 6.5815e-11, 8.9119e-10,\n 8.3103e-12, 1.2761e-12, 9.1461e-11, 6.7872e-10, 1.0244e-12, 7.5936e-11,\n 1.9041e-11, 4.8538e-10, 2.3060e-10, 1.1634e-10, 1.4143e-10, 2.6723e-09,\n 4.4000e-11, 6.4974e-10, 6.0892e-10, 8.9904e-11, 1.5549e-09, 8.3573e-11,\n 1.0139e-10, 7.2037e-10, 3.0528e-10, 3.7339e-10, 7.6367e-10, 6.7124e-11,\n 1.8289e-11, 2.0840e-10, 3.2203e-12, 1.3071e-09, 5.2690e-11, 2.1118e-10,\n 3.8233e-11, 4.7360e-12, 1.0743e-09, 1.1504e-10, 1.3261e-09, 3.2677e-10,\n 2.7414e-11, 5.6093e-11, 1.0599e-10, 5.2924e-11, 6.4400e-11, 1.6574e-10,\n 2.8372e-10, 1.7094e-10, 2.1399e-11, 2.2187e-14, 1.4066e-12, 1.0553e-10,\n 8.1893e-10, 5.6675e-10, 5.0968e-11, 5.0969e-10, 2.2208e-13, 3.3535e-09,\n 1.7052e-10, 3.0045e-10, 3.8258e-10, 1.0793e-09, 4.5523e-10, 1.1130e-09,\n 3.8140e-10, 6.5415e-13, 1.5090e-11, 4.0560e-11, 1.2106e-10, 1.1785e-09,\n 2.5912e-10, 8.9330e-11, 7.3604e-11, 2.9705e-11, 1.1801e-09, 1.2077e-11,\n 4.2920e-11, 9.6323e-11, 1.5386e-10, 3.7455e-10, 1.2430e-10, 4.1111e-11,\n 8.1850e-11, 1.0952e-09, 2.7214e-11, 1.8147e-09, 3.7600e-10, 4.9534e-10,\n 4.0664e-11, 6.6865e-10, 5.3135e-10, 2.9634e-10, 3.1721e-09, 4.3411e-10,\n 5.9806e-10, 8.8899e-11, 1.1098e-11, 6.9434e-11, 5.8382e-10, 2.4338e-10,\n 3.1419e-10, 8.0057e-12, 1.8930e-11, 2.7636e-11, 2.6618e-11, 1.8629e-10,\n 1.5895e-11, 8.1691e-11, 5.9509e-11, 9.0105e-11, 9.3842e-10, 8.2597e-14,\n 7.3887e-11, 2.0644e-10, 3.5127e-11, 1.5004e-09, 1.2786e-10, 1.5815e-10,\n 2.7894e-10, 2.6667e-11, 1.2857e-12, 4.1442e-09, 2.9291e-11, 3.3736e-11,\n 3.5331e-10, 6.7891e-10, 3.2436e-12, 2.4949e-10, 6.6971e-10, 5.6499e-12,\n 5.2977e-10, 1.4950e-09, 2.2074e-10, 3.5542e-11, 2.9742e-14, 2.0979e-09,\n 2.3286e-11, 8.3823e-10, 4.3327e-11, 1.3367e-10, 7.2954e-12, 1.8562e-10,\n 2.4408e-10, 1.0046e-12, 3.3909e-11, 1.6341e-11, 4.8991e-11, 4.0902e-10,\n 9.2506e-11, 1.3889e-09, 2.7647e-10, 1.9190e-11, 2.6634e-13, 3.2972e-11,\n 2.7787e-11, 1.4231e-10, 1.0226e-11, 5.6675e-12, 2.3780e-10, 2.1820e-10,\n 7.3911e-10, 2.5406e-10, 2.9269e-11, 4.0628e-10, 8.3612e-11, 7.1657e-11,\n 4.0865e-10, 1.7812e-13, 5.6721e-10, 4.9729e-10, 1.5191e-10, 6.0491e-13,\n 3.0285e-12, 1.7531e-09, 4.2234e-12, 7.6834e-10, 1.9844e-10, 2.5766e-11,\n 2.6376e-10, 2.6655e-10, 6.7354e-10, 3.9561e-11], device='cuda:0')" }, "26": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([[-5.6052e-45, -5.6052e-45, 0.0000e+00, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [-5.6052e-45, -5.6052e-45, 0.0000e+00, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n ...,\n [ 5.6052e-45, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [-5.6052e-45, -5.6052e-45, 0.0000e+00, ..., -5.6052e-45,\n -5.6052e-45, 5.6052e-45]], device='cuda:0')", - "exp_avg_sq": "tensor([[2.0234e-10, 2.7969e-10, 0.0000e+00, ..., 5.9202e-10, 9.5464e-10,\n 1.1661e-10],\n [1.8706e-11, 2.6006e-11, 0.0000e+00, ..., 7.2681e-13, 2.3430e-11,\n 7.4823e-13],\n [1.5120e-10, 1.1815e-10, 0.0000e+00, ..., 1.2162e-10, 1.3135e-10,\n 8.5434e-11],\n ...,\n [7.5290e-11, 6.1873e-11, 0.0000e+00, ..., 4.3647e-11, 4.2628e-10,\n 1.2303e-11],\n [4.4183e-10, 1.3673e-10, 0.0000e+00, ..., 1.1783e-10, 9.6572e-10,\n 2.1387e-10],\n [8.9350e-11, 1.5708e-11, 0.0000e+00, ..., 2.1673e-11, 1.8618e-10,\n 1.9573e-12]], device='cuda:0')" + "exp_avg_sq": "tensor([[5.7820e-11, 7.9925e-11, 0.0000e+00, ..., 1.6917e-10, 2.7280e-10,\n 3.3323e-11],\n [5.3454e-12, 7.4315e-12, 0.0000e+00, ..., 2.0769e-13, 6.6953e-12,\n 2.1381e-13],\n [4.3205e-11, 3.3764e-11, 0.0000e+00, ..., 3.4753e-11, 3.7534e-11,\n 2.4413e-11],\n ...,\n [2.1515e-11, 1.7681e-11, 0.0000e+00, ..., 1.2472e-11, 1.2181e-10,\n 3.5156e-12],\n [1.2626e-10, 3.9072e-11, 0.0000e+00, ..., 3.3672e-11, 2.7596e-10,\n 6.1115e-11],\n [2.5532e-11, 4.4888e-12, 0.0000e+00, ..., 6.1932e-12, 5.3201e-11,\n 5.5933e-13]], device='cuda:0')" }, "27": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([3.4292e-07, 1.5379e-08, 1.2207e-07, 1.5861e-07, 4.8299e-08, 1.2175e-08,\n 1.6021e-07, 3.8991e-09, 7.1899e-08, 1.0973e-08, 3.6453e-07, 1.2616e-06,\n 2.1478e-07, 2.5258e-07, 2.6353e-07, 1.0908e-06, 1.2225e-07, 1.2120e-07,\n 4.0910e-07, 4.7564e-07, 1.5253e-09, 5.1103e-08, 1.0307e-07, 1.2779e-07,\n 1.1836e-07, 2.1941e-07, 1.7443e-07, 3.4260e-08, 5.0063e-09, 5.7023e-08,\n 3.0303e-07, 2.0021e-07, 3.1393e-08, 1.5398e-06, 5.5774e-07, 3.4184e-07,\n 1.0169e-08, 1.9274e-10, 1.4339e-07, 4.6304e-06, 6.9076e-07, 9.7114e-08,\n 3.0442e-08, 1.9053e-08, 7.9862e-07, 1.4064e-06, 2.1907e-07, 4.8531e-07,\n 6.0459e-08, 3.8851e-07, 7.4923e-09, 8.2714e-07, 2.1089e-07, 1.8999e-07,\n 3.0194e-08, 2.6338e-09, 3.6017e-07, 1.3366e-08, 2.3938e-07, 2.0948e-08,\n 1.2301e-07, 2.3550e-07, 5.9367e-07, 3.1640e-07, 8.9238e-09, 2.2080e-08,\n 2.3738e-07, 1.5454e-08, 1.5052e-07, 2.8601e-08, 5.2214e-10, 1.4587e-07,\n 6.1653e-07, 4.5282e-08, 1.2482e-07, 8.3442e-10, 8.1062e-09, 5.0358e-07,\n 3.2498e-08, 1.0731e-11, 2.9807e-07, 3.5872e-07, 5.1369e-10, 3.0450e-08,\n 2.1253e-07, 9.0778e-07, 7.2971e-08, 3.1344e-07, 6.2089e-08, 5.1776e-07,\n 3.6666e-07, 2.2553e-07, 5.7323e-07, 1.0459e-07, 1.8470e-07, 2.8309e-08,\n 9.6295e-08, 6.5926e-07, 1.5996e-07, 1.3314e-06, 1.1794e-06, 2.8858e-08,\n 5.1294e-10, 1.5840e-07, 8.9022e-09, 1.0445e-06, 8.3708e-08, 7.6448e-08,\n 2.2779e-07, 1.4205e-09, 1.1291e-07, 7.3680e-10, 7.3589e-07, 7.2085e-07,\n 5.6823e-08, 6.3016e-07, 5.3809e-08, 1.0278e-07, 1.5227e-07, 1.5640e-07,\n 5.4506e-08, 1.2401e-08, 2.2363e-08, 2.9666e-11, 2.9060e-09, 1.3997e-07,\n 1.1895e-06, 1.9646e-07, 4.9678e-07, 2.2589e-07, 1.2475e-09, 1.4334e-06,\n 2.3111e-07, 2.1361e-07, 5.2165e-07, 3.0438e-07, 3.3921e-07, 4.5517e-07,\n 3.0918e-07, 1.1912e-09, 9.0227e-09, 1.6624e-08, 1.2241e-07, 1.5577e-06,\n 2.9761e-07, 1.8181e-08, 2.8067e-07, 3.3905e-08, 2.8656e-07, 7.0887e-09,\n 1.3810e-08, 4.8559e-07, 3.5425e-07, 4.4332e-07, 9.4137e-08, 1.5870e-08,\n 2.8228e-07, 8.3567e-07, 1.4661e-07, 1.8473e-07, 2.2823e-07, 2.6203e-07,\n 1.4225e-07, 1.8994e-07, 1.7784e-07, 1.6739e-07, 1.2016e-06, 9.2548e-07,\n 1.3509e-07, 5.0971e-07, 5.0747e-08, 3.6620e-08, 7.5627e-07, 7.1954e-08,\n 5.4010e-07, 4.0173e-08, 1.5457e-08, 4.8767e-08, 5.2153e-08, 5.7991e-08,\n 6.8355e-08, 5.8835e-08, 2.1431e-08, 1.5267e-07, 2.3493e-06, 3.5886e-10,\n 1.1647e-07, 9.7663e-08, 6.8948e-08, 7.0462e-07, 2.0714e-07, 2.5403e-08,\n 1.0476e-07, 1.4443e-07, 1.0970e-09, 1.2797e-06, 6.7338e-08, 1.0255e-08,\n 3.1984e-07, 1.4192e-06, 1.1884e-12, 1.3878e-07, 6.8884e-07, 1.5330e-07,\n 1.2801e-07, 9.9941e-07, 2.4717e-07, 5.3252e-08, 5.3080e-11, 1.1986e-06,\n 2.4347e-08, 1.2010e-06, 7.3789e-10, 2.6632e-07, 3.8223e-09, 3.1999e-07,\n 4.5041e-07, 7.6569e-12, 3.8076e-08, 5.2944e-08, 4.5068e-08, 1.4657e-06,\n 7.7488e-08, 3.0744e-07, 6.9219e-08, 2.8481e-09, 1.5856e-09, 5.6487e-09,\n 5.1789e-08, 1.0319e-07, 8.6511e-09, 1.2694e-09, 5.3993e-07, 3.8538e-08,\n 2.8329e-08, 1.9755e-07, 1.6804e-07, 2.4726e-07, 1.0226e-07, 7.9346e-08,\n 2.6814e-08, 6.0742e-11, 2.0359e-07, 1.3201e-07, 5.9060e-08, 1.8866e-09,\n 1.2189e-09, 1.0085e-06, 4.6859e-08, 1.1628e-06, 3.6963e-07, 4.0348e-08,\n 5.5499e-07, 1.5686e-07, 3.2767e-07, 6.2278e-08], device='cuda:0')" + "exp_avg_sq": "tensor([9.7991e-08, 4.3947e-09, 3.4882e-08, 4.5325e-08, 1.3802e-08, 3.4790e-09,\n 4.5782e-08, 1.1142e-09, 2.0546e-08, 3.1357e-09, 1.0417e-07, 3.6052e-07,\n 6.1376e-08, 7.2178e-08, 7.5306e-08, 3.1171e-07, 3.4935e-08, 3.4634e-08,\n 1.1690e-07, 1.3592e-07, 4.3586e-10, 1.4603e-08, 2.9453e-08, 3.6517e-08,\n 3.3821e-08, 6.2698e-08, 4.9844e-08, 9.7901e-09, 1.4306e-09, 1.6295e-08,\n 8.6594e-08, 5.7211e-08, 8.9708e-09, 4.4002e-07, 1.5938e-07, 9.7684e-08,\n 2.9059e-09, 5.5076e-11, 4.0976e-08, 1.3232e-06, 1.9739e-07, 2.7751e-08,\n 8.6989e-09, 5.4445e-09, 2.2821e-07, 4.0188e-07, 6.2600e-08, 1.3868e-07,\n 1.7277e-08, 1.1102e-07, 2.1410e-09, 2.3636e-07, 6.0264e-08, 5.4292e-08,\n 8.6282e-09, 7.5263e-10, 1.0292e-07, 3.8194e-09, 6.8404e-08, 5.9862e-09,\n 3.5152e-08, 6.7296e-08, 1.6965e-07, 9.0413e-08, 2.5500e-09, 6.3095e-09,\n 6.7832e-08, 4.4160e-09, 4.3013e-08, 8.1728e-09, 1.4921e-10, 4.1684e-08,\n 1.7618e-07, 1.2940e-08, 3.5669e-08, 2.3844e-10, 2.3164e-09, 1.4390e-07,\n 9.2864e-09, 3.0665e-12, 8.5176e-08, 1.0251e-07, 1.4679e-10, 8.7012e-09,\n 6.0731e-08, 2.5940e-07, 2.0852e-08, 8.9567e-08, 1.7743e-08, 1.4795e-07,\n 1.0478e-07, 6.4446e-08, 1.6381e-07, 2.9888e-08, 5.2780e-08, 8.0896e-09,\n 2.7517e-08, 1.8839e-07, 4.5710e-08, 3.8046e-07, 3.3703e-07, 8.2463e-09,\n 1.4658e-10, 4.5265e-08, 2.5439e-09, 2.9846e-07, 2.3920e-08, 2.1846e-08,\n 6.5094e-08, 4.0592e-10, 3.2265e-08, 2.1055e-10, 2.1029e-07, 2.0599e-07,\n 1.6238e-08, 1.8007e-07, 1.5376e-08, 2.9369e-08, 4.3512e-08, 4.4692e-08,\n 1.5575e-08, 3.5438e-09, 6.3904e-09, 8.4773e-12, 8.3041e-10, 3.9997e-08,\n 3.3991e-07, 5.6140e-08, 1.4196e-07, 6.4549e-08, 3.5647e-10, 4.0962e-07,\n 6.6040e-08, 6.1042e-08, 1.4907e-07, 8.6978e-08, 9.6932e-08, 1.3007e-07,\n 8.8351e-08, 3.4040e-10, 2.5783e-09, 4.7506e-09, 3.4979e-08, 4.4511e-07,\n 8.5044e-08, 5.1952e-09, 8.0203e-08, 9.6887e-09, 8.1886e-08, 2.0256e-09,\n 3.9463e-09, 1.3876e-07, 1.0123e-07, 1.2668e-07, 2.6900e-08, 4.5351e-09,\n 8.0664e-08, 2.3880e-07, 4.1894e-08, 5.2787e-08, 6.5218e-08, 7.4877e-08,\n 4.0651e-08, 5.4278e-08, 5.0819e-08, 4.7833e-08, 3.4336e-07, 2.6446e-07,\n 3.8604e-08, 1.4565e-07, 1.4501e-08, 1.0464e-08, 2.1611e-07, 2.0561e-08,\n 1.5434e-07, 1.1480e-08, 4.4169e-09, 1.3936e-08, 1.4903e-08, 1.6571e-08,\n 1.9533e-08, 1.6813e-08, 6.1242e-09, 4.3627e-08, 6.7132e-07, 1.0255e-10,\n 3.3282e-08, 2.7908e-08, 1.9702e-08, 2.0135e-07, 5.9192e-08, 7.2591e-09,\n 2.9935e-08, 4.1273e-08, 3.1348e-10, 3.6569e-07, 1.9242e-08, 2.9304e-09,\n 9.1396e-08, 4.0555e-07, 3.3960e-13, 3.9658e-08, 1.9684e-07, 4.3807e-08,\n 3.6580e-08, 2.8559e-07, 7.0631e-08, 1.5217e-08, 1.5168e-11, 3.4251e-07,\n 6.9574e-09, 3.4320e-07, 2.1086e-10, 7.6103e-08, 1.0923e-09, 9.1440e-08,\n 1.2871e-07, 2.1880e-12, 1.0881e-08, 1.5129e-08, 1.2879e-08, 4.1884e-07,\n 2.2143e-08, 8.7854e-08, 1.9780e-08, 8.1388e-10, 4.5310e-10, 1.6142e-09,\n 1.4799e-08, 2.9487e-08, 2.4721e-09, 3.6273e-10, 1.5429e-07, 1.1012e-08,\n 8.0952e-09, 5.6452e-08, 4.8020e-08, 7.0656e-08, 2.9222e-08, 2.2674e-08,\n 7.6623e-09, 1.7357e-11, 5.8176e-08, 3.7723e-08, 1.6877e-08, 5.3911e-10,\n 3.4832e-10, 2.8819e-07, 1.3390e-08, 3.3227e-07, 1.0563e-07, 1.1530e-08,\n 1.5859e-07, 4.4823e-08, 9.3635e-08, 1.7796e-08], device='cuda:0')" }, "28": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([6.1357e-10, 7.6116e-11, 2.2349e-10, 5.1826e-10, 9.2447e-11, 5.7124e-11,\n 3.2980e-10, 2.1854e-11, 1.1180e-10, 1.3604e-11, 6.8368e-10, 3.3352e-09,\n 4.6885e-10, 6.3183e-10, 1.0134e-09, 4.5225e-09, 3.6887e-10, 2.9798e-10,\n 1.3793e-09, 9.4312e-10, 1.6680e-13, 1.3533e-10, 3.8361e-10, 2.6008e-10,\n 4.2203e-10, 5.2920e-10, 3.6310e-10, 1.0808e-10, 1.2863e-11, 9.9823e-11,\n 8.2028e-10, 4.7638e-10, 7.3883e-11, 3.8638e-09, 1.4590e-09, 1.0320e-09,\n 2.1786e-11, 8.2307e-14, 3.4790e-10, 2.8307e-08, 1.5500e-09, 3.1066e-10,\n 1.0166e-10, 5.6851e-11, 2.7035e-09, 6.2059e-09, 6.7633e-10, 2.2107e-09,\n 6.5660e-11, 8.0579e-10, 7.4653e-12, 2.0891e-09, 4.0478e-10, 4.1982e-10,\n 4.2732e-11, 9.9821e-12, 1.1542e-09, 1.1361e-11, 9.7795e-10, 7.2760e-11,\n 2.8495e-10, 1.6909e-09, 2.8306e-09, 1.1081e-09, 4.1752e-12, 2.5547e-11,\n 6.4271e-10, 1.3645e-11, 2.4645e-10, 9.2796e-11, 2.0463e-11, 4.3996e-10,\n 1.4336e-09, 1.0685e-10, 2.9925e-10, 8.7571e-12, 2.9134e-11, 1.4443e-09,\n 4.6088e-11, 8.2071e-13, 1.1077e-09, 7.7190e-10, 8.9657e-13, 1.4793e-10,\n 8.4217e-10, 2.6415e-09, 2.1130e-10, 1.3375e-09, 1.5195e-10, 1.1864e-09,\n 2.0778e-09, 5.6486e-10, 1.4163e-09, 1.5316e-10, 3.9951e-10, 2.7661e-11,\n 1.9439e-10, 1.1710e-09, 3.7646e-10, 3.7258e-09, 7.7219e-09, 1.3167e-10,\n 2.9453e-12, 3.6112e-10, 4.7184e-12, 2.3098e-09, 2.5524e-10, 2.0824e-10,\n 6.8160e-10, 5.8898e-15, 2.9680e-10, 1.4896e-11, 2.2516e-09, 1.4207e-09,\n 7.9371e-11, 1.6591e-09, 1.8002e-10, 1.9715e-10, 3.8195e-10, 4.1146e-10,\n 7.1506e-11, 5.0900e-11, 4.1398e-11, 1.7554e-13, 1.1573e-11, 2.6390e-10,\n 3.0294e-09, 6.1013e-10, 2.5281e-09, 5.1205e-10, 2.6915e-13, 3.7432e-09,\n 1.0725e-09, 4.1514e-10, 1.6262e-09, 5.5245e-10, 5.8051e-10, 7.7324e-10,\n 6.6626e-10, 6.5475e-14, 4.4982e-11, 1.0124e-10, 1.4560e-10, 4.4484e-09,\n 7.6407e-10, 5.3760e-11, 1.2731e-09, 8.9191e-11, 5.0683e-10, 1.4357e-12,\n 3.7796e-11, 1.2618e-09, 2.1844e-09, 1.7964e-09, 9.4807e-11, 1.3618e-10,\n 1.3369e-09, 2.2281e-09, 3.2231e-10, 3.9726e-10, 3.1078e-10, 4.3808e-10,\n 6.9766e-10, 3.5022e-10, 5.5325e-10, 3.9174e-10, 2.1765e-09, 3.0525e-09,\n 2.7795e-10, 1.2003e-09, 9.3548e-11, 5.3298e-11, 1.6596e-09, 1.4156e-10,\n 1.5795e-09, 2.4177e-10, 3.3510e-11, 1.1331e-10, 1.0172e-10, 2.0379e-10,\n 9.2085e-11, 1.2832e-10, 2.4162e-11, 2.8374e-10, 7.2241e-09, 2.4422e-12,\n 1.8499e-10, 1.7751e-10, 7.6793e-11, 1.5457e-09, 3.0601e-10, 4.2227e-11,\n 3.6811e-10, 6.6436e-10, 4.5912e-12, 2.8421e-09, 2.2770e-10, 1.5748e-11,\n 1.4846e-09, 3.5159e-09, 4.6932e-13, 3.1615e-10, 2.5123e-09, 5.0447e-10,\n 4.7339e-10, 1.9874e-09, 4.8110e-10, 1.3828e-10, 1.9403e-13, 2.5626e-09,\n 3.8373e-11, 3.2590e-09, 8.5017e-12, 7.1237e-10, 4.0002e-12, 7.2201e-10,\n 6.3470e-10, 3.4451e-12, 1.2764e-10, 9.3190e-11, 1.1396e-10, 6.6383e-09,\n 2.8101e-10, 8.0463e-10, 1.9227e-10, 3.0841e-11, 4.3818e-14, 3.2399e-11,\n 7.1311e-11, 2.3551e-10, 3.3315e-11, 2.4189e-14, 2.0119e-09, 1.3016e-10,\n 1.0294e-10, 7.1375e-10, 6.1069e-10, 5.1120e-10, 4.6088e-10, 1.7784e-10,\n 1.3840e-10, 1.2400e-12, 6.8334e-10, 2.2100e-10, 1.0772e-10, 1.8963e-12,\n 2.1963e-12, 3.3529e-09, 1.0730e-10, 3.7964e-09, 6.2923e-10, 6.9562e-11,\n 2.2586e-09, 2.8443e-10, 5.9109e-10, 4.0259e-10], device='cuda:0')" + "exp_avg_sq": "tensor([1.7533e-10, 2.1751e-11, 6.3863e-11, 1.4810e-10, 2.6418e-11, 1.6324e-11,\n 9.4243e-11, 6.2449e-12, 3.1948e-11, 3.8874e-12, 1.9537e-10, 9.5306e-10,\n 1.3398e-10, 1.8055e-10, 2.8960e-10, 1.2923e-09, 1.0541e-10, 8.5149e-11,\n 3.9413e-10, 2.6950e-10, 4.7665e-14, 3.8671e-11, 1.0962e-10, 7.4320e-11,\n 1.2060e-10, 1.5122e-10, 1.0376e-10, 3.0885e-11, 3.6758e-12, 2.8525e-11,\n 2.3440e-10, 1.3613e-10, 2.1113e-11, 1.1041e-09, 4.1692e-10, 2.9489e-10,\n 6.2254e-12, 2.3520e-14, 9.9415e-11, 8.0888e-09, 4.4292e-10, 8.8774e-11,\n 2.9050e-11, 1.6246e-11, 7.7254e-10, 1.7734e-09, 1.9327e-10, 6.3171e-10,\n 1.8763e-11, 2.3026e-10, 2.1333e-12, 5.9697e-10, 1.1567e-10, 1.1997e-10,\n 1.2211e-11, 2.8525e-12, 3.2983e-10, 3.2465e-12, 2.7946e-10, 2.0792e-11,\n 8.1425e-11, 4.8319e-10, 8.0888e-10, 3.1666e-10, 1.1931e-12, 7.3003e-12,\n 1.8366e-10, 3.8991e-12, 7.0424e-11, 2.6517e-11, 5.8474e-12, 1.2572e-10,\n 4.0965e-10, 3.0533e-11, 8.5512e-11, 2.5024e-12, 8.3254e-12, 4.1272e-10,\n 1.3170e-11, 2.3453e-13, 3.1652e-10, 2.2058e-10, 2.5620e-13, 4.2271e-11,\n 2.4066e-10, 7.5482e-10, 6.0379e-11, 3.8222e-10, 4.3421e-11, 3.3902e-10,\n 5.9375e-10, 1.6141e-10, 4.0471e-10, 4.3768e-11, 1.1416e-10, 7.9044e-12,\n 5.5549e-11, 3.3462e-10, 1.0758e-10, 1.0647e-09, 2.2066e-09, 3.7626e-11,\n 8.4164e-13, 1.0319e-10, 1.3483e-12, 6.6003e-10, 7.2937e-11, 5.9506e-11,\n 1.9477e-10, 1.6831e-15, 8.4814e-11, 4.2568e-12, 6.4342e-10, 4.0598e-10,\n 2.2681e-11, 4.7411e-10, 5.1443e-11, 5.6337e-11, 1.0915e-10, 1.1758e-10,\n 2.0433e-11, 1.4545e-11, 1.1830e-11, 5.0162e-14, 3.3071e-12, 7.5412e-11,\n 8.6566e-10, 1.7435e-10, 7.2243e-10, 1.4632e-10, 7.6911e-14, 1.0696e-09,\n 3.0647e-10, 1.1863e-10, 4.6470e-10, 1.5787e-10, 1.6588e-10, 2.2096e-10,\n 1.9039e-10, 1.8710e-14, 1.2854e-11, 2.8932e-11, 4.1608e-11, 1.2712e-09,\n 2.1834e-10, 1.5362e-11, 3.6379e-10, 2.5487e-11, 1.4483e-10, 4.1026e-13,\n 1.0800e-11, 3.6055e-10, 6.2421e-10, 5.1333e-10, 2.7092e-11, 3.8915e-11,\n 3.8204e-10, 6.3669e-10, 9.2102e-11, 1.1352e-10, 8.8809e-11, 1.2518e-10,\n 1.9936e-10, 1.0008e-10, 1.5809e-10, 1.1194e-10, 6.2194e-10, 8.7227e-10,\n 7.9425e-11, 3.4301e-10, 2.6732e-11, 1.5230e-11, 4.7424e-10, 4.0453e-11,\n 4.5135e-10, 6.9089e-11, 9.5759e-12, 3.2379e-11, 2.9066e-11, 5.8234e-11,\n 2.6314e-11, 3.6669e-11, 6.9043e-12, 8.1080e-11, 2.0643e-09, 6.9788e-13,\n 5.2862e-11, 5.0726e-11, 2.1944e-11, 4.4169e-10, 8.7444e-11, 1.2067e-11,\n 1.0519e-10, 1.8985e-10, 1.3120e-12, 8.1215e-10, 6.5068e-11, 4.5002e-12,\n 4.2424e-10, 1.0047e-09, 1.3411e-13, 9.0344e-11, 7.1792e-10, 1.4416e-10,\n 1.3527e-10, 5.6792e-10, 1.3748e-10, 3.9515e-11, 5.5446e-14, 7.3230e-10,\n 1.0966e-11, 9.3127e-10, 2.4294e-12, 2.0357e-10, 1.1431e-12, 2.0632e-10,\n 1.8137e-10, 9.8445e-13, 3.6474e-11, 2.6630e-11, 3.2565e-11, 1.8969e-09,\n 8.0300e-11, 2.2993e-10, 5.4943e-11, 8.8130e-12, 1.2521e-14, 9.2584e-12,\n 2.0378e-11, 6.7298e-11, 9.5200e-12, 6.9121e-15, 5.7492e-10, 3.7194e-11,\n 2.9416e-11, 2.0396e-10, 1.7451e-10, 1.4608e-10, 1.3170e-10, 5.0818e-11,\n 3.9549e-11, 3.5435e-13, 1.9527e-10, 6.3153e-11, 3.0781e-11, 5.4188e-13,\n 6.2762e-13, 9.5813e-10, 3.0663e-11, 1.0849e-09, 1.7981e-10, 1.9878e-11,\n 6.4541e-10, 8.1277e-11, 1.6891e-10, 1.1504e-10], device='cuda:0')" }, "29": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([1.3685e-09, 7.5201e-11, 4.0876e-10, 6.9613e-10, 1.1196e-10, 7.7763e-11,\n 4.9385e-10, 3.5380e-11, 2.0388e-10, 2.2288e-11, 1.2031e-09, 4.6042e-09,\n 7.1720e-10, 1.0280e-09, 1.1987e-09, 4.6856e-09, 6.2974e-10, 4.0899e-10,\n 1.3451e-09, 2.0523e-09, 2.3125e-14, 2.4105e-10, 4.9941e-10, 5.9053e-10,\n 5.7631e-10, 7.2841e-10, 7.7883e-10, 1.8352e-10, 3.1721e-11, 1.7787e-10,\n 1.3700e-09, 8.6115e-10, 1.6532e-10, 5.5894e-09, 2.4122e-09, 1.5470e-09,\n 1.7479e-11, 3.7817e-13, 6.3789e-10, 1.6849e-08, 2.4104e-09, 4.8971e-10,\n 1.6528e-10, 1.1222e-10, 2.7781e-09, 4.9319e-09, 1.0066e-09, 2.1939e-09,\n 1.6399e-10, 1.3557e-09, 1.3034e-11, 3.5164e-09, 6.9390e-10, 6.5187e-10,\n 7.7787e-11, 2.5677e-11, 1.5680e-09, 3.4785e-11, 1.1276e-09, 1.1811e-10,\n 3.6005e-10, 1.1561e-09, 2.5725e-09, 1.4086e-09, 1.4716e-11, 6.3426e-11,\n 1.0724e-09, 3.3479e-11, 6.6313e-10, 1.4122e-10, 1.3214e-11, 6.9884e-10,\n 2.0594e-09, 2.4538e-10, 3.8339e-10, 7.3580e-12, 4.7086e-11, 2.1978e-09,\n 6.8074e-11, 9.5487e-13, 1.3197e-09, 1.6055e-09, 3.4425e-12, 1.7231e-10,\n 1.0141e-09, 3.9285e-09, 3.3127e-10, 1.3661e-09, 2.1355e-10, 2.2830e-09,\n 1.6922e-09, 1.0054e-09, 2.0085e-09, 3.3893e-10, 8.1020e-10, 8.2801e-11,\n 3.1520e-10, 2.4298e-09, 7.6286e-10, 5.3641e-09, 5.0923e-09, 1.7323e-10,\n 5.0588e-12, 4.8789e-10, 7.0504e-12, 3.8158e-09, 2.1471e-10, 3.6814e-10,\n 7.2640e-10, 1.4271e-13, 4.0219e-10, 4.0558e-12, 3.0712e-09, 2.5710e-09,\n 1.8134e-10, 2.1985e-09, 2.6174e-10, 3.2513e-10, 6.9180e-10, 6.6035e-10,\n 1.7331e-10, 4.1173e-11, 4.7700e-11, 4.5776e-13, 2.7306e-11, 4.4251e-10,\n 5.0469e-09, 9.0766e-10, 2.1868e-09, 7.3196e-10, 1.6368e-14, 5.7708e-09,\n 1.0956e-09, 9.6138e-10, 1.7770e-09, 1.0437e-09, 1.4771e-09, 1.6762e-09,\n 1.3564e-09, 1.4086e-13, 6.7037e-11, 1.1839e-10, 4.2122e-10, 5.4551e-09,\n 1.2529e-09, 1.0167e-10, 1.2824e-09, 1.7863e-10, 9.6593e-10, 1.1710e-11,\n 8.2755e-11, 1.6407e-09, 1.6092e-09, 1.4716e-09, 2.8734e-10, 1.1537e-10,\n 8.6409e-10, 3.0519e-09, 4.5327e-10, 8.5210e-10, 8.0976e-10, 8.8896e-10,\n 3.8526e-10, 8.1250e-10, 8.0056e-10, 5.4574e-10, 4.4148e-09, 3.2040e-09,\n 6.5334e-10, 2.2760e-09, 1.2647e-10, 1.2471e-10, 2.7020e-09, 3.3285e-10,\n 1.8105e-09, 2.4726e-10, 4.4562e-11, 2.4575e-10, 1.3402e-10, 2.9268e-10,\n 2.0309e-10, 1.7133e-10, 5.7228e-11, 4.7482e-10, 9.2146e-09, 4.5782e-12,\n 3.8935e-10, 3.0638e-10, 1.9926e-10, 2.9080e-09, 7.5911e-10, 7.8716e-11,\n 4.9382e-10, 7.2591e-10, 1.6838e-11, 5.5413e-09, 1.8213e-10, 2.6992e-11,\n 1.4897e-09, 5.3045e-09, 4.3077e-12, 6.3591e-10, 2.3764e-09, 4.2166e-10,\n 6.2755e-10, 3.8968e-09, 8.3113e-10, 1.6256e-10, 1.6535e-12, 4.3664e-09,\n 5.3242e-11, 4.8353e-09, 8.0175e-12, 1.1712e-09, 3.1695e-12, 1.1855e-09,\n 1.5564e-09, 4.0490e-12, 2.1136e-10, 1.6904e-10, 9.4946e-11, 6.0819e-09,\n 3.8507e-10, 1.4228e-09, 3.6179e-10, 4.0544e-11, 6.8266e-14, 5.2631e-11,\n 1.3619e-10, 4.5937e-10, 5.9480e-11, 7.2016e-13, 2.3813e-09, 1.9199e-10,\n 1.3793e-10, 9.1461e-10, 4.8923e-10, 1.1198e-09, 5.1198e-10, 2.3357e-10,\n 1.0897e-10, 2.9078e-12, 9.3567e-10, 4.5923e-10, 1.9727e-10, 2.3068e-12,\n 9.5820e-13, 3.4940e-09, 1.1141e-10, 4.8012e-09, 1.2457e-09, 1.0951e-10,\n 1.8249e-09, 5.0482e-10, 1.2479e-09, 3.4591e-10], device='cuda:0')" + "exp_avg_sq": "tensor([3.9107e-10, 2.1489e-11, 1.1681e-10, 1.9892e-10, 3.1992e-11, 2.2221e-11,\n 1.4112e-10, 1.0110e-11, 5.8261e-11, 6.3689e-12, 3.4380e-10, 1.3157e-09,\n 2.0495e-10, 2.9375e-10, 3.4255e-10, 1.3389e-09, 1.7995e-10, 1.1687e-10,\n 3.8436e-10, 5.8645e-10, 6.6082e-15, 6.8883e-11, 1.4271e-10, 1.6875e-10,\n 1.6469e-10, 2.0815e-10, 2.2256e-10, 5.2443e-11, 9.0644e-12, 5.0827e-11,\n 3.9149e-10, 2.4608e-10, 4.7242e-11, 1.5972e-09, 6.8931e-10, 4.4208e-10,\n 4.9947e-12, 1.0806e-13, 1.8228e-10, 4.8146e-09, 6.8878e-10, 1.3994e-10,\n 4.7230e-11, 3.2067e-11, 7.9386e-10, 1.4093e-09, 2.8766e-10, 6.2692e-10,\n 4.6860e-11, 3.8741e-10, 3.7245e-12, 1.0048e-09, 1.9829e-10, 1.8628e-10,\n 2.2228e-11, 7.3375e-12, 4.4806e-10, 9.9401e-12, 3.2223e-10, 3.3751e-11,\n 1.0289e-10, 3.3036e-10, 7.3511e-10, 4.0253e-10, 4.2051e-12, 1.8124e-11,\n 3.0645e-10, 9.5670e-12, 1.8949e-10, 4.0356e-11, 3.7761e-12, 1.9970e-10,\n 5.8850e-10, 7.0120e-11, 1.0956e-10, 2.1026e-12, 1.3455e-11, 6.2804e-10,\n 1.9453e-11, 2.7286e-13, 3.7710e-10, 4.5879e-10, 9.8373e-13, 4.9239e-11,\n 2.8980e-10, 1.1226e-09, 9.4663e-11, 3.9038e-10, 6.1023e-11, 6.5240e-10,\n 4.8355e-10, 2.8729e-10, 5.7393e-10, 9.6852e-11, 2.3152e-10, 2.3661e-11,\n 9.0070e-11, 6.9434e-10, 2.1799e-10, 1.5328e-09, 1.4552e-09, 4.9502e-11,\n 1.4456e-12, 1.3942e-10, 2.0147e-12, 1.0904e-09, 6.1354e-11, 1.0520e-10,\n 2.0758e-10, 4.0779e-14, 1.1493e-10, 1.1590e-12, 8.7763e-10, 7.3468e-10,\n 5.1819e-11, 6.2825e-10, 7.4795e-11, 9.2910e-11, 1.9769e-10, 1.8870e-10,\n 4.9525e-11, 1.1766e-11, 1.3631e-11, 1.3081e-13, 7.8030e-12, 1.2645e-10,\n 1.4422e-09, 2.5937e-10, 6.2489e-10, 2.0916e-10, 4.6773e-15, 1.6491e-09,\n 3.1307e-10, 2.7472e-10, 5.0778e-10, 2.9824e-10, 4.2208e-10, 4.7899e-10,\n 3.8760e-10, 4.0253e-14, 1.9156e-11, 3.3830e-11, 1.2037e-10, 1.5588e-09,\n 3.5803e-10, 2.9053e-11, 3.6645e-10, 5.1045e-11, 2.7602e-10, 3.3461e-12,\n 2.3648e-11, 4.6885e-10, 4.5986e-10, 4.2052e-10, 8.2109e-11, 3.2967e-11,\n 2.4692e-10, 8.7211e-10, 1.2952e-10, 2.4349e-10, 2.3140e-10, 2.5403e-10,\n 1.1009e-10, 2.3218e-10, 2.2877e-10, 1.5595e-10, 1.2616e-09, 9.1557e-10,\n 1.8670e-10, 6.5040e-10, 3.6139e-11, 3.5636e-11, 7.7212e-10, 9.5116e-11,\n 5.1735e-10, 7.0657e-11, 1.2734e-11, 7.0225e-11, 3.8298e-11, 8.3636e-11,\n 5.8034e-11, 4.8960e-11, 1.6353e-11, 1.3568e-10, 2.6331e-09, 1.3082e-12,\n 1.1126e-10, 8.7550e-11, 5.6940e-11, 8.3099e-10, 2.1692e-10, 2.2494e-11,\n 1.4111e-10, 2.0744e-10, 4.8116e-12, 1.5835e-09, 5.2046e-11, 7.7132e-12,\n 4.2569e-10, 1.5158e-09, 1.2309e-12, 1.8172e-10, 6.7908e-10, 1.2049e-10,\n 1.7933e-10, 1.1135e-09, 2.3750e-10, 4.6452e-11, 4.7250e-13, 1.2477e-09,\n 1.5214e-11, 1.3817e-09, 2.2911e-12, 3.3467e-10, 9.0570e-13, 3.3876e-10,\n 4.4476e-10, 1.1570e-12, 6.0398e-11, 4.8304e-11, 2.7131e-11, 1.7380e-09,\n 1.1004e-10, 4.0659e-10, 1.0338e-10, 1.1586e-11, 1.9508e-14, 1.5040e-11,\n 3.8918e-11, 1.3127e-10, 1.6997e-11, 2.0579e-13, 6.8048e-10, 5.4863e-11,\n 3.9413e-11, 2.6136e-10, 1.3980e-10, 3.1998e-10, 1.4630e-10, 6.6744e-11,\n 3.1139e-11, 8.3092e-13, 2.6738e-10, 1.3123e-10, 5.6371e-11, 6.5920e-13,\n 2.7381e-13, 9.9844e-10, 3.1836e-11, 1.3720e-09, 3.5598e-10, 3.1293e-11,\n 5.2147e-10, 1.4426e-10, 3.5660e-10, 9.8846e-11], device='cuda:0')" }, "30": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([[-5.6052e-45, -5.6052e-45, 0.0000e+00, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [-5.6052e-45, -5.6052e-45, 0.0000e+00, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n ...,\n [ 5.6052e-45, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [-5.6052e-45, -5.6052e-45, 0.0000e+00, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.2601e-10, 4.2558e-10, 0.0000e+00, ..., 1.4640e-09, 8.3305e-10,\n 8.9057e-10],\n [2.0131e-10, 7.5190e-11, 0.0000e+00, ..., 6.2886e-11, 7.2841e-10,\n 3.2847e-11],\n [1.6396e-10, 2.8932e-10, 0.0000e+00, ..., 2.0125e-10, 1.3653e-09,\n 8.0719e-11],\n ...,\n [3.9639e-11, 1.0792e-10, 0.0000e+00, ..., 1.3726e-10, 2.0400e-10,\n 1.2594e-12],\n [9.0506e-10, 8.0456e-10, 0.0000e+00, ..., 1.1700e-09, 4.9905e-09,\n 1.0174e-09],\n [7.3065e-12, 9.7694e-12, 0.0000e+00, ..., 3.7683e-11, 8.8126e-11,\n 9.1702e-13]], device='cuda:0')" + "exp_avg_sq": "tensor([[3.6009e-11, 1.2161e-10, 0.0000e+00, ..., 4.1834e-10, 2.3805e-10,\n 2.5449e-10],\n [5.7525e-11, 2.1486e-11, 0.0000e+00, ..., 1.7970e-11, 2.0815e-10,\n 9.3863e-12],\n [4.6852e-11, 8.2675e-11, 0.0000e+00, ..., 5.7510e-11, 3.9014e-10,\n 2.3066e-11],\n ...,\n [1.1327e-11, 3.0838e-11, 0.0000e+00, ..., 3.9222e-11, 5.8295e-11,\n 3.5988e-13],\n [2.5863e-10, 2.2991e-10, 0.0000e+00, ..., 3.3432e-10, 1.4261e-09,\n 2.9072e-10],\n [2.0879e-12, 2.7917e-12, 0.0000e+00, ..., 1.0768e-11, 2.5183e-11,\n 2.6205e-13]], device='cuda:0')" }, "31": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([7.8087e-07, 2.4273e-07, 3.0864e-07, 9.6051e-09, 8.7272e-08, 9.2314e-09,\n 9.4683e-08, 8.0973e-09, 5.5327e-08, 3.5461e-08, 7.2848e-07, 7.1804e-07,\n 9.8548e-08, 7.1144e-08, 1.3920e-07, 1.8832e-08, 3.5354e-07, 1.0517e-07,\n 3.5314e-07, 9.8322e-07, 1.6173e-09, 2.8185e-07, 6.2106e-08, 5.7699e-07,\n 4.2558e-08, 9.3457e-08, 6.1794e-07, 9.9897e-09, 3.5086e-08, 1.0713e-07,\n 1.8584e-07, 6.1174e-07, 8.0845e-10, 1.1010e-06, 6.3347e-07, 7.5201e-07,\n 2.8523e-08, 1.0167e-09, 4.0864e-07, 2.6982e-06, 1.1340e-06, 3.1934e-08,\n 2.9023e-08, 2.9254e-08, 2.0622e-07, 4.7072e-07, 2.8689e-07, 1.3259e-07,\n 2.8880e-08, 3.9887e-07, 1.0691e-08, 9.9886e-07, 5.0062e-07, 3.3355e-07,\n 1.4765e-08, 3.0665e-08, 2.2449e-07, 2.6263e-08, 1.6643e-07, 5.2623e-08,\n 6.4303e-08, 1.6642e-07, 2.7394e-07, 6.7383e-07, 7.0593e-09, 1.6795e-07,\n 1.6176e-07, 3.9108e-08, 3.2224e-07, 4.6783e-08, 3.2810e-10, 1.3037e-07,\n 5.3505e-07, 5.2145e-08, 3.1930e-08, 4.1332e-10, 8.3934e-08, 9.3905e-07,\n 3.6742e-08, 2.6368e-10, 3.8263e-07, 1.2219e-06, 5.4614e-10, 8.9877e-09,\n 1.5176e-07, 9.0462e-07, 4.0843e-08, 1.8554e-07, 1.3573e-07, 3.4745e-08,\n 1.9221e-07, 6.9305e-07, 2.1111e-07, 2.4494e-07, 7.8548e-07, 1.9578e-07,\n 3.4715e-07, 1.1778e-06, 1.8359e-07, 1.4248e-06, 1.3272e-07, 2.7181e-08,\n 5.2954e-11, 1.4054e-07, 1.5427e-09, 1.6983e-06, 4.8613e-08, 1.9212e-07,\n 1.4609e-07, 1.0155e-09, 1.2877e-06, 5.6962e-08, 4.0000e-07, 6.0240e-07,\n 1.0525e-07, 8.3871e-07, 1.8079e-08, 6.6598e-08, 1.0419e-07, 5.5920e-07,\n 5.8316e-08, 4.4652e-08, 2.4860e-08, 4.1747e-10, 4.1325e-09, 3.0046e-07,\n 5.0280e-07, 7.4075e-08, 2.4916e-07, 5.1063e-07, 5.5753e-09, 1.1449e-06,\n 2.7960e-07, 4.1858e-08, 4.8466e-07, 1.8336e-07, 9.1748e-07, 1.7779e-07,\n 1.4011e-07, 1.0080e-10, 3.5728e-10, 3.4536e-09, 6.2834e-07, 2.3296e-06,\n 4.0792e-07, 8.0309e-08, 9.8551e-08, 1.5591e-07, 1.3769e-07, 1.1766e-08,\n 3.6021e-08, 3.1910e-07, 2.3141e-07, 5.6637e-08, 6.9315e-08, 4.1000e-08,\n 1.1026e-07, 1.2410e-06, 5.4911e-08, 5.9485e-08, 1.9658e-07, 6.1541e-08,\n 5.8599e-08, 1.7556e-07, 5.5720e-08, 2.6147e-07, 1.1989e-06, 8.6739e-08,\n 4.0947e-07, 7.4244e-08, 3.7634e-09, 1.3597e-07, 2.5130e-07, 3.9310e-08,\n 5.0864e-07, 8.6468e-08, 6.0041e-08, 5.7348e-08, 6.3926e-08, 1.7284e-07,\n 2.0870e-07, 6.2268e-08, 9.3162e-08, 1.5472e-07, 1.3670e-06, 7.9890e-10,\n 1.0718e-07, 4.0291e-07, 7.1547e-08, 1.3129e-06, 2.2808e-07, 1.3430e-07,\n 7.7213e-08, 2.6924e-08, 1.1813e-10, 3.6409e-06, 5.0872e-08, 1.0860e-08,\n 1.0672e-07, 4.9926e-07, 5.7301e-10, 1.4811e-06, 1.0209e-07, 1.2740e-07,\n 1.2517e-07, 3.4393e-07, 6.2440e-07, 1.2715e-08, 2.9080e-10, 2.3835e-07,\n 6.8855e-08, 1.7504e-07, 6.5233e-09, 4.3563e-07, 8.6268e-09, 9.4862e-07,\n 1.1007e-06, 4.3325e-09, 1.0840e-08, 1.1027e-07, 4.4915e-08, 6.1457e-07,\n 5.7153e-08, 7.4411e-07, 1.0036e-07, 1.3777e-11, 1.2763e-10, 9.8398e-09,\n 1.5272e-08, 1.6082e-07, 4.2331e-09, 1.4280e-08, 1.6615e-07, 9.6387e-08,\n 6.5777e-07, 5.0627e-09, 4.6227e-08, 3.0795e-07, 8.6034e-08, 1.0078e-07,\n 1.8275e-07, 7.4527e-10, 7.2365e-08, 2.5818e-07, 1.0609e-07, 2.5597e-09,\n 5.9656e-09, 4.9311e-07, 3.4177e-08, 1.1036e-07, 8.8750e-07, 1.3595e-08,\n 5.5350e-07, 4.8941e-08, 1.3178e-06, 3.1519e-08], device='cuda:0')" + "exp_avg_sq": "tensor([2.2314e-07, 6.9363e-08, 8.8195e-08, 2.7447e-09, 2.4939e-08, 2.6379e-09,\n 2.7056e-08, 2.3139e-09, 1.5810e-08, 1.0133e-08, 2.0817e-07, 2.0519e-07,\n 2.8161e-08, 2.0330e-08, 3.9778e-08, 5.3814e-09, 1.0103e-07, 3.0052e-08,\n 1.0091e-07, 2.8096e-07, 4.6215e-10, 8.0541e-08, 1.7747e-08, 1.6488e-07,\n 1.2161e-08, 2.6706e-08, 1.7658e-07, 2.8546e-09, 1.0026e-08, 3.0613e-08,\n 5.3106e-08, 1.7481e-07, 2.3102e-10, 3.1462e-07, 1.8102e-07, 2.1489e-07,\n 8.1508e-09, 2.9052e-10, 1.1677e-07, 7.7103e-07, 3.2405e-07, 9.1255e-09,\n 8.2935e-09, 8.3596e-09, 5.8928e-08, 1.3451e-07, 8.1981e-08, 3.7889e-08,\n 8.2526e-09, 1.1398e-07, 3.0550e-09, 2.8543e-07, 1.4306e-07, 9.5313e-08,\n 4.2193e-09, 8.7629e-09, 6.4150e-08, 7.5049e-09, 4.7560e-08, 1.5037e-08,\n 1.8375e-08, 4.7557e-08, 7.8280e-08, 1.9255e-07, 2.0172e-09, 4.7994e-08,\n 4.6225e-08, 1.1176e-08, 9.2082e-08, 1.3369e-08, 9.3757e-11, 3.7254e-08,\n 1.5289e-07, 1.4901e-08, 9.1244e-09, 1.1811e-10, 2.3985e-08, 2.6834e-07,\n 1.0499e-08, 7.5349e-11, 1.0934e-07, 3.4918e-07, 1.5606e-10, 2.5683e-09,\n 4.3367e-08, 2.5850e-07, 1.1671e-08, 5.3019e-08, 3.8787e-08, 9.9286e-09,\n 5.4925e-08, 1.9804e-07, 6.0327e-08, 6.9993e-08, 2.2446e-07, 5.5946e-08,\n 9.9201e-08, 3.3657e-07, 5.2463e-08, 4.0714e-07, 3.7926e-08, 7.7672e-09,\n 1.5132e-11, 4.0159e-08, 4.4083e-10, 4.8531e-07, 1.3892e-08, 5.4901e-08,\n 4.1745e-08, 2.9018e-10, 3.6798e-07, 1.6277e-08, 1.1430e-07, 1.7214e-07,\n 3.0077e-08, 2.3967e-07, 5.1661e-09, 1.9031e-08, 2.9775e-08, 1.5980e-07,\n 1.6664e-08, 1.2760e-08, 7.1039e-09, 1.1929e-10, 1.1809e-09, 8.5859e-08,\n 1.4368e-07, 2.1167e-08, 7.1201e-08, 1.4592e-07, 1.5932e-09, 3.2717e-07,\n 7.9899e-08, 1.1961e-08, 1.3850e-07, 5.2397e-08, 2.6218e-07, 5.0804e-08,\n 4.0036e-08, 2.8804e-11, 1.0210e-10, 9.8689e-10, 1.7955e-07, 6.6570e-07,\n 1.1657e-07, 2.2949e-08, 2.8162e-08, 4.4553e-08, 3.9345e-08, 3.3621e-09,\n 1.0293e-08, 9.1186e-08, 6.6127e-08, 1.6184e-08, 1.9807e-08, 1.1716e-08,\n 3.1507e-08, 3.5461e-07, 1.5691e-08, 1.6998e-08, 5.6174e-08, 1.7586e-08,\n 1.6745e-08, 5.0168e-08, 1.5922e-08, 7.4717e-08, 3.4260e-07, 2.4786e-08,\n 1.1701e-07, 2.1216e-08, 1.0754e-09, 3.8855e-08, 7.1811e-08, 1.1233e-08,\n 1.4535e-07, 2.4709e-08, 1.7157e-08, 1.6388e-08, 1.8267e-08, 4.9390e-08,\n 5.9638e-08, 1.7793e-08, 2.6622e-08, 4.4211e-08, 3.9065e-07, 2.2829e-10,\n 3.0629e-08, 1.1514e-07, 2.0445e-08, 3.7517e-07, 6.5176e-08, 3.8376e-08,\n 2.2064e-08, 7.6938e-09, 3.3755e-11, 1.0404e-06, 1.4537e-08, 3.1032e-09,\n 3.0495e-08, 1.4267e-07, 1.6374e-10, 4.2323e-07, 2.9174e-08, 3.6406e-08,\n 3.5767e-08, 9.8282e-08, 1.7843e-07, 3.6335e-09, 8.3099e-11, 6.8111e-08,\n 1.9676e-08, 5.0019e-08, 1.8641e-09, 1.2448e-07, 2.4652e-09, 2.7107e-07,\n 3.1453e-07, 1.2381e-09, 3.0975e-09, 3.1511e-08, 1.2835e-08, 1.7562e-07,\n 1.6332e-08, 2.1264e-07, 2.8680e-08, 3.9369e-12, 3.6470e-11, 2.8118e-09,\n 4.3642e-09, 4.5956e-08, 1.2096e-09, 4.0805e-09, 4.7478e-08, 2.7543e-08,\n 1.8796e-07, 1.4467e-09, 1.3210e-08, 8.7999e-08, 2.4585e-08, 2.8799e-08,\n 5.2223e-08, 2.1297e-10, 2.0679e-08, 7.3777e-08, 3.0317e-08, 7.3146e-10,\n 1.7047e-09, 1.4091e-07, 9.7664e-09, 3.1535e-08, 2.5361e-07, 3.8848e-09,\n 1.5817e-07, 1.3985e-08, 3.7656e-07, 9.0068e-09], device='cuda:0')" }, "32": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([1.6831e-09, 7.7800e-10, 8.6211e-10, 5.1025e-11, 1.9332e-10, 2.7403e-11,\n 1.3810e-10, 6.5898e-11, 7.5491e-11, 8.5115e-11, 1.3411e-09, 1.1806e-09,\n 2.0397e-10, 1.8028e-10, 4.5964e-10, 1.3674e-10, 8.4367e-10, 1.8302e-10,\n 8.6407e-10, 2.3379e-09, 7.7815e-13, 9.8946e-10, 2.9600e-10, 1.2153e-09,\n 1.4285e-10, 1.0815e-10, 2.9509e-09, 3.0758e-11, 1.2202e-10, 2.0562e-10,\n 2.8117e-10, 4.4558e-09, 7.6024e-12, 1.8825e-09, 1.8839e-09, 2.8640e-09,\n 3.5720e-11, 1.1518e-12, 1.1922e-09, 6.8297e-09, 3.3538e-09, 5.3110e-11,\n 7.7332e-11, 1.3318e-10, 2.6734e-10, 8.4428e-10, 1.1835e-09, 3.1928e-10,\n 3.6861e-11, 9.2630e-10, 1.1676e-11, 3.0601e-09, 1.0404e-09, 7.2657e-10,\n 1.7277e-11, 2.1867e-10, 4.8365e-10, 2.9064e-11, 8.4259e-10, 1.8144e-10,\n 1.4788e-10, 6.8256e-10, 8.6972e-10, 3.0259e-09, 8.5107e-12, 4.8876e-10,\n 4.4075e-10, 6.2711e-11, 7.4847e-10, 1.0143e-10, 9.5036e-14, 4.7546e-10,\n 1.3677e-09, 1.7242e-10, 4.8273e-11, 1.5630e-12, 5.9665e-10, 2.7297e-09,\n 7.6509e-11, 1.4123e-14, 2.2336e-09, 6.7254e-09, 3.3165e-12, 5.0481e-11,\n 6.8288e-10, 2.6417e-09, 1.5640e-10, 7.2399e-10, 2.8911e-10, 1.8246e-10,\n 5.2308e-10, 1.6607e-09, 3.0571e-10, 6.4686e-10, 2.6851e-09, 3.9064e-10,\n 9.2850e-10, 2.6958e-09, 4.6798e-10, 4.1268e-09, 2.9345e-10, 1.6471e-10,\n 7.5557e-12, 2.6921e-10, 7.8305e-13, 5.2288e-09, 9.6588e-11, 9.7357e-10,\n 2.3478e-10, 3.7353e-16, 3.9439e-09, 1.6772e-10, 1.0080e-09, 1.0484e-09,\n 1.8741e-10, 2.6348e-09, 5.6579e-11, 1.0897e-10, 2.0757e-10, 2.4170e-09,\n 7.1065e-11, 6.3977e-11, 2.6377e-11, 1.9531e-11, 3.8357e-11, 9.9468e-10,\n 9.0030e-10, 1.8637e-10, 7.6683e-10, 1.9252e-09, 2.9054e-13, 2.1488e-09,\n 1.0789e-09, 1.2813e-10, 8.0332e-10, 2.6387e-10, 3.1312e-09, 4.0109e-10,\n 3.5486e-10, 1.5522e-14, 1.4681e-13, 1.6997e-11, 1.5100e-09, 9.6412e-09,\n 1.4138e-09, 3.0532e-10, 2.4186e-10, 9.3105e-10, 1.6296e-10, 1.1726e-11,\n 2.3616e-10, 6.5155e-10, 9.1573e-10, 9.2390e-11, 9.1397e-11, 3.2764e-10,\n 1.9512e-10, 3.1527e-09, 8.2959e-11, 2.5895e-10, 4.3565e-10, 9.1498e-11,\n 7.0917e-11, 5.9831e-10, 1.3364e-10, 5.6772e-10, 2.2469e-09, 1.6621e-10,\n 1.9716e-09, 2.4454e-10, 1.1857e-12, 2.5520e-10, 4.0830e-10, 1.1999e-10,\n 1.4437e-09, 5.9943e-10, 8.8445e-11, 1.8510e-10, 1.3354e-10, 3.6953e-10,\n 4.7275e-10, 9.5648e-11, 1.9874e-10, 2.5353e-10, 3.1739e-09, 5.1548e-14,\n 1.8066e-10, 1.4697e-09, 8.1846e-11, 2.8644e-09, 3.8391e-10, 3.3062e-10,\n 2.9876e-10, 8.4625e-11, 3.0474e-13, 1.4916e-08, 5.4649e-11, 8.0472e-12,\n 2.2269e-10, 9.5686e-10, 1.9745e-11, 6.5039e-09, 1.7913e-10, 3.5164e-10,\n 2.4962e-10, 7.6769e-10, 1.6107e-09, 2.2344e-11, 2.0981e-14, 4.1481e-10,\n 1.4201e-10, 3.9657e-10, 2.0587e-11, 1.5256e-09, 6.2294e-12, 2.7495e-09,\n 2.5936e-09, 5.2496e-13, 3.0171e-11, 2.3636e-10, 8.7519e-11, 1.4940e-09,\n 1.9320e-10, 1.5824e-09, 2.7501e-10, 1.2494e-12, 9.2642e-13, 3.5999e-11,\n 2.1698e-11, 4.8555e-10, 1.2675e-11, 8.2060e-12, 5.7499e-10, 3.3720e-10,\n 2.0734e-09, 3.4339e-11, 6.8304e-11, 6.5773e-10, 3.6760e-10, 1.8458e-10,\n 3.5899e-10, 4.0571e-13, 2.0545e-10, 7.3601e-10, 1.7349e-10, 2.0216e-13,\n 2.5639e-12, 8.0680e-10, 5.4005e-11, 3.7704e-10, 2.6099e-09, 1.2708e-11,\n 1.9847e-09, 9.2230e-11, 3.5142e-09, 9.9344e-11], device='cuda:0')" + "exp_avg_sq": "tensor([4.8096e-10, 2.2232e-10, 2.4636e-10, 1.4581e-11, 5.5242e-11, 7.8308e-12,\n 3.9463e-11, 1.8831e-11, 2.1572e-11, 2.4322e-11, 3.8324e-10, 3.3736e-10,\n 5.8286e-11, 5.1517e-11, 1.3135e-10, 3.9075e-11, 2.4109e-10, 5.2299e-11,\n 2.4692e-10, 6.6808e-10, 2.2236e-13, 2.8275e-10, 8.4585e-11, 3.4729e-10,\n 4.0820e-11, 3.0905e-11, 8.4324e-10, 8.7893e-12, 3.4869e-11, 5.8758e-11,\n 8.0348e-11, 1.2733e-09, 2.1725e-12, 5.3793e-10, 5.3834e-10, 8.1841e-10,\n 1.0207e-11, 3.2913e-13, 3.4067e-10, 1.9517e-09, 9.5838e-10, 1.5177e-11,\n 2.2098e-11, 3.8056e-11, 7.6394e-11, 2.4126e-10, 3.3818e-10, 9.1236e-11,\n 1.0533e-11, 2.6470e-10, 3.3366e-12, 8.7444e-10, 2.9732e-10, 2.0762e-10,\n 4.9371e-12, 6.2487e-11, 1.3821e-10, 8.3051e-12, 2.4078e-10, 5.1849e-11,\n 4.2258e-11, 1.9505e-10, 2.4853e-10, 8.6468e-10, 2.4320e-12, 1.3967e-10,\n 1.2595e-10, 1.7920e-11, 2.1388e-10, 2.8984e-11, 2.7157e-14, 1.3587e-10,\n 3.9083e-10, 4.9271e-11, 1.3794e-11, 4.4663e-13, 1.7050e-10, 7.8002e-10,\n 2.1863e-11, 4.0358e-15, 6.3827e-10, 1.9218e-09, 9.4772e-13, 1.4425e-11,\n 1.9514e-10, 7.5489e-10, 4.4691e-11, 2.0689e-10, 8.2615e-11, 5.2139e-11,\n 1.4947e-10, 4.7456e-10, 8.7358e-11, 1.8484e-10, 7.6729e-10, 1.1163e-10,\n 2.6533e-10, 7.7035e-10, 1.3373e-10, 1.1793e-09, 8.3857e-11, 4.7067e-11,\n 2.1591e-12, 7.6928e-11, 2.2376e-13, 1.4942e-09, 2.7601e-11, 2.7821e-10,\n 6.7091e-11, 1.0674e-16, 1.1270e-09, 4.7928e-11, 2.8804e-10, 2.9959e-10,\n 5.3553e-11, 7.5291e-10, 1.6168e-11, 3.1139e-11, 5.9314e-11, 6.9067e-10,\n 2.0307e-11, 1.8282e-11, 7.5374e-12, 5.5813e-12, 1.0961e-11, 2.8424e-10,\n 2.5727e-10, 5.3257e-11, 2.1913e-10, 5.5014e-10, 8.3024e-14, 6.1402e-10,\n 3.0830e-10, 3.6615e-11, 2.2955e-10, 7.5403e-11, 8.9478e-10, 1.1461e-10,\n 1.0140e-10, 4.4354e-15, 4.1952e-14, 4.8571e-12, 4.3149e-10, 2.7551e-09,\n 4.0399e-10, 8.7249e-11, 6.9112e-11, 2.6605e-10, 4.6568e-11, 3.3508e-12,\n 6.7486e-11, 1.8618e-10, 2.6168e-10, 2.6401e-11, 2.6117e-11, 9.3626e-11,\n 5.5757e-11, 9.0092e-10, 2.3706e-11, 7.3997e-11, 1.2449e-10, 2.6146e-11,\n 2.0265e-11, 1.7097e-10, 3.8188e-11, 1.6223e-10, 6.4208e-10, 4.7495e-11,\n 5.6340e-10, 6.9881e-11, 3.3882e-13, 7.2924e-11, 1.1667e-10, 3.4287e-11,\n 4.1255e-10, 1.7129e-10, 2.5274e-11, 5.2893e-11, 3.8161e-11, 1.0560e-10,\n 1.3509e-10, 2.7332e-11, 5.6793e-11, 7.2448e-11, 9.0696e-10, 1.4730e-14,\n 5.1625e-11, 4.1997e-10, 2.3388e-11, 8.1853e-10, 1.0971e-10, 9.4477e-11,\n 8.5373e-11, 2.4182e-11, 8.7083e-14, 4.2624e-09, 1.5616e-11, 2.2995e-12,\n 6.3635e-11, 2.7343e-10, 5.6422e-12, 1.8585e-09, 5.1187e-11, 1.0049e-10,\n 7.1329e-11, 2.1937e-10, 4.6028e-10, 6.3850e-12, 5.9954e-15, 1.1853e-10,\n 4.0581e-11, 1.1332e-10, 5.8830e-12, 4.3595e-10, 1.7801e-12, 7.8570e-10,\n 7.4113e-10, 1.5001e-13, 8.6217e-12, 6.7543e-11, 2.5009e-11, 4.2693e-10,\n 5.5207e-11, 4.5218e-10, 7.8587e-11, 3.5702e-13, 2.6473e-13, 1.0287e-11,\n 6.2003e-12, 1.3875e-10, 3.6220e-12, 2.3449e-12, 1.6431e-10, 9.6357e-11,\n 5.9248e-10, 9.8125e-12, 1.9518e-11, 1.8795e-10, 1.0505e-10, 5.2746e-11,\n 1.0258e-10, 1.1594e-13, 5.8709e-11, 2.1032e-10, 4.9575e-11, 5.7768e-14,\n 7.3265e-13, 2.3055e-10, 1.5432e-11, 1.0774e-10, 7.4580e-10, 3.6314e-12,\n 5.6714e-10, 2.6355e-11, 1.0042e-09, 2.8388e-11], device='cuda:0')" }, "33": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([3.2281e-09, 1.1382e-09, 9.0442e-10, 5.8763e-11, 1.8876e-10, 7.3388e-11,\n 2.5988e-10, 8.5104e-11, 1.2730e-10, 5.9210e-11, 2.3783e-09, 2.5553e-09,\n 3.0680e-10, 3.4005e-10, 7.2304e-10, 7.8269e-11, 1.5844e-09, 2.6955e-10,\n 1.1357e-09, 4.0995e-09, 6.4140e-13, 1.3071e-09, 3.6540e-10, 2.4245e-09,\n 2.3674e-10, 2.9544e-10, 2.7736e-09, 5.8569e-11, 2.1247e-10, 2.8088e-10,\n 8.8237e-10, 2.8641e-09, 8.5501e-12, 3.7704e-09, 2.6862e-09, 3.3253e-09,\n 6.0033e-11, 2.0674e-12, 1.8212e-09, 9.4490e-09, 3.7966e-09, 1.8806e-10,\n 1.7206e-10, 1.8242e-10, 6.9541e-10, 1.5712e-09, 1.3852e-09, 6.4710e-10,\n 7.6603e-11, 1.3504e-09, 1.0054e-11, 4.2472e-09, 1.7016e-09, 1.0119e-09,\n 2.9374e-11, 2.3443e-10, 9.9679e-10, 6.4544e-11, 8.5966e-10, 2.8836e-10,\n 1.4313e-10, 8.5008e-10, 1.3057e-09, 2.9993e-09, 1.1741e-11, 4.3942e-10,\n 8.0442e-10, 8.3815e-11, 1.3935e-09, 2.3833e-10, 2.3775e-15, 6.5324e-10,\n 1.7262e-09, 2.9622e-10, 8.8615e-11, 7.1837e-12, 4.8630e-10, 3.8813e-09,\n 5.6040e-11, 3.8488e-13, 1.8122e-09, 5.3097e-09, 4.5287e-12, 8.1004e-11,\n 7.7050e-10, 3.7803e-09, 2.3358e-10, 9.1869e-10, 4.4226e-10, 1.5315e-10,\n 8.8428e-10, 3.0621e-09, 6.8722e-10, 7.8098e-10, 3.4265e-09, 5.6370e-10,\n 1.0613e-09, 4.0240e-09, 9.1710e-10, 5.9857e-09, 6.3995e-10, 1.7817e-10,\n 1.6012e-11, 3.9277e-10, 4.7452e-13, 5.7308e-09, 9.9642e-11, 9.5649e-10,\n 4.2729e-10, 4.9480e-13, 4.3349e-09, 3.0969e-10, 1.8020e-09, 2.0712e-09,\n 2.6169e-10, 2.7833e-09, 9.6641e-11, 1.8819e-10, 4.7053e-10, 2.5713e-09,\n 1.7726e-10, 1.2773e-10, 4.6830e-11, 1.0337e-11, 4.7666e-11, 8.8332e-10,\n 2.1155e-09, 3.7327e-10, 1.1537e-09, 1.5389e-09, 1.5682e-12, 4.6910e-09,\n 1.3134e-09, 2.1943e-10, 1.5910e-09, 5.9262e-10, 3.8814e-09, 6.1334e-10,\n 6.5852e-10, 6.3209e-13, 1.3686e-12, 3.0017e-11, 1.9284e-09, 7.8514e-09,\n 1.7922e-09, 4.2469e-10, 4.6970e-10, 8.3466e-10, 4.5728e-10, 1.2476e-11,\n 2.4300e-10, 1.0146e-09, 1.1258e-09, 1.7123e-10, 2.1254e-10, 2.9658e-10,\n 3.0983e-10, 4.2385e-09, 1.5665e-10, 2.6689e-10, 6.5681e-10, 1.8052e-10,\n 1.4408e-10, 8.0680e-10, 2.8647e-10, 8.1910e-10, 4.2218e-09, 2.8684e-10,\n 1.9455e-09, 3.5854e-10, 6.5442e-12, 3.8775e-10, 8.8646e-10, 2.1466e-10,\n 1.6462e-09, 5.2842e-10, 1.5110e-10, 3.2686e-10, 1.3440e-10, 7.8529e-10,\n 5.7177e-10, 1.6392e-10, 2.2328e-10, 4.4455e-10, 5.7737e-09, 4.2655e-13,\n 3.3726e-10, 1.1613e-09, 1.9676e-10, 5.3242e-09, 7.6892e-10, 3.8983e-10,\n 3.9645e-10, 1.7235e-10, 2.8918e-13, 1.4708e-08, 1.1965e-10, 1.9965e-11,\n 5.2889e-10, 1.7906e-09, 3.4677e-11, 6.3762e-09, 3.6187e-10, 3.2140e-10,\n 5.9483e-10, 1.2972e-09, 1.9193e-09, 3.6342e-11, 9.9455e-13, 8.5450e-10,\n 1.4922e-10, 7.3322e-10, 4.9443e-11, 1.9179e-09, 6.9934e-12, 3.1303e-09,\n 3.7256e-09, 6.4230e-13, 6.1503e-11, 2.7560e-10, 6.8772e-11, 2.6803e-09,\n 3.2166e-10, 3.1873e-09, 4.8110e-10, 3.3111e-12, 7.5852e-13, 7.6003e-11,\n 3.6292e-11, 7.7822e-10, 2.8149e-11, 1.6705e-11, 7.5451e-10, 4.9786e-10,\n 2.8566e-09, 2.5949e-11, 1.1123e-10, 1.3819e-09, 4.7050e-10, 2.6682e-10,\n 8.2127e-10, 4.2365e-13, 3.8645e-10, 9.1147e-10, 3.1574e-10, 9.1802e-15,\n 3.5643e-12, 1.6014e-09, 7.6213e-11, 5.0744e-10, 2.9833e-09, 2.7653e-11,\n 1.6532e-09, 1.5192e-10, 4.5810e-09, 1.9403e-10], device='cuda:0')" + "exp_avg_sq": "tensor([9.2247e-10, 3.2524e-10, 2.5845e-10, 1.6792e-11, 5.3938e-11, 2.0971e-11,\n 7.4263e-11, 2.4319e-11, 3.6377e-11, 1.6920e-11, 6.7963e-10, 7.3020e-10,\n 8.7671e-11, 9.7171e-11, 2.0661e-10, 2.2366e-11, 4.5275e-10, 7.7025e-11,\n 3.2452e-10, 1.1715e-09, 1.8328e-13, 3.7351e-10, 1.0441e-10, 6.9282e-10,\n 6.7651e-11, 8.4423e-11, 7.9257e-10, 1.6737e-11, 6.0716e-11, 8.0264e-11,\n 2.5215e-10, 8.1843e-10, 2.4433e-12, 1.0774e-09, 7.6761e-10, 9.5022e-10,\n 1.7155e-11, 5.9077e-13, 5.2043e-10, 2.7001e-09, 1.0849e-09, 5.3741e-11,\n 4.9168e-11, 5.2129e-11, 1.9872e-10, 4.4898e-10, 3.9583e-10, 1.8491e-10,\n 2.1890e-11, 3.8588e-10, 2.8729e-12, 1.2137e-09, 4.8623e-10, 2.8915e-10,\n 8.3939e-12, 6.6989e-11, 2.8484e-10, 1.8444e-11, 2.4565e-10, 8.2402e-11,\n 4.0900e-11, 2.4292e-10, 3.7312e-10, 8.5706e-10, 3.3551e-12, 1.2557e-10,\n 2.2987e-10, 2.3951e-11, 3.9821e-10, 6.8106e-11, 6.7940e-16, 1.8667e-10,\n 4.9329e-10, 8.4646e-11, 2.5322e-11, 2.0528e-12, 1.3896e-10, 1.1091e-09,\n 1.6014e-11, 1.0998e-13, 5.1784e-10, 1.5173e-09, 1.2941e-12, 2.3148e-11,\n 2.2018e-10, 1.0803e-09, 6.6747e-11, 2.6252e-10, 1.2638e-10, 4.3763e-11,\n 2.5269e-10, 8.7501e-10, 1.9638e-10, 2.2317e-10, 9.7914e-10, 1.6108e-10,\n 3.0327e-10, 1.1499e-09, 2.6207e-10, 1.7105e-09, 1.8287e-10, 5.0912e-11,\n 4.5757e-12, 1.1224e-10, 1.3560e-13, 1.6376e-09, 2.8473e-11, 2.7333e-10,\n 1.2210e-10, 1.4139e-13, 1.2387e-09, 8.8496e-11, 5.1492e-10, 5.9186e-10,\n 7.4779e-11, 7.9536e-10, 2.7616e-11, 5.3777e-11, 1.3446e-10, 7.3476e-10,\n 5.0653e-11, 3.6500e-11, 1.3382e-11, 2.9540e-12, 1.3621e-11, 2.5242e-10,\n 6.0453e-10, 1.0666e-10, 3.2968e-10, 4.3975e-10, 4.4811e-13, 1.3405e-09,\n 3.7532e-10, 6.2703e-11, 4.5465e-10, 1.6935e-10, 1.1091e-09, 1.7527e-10,\n 1.8818e-10, 1.8062e-13, 3.9109e-13, 8.5775e-12, 5.5105e-10, 2.2436e-09,\n 5.1212e-10, 1.2136e-10, 1.3422e-10, 2.3851e-10, 1.3067e-10, 3.5650e-12,\n 6.9440e-11, 2.8994e-10, 3.2169e-10, 4.8930e-11, 6.0736e-11, 8.4750e-11,\n 8.8537e-11, 1.2112e-09, 4.4764e-11, 7.6266e-11, 1.8769e-10, 5.1584e-11,\n 4.1172e-11, 2.3055e-10, 8.1861e-11, 2.3406e-10, 1.2064e-09, 8.1967e-11,\n 5.5593e-10, 1.0245e-10, 1.8700e-12, 1.1080e-10, 2.5331e-10, 6.1340e-11,\n 4.7040e-10, 1.5100e-10, 4.3178e-11, 9.3402e-11, 3.8407e-11, 2.2440e-10,\n 1.6339e-10, 4.6842e-11, 6.3804e-11, 1.2703e-10, 1.6499e-09, 1.2189e-13,\n 9.6375e-11, 3.3185e-10, 5.6225e-11, 1.5214e-09, 2.1972e-10, 1.1140e-10,\n 1.1329e-10, 4.9250e-11, 8.2635e-14, 4.2029e-09, 3.4192e-11, 5.7051e-12,\n 1.5113e-10, 5.1168e-10, 9.9092e-12, 1.8221e-09, 1.0341e-10, 9.1843e-11,\n 1.6998e-10, 3.7070e-10, 5.4846e-10, 1.0385e-11, 2.8420e-13, 2.4418e-10,\n 4.2640e-11, 2.0952e-10, 1.4129e-11, 5.4804e-10, 1.9984e-12, 8.9452e-10,\n 1.0646e-09, 1.8354e-13, 1.7575e-11, 7.8756e-11, 1.9652e-11, 7.6592e-10,\n 9.1917e-11, 9.1079e-10, 1.3748e-10, 9.4617e-13, 2.1675e-13, 2.1719e-11,\n 1.0371e-11, 2.2238e-10, 8.0438e-12, 4.7735e-12, 2.1561e-10, 1.4227e-10,\n 8.1629e-10, 7.4152e-12, 3.1786e-11, 3.9488e-10, 1.3445e-10, 7.6245e-11,\n 2.3468e-10, 1.2106e-13, 1.1043e-10, 2.6046e-10, 9.0226e-11, 2.6233e-15,\n 1.0185e-12, 4.5760e-10, 2.1778e-11, 1.4501e-10, 8.5250e-10, 7.9020e-12,\n 4.7242e-10, 4.3412e-11, 1.3091e-09, 5.5445e-11], device='cuda:0')" }, "34": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([[-5.6052e-45, -5.6052e-45, -5.6052e-45, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n ...,\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.1506e-11, 4.1548e-11, 2.5346e-11, ..., 5.7366e-12, 3.1409e-11,\n 1.8874e-11],\n [8.1894e-13, 1.2252e-13, 1.3353e-12, ..., 3.5635e-12, 4.1426e-12,\n 1.4765e-13],\n [3.9919e-12, 1.7556e-11, 5.5357e-12, ..., 2.8327e-12, 1.7058e-11,\n 3.2640e-12],\n ...,\n [9.9888e-11, 2.3694e-10, 2.4555e-10, ..., 4.3631e-11, 3.6840e-10,\n 5.3317e-10],\n [6.6701e-10, 1.0550e-09, 1.6283e-09, ..., 3.6158e-10, 1.5932e-09,\n 2.9252e-09],\n [9.1514e-09, 2.0298e-08, 2.4256e-08, ..., 5.1638e-09, 3.0552e-08,\n 4.7933e-08]], device='cuda:0')" + "exp_avg_sq": "tensor([[3.2880e-12, 1.1873e-11, 7.2428e-12, ..., 1.6393e-12, 8.9753e-12,\n 5.3934e-12],\n [2.3402e-13, 3.5010e-14, 3.8158e-13, ..., 1.0183e-12, 1.1838e-12,\n 4.2191e-14],\n [1.1407e-12, 5.0168e-12, 1.5819e-12, ..., 8.0945e-13, 4.8744e-12,\n 9.3270e-13],\n ...,\n [2.8544e-11, 6.7706e-11, 7.0167e-11, ..., 1.2468e-11, 1.0527e-10,\n 1.5236e-10],\n [1.9060e-10, 3.0147e-10, 4.6530e-10, ..., 1.0333e-10, 4.5527e-10,\n 8.3589e-10],\n [2.6151e-09, 5.8004e-09, 6.9314e-09, ..., 1.4756e-09, 8.7304e-09,\n 1.3697e-08]], device='cuda:0')" }, "35": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([2.3199e-10, 7.4555e-12, 6.6557e-11, 4.3799e-12, 2.9327e-11, 4.3356e-11,\n 6.4021e-12, 5.5383e-12, 2.6980e-13, 4.5201e-11, 3.6663e-11, 6.3410e-11,\n 1.5749e-12, 1.3333e-12, 1.6222e-11, 1.0640e-11, 3.5097e-12, 1.3891e-10,\n 4.7405e-11, 6.0717e-13, 2.3904e-12, 1.9398e-11, 4.4788e-11, 1.7277e-12,\n 8.3677e-11, 9.8135e-12, 1.3895e-11, 1.4168e-12, 3.2293e-12, 6.6056e-12,\n 6.5930e-12, 1.0966e-10, 1.5891e-11, 8.5065e-13, 3.2088e-14, 3.2784e-11,\n 1.1666e-11, 4.7001e-12, 1.4054e-13, 6.2663e-15, 1.9039e-11, 2.3826e-12,\n 1.1634e-11, 5.4932e-12, 9.3972e-13, 1.0698e-11, 1.3546e-11, 1.3744e-11,\n 7.2567e-12, 3.0836e-11, 1.1966e-11, 3.4035e-12, 4.4410e-12, 2.6153e-12,\n 9.8106e-14, 1.3339e-13, 3.0530e-12, 2.4505e-13, 9.8229e-12, 4.1776e-13,\n 3.4693e-11, 1.1570e-11, 2.7304e-14, 2.1049e-11, 4.0708e-11, 1.3865e-11,\n 2.0789e-11, 4.1163e-12, 5.2718e-13, 3.2496e-13, 2.1300e-11, 2.2500e-12,\n 1.7903e-13, 8.7959e-12, 5.1251e-12, 1.3815e-11, 1.0161e-11, 2.6315e-12,\n 4.3581e-17, 9.5428e-12, 7.2675e-12, 6.3967e-11, 9.7885e-13, 9.0434e-12,\n 1.2302e-12, 6.2073e-12, 3.6184e-12, 1.0013e-14, 6.1569e-12, 7.9909e-13,\n 2.8582e-14, 4.6372e-12, 4.1467e-12, 1.1818e-10, 3.8936e-11, 2.7077e-11,\n 2.9499e-12, 1.3920e-10, 9.6151e-12, 1.7831e-11, 1.6573e-12, 1.9064e-12,\n 9.3897e-13, 1.0198e-11, 1.2396e-13, 3.5564e-11, 5.8400e-11, 6.4913e-12,\n 7.3709e-12, 1.1081e-12, 1.1209e-11, 1.2729e-10, 9.6290e-13, 1.4583e-10,\n 2.0228e-11, 9.0112e-11, 8.8429e-18, 8.2660e-11, 6.0213e-11, 2.9800e-11,\n 1.1405e-10, 5.1022e-13, 5.9541e-11, 2.2451e-13, 3.2495e-11, 5.4124e-11,\n 4.9051e-11, 3.5528e-11, 6.5580e-11, 9.9242e-11, 1.1413e-11, 2.2260e-11,\n 6.7038e-13, 3.8252e-11, 1.0990e-11, 1.0052e-11, 2.3242e-11, 3.1471e-11,\n 2.6442e-14, 6.1399e-12, 7.7194e-11, 2.8408e-13, 1.6236e-11, 5.9904e-12,\n 3.7500e-11, 1.5599e-10, 3.1017e-11, 3.9836e-13, 9.8380e-11, 1.4919e-12,\n 3.1730e-12, 2.5776e-11, 1.0703e-10, 8.7970e-12, 3.7796e-11, 5.6704e-11,\n 4.1755e-13, 1.0530e-13, 3.8598e-12, 1.8976e-11, 1.8104e-11, 5.0208e-12,\n 7.5581e-11, 3.9094e-11, 4.3534e-12, 4.2006e-14, 8.3885e-13, 1.1884e-13,\n 2.3731e-12, 1.0686e-11, 2.1609e-12, 1.3109e-12, 9.0497e-12, 1.0963e-11,\n 9.9788e-13, 2.7997e-13, 1.0490e-11, 2.4220e-12, 1.5042e-12, 2.1438e-13,\n 1.7479e-11, 5.0841e-14, 3.7804e-12, 1.6320e-11, 4.6461e-12, 2.1664e-12,\n 1.6189e-11, 5.4350e-12, 5.4537e-11, 1.7806e-12, 1.5208e-11, 1.2535e-12,\n 1.3775e-11, 1.3678e-10, 1.9305e-11, 4.1139e-11, 2.1043e-11, 6.1509e-11,\n 7.0352e-15, 7.4696e-13, 3.5779e-12, 5.3122e-12, 9.1583e-12, 7.5444e-11,\n 1.8940e-12, 5.6709e-12, 4.2408e-13, 6.3873e-12, 7.3573e-13, 1.7085e-11,\n 2.9237e-15, 1.1489e-11, 7.7622e-13, 8.4620e-12, 7.7858e-11, 3.9469e-11,\n 5.7982e-11, 1.5681e-11, 1.9843e-11, 6.1731e-11, 1.5677e-11, 2.9936e-13,\n 7.9370e-12, 7.0600e-11, 4.5160e-13, 9.3206e-12, 5.6441e-11, 2.0067e-11,\n 1.7702e-11, 1.7126e-11, 1.4822e-11, 4.4132e-13, 6.5098e-12, 3.4388e-12,\n 1.5953e-11, 4.3784e-12, 1.1025e-13, 1.0114e-12, 4.0843e-12, 4.3014e-13,\n 3.9567e-11, 1.0913e-11, 1.6707e-12, 1.3226e-11, 4.0508e-13, 1.5125e-11,\n 4.0626e-11, 5.2128e-12, 1.7272e-11, 5.3843e-12, 4.1076e-12, 4.2243e-11,\n 1.6407e-11, 4.6744e-11, 4.3727e-11, 9.4784e-12, 1.5038e-27, 6.6474e-29,\n 2.7120e-26, 1.9068e-27, 3.0364e-26, 1.1934e-27, 7.8278e-28, 1.6582e-26,\n 1.0867e-26, 1.5817e-26, 4.7750e-28, 5.9144e-28, 1.6587e-26, 1.4907e-26,\n 8.2181e-28, 1.8109e-27, 6.6713e-27, 4.2502e-30, 4.2172e-26, 6.6752e-27,\n 2.1966e-27, 1.5004e-26, 6.6731e-27, 6.8503e-27, 5.0088e-27, 1.5998e-27,\n 3.9306e-27, 4.4984e-28, 3.3989e-26, 6.9859e-27, 6.0684e-27, 5.4048e-27,\n 2.9255e-26, 8.8619e-27, 9.5594e-28, 3.3874e-27, 5.3830e-27, 7.6809e-29,\n 7.3938e-28, 2.3193e-28, 1.5968e-29, 7.1364e-28, 3.5339e-27, 5.0731e-29,\n 5.7959e-27, 1.7039e-27, 9.0042e-27, 4.6524e-29, 7.3644e-27, 6.3441e-29,\n 1.9414e-27, 2.2175e-31, 2.7263e-27, 1.2979e-26, 3.7659e-27, 7.2278e-28,\n 2.8196e-26, 3.6274e-27, 5.2393e-26, 4.2159e-27, 1.2112e-28, 6.7370e-28,\n 6.1160e-27, 3.6262e-27, 9.3428e-28, 4.8691e-27, 1.7820e-26, 2.3351e-26,\n 1.2435e-26, 5.1101e-28, 4.3508e-26, 3.1715e-26, 3.3329e-26, 5.5628e-26,\n 1.1509e-25, 3.6188e-27, 7.1719e-26, 1.5252e-27, 5.2822e-27, 1.0534e-27,\n 1.5524e-26, 6.1490e-26, 3.6739e-26, 2.3938e-26, 6.3691e-27, 1.1519e-26,\n 5.6060e-26, 1.2608e-25, 7.0608e-27, 1.9339e-26, 1.1132e-26, 5.8293e-26,\n 6.1164e-27, 3.5740e-28, 4.0559e-27, 3.8723e-27, 3.6012e-26, 1.2243e-28,\n 7.3423e-27, 5.3094e-29, 4.1746e-27, 1.9711e-26, 2.8980e-26, 2.4836e-27,\n 6.5518e-28, 2.8619e-27, 4.9520e-27, 2.8400e-28, 1.7444e-28, 8.1542e-28,\n 9.4262e-28, 1.6119e-28, 3.7307e-27, 4.6631e-27, 6.5961e-27, 5.9824e-27,\n 4.2545e-27, 5.6182e-29, 1.8977e-30, 1.0933e-28, 7.7434e-29, 8.0438e-27,\n 1.1153e-26, 8.1630e-27, 3.8543e-27, 1.0458e-26, 5.1833e-29, 1.7525e-26,\n 1.8658e-28, 2.1431e-27, 8.7294e-28, 2.7817e-26, 1.1119e-26, 1.7384e-26,\n 1.4463e-26, 3.0634e-28, 2.7248e-28, 3.2164e-27, 6.4395e-27, 1.9931e-26,\n 2.1068e-27, 9.9949e-28, 5.7349e-28, 8.6865e-29, 3.2696e-26, 3.0753e-27,\n 9.2427e-27, 6.9640e-28, 1.0950e-27, 7.0826e-26, 1.2040e-26, 1.3690e-28,\n 9.7714e-27, 1.1599e-26, 3.3666e-26, 2.0012e-26, 2.5634e-26, 1.2263e-26,\n 1.8506e-28, 1.4431e-25, 3.2695e-26, 4.0098e-28, 3.9548e-26, 1.0047e-26,\n 8.7440e-27, 5.1492e-28, 5.9364e-27, 2.8297e-27, 2.9476e-27, 1.1585e-26,\n 9.2656e-27, 1.1026e-30, 8.5968e-27, 8.6894e-28, 1.3378e-28, 2.0446e-27,\n 2.0936e-26, 1.4815e-26, 7.6602e-29, 1.1669e-26, 3.3875e-27, 2.4330e-27,\n 4.0771e-26, 4.0836e-26, 4.3222e-27, 2.2212e-28, 4.1726e-26, 3.8661e-26,\n 1.8028e-26, 2.2929e-27, 2.3398e-27, 9.2923e-27, 1.0897e-26, 1.2204e-26,\n 6.0053e-26, 4.2270e-27, 1.3241e-27, 2.0906e-26, 4.0354e-27, 2.0168e-26,\n 3.1829e-27, 3.4502e-27, 3.2721e-27, 1.5586e-26, 3.0411e-26, 2.1930e-26,\n 1.4608e-28, 5.5392e-28, 7.1180e-27, 2.9409e-30, 2.1587e-27, 1.3544e-26,\n 1.4456e-27, 4.1948e-27, 6.1890e-27, 1.2353e-26, 3.8594e-27, 6.1407e-27,\n 1.2709e-26, 6.9553e-27, 6.0059e-28, 1.3925e-27, 4.4361e-28, 8.0621e-28,\n 5.4794e-27, 3.6338e-27, 1.0249e-26, 3.5201e-26, 2.2194e-26, 1.5231e-26,\n 1.0901e-26, 1.3019e-26, 4.7020e-28, 2.6969e-26, 8.1618e-28, 1.2299e-27,\n 5.8018e-27, 3.3642e-27, 6.1297e-27, 1.2383e-27, 9.4999e-28, 1.9069e-31,\n 5.7545e-28, 4.7618e-28, 8.8578e-27, 1.5747e-26, 1.2250e-27, 1.0584e-26,\n 4.6566e-27, 1.3917e-26, 1.1500e-28, 3.9081e-27, 1.1139e-27, 1.6488e-26,\n 2.5570e-27, 1.4744e-27, 3.0598e-08, 4.7900e-07, 8.4316e-08, 8.0408e-08,\n 6.3620e-07, 5.6528e-07, 7.7109e-09, 4.1100e-09, 1.0304e-08, 1.0313e-07,\n 1.3660e-10, 2.7494e-08, 4.9909e-07, 1.6788e-08, 2.7624e-08, 1.2648e-07,\n 1.8569e-07, 3.8142e-08, 3.3998e-07, 1.2828e-07, 1.4227e-10, 3.9955e-10,\n 6.9000e-08, 4.5457e-07, 3.8296e-07, 1.3342e-09, 1.7250e-07, 9.2776e-09,\n 4.8536e-08, 9.8469e-09, 2.1466e-07, 1.7924e-07, 6.9473e-07, 7.4573e-09,\n 1.4213e-07, 2.0403e-07, 1.2036e-07, 1.2182e-07, 1.0744e-07, 4.2463e-09,\n 6.5794e-07, 2.1546e-09, 1.0644e-08, 1.8740e-08, 3.1862e-07, 2.7247e-08,\n 5.9585e-08, 2.4864e-08, 4.7825e-09, 4.5539e-08, 9.1284e-08, 1.6306e-08,\n 2.6626e-07, 2.5130e-07, 2.7083e-08, 1.8979e-08, 2.9367e-08, 2.1285e-09,\n 1.7196e-07, 6.5230e-08, 1.3904e-07, 1.3761e-07, 1.4488e-08, 3.7517e-09,\n 7.7458e-12, 3.7763e-10, 1.1664e-09, 3.4883e-09, 5.0308e-08, 3.5852e-09,\n 4.4799e-08, 1.1683e-07, 1.6814e-07, 2.6801e-08, 6.0736e-08, 7.6687e-10,\n 3.4251e-08, 1.0143e-06, 5.3217e-09, 5.1130e-08, 6.8645e-08, 4.6946e-07,\n 1.7230e-10, 1.3315e-07, 3.8367e-08, 1.8530e-07, 1.1831e-07, 1.3738e-07,\n 1.1225e-11, 4.3820e-07, 2.9457e-07, 2.6816e-10, 1.3192e-08, 3.1733e-07,\n 1.4739e-08, 6.1560e-08, 8.7007e-10, 2.1865e-07, 4.2099e-08, 2.0462e-07,\n 1.0665e-07, 3.6984e-07, 4.6579e-09, 8.9426e-08, 1.8335e-07, 6.6813e-07,\n 2.9545e-07, 5.2146e-10, 1.1202e-08, 1.1218e-07, 7.0116e-08, 1.2948e-08,\n 1.9755e-08, 2.3354e-08, 6.9260e-08, 1.0301e-08, 8.7271e-08, 1.7320e-09,\n 2.6805e-07, 3.3879e-09, 2.6264e-07, 3.0088e-07, 9.2707e-08, 2.7320e-09,\n 3.1684e-08, 3.7091e-09, 8.6737e-08, 2.0557e-07, 2.0527e-08, 5.3250e-09,\n 2.1374e-07, 1.0605e-08, 1.2483e-07, 7.0338e-08, 3.6236e-08, 2.4680e-10,\n 5.0354e-09, 5.3333e-07, 6.8729e-09, 3.7001e-07, 6.3890e-11, 1.8315e-07,\n 6.9394e-08, 6.9321e-08, 2.0459e-08, 2.4927e-07, 9.3773e-08, 3.6419e-08,\n 1.5117e-07, 4.7180e-10, 3.4357e-07, 2.1641e-08, 2.6644e-09, 3.5726e-08,\n 7.0635e-08, 2.7704e-07, 5.3387e-08, 1.2690e-07, 1.9683e-09, 3.1918e-07,\n 9.3320e-08, 9.5370e-08, 7.4840e-08, 5.8532e-07, 5.3307e-10, 1.6560e-08,\n 5.0789e-07, 3.9640e-08, 4.6829e-08, 2.4362e-09, 8.1148e-08, 2.0995e-07,\n 1.8243e-07, 2.7743e-07, 6.7319e-09, 8.2571e-08, 4.8186e-08, 4.3119e-08,\n 4.4069e-08, 2.7761e-08, 7.0244e-08, 1.7623e-07, 1.3535e-08, 3.0051e-07,\n 3.3560e-07, 8.1546e-08, 8.0089e-08, 7.6227e-08, 5.5132e-10, 1.6312e-08,\n 1.6613e-08, 3.5827e-08, 1.6207e-07, 2.0528e-08, 8.9322e-09, 1.2355e-07,\n 3.4095e-11, 3.0037e-07, 2.5399e-09, 3.2911e-07, 1.4002e-07, 3.9358e-07,\n 1.0159e-08, 6.7321e-08, 2.6436e-07, 9.9473e-08, 9.6577e-09, 2.5516e-07,\n 1.9440e-08, 1.2368e-07, 6.2826e-10, 1.9055e-08, 5.5657e-07, 1.2142e-07,\n 5.5860e-07, 6.2851e-08, 8.9740e-08, 2.2228e-10, 3.1040e-08, 7.8274e-08,\n 7.1977e-10, 5.2690e-09, 6.7617e-08, 2.1650e-08, 3.1107e-08, 1.7719e-08,\n 2.0075e-07, 2.0110e-08, 5.0782e-09, 1.6763e-09, 5.5506e-10, 9.6669e-10,\n 1.2113e-08, 1.6602e-07, 1.1583e-08, 2.3683e-08, 2.9600e-09, 2.8265e-07,\n 1.3861e-07, 2.6876e-07, 6.1154e-07, 1.2666e-08, 3.3101e-07, 1.3509e-07,\n 2.9998e-08, 1.1235e-06, 2.4246e-08, 1.1562e-07, 3.8766e-08, 2.7818e-08,\n 1.4471e-08, 1.2431e-07, 6.6965e-08, 3.2344e-09, 1.6751e-08, 2.8784e-07],\n device='cuda:0')" + "exp_avg_sq": "tensor([6.6292e-11, 2.1305e-12, 1.9019e-11, 1.2516e-12, 8.3804e-12, 1.2389e-11,\n 1.8295e-12, 1.5826e-12, 7.7098e-14, 1.2917e-11, 1.0477e-11, 1.8120e-11,\n 4.5005e-13, 3.8100e-13, 4.6356e-12, 3.0403e-12, 1.0029e-12, 3.9695e-11,\n 1.3547e-11, 1.7350e-13, 6.8309e-13, 5.5430e-12, 1.2799e-11, 4.9370e-13,\n 2.3911e-11, 2.8043e-12, 3.9706e-12, 4.0486e-13, 9.2279e-13, 1.8876e-12,\n 1.8840e-12, 3.1335e-11, 4.5411e-12, 2.4308e-13, 9.1695e-15, 9.3683e-12,\n 3.3337e-12, 1.3431e-12, 4.0159e-14, 1.7907e-15, 5.4407e-12, 6.8084e-13,\n 3.3245e-12, 1.5697e-12, 2.6853e-13, 3.0570e-12, 3.8709e-12, 3.9274e-12,\n 2.0737e-12, 8.8116e-12, 3.4195e-12, 9.7259e-13, 1.2691e-12, 7.4734e-13,\n 2.8035e-14, 3.8117e-14, 8.7241e-13, 7.0024e-14, 2.8070e-12, 1.1938e-13,\n 9.9138e-12, 3.3064e-12, 7.8024e-15, 6.0150e-12, 1.1633e-11, 3.9621e-12,\n 5.9405e-12, 1.1763e-12, 1.5065e-13, 9.2859e-14, 6.0865e-12, 6.4294e-13,\n 5.1159e-14, 2.5135e-12, 1.4646e-12, 3.9478e-12, 2.9036e-12, 7.5196e-13,\n 1.2454e-17, 2.7269e-12, 2.0767e-12, 1.8279e-11, 2.7971e-13, 2.5842e-12,\n 3.5155e-13, 1.7738e-12, 1.0340e-12, 2.8612e-15, 1.7594e-12, 2.2835e-13,\n 8.1674e-15, 1.3251e-12, 1.1849e-12, 3.3769e-11, 1.1126e-11, 7.7374e-12,\n 8.4295e-13, 3.9777e-11, 2.7476e-12, 5.0955e-12, 4.7358e-13, 5.4478e-13,\n 2.6832e-13, 2.9142e-12, 3.5422e-14, 1.0163e-11, 1.6688e-11, 1.8549e-12,\n 2.1063e-12, 3.1665e-13, 3.2031e-12, 3.6374e-11, 2.7516e-13, 4.1672e-11,\n 5.7802e-12, 2.5750e-11, 2.5269e-18, 2.3621e-11, 1.7206e-11, 8.5157e-12,\n 3.2590e-11, 1.4580e-13, 1.7014e-11, 6.4156e-14, 9.2857e-12, 1.5466e-11,\n 1.4017e-11, 1.0152e-11, 1.8740e-11, 2.8359e-11, 3.2614e-12, 6.3610e-12,\n 1.9157e-13, 1.0931e-11, 3.1405e-12, 2.8724e-12, 6.6417e-12, 8.9929e-12,\n 7.5560e-15, 1.7545e-12, 2.2059e-11, 8.1178e-14, 4.6396e-12, 1.7118e-12,\n 1.0716e-11, 4.4576e-11, 8.8635e-12, 1.1384e-13, 2.8113e-11, 4.2633e-13,\n 9.0670e-13, 7.3658e-12, 3.0584e-11, 2.5138e-12, 1.0800e-11, 1.6203e-11,\n 1.1932e-13, 3.0090e-14, 1.1030e-12, 5.4225e-12, 5.1732e-12, 1.4347e-12,\n 2.1598e-11, 1.1172e-11, 1.2440e-12, 1.2004e-14, 2.3971e-13, 3.3961e-14,\n 6.7812e-13, 3.0537e-12, 6.1748e-13, 3.7459e-13, 2.5860e-12, 3.1327e-12,\n 2.8515e-13, 8.0004e-14, 2.9977e-12, 6.9211e-13, 4.2983e-13, 6.1260e-14,\n 4.9947e-12, 1.4528e-14, 1.0803e-12, 4.6635e-12, 1.3277e-12, 6.1907e-13,\n 4.6262e-12, 1.5531e-12, 1.5584e-11, 5.0881e-13, 4.3458e-12, 3.5820e-13,\n 3.9362e-12, 3.9086e-11, 5.5167e-12, 1.1756e-11, 6.0132e-12, 1.7577e-11,\n 2.0103e-15, 2.1345e-13, 1.0224e-12, 1.5180e-12, 2.6171e-12, 2.1559e-11,\n 5.4123e-13, 1.6205e-12, 1.2118e-13, 1.8252e-12, 2.1024e-13, 4.8822e-12,\n 8.3547e-16, 3.2830e-12, 2.2181e-13, 2.4181e-12, 2.2249e-11, 1.1279e-11,\n 1.6569e-11, 4.4809e-12, 5.6704e-12, 1.7640e-11, 4.4797e-12, 8.5543e-14,\n 2.2681e-12, 2.0175e-11, 1.2905e-13, 2.6634e-12, 1.6128e-11, 5.7342e-12,\n 5.0584e-12, 4.8939e-12, 4.2356e-12, 1.2611e-13, 1.8602e-12, 9.8265e-13,\n 4.5586e-12, 1.2512e-12, 3.1506e-14, 2.8903e-13, 1.1671e-12, 1.2291e-13,\n 1.1306e-11, 3.1185e-12, 4.7742e-13, 3.7793e-12, 1.1575e-13, 4.3222e-12,\n 1.1609e-11, 1.4896e-12, 4.9357e-12, 1.5386e-12, 1.1738e-12, 1.2071e-11,\n 4.6883e-12, 1.3358e-11, 1.2495e-11, 2.7085e-12, 4.2971e-28, 1.8996e-29,\n 7.7498e-27, 5.4489e-28, 8.6767e-27, 3.4101e-28, 2.2369e-28, 4.7386e-27,\n 3.1053e-27, 4.5199e-27, 1.3645e-28, 1.6901e-28, 4.7399e-27, 4.2599e-27,\n 2.3484e-28, 5.1748e-28, 1.9064e-27, 1.2145e-30, 1.2051e-26, 1.9075e-27,\n 6.2769e-28, 4.2876e-27, 1.9069e-27, 1.9575e-27, 1.4313e-27, 4.5716e-28,\n 1.1232e-27, 1.2854e-28, 9.7128e-27, 1.9963e-27, 1.7341e-27, 1.5445e-27,\n 8.3598e-27, 2.5323e-27, 2.7317e-28, 9.6797e-28, 1.5382e-27, 2.1949e-29,\n 2.1128e-28, 6.6276e-29, 4.5629e-30, 2.0393e-28, 1.0098e-27, 1.4497e-29,\n 1.6562e-27, 4.8692e-28, 2.5730e-27, 1.3295e-29, 2.1044e-27, 1.8129e-29,\n 5.5476e-28, 6.3368e-32, 7.7907e-28, 3.7087e-27, 1.0761e-27, 2.0654e-28,\n 8.0572e-27, 1.0365e-27, 1.4972e-26, 1.2047e-27, 3.4610e-29, 1.9252e-28,\n 1.7477e-27, 1.0362e-27, 2.6698e-28, 1.3914e-27, 5.0923e-27, 6.6727e-27,\n 3.5535e-27, 1.4602e-28, 1.2433e-26, 9.0627e-27, 9.5241e-27, 1.5896e-26,\n 3.2887e-26, 1.0341e-27, 2.0494e-26, 4.3585e-28, 1.5094e-27, 3.0101e-28,\n 4.4360e-27, 1.7571e-26, 1.0498e-26, 6.8405e-27, 1.8200e-27, 3.2918e-27,\n 1.6019e-26, 3.6028e-26, 2.0177e-27, 5.5263e-27, 3.1811e-27, 1.6658e-26,\n 1.7478e-27, 1.0213e-28, 1.1590e-27, 1.1065e-27, 1.0291e-26, 3.4986e-29,\n 2.0981e-27, 1.5172e-29, 1.1929e-27, 5.6327e-27, 8.2812e-27, 7.0971e-28,\n 1.8722e-28, 8.1782e-28, 1.4151e-27, 8.1155e-29, 4.9848e-29, 2.3301e-28,\n 2.6936e-28, 4.6060e-29, 1.0661e-27, 1.3325e-27, 1.8849e-27, 1.7095e-27,\n 1.2158e-27, 1.6055e-29, 5.4228e-31, 3.1242e-29, 2.2127e-29, 2.2986e-27,\n 3.1870e-27, 2.3327e-27, 1.1014e-27, 2.9883e-27, 1.4812e-29, 5.0079e-27,\n 5.3317e-29, 6.1239e-28, 2.4945e-28, 7.9488e-27, 3.1772e-27, 4.9677e-27,\n 4.1329e-27, 8.7538e-29, 7.7862e-29, 9.1911e-28, 1.8401e-27, 5.6954e-27,\n 6.0203e-28, 2.8561e-28, 1.6388e-28, 2.4822e-29, 9.3432e-27, 8.7879e-28,\n 2.6412e-27, 1.9900e-28, 3.1291e-28, 2.0239e-26, 3.4405e-27, 3.9119e-29,\n 2.7922e-27, 3.3145e-27, 9.6203e-27, 5.7187e-27, 7.3250e-27, 3.5042e-27,\n 5.2882e-29, 4.1237e-26, 9.3429e-27, 1.1458e-28, 1.1301e-26, 2.8710e-27,\n 2.4987e-27, 1.4714e-28, 1.6964e-27, 8.0862e-28, 8.4230e-28, 3.3106e-27,\n 2.6477e-27, 3.1508e-31, 2.4566e-27, 2.4831e-28, 3.8229e-29, 5.8427e-28,\n 5.9826e-27, 4.2336e-27, 2.1890e-29, 3.3346e-27, 9.6800e-28, 6.9525e-28,\n 1.1651e-26, 1.1669e-26, 1.2351e-27, 6.3473e-29, 1.1924e-26, 1.1048e-26,\n 5.1515e-27, 6.5521e-28, 6.6861e-28, 2.6554e-27, 3.1139e-27, 3.4873e-27,\n 1.7161e-26, 1.2079e-27, 3.7838e-28, 5.9741e-27, 1.1531e-27, 5.7633e-27,\n 9.0955e-28, 9.8591e-28, 9.3502e-28, 4.4538e-27, 8.6901e-27, 6.2667e-27,\n 4.1744e-29, 1.5829e-28, 2.0340e-27, 8.4039e-31, 6.1687e-28, 3.8704e-27,\n 4.1310e-28, 1.1987e-27, 1.7685e-27, 3.5299e-27, 1.1029e-27, 1.7547e-27,\n 3.6318e-27, 1.9875e-27, 1.7162e-28, 3.9793e-28, 1.2677e-28, 2.3038e-28,\n 1.5658e-27, 1.0384e-27, 2.9288e-27, 1.0059e-26, 6.3423e-27, 4.3524e-27,\n 3.1149e-27, 3.7202e-27, 1.3436e-28, 7.7065e-27, 2.3323e-28, 3.5146e-28,\n 1.6579e-27, 9.6135e-28, 1.7516e-27, 3.5384e-28, 2.7147e-28, 5.4490e-32,\n 1.6444e-28, 1.3607e-28, 2.5312e-27, 4.4998e-27, 3.5006e-28, 3.0244e-27,\n 1.3307e-27, 3.9770e-27, 3.2862e-29, 1.1168e-27, 3.1830e-28, 4.7117e-27,\n 7.3069e-28, 4.2133e-28, 8.7437e-09, 1.3688e-07, 2.4094e-08, 2.2977e-08,\n 1.8180e-07, 1.6153e-07, 2.2035e-09, 1.1745e-09, 2.9444e-09, 2.9470e-08,\n 3.9033e-11, 7.8566e-09, 1.4262e-07, 4.7973e-09, 7.8937e-09, 3.6144e-08,\n 5.3062e-08, 1.0899e-08, 9.7153e-08, 3.6656e-08, 4.0655e-11, 1.1417e-10,\n 1.9717e-08, 1.2990e-07, 1.0943e-07, 3.8125e-10, 4.9293e-08, 2.6511e-09,\n 1.3870e-08, 2.8138e-09, 6.1340e-08, 5.1220e-08, 1.9852e-07, 2.1310e-09,\n 4.0615e-08, 5.8303e-08, 3.4394e-08, 3.4810e-08, 3.0701e-08, 1.2134e-09,\n 1.8801e-07, 6.1570e-10, 3.0415e-09, 5.3551e-09, 9.1049e-08, 7.7861e-09,\n 1.7027e-08, 7.1051e-09, 1.3666e-09, 1.3013e-08, 2.6085e-08, 4.6595e-09,\n 7.6086e-08, 7.1810e-08, 7.7391e-09, 5.4233e-09, 8.3919e-09, 6.0823e-10,\n 4.9138e-08, 1.8640e-08, 3.9732e-08, 3.9322e-08, 4.1401e-09, 1.0721e-09,\n 2.2134e-12, 1.0791e-10, 3.3331e-10, 9.9681e-10, 1.4376e-08, 1.0245e-09,\n 1.2802e-08, 3.3385e-08, 4.8047e-08, 7.6586e-09, 1.7356e-08, 2.1914e-10,\n 9.7875e-09, 2.8983e-07, 1.5207e-09, 1.4611e-08, 1.9616e-08, 1.3415e-07,\n 4.9236e-11, 3.8050e-08, 1.0964e-08, 5.2952e-08, 3.3807e-08, 3.9257e-08,\n 3.2077e-12, 1.2522e-07, 8.4175e-08, 7.6630e-11, 3.7698e-09, 9.0680e-08,\n 4.2118e-09, 1.7591e-08, 2.4863e-10, 6.2481e-08, 1.2030e-08, 5.8472e-08,\n 3.0477e-08, 1.0568e-07, 1.3310e-09, 2.5554e-08, 5.2393e-08, 1.9092e-07,\n 8.4427e-08, 1.4901e-10, 3.2011e-09, 3.2056e-08, 2.0036e-08, 3.7000e-09,\n 5.6451e-09, 6.6735e-09, 1.9792e-08, 2.9435e-09, 2.4938e-08, 4.9493e-10,\n 7.6599e-08, 9.6813e-10, 7.5052e-08, 8.5979e-08, 2.6492e-08, 7.8070e-10,\n 9.0539e-09, 1.0599e-09, 2.4786e-08, 5.8745e-08, 5.8657e-09, 1.5217e-09,\n 6.1077e-08, 3.0306e-09, 3.5671e-08, 2.0100e-08, 1.0355e-08, 7.0525e-11,\n 1.4389e-09, 1.5240e-07, 1.9640e-09, 1.0573e-07, 1.8257e-11, 5.2335e-08,\n 1.9830e-08, 1.9809e-08, 5.8465e-09, 7.1232e-08, 2.6796e-08, 1.0407e-08,\n 4.3197e-08, 1.3482e-10, 9.8179e-08, 6.1840e-09, 7.6138e-10, 1.0209e-08,\n 2.0185e-08, 7.9166e-08, 1.5256e-08, 3.6262e-08, 5.6247e-10, 9.1207e-08,\n 2.6667e-08, 2.7253e-08, 2.1386e-08, 1.6726e-07, 1.5233e-10, 4.7321e-09,\n 1.4513e-07, 1.1327e-08, 1.3382e-08, 6.9615e-10, 2.3189e-08, 5.9993e-08,\n 5.2130e-08, 7.9279e-08, 1.9237e-09, 2.3595e-08, 1.3769e-08, 1.2322e-08,\n 1.2593e-08, 7.9328e-09, 2.0073e-08, 5.0358e-08, 3.8677e-09, 8.5873e-08,\n 9.5899e-08, 2.3302e-08, 2.2886e-08, 2.1782e-08, 1.5754e-10, 4.6612e-09,\n 4.7472e-09, 1.0238e-08, 4.6314e-08, 5.8662e-09, 2.5525e-09, 3.5307e-08,\n 9.7430e-12, 8.5833e-08, 7.2580e-10, 9.4045e-08, 4.0011e-08, 1.1247e-07,\n 2.9031e-09, 1.9237e-08, 7.5542e-08, 2.8425e-08, 2.7598e-09, 7.2913e-08,\n 5.5552e-09, 3.5342e-08, 1.7953e-10, 5.4452e-09, 1.5904e-07, 3.4697e-08,\n 1.5962e-07, 1.7960e-08, 2.5644e-08, 6.3518e-11, 8.8698e-09, 2.2368e-08,\n 2.0568e-10, 1.5056e-09, 1.9322e-08, 6.1867e-09, 8.8891e-09, 5.0632e-09,\n 5.7367e-08, 5.7464e-09, 1.4511e-09, 4.7901e-10, 1.5861e-10, 2.7624e-10,\n 3.4613e-09, 4.7442e-08, 3.3098e-09, 6.7676e-09, 8.4584e-10, 8.0770e-08,\n 3.9607e-08, 7.6802e-08, 1.7475e-07, 3.6194e-09, 9.4587e-08, 3.8602e-08,\n 8.5721e-09, 3.2106e-07, 6.9286e-09, 3.3040e-08, 1.1078e-08, 7.9492e-09,\n 4.1353e-09, 3.5522e-08, 1.9136e-08, 9.2426e-10, 4.7868e-09, 8.2253e-08],\n device='cuda:0')" }, "36": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([[ 5.6052e-45, -5.6052e-45, 5.6052e-45, ..., -5.6052e-45,\n 5.6052e-45, -5.6052e-45],\n [-5.6052e-45, -5.6052e-45, -5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., -5.6052e-45,\n 5.6052e-45, -5.6052e-45],\n ...,\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., -5.6052e-45,\n 5.6052e-45, -5.6052e-45],\n [-5.6052e-45, 5.6052e-45, -5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, 5.6052e-45],\n [-5.6052e-45, 5.6052e-45, -5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, 5.6052e-45]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.0249e-08, 5.7233e-12, 5.7234e-08, ..., 1.4822e-10, 1.4921e-09,\n 4.1247e-09],\n [1.3893e-08, 1.4479e-12, 8.1180e-08, ..., 2.3234e-10, 2.2049e-09,\n 5.8231e-09],\n [7.8449e-09, 9.9788e-13, 4.5882e-08, ..., 1.2123e-10, 1.3058e-09,\n 2.9669e-09],\n ...,\n [5.9362e-09, 8.4729e-13, 3.6001e-08, ..., 5.5682e-11, 8.9820e-10,\n 2.3880e-09],\n [7.7985e-10, 2.0857e-12, 4.2957e-09, ..., 1.4941e-11, 8.4439e-11,\n 3.2375e-10],\n [4.1263e-09, 4.9721e-12, 2.4213e-08, ..., 7.0467e-11, 5.6390e-10,\n 1.6570e-09]], device='cuda:0')" + "exp_avg_sq": "tensor([[2.9288e-09, 1.6355e-12, 1.6355e-08, ..., 4.2355e-11, 4.2639e-10,\n 1.1787e-09],\n [3.9699e-09, 4.1374e-13, 2.3198e-08, ..., 6.6393e-11, 6.3007e-10,\n 1.6640e-09],\n [2.2417e-09, 2.8515e-13, 1.3111e-08, ..., 3.4642e-11, 3.7315e-10,\n 8.4782e-10],\n ...,\n [1.6963e-09, 2.4212e-13, 1.0288e-08, ..., 1.5912e-11, 2.5667e-10,\n 6.8239e-10],\n [2.2285e-10, 5.9601e-13, 1.2275e-09, ..., 4.2694e-12, 2.4129e-11,\n 9.2513e-11],\n [1.1791e-09, 1.4208e-12, 6.9190e-09, ..., 2.0137e-11, 1.6114e-10,\n 4.7349e-10]], device='cuda:0')" }, "37": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([3.7079e-07, 5.1584e-07, 2.7586e-07, 1.4646e-06, 4.4450e-09, 4.2481e-08,\n 7.7796e-07, 1.4608e-06, 1.1815e-06, 5.8837e-07, 2.5502e-08, 1.2285e-06,\n 6.0692e-08, 3.0266e-09, 1.5976e-07, 1.5264e-09, 6.6728e-07, 1.2636e-06,\n 3.5340e-07, 7.5499e-08, 1.3760e-07, 1.0127e-07, 4.8209e-07, 6.1799e-09,\n 1.6064e-08, 1.3631e-07, 1.4055e-08, 2.0549e-07, 9.7382e-08, 3.0573e-07,\n 6.6853e-10, 1.6406e-07, 1.6601e-07, 5.7817e-08, 7.4436e-09, 3.2548e-07,\n 1.1145e-08, 1.8715e-06, 3.1384e-07, 2.2336e-08, 7.0505e-07, 4.5927e-08,\n 1.3495e-07, 1.2203e-08, 2.4625e-07, 4.2233e-07, 5.3029e-09, 5.5168e-10,\n 3.4644e-07, 3.1476e-07, 4.4965e-07, 1.8740e-08, 1.9542e-07, 7.3247e-07,\n 9.3572e-09, 8.1845e-08, 7.4036e-07, 2.9373e-09, 3.3975e-08, 6.0997e-08,\n 2.5605e-07, 2.0041e-07, 4.1230e-08, 3.9175e-07, 2.4628e-09, 6.1749e-07,\n 8.4105e-08, 1.6079e-07, 1.8106e-07, 6.6298e-08, 3.1147e-08, 6.1021e-07,\n 7.4765e-07, 1.5664e-06, 1.0108e-08, 1.2230e-09, 6.1682e-08, 5.2665e-07,\n 7.9803e-09, 1.5146e-09, 1.2380e-07, 7.8499e-09, 3.1387e-07, 5.3916e-08,\n 5.9328e-08, 3.0667e-08, 5.9924e-08, 8.6155e-08, 8.8815e-08, 1.0362e-06,\n 9.0299e-10, 6.1603e-07, 2.5546e-07, 6.8638e-07, 1.6728e-06, 1.0662e-06,\n 1.0827e-07, 1.4597e-06, 5.2471e-08, 1.8449e-08, 6.4971e-07, 2.2335e-10,\n 2.8323e-07, 1.7111e-07, 1.6131e-08, 4.5271e-07, 2.7209e-10, 2.6661e-06,\n 5.5236e-07, 5.7303e-07, 2.0251e-09, 1.6575e-08, 5.1330e-08, 5.1206e-07,\n 1.6141e-07, 4.3030e-07, 9.0993e-10, 5.3185e-07, 6.6878e-07, 2.6811e-07,\n 8.9415e-07, 3.5323e-07, 2.7609e-07, 3.9567e-08, 1.8706e-06, 9.8093e-07,\n 2.4932e-07, 1.0304e-07, 2.1361e-07, 7.6028e-10, 2.2244e-06, 2.1618e-07,\n 2.8082e-09, 8.2931e-10, 1.4448e-06, 5.5733e-07, 8.3920e-08, 3.6876e-08,\n 3.4101e-09, 1.1166e-09, 9.2110e-07, 1.9159e-09, 1.3821e-06, 2.6893e-08,\n 3.1201e-08, 6.6141e-08, 5.5035e-07, 4.6162e-07, 1.3407e-07, 1.6144e-09,\n 2.6581e-08, 8.4818e-09, 1.0065e-09, 9.1002e-07, 8.0521e-08, 1.1939e-06,\n 2.8644e-08, 6.6924e-07, 9.0549e-08, 9.0694e-07, 2.3708e-07, 3.8590e-08,\n 4.0714e-07, 3.5420e-07, 4.9141e-07, 1.0830e-06, 2.2755e-06, 1.0751e-09,\n 4.2696e-09, 2.0706e-07, 1.1973e-07, 1.8229e-09, 7.3427e-07, 9.4512e-08,\n 1.0616e-06, 1.3238e-09, 4.9609e-07, 5.5358e-08, 2.3884e-07, 2.3970e-07,\n 2.5555e-06, 2.5445e-07, 1.6431e-07, 2.4427e-10, 4.7873e-08, 7.4204e-07,\n 3.0310e-09, 2.7110e-07, 6.9066e-07, 6.1538e-08, 7.0565e-07, 6.1359e-08,\n 1.7789e-07, 1.6283e-07, 1.0759e-06, 5.7575e-07, 6.7742e-07, 4.9729e-09,\n 2.1580e-07, 1.3821e-06, 1.4777e-09, 1.2820e-07, 6.6082e-08, 9.6904e-07,\n 9.4674e-07, 1.3541e-07, 4.4791e-07, 9.2397e-08, 6.6184e-09, 2.5246e-09,\n 2.9983e-09, 1.5078e-07, 7.7688e-07, 6.9147e-07, 2.4952e-07, 8.6800e-07,\n 1.0543e-07, 5.2152e-07, 3.7991e-09, 1.3685e-08, 2.6132e-10, 7.3657e-08,\n 8.1701e-07, 6.0887e-08, 2.0534e-08, 4.2051e-07, 1.1838e-07, 3.6964e-07,\n 2.6286e-08, 5.2140e-07, 7.2970e-09, 2.8531e-07, 8.8684e-07, 3.7408e-09,\n 7.3846e-07, 1.6570e-08, 1.7534e-07, 6.5818e-10, 1.3923e-06, 2.8062e-08,\n 1.2762e-07, 2.8810e-08, 3.3707e-08, 1.9743e-07, 2.9055e-07, 2.0135e-07,\n 7.2784e-07, 2.7678e-07, 1.1783e-06, 1.1322e-07, 6.1281e-07, 1.3285e-11,\n 4.8302e-08, 2.2758e-07, 2.9418e-08, 1.5659e-07], device='cuda:0')" + "exp_avg_sq": "tensor([1.0596e-07, 1.4741e-07, 7.8830e-08, 4.1853e-07, 1.2702e-09, 1.2139e-08,\n 2.2231e-07, 4.1743e-07, 3.3763e-07, 1.6813e-07, 7.2874e-09, 3.5106e-07,\n 1.7343e-08, 8.6488e-10, 4.5652e-08, 4.3618e-10, 1.9068e-07, 3.6110e-07,\n 1.0099e-07, 2.1574e-08, 3.9319e-08, 2.8939e-08, 1.3776e-07, 1.7660e-09,\n 4.5904e-09, 3.8951e-08, 4.0163e-09, 5.8721e-08, 2.7828e-08, 8.7366e-08,\n 1.9104e-10, 4.6882e-08, 4.7438e-08, 1.6522e-08, 2.1271e-09, 9.3008e-08,\n 3.1847e-09, 5.3480e-07, 8.9682e-08, 6.3826e-09, 2.0147e-07, 1.3124e-08,\n 3.8563e-08, 3.4871e-09, 7.0367e-08, 1.2068e-07, 1.5154e-09, 1.5765e-10,\n 9.8998e-08, 8.9946e-08, 1.2849e-07, 5.3552e-09, 5.5843e-08, 2.0931e-07,\n 2.6739e-09, 2.3388e-08, 2.1157e-07, 8.3936e-10, 9.7088e-09, 1.7430e-08,\n 7.3168e-08, 5.7269e-08, 1.1782e-08, 1.1194e-07, 7.0376e-10, 1.7645e-07,\n 2.4034e-08, 4.5948e-08, 5.1739e-08, 1.8945e-08, 8.9005e-09, 1.7437e-07,\n 2.1365e-07, 4.4761e-07, 2.8885e-09, 3.4947e-10, 1.7626e-08, 1.5049e-07,\n 2.2804e-09, 4.3280e-10, 3.5377e-08, 2.2432e-09, 8.9692e-08, 1.5407e-08,\n 1.6953e-08, 8.7634e-09, 1.7124e-08, 2.4619e-08, 2.5380e-08, 2.9609e-07,\n 2.5804e-10, 1.7603e-07, 7.2999e-08, 1.9614e-07, 4.7801e-07, 3.0467e-07,\n 3.0939e-08, 4.1712e-07, 1.4994e-08, 5.2718e-09, 1.8566e-07, 6.3825e-11,\n 8.0936e-08, 4.8897e-08, 4.6095e-09, 1.2937e-07, 7.7751e-11, 7.6186e-07,\n 1.5784e-07, 1.6375e-07, 5.7869e-10, 4.7365e-09, 1.4668e-08, 1.4632e-07,\n 4.6125e-08, 1.2296e-07, 2.6002e-10, 1.5198e-07, 1.9111e-07, 7.6615e-08,\n 2.5551e-07, 1.0094e-07, 7.8895e-08, 1.1307e-08, 5.3453e-07, 2.8031e-07,\n 7.1245e-08, 2.9443e-08, 6.1040e-08, 2.1726e-10, 6.3563e-07, 6.1775e-08,\n 8.0248e-10, 2.3698e-10, 4.1287e-07, 1.5926e-07, 2.3981e-08, 1.0538e-08,\n 9.7446e-10, 3.1907e-10, 2.6321e-07, 5.4747e-10, 3.9494e-07, 7.6848e-09,\n 8.9159e-09, 1.8900e-08, 1.5727e-07, 1.3191e-07, 3.8310e-08, 4.6134e-10,\n 7.5958e-09, 2.4237e-09, 2.8762e-10, 2.6005e-07, 2.3010e-08, 3.4115e-07,\n 8.1853e-09, 1.9124e-07, 2.5875e-08, 2.5916e-07, 6.7748e-08, 1.1027e-08,\n 1.1634e-07, 1.0121e-07, 1.4043e-07, 3.0949e-07, 6.5025e-07, 3.0722e-10,\n 1.2201e-09, 5.9169e-08, 3.4215e-08, 5.2092e-10, 2.0982e-07, 2.7008e-08,\n 3.0335e-07, 3.7829e-10, 1.4176e-07, 1.5819e-08, 6.8252e-08, 6.8496e-08,\n 7.3026e-07, 7.2711e-08, 4.6951e-08, 6.9802e-11, 1.3680e-08, 2.1204e-07,\n 8.6614e-10, 7.7468e-08, 1.9736e-07, 1.7585e-08, 2.0165e-07, 1.7534e-08,\n 5.0833e-08, 4.6529e-08, 3.0745e-07, 1.6453e-07, 1.9358e-07, 1.4210e-09,\n 6.1668e-08, 3.9495e-07, 4.2227e-10, 3.6633e-08, 1.8884e-08, 2.7691e-07,\n 2.7054e-07, 3.8696e-08, 1.2799e-07, 2.6403e-08, 1.8913e-09, 7.2143e-10,\n 8.5678e-10, 4.3086e-08, 2.2200e-07, 1.9759e-07, 7.1302e-08, 2.4804e-07,\n 3.0127e-08, 1.4903e-07, 1.0856e-09, 3.9105e-09, 7.4674e-11, 2.1048e-08,\n 2.3347e-07, 1.7399e-08, 5.8677e-09, 1.2017e-07, 3.3827e-08, 1.0563e-07,\n 7.5115e-09, 1.4900e-07, 2.0852e-09, 8.1531e-08, 2.5342e-07, 1.0690e-09,\n 2.1102e-07, 4.7351e-09, 5.0106e-08, 1.8808e-10, 3.9786e-07, 8.0190e-09,\n 3.6470e-08, 8.2328e-09, 9.6320e-09, 5.6418e-08, 8.3028e-08, 5.7536e-08,\n 2.0799e-07, 7.9092e-08, 3.3670e-07, 3.2353e-08, 1.7512e-07, 3.7963e-12,\n 1.3803e-08, 6.5032e-08, 8.4063e-09, 4.4746e-08], device='cuda:0')" }, "38": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([[-5.6052e-45, -5.6052e-45, -5.6052e-45, ..., -5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, -5.6052e-45]], device='cuda:0')", - "exp_avg_sq": "tensor([[2.5943e-06, 6.8672e-07, 9.8298e-07, ..., 1.2520e-07, 7.6421e-07,\n 1.4112e-06],\n [2.6787e-07, 6.9628e-08, 1.0424e-07, ..., 1.3882e-08, 7.7159e-08,\n 1.3990e-07],\n [2.9181e-07, 7.5233e-08, 1.0861e-07, ..., 1.3366e-08, 8.6616e-08,\n 1.6180e-07],\n [3.0573e-07, 8.4408e-08, 1.1494e-07, ..., 1.4498e-08, 9.1272e-08,\n 1.6946e-07]], device='cuda:0')" + "exp_avg_sq": "tensor([[7.4134e-07, 1.9624e-07, 2.8089e-07, ..., 3.5777e-08, 2.1838e-07,\n 4.0326e-07],\n [7.6546e-08, 1.9897e-08, 2.9788e-08, ..., 3.9668e-09, 2.2049e-08,\n 3.9978e-08],\n [8.3387e-08, 2.1498e-08, 3.1037e-08, ..., 3.8195e-09, 2.4751e-08,\n 4.6237e-08],\n [8.7365e-08, 2.4120e-08, 3.2844e-08, ..., 4.1428e-09, 2.6082e-08,\n 4.8425e-08]], device='cuda:0')" }, "39": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([6.5055e-05, 6.7032e-06, 7.2531e-06, 7.7478e-06], device='cuda:0')" + "exp_avg_sq": "tensor([1.8590e-05, 1.9155e-06, 2.0726e-06, 2.2140e-06], device='cuda:0')" }, "40": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([[-5.6052e-45, -5.6052e-45, -5.6052e-45, ..., -5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, -5.6052e-45]], device='cuda:0')", - "exp_avg_sq": "tensor([[2.5943e-06, 6.8672e-07, 9.8298e-07, ..., 1.2520e-07, 7.6421e-07,\n 1.4112e-06],\n [2.6787e-07, 6.9628e-08, 1.0424e-07, ..., 1.3882e-08, 7.7159e-08,\n 1.3990e-07],\n [2.9181e-07, 7.5233e-08, 1.0861e-07, ..., 1.3366e-08, 8.6616e-08,\n 1.6180e-07],\n [3.0573e-07, 8.4408e-08, 1.1494e-07, ..., 1.4498e-08, 9.1272e-08,\n 1.6946e-07]], device='cuda:0')" + "exp_avg_sq": "tensor([[7.4134e-07, 1.9624e-07, 2.8089e-07, ..., 3.5777e-08, 2.1838e-07,\n 4.0326e-07],\n [7.6546e-08, 1.9897e-08, 2.9788e-08, ..., 3.9668e-09, 2.2049e-08,\n 3.9978e-08],\n [8.3387e-08, 2.1498e-08, 3.1037e-08, ..., 3.8195e-09, 2.4751e-08,\n 4.6237e-08],\n [8.7365e-08, 2.4120e-08, 3.2844e-08, ..., 4.1428e-09, 2.6082e-08,\n 4.8425e-08]], device='cuda:0')" }, "41": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([6.5055e-05, 6.7032e-06, 7.2531e-06, 7.7478e-06], device='cuda:0')" + "exp_avg_sq": "tensor([1.8590e-05, 1.9155e-06, 2.0726e-06, 2.2140e-06], device='cuda:0')" }, "42": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([[-5.6052e-45, -5.6052e-45, -5.6052e-45, ..., -5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, -5.6052e-45]], device='cuda:0')", - "exp_avg_sq": "tensor([[2.5943e-06, 6.8672e-07, 9.8298e-07, ..., 1.2520e-07, 7.6421e-07,\n 1.4112e-06],\n [2.6787e-07, 6.9628e-08, 1.0424e-07, ..., 1.3882e-08, 7.7159e-08,\n 1.3990e-07],\n [2.9181e-07, 7.5233e-08, 1.0861e-07, ..., 1.3366e-08, 8.6616e-08,\n 1.6180e-07],\n [3.0573e-07, 8.4408e-08, 1.1494e-07, ..., 1.4498e-08, 9.1272e-08,\n 1.6946e-07]], device='cuda:0')" + "exp_avg_sq": "tensor([[7.4134e-07, 1.9624e-07, 2.8089e-07, ..., 3.5777e-08, 2.1838e-07,\n 4.0326e-07],\n [7.6546e-08, 1.9897e-08, 2.9788e-08, ..., 3.9668e-09, 2.2049e-08,\n 3.9978e-08],\n [8.3387e-08, 2.1498e-08, 3.1037e-08, ..., 3.8195e-09, 2.4751e-08,\n 4.6237e-08],\n [8.7365e-08, 2.4120e-08, 3.2844e-08, ..., 4.1428e-09, 2.6082e-08,\n 4.8425e-08]], device='cuda:0')" }, "43": { - "step": "tensor(1252.)", + "step": "tensor(2504.)", "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([6.5055e-05, 6.7032e-06, 7.2531e-06, 7.7478e-06], device='cuda:0')" + "exp_avg_sq": "tensor([1.8590e-05, 1.9155e-06, 2.0726e-06, 2.2140e-06], device='cuda:0')" + }, + "8": { + "step": "tensor(1252.)", + "exp_avg": "tensor([[ 9.5150e-07, -3.8557e-06, 0.0000e+00, ..., -4.2832e-06,\n 4.4121e-06, 1.9494e-06],\n [ 7.1172e-07, -1.6450e-06, 0.0000e+00, ..., -8.8052e-07,\n -3.8274e-06, -1.6891e-06],\n [-1.8110e-06, 3.2072e-07, 0.0000e+00, ..., -2.7184e-07,\n -4.4957e-06, -1.4556e-06],\n ...,\n [ 6.7636e-08, -2.5984e-07, 0.0000e+00, ..., 3.8079e-07,\n 2.7812e-07, 1.9362e-06],\n [ 1.1224e-06, 4.5980e-08, 0.0000e+00, ..., -5.3150e-07,\n 7.8061e-07, -7.7626e-07],\n [-2.8625e-06, 4.3572e-07, 0.0000e+00, ..., 7.9463e-07,\n 3.0544e-08, 1.9235e-09]], device='cuda:0')", + "exp_avg_sq": "tensor([[8.1393e-11, 3.1440e-11, 0.0000e+00, ..., 1.2101e-10, 1.8385e-10,\n 6.7095e-11],\n [1.0556e-10, 1.4296e-10, 0.0000e+00, ..., 2.0345e-10, 1.3139e-09,\n 3.1743e-11],\n [4.6596e-10, 5.7350e-11, 0.0000e+00, ..., 1.3298e-10, 4.8543e-10,\n 1.8508e-10],\n ...,\n [1.4537e-10, 5.4636e-11, 0.0000e+00, ..., 5.9882e-11, 2.3956e-10,\n 7.1686e-11],\n [2.3094e-10, 2.3896e-10, 0.0000e+00, ..., 3.0046e-10, 1.6661e-10,\n 1.9870e-11],\n [6.5843e-11, 2.6893e-11, 0.0000e+00, ..., 7.8340e-11, 8.1009e-11,\n 1.3380e-10]], device='cuda:0')" + }, + "9": { + "step": "tensor(1252.)", + "exp_avg": "tensor([ 3.5455e-05, 1.1664e-05, -7.1505e-05, ..., -8.4824e-06,\n -1.8059e-05, 4.3417e-07], device='cuda:0')", + "exp_avg_sq": "tensor([1.4199e-08, 3.8053e-08, 2.7986e-08, ..., 1.7082e-08, 1.9369e-08,\n 6.2494e-09], device='cuda:0')" + }, + "10": { + "step": "tensor(1252.)", + "exp_avg": "tensor([[-6.0034e-08, 8.3175e-07, 5.6886e-07, ..., -1.3715e-07,\n 3.5541e-07, -2.4610e-07],\n [-2.7396e-07, 5.3619e-07, -2.5902e-06, ..., -1.7347e-06,\n 8.8987e-07, -5.4779e-07],\n [-4.0385e-07, -4.3323e-07, 8.1043e-07, ..., 2.3942e-07,\n 1.1977e-06, 7.2821e-07],\n ...,\n [-1.1566e-06, 7.1615e-07, -3.2375e-06, ..., -3.3065e-07,\n 1.0924e-06, 5.3783e-07],\n [ 2.4937e-06, 1.2662e-06, 2.0117e-07, ..., 7.8634e-08,\n 4.8693e-07, -1.5626e-07],\n [ 9.2190e-07, 5.1389e-07, -1.2902e-06, ..., -4.0980e-07,\n 1.0450e-06, -1.0818e-07]], device='cuda:0')", + "exp_avg_sq": "tensor([[6.9890e-11, 7.0728e-11, 4.1142e-11, ..., 2.0425e-11, 4.8474e-11,\n 1.0066e-10],\n [1.9791e-11, 1.2321e-10, 3.9717e-11, ..., 3.1398e-11, 3.0796e-11,\n 1.6469e-11],\n [3.0407e-11, 1.0305e-10, 4.5896e-11, ..., 4.8042e-11, 5.0357e-11,\n 4.0974e-11],\n ...,\n [1.0168e-10, 1.1815e-10, 5.8931e-11, ..., 5.4174e-11, 3.8480e-11,\n 2.0247e-10],\n [5.3361e-11, 1.0527e-10, 5.0470e-11, ..., 3.9453e-11, 3.5202e-11,\n 2.4136e-11],\n [4.3872e-11, 9.0027e-11, 6.3022e-11, ..., 6.6265e-11, 4.6989e-11,\n 3.8512e-11]], device='cuda:0')" } }, "param_groups": [ { - "lr": 0.00904518046337755, + "lr": 0.00793913236883622, "name": "shared", "betas": [ 0.9, @@ -212,7 +227,7 @@ ] }, { - "lr": 0.00904518046337755, + "lr": 0.00793913236883622, "name": "scale_384", "betas": [ 0.9, @@ -235,7 +250,7 @@ ] }, { - "lr": 0.00904518046337755, + "lr": 0.00793913236883622, "name": "scale_768", "betas": [ 0.9, @@ -258,7 +273,7 @@ ] }, { - "lr": 0.00904518046337755, + "lr": 0.00793913236883622, "name": "scale_1024", "betas": [ 0.9, @@ -281,7 +296,7 @@ ] }, { - "lr": 0.00904518046337755, + "lr": 0.00793913236883622, "name": "scale_1280", "betas": [ 0.9, @@ -304,7 +319,7 @@ ] }, { - "lr": 0.004522637977440181, + "lr": 0.003969669238105037, "name": "fusion", "betas": [ 0.9, @@ -360,7 +375,7 @@ "T_i": 10, "T_mult": 2, "eta_min": 1e-06, - "T_cur": 2, + "T_cur": 3, "base_lrs": [ 0.01, 0.01, @@ -369,25 +384,26 @@ 0.01, 0.005 ], - "last_epoch": 2, + "last_epoch": 3, "_step_count": 0, "_is_initial": false, "_get_lr_called_within_step": false, "_last_lr": [ - 0.00904518046337755, - 0.00904518046337755, - 0.00904518046337755, - 0.00904518046337755, - 0.00904518046337755, - 0.004522637977440181 + 0.00793913236883622, + 0.00793913236883622, + 0.00793913236883622, + 0.00793913236883622, + 0.00793913236883622, + 0.003969669238105037 ] }, "metrics": { - "best_val_acc": 80.924, - "best_epoch": 1, + "best_val_acc": 81.482, + "best_epoch": 2, "scale_accuracies": { - "384": 80.924, - "768": 80.638 + "384": 81.482, + "768": 81.574, + "1024": 81.072 } }, "train_config": {