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" @@ -1,226 +1,226 @@ { - "epoch": 4, + "epoch": 5, "optimizer_state_dict": { "state": { "0": { - "step": "tensor(6260.)", - "exp_avg": "tensor([[ 7.8277e-05, 4.9578e-06, 2.3352e-05, ..., -1.4571e-05,\n -4.8797e-05, 3.8431e-05],\n [ 8.8573e-07, 2.5922e-05, -5.2547e-05, ..., 2.1645e-05,\n -4.8459e-05, 3.4791e-05],\n [-4.6154e-08, 4.1890e-08, 1.3520e-08, ..., -2.1759e-08,\n 5.8983e-09, 6.1123e-09],\n ...,\n [-5.7400e-05, -1.7486e-05, -1.8404e-05, ..., -6.6152e-06,\n 1.5364e-05, -1.8938e-05],\n [-2.3509e-05, 4.3610e-05, -2.1072e-05, ..., -6.2180e-05,\n -1.5057e-05, 3.1203e-05],\n [ 1.3180e-05, -9.1399e-06, 2.6388e-05, ..., -1.5124e-05,\n -1.3041e-05, 3.6407e-05]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.3581e-08, 1.6482e-08, 7.6694e-09, ..., 9.3753e-09, 9.0039e-09,\n 5.7178e-09],\n [1.3218e-08, 1.1571e-08, 1.1668e-08, ..., 9.4941e-09, 7.0953e-09,\n 5.6897e-09],\n [7.1313e-13, 6.7727e-13, 5.5255e-13, ..., 8.2565e-13, 1.6906e-13,\n 4.4071e-13],\n ...,\n [1.3172e-08, 1.1464e-08, 9.6745e-09, ..., 7.9186e-09, 7.1973e-09,\n 5.6510e-09],\n [1.5851e-08, 1.3371e-08, 9.1543e-09, ..., 1.0205e-08, 8.3490e-09,\n 6.9020e-09],\n [3.4848e-09, 5.3274e-09, 3.8560e-09, ..., 2.3512e-09, 2.4853e-09,\n 2.4110e-09]], device='cuda:0')" + "step": "tensor(7512.)", + "exp_avg": "tensor([[-3.3985e-05, -4.9093e-05, 6.0957e-06, ..., 1.5105e-05,\n -1.4398e-05, -2.0035e-05],\n [-1.6476e-05, -1.9554e-05, -1.5610e-05, ..., 1.8453e-05,\n 3.4238e-05, -5.9351e-06],\n [ 4.7787e-09, 9.2635e-09, 3.0895e-09, ..., -1.6027e-09,\n 3.1980e-09, -3.7626e-09],\n ...,\n [-5.8897e-06, -9.3333e-06, -1.4377e-05, ..., 1.6443e-05,\n 2.3186e-05, 3.8037e-05],\n [ 3.4363e-05, -9.3591e-06, 8.2332e-06, ..., -4.1125e-05,\n 1.8069e-05, 9.5337e-06],\n [-1.5717e-05, -1.0244e-05, -1.6574e-06, ..., -6.2574e-06,\n -7.6530e-06, 5.0638e-06]], device='cuda:0')", + "exp_avg_sq": "tensor([[1.3585e-08, 1.5479e-08, 7.2596e-09, ..., 9.1910e-09, 8.2693e-09,\n 5.7905e-09],\n [1.2019e-08, 1.1002e-08, 1.2588e-08, ..., 8.9772e-09, 6.8231e-09,\n 5.6990e-09],\n [1.8747e-12, 1.1074e-12, 5.9346e-13, ..., 1.4549e-12, 1.3279e-12,\n 1.3741e-12],\n ...,\n [1.2980e-08, 1.0988e-08, 9.2983e-09, ..., 7.6484e-09, 6.6305e-09,\n 5.5716e-09],\n [1.5679e-08, 1.4097e-08, 1.0239e-08, ..., 1.1735e-08, 8.7324e-09,\n 6.9065e-09],\n [3.7470e-09, 5.5521e-09, 3.9332e-09, ..., 2.4848e-09, 2.7056e-09,\n 2.2699e-09]], device='cuda:0')" }, "1": { - "step": "tensor(6260.)", - "exp_avg": "tensor([ 2.4973e-03, -1.4545e-03, 8.9858e-07, ..., -3.1942e-04,\n 8.1322e-04, -1.3398e-03], device='cuda:0')", - "exp_avg_sq": "tensor([1.7700e-05, 1.6147e-05, 1.1544e-09, ..., 1.6169e-05, 1.7467e-05,\n 6.0526e-06], device='cuda:0')" + "step": "tensor(7512.)", + "exp_avg": "tensor([-1.4362e-03, -6.5555e-04, -2.2324e-07, ..., 1.0182e-04,\n -5.1545e-04, -1.8445e-04], device='cuda:0')", + "exp_avg_sq": "tensor([1.7014e-05, 1.5400e-05, 2.2717e-09, ..., 1.6007e-05, 1.8350e-05,\n 5.9587e-06], device='cuda:0')" }, "2": { - "step": "tensor(6260.)", - "exp_avg": "tensor([[-1.9668e-06, 1.8664e-06, 5.6052e-45, ..., -9.4905e-06,\n -2.9698e-07, 1.2132e-06],\n [ 5.6052e-45, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [-7.2952e-08, -2.9029e-06, 0.0000e+00, ..., -9.9339e-07,\n -8.3676e-06, -1.6951e-08],\n ...,\n [ 0.0000e+00, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 0.0000e+00],\n [ 2.0413e-06, -9.1185e-07, -5.6052e-45, ..., 4.2819e-06,\n -1.9084e-06, 4.0816e-06],\n [ 9.7527e-07, 2.1144e-06, -5.6052e-45, ..., -4.1452e-07,\n 1.3256e-05, -5.2286e-08]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.7336e-09, 4.6745e-10, 2.2801e-13, ..., 2.4491e-09, 1.2078e-09,\n 3.7316e-10],\n [2.4385e-13, 4.4660e-12, 0.0000e+00, ..., 8.3785e-13, 3.0998e-16,\n 3.0858e-12],\n [3.2643e-10, 5.8905e-10, 0.0000e+00, ..., 4.1421e-10, 1.6597e-09,\n 3.9613e-11],\n ...,\n [0.0000e+00, 1.0455e-18, 0.0000e+00, ..., 3.0629e-20, 3.6581e-20,\n 0.0000e+00],\n [5.6132e-09, 9.2022e-10, 3.2857e-14, ..., 2.3175e-09, 6.9180e-10,\n 1.4418e-09],\n [1.7004e-09, 1.2668e-09, 1.7515e-13, ..., 6.6202e-10, 5.6460e-09,\n 6.5041e-10]], device='cuda:0')" + "step": "tensor(7512.)", + "exp_avg": "tensor([[ 2.1128e-07, -4.1371e-06, 1.6816e-44, ..., -2.9873e-06,\n -3.0646e-06, -6.9167e-07],\n [-6.0517e-27, 5.6052e-45, 0.0000e+00, ..., -1.0318e-11,\n -1.2754e-14, 5.6052e-45],\n [ 2.6595e-06, -2.8459e-06, 0.0000e+00, ..., 5.5072e-07,\n -3.5839e-07, 2.7139e-07],\n ...,\n [ 0.0000e+00, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 0.0000e+00],\n [ 1.7043e-05, 1.4572e-06, -5.6052e-45, ..., -1.2960e-05,\n 4.1751e-06, 2.0426e-06],\n [ 5.8075e-06, -2.8900e-07, -5.6052e-45, ..., 4.2279e-07,\n -1.9386e-07, 6.5403e-07]], device='cuda:0')", + "exp_avg_sq": "tensor([[1.3425e-09, 3.0900e-10, 6.5172e-14, ..., 1.8708e-09, 8.8117e-10,\n 4.1461e-10],\n [6.9998e-14, 1.2762e-12, 0.0000e+00, ..., 2.3942e-13, 9.0757e-17,\n 8.8180e-13],\n [2.2332e-10, 6.8630e-10, 0.0000e+00, ..., 9.4143e-10, 1.9582e-09,\n 3.5520e-11],\n ...,\n [0.0000e+00, 2.9877e-19, 0.0000e+00, ..., 8.7525e-21, 1.0453e-20,\n 0.0000e+00],\n [6.9411e-09, 7.0666e-10, 9.3892e-15, ..., 2.2470e-09, 1.1550e-09,\n 1.1908e-09],\n [1.2712e-09, 1.1782e-09, 5.0051e-14, ..., 7.7378e-10, 6.2121e-09,\n 3.0098e-10]], device='cuda:0')" }, "3": { - "step": "tensor(6260.)", - "exp_avg": "tensor([ 3.1521e-05, -1.2311e-23, 1.9869e-04, -7.6081e-05, 9.2364e-05,\n 3.6461e-21, 1.5613e-04, -2.0859e-04, -3.8042e-04, 3.2588e-04,\n 1.8848e-04, 1.0349e-36, 9.5781e-05, 2.4403e-05, 7.2596e-05,\n 2.0570e-04, 1.0269e-04, -2.7391e-05, 9.5326e-06, 5.0980e-05,\n -5.2700e-11, 8.5553e-05, -1.4996e-06, -3.7008e-06, -9.6250e-06,\n 9.8987e-05, 7.0346e-05, 5.6052e-45, -8.5273e-05, -1.5279e-05,\n 4.9009e-05, -3.8952e-06, -9.9135e-05, -3.3720e-06, 2.0738e-04,\n 5.9665e-05, 2.4778e-12, 2.9097e-05, -2.2584e-04, -1.2372e-04,\n -2.7847e-04, 7.7207e-05, -1.4023e-04, -1.9990e-04, 3.2947e-05,\n 9.3529e-05, -5.7796e-04, -8.3135e-05, -1.1395e-04, 1.1283e-05,\n 5.6052e-45, 4.5786e-06, 6.0989e-05, 7.4132e-05, 1.6638e-05,\n 1.0488e-04, 3.6284e-06, -4.1525e-05, 6.5765e-05, 3.4452e-05,\n -1.2250e-04, -1.6437e-04, -1.7052e-04, -1.3674e-04, -2.7126e-05,\n 2.8745e-05, 2.6645e-05, 2.6426e-04, -9.0528e-05, -2.3231e-05,\n 1.5187e-04, -9.4069e-27, 1.4530e-04, -1.8770e-05, 3.1416e-05,\n 3.7059e-06, 5.6052e-45, 5.8168e-04, -1.5514e-04, -2.7781e-05,\n -6.9010e-05, -1.8606e-04, -5.1562e-05, -1.3361e-33, -4.8092e-05,\n 4.0571e-05, 3.9434e-05, -3.3354e-04, -6.4586e-05, -6.3396e-05,\n 1.4266e-05, 3.2508e-05, -7.4531e-06, 2.6662e-05, 6.6810e-05,\n -3.4404e-04, -4.5648e-05, -1.1256e-05, -5.0714e-04, 1.7754e-04,\n 4.2586e-05, -9.2744e-05, 2.0982e-04, 3.1810e-05, 1.5493e-04,\n -4.9194e-04, -9.5192e-06, -4.9092e-05, 1.8848e-04, 1.5955e-04,\n -1.8737e-04, -1.4154e-04, 1.8291e-05, -5.5272e-06, 2.6245e-05,\n -2.0715e-04, -4.3237e-04, 1.2184e-04, -3.0999e-05, 5.6052e-45,\n -7.0565e-06, -1.2552e-04, -6.1019e-06, 5.6052e-45, -7.7955e-05,\n -1.7411e-04, 2.6180e-05, -2.3926e-14, -1.5691e-05, -2.9317e-05,\n -7.4496e-05, -2.7922e-05, 5.6052e-45, -4.2639e-05, -6.4004e-08,\n 1.6341e-04, 1.2682e-36, 4.4140e-05, -4.8538e-05, 5.6490e-05,\n 5.6052e-45, 5.6052e-45, -1.1404e-04, -1.7011e-09, 5.6052e-45,\n 5.6052e-45, 1.4518e-05, -1.0252e-04, 5.6052e-45, 7.7463e-05,\n -2.3733e-05, -9.5883e-06, 4.6006e-05, 1.8538e-05, 4.7356e-05,\n -8.8006e-06, 1.7081e-05, 7.8186e-05, 1.9898e-17, -1.0368e-04,\n -4.6848e-05, 8.1403e-05, 1.4906e-04, 3.2709e-05, 9.1180e-07,\n 1.8322e-04, -1.4427e-08, 5.6052e-45, 3.2181e-05, -1.4238e-04,\n 2.7206e-04, 3.6773e-04, 9.1961e-05, -8.8061e-05, 2.8030e-04,\n 2.4392e-05, -1.8160e-05, 2.8750e-05, 3.0516e-04, -2.0983e-04,\n 1.3941e-04, 7.4210e-05, -5.1194e-05, -5.7053e-05, 5.6052e-45,\n 1.2510e-40, 1.5681e-04, 1.4842e-04, 3.7542e-05, 9.3564e-05,\n 2.6983e-05, -7.0105e-05, -8.6893e-05, 5.4492e-05, -6.9348e-05,\n -1.6202e-05, 2.3960e-04, 1.0463e-04, -3.4105e-05, 5.6052e-45,\n 1.9414e-05, 6.7005e-05, 5.8286e-05, 1.7512e-04, 1.6003e-04,\n 1.6040e-05, -3.8507e-04, 4.6186e-05, -8.6852e-07, 5.6052e-45,\n -1.1705e-04, 6.1573e-28, -6.7348e-05, 9.9849e-05, -1.7887e-04,\n 8.5411e-05, -3.1872e-04, 5.7193e-05, -8.4574e-05, 1.1582e-05,\n -7.0984e-05, -4.8982e-05, -2.1179e-05, 6.4510e-05, 1.3626e-04,\n -5.6052e-45, -1.6633e-04, 1.4200e-04, 4.6956e-05, 2.9281e-05,\n -5.5499e-05, 6.6239e-06, -1.2351e-04, -1.0753e-04, 1.0038e-30,\n 2.1582e-04, 3.4709e-04, 5.6052e-45, 1.4857e-04, 1.9096e-04,\n 4.9188e-05, 5.6052e-45, 1.7566e-04, 8.0078e-05, -3.0343e-05,\n 1.8349e-04, -4.2999e-05, -2.2348e-04, -1.4158e-04, 9.6752e-36,\n -1.9542e-04, -2.5392e-04, -1.0916e-04, -1.3869e-04, -2.5880e-05,\n 1.1429e-04, 4.1407e-40, 9.7374e-05, 1.9185e-06, 1.3247e-04,\n 5.6052e-45, 3.0778e-04, 3.9682e-05, 3.3542e-04, -1.8589e-05,\n -4.2212e-05, -1.8472e-04, 9.9252e-06, 3.6182e-04, -4.3417e-05,\n -4.5407e-05, 3.5957e-05, 2.6875e-05, -8.2606e-05, -2.1872e-04,\n -1.3050e-04, -4.2799e-05, 1.2146e-11, 6.1240e-05, -9.0383e-05,\n 1.6296e-23, 1.2110e-05, 4.9330e-05, 9.5203e-05, 1.5581e-04,\n -1.1960e-04, -5.1113e-06, 9.0011e-05, -9.0278e-05, 2.4280e-05,\n 5.2265e-05, -4.0891e-05, 2.2373e-05, -3.3726e-05, -2.3673e-04,\n 4.1865e-05, -2.5736e-04, -1.1135e-04, 2.0706e-04, 7.1923e-05,\n -8.1276e-05, -1.7075e-04, 3.2565e-05, -2.4586e-05, -2.6026e-05,\n 5.6052e-45, -2.0990e-09, -1.1199e-04, -1.8375e-04, 5.6440e-05,\n -1.1075e-04, 5.6052e-45, 1.2897e-04, 1.0463e-04, 3.4039e-04,\n 5.6052e-45, 8.2677e-05, 2.9165e-05, 8.2419e-06, -1.9874e-05,\n 5.6985e-05, -2.3537e-06, 2.9127e-05, -6.5033e-05, -6.7711e-05,\n 5.6052e-45, -2.3083e-04, 1.8713e-04, 2.6040e-04, 1.0097e-04,\n 1.4297e-04, -4.5225e-05, 8.7781e-05, 6.1370e-05, 4.8244e-05,\n -5.0370e-05, -2.7739e-05, -8.9205e-06, 5.6052e-45, -5.1096e-08,\n -1.4312e-04, 1.5349e-04, -1.2869e-05, 6.4124e-05, 8.5775e-07,\n -4.4252e-05, 1.0077e-04, 9.8495e-05, 1.6101e-05, 1.3111e-04,\n -1.2295e-05, -3.1138e-05, 2.0143e-06, 5.6052e-45, -5.1130e-05,\n -1.6264e-04, -8.6253e-06, -6.4645e-05, 2.3816e-05, -1.9372e-05,\n 1.7761e-04, 1.8192e-04, 5.6052e-45, 7.8429e-05, -5.5876e-05,\n -1.8691e-04, -1.0389e-06, -2.8513e-04, 3.8670e-05, -1.4340e-06,\n 4.6800e-05, 5.1832e-05, 4.5576e-06, 1.4786e-05, 5.1367e-05,\n 8.2881e-05, -1.5044e-18, -1.1618e-04, 5.6052e-45, 1.6986e-04,\n 5.3990e-05, 2.2211e-04, -1.2810e-05, 1.3811e-04, 5.6052e-45,\n -8.7641e-05, 8.9677e-06, 4.6456e-05, 1.3708e-04, 9.9589e-05,\n 1.8055e-05, 1.0672e-04, -1.6417e-04, -2.4381e-05, 6.0763e-05,\n 8.5640e-05, 8.0096e-05, -5.8416e-14, 5.6052e-45, 7.8898e-05,\n -4.5476e-05, -1.2014e-04, -3.8826e-04, 9.1430e-05, 6.2766e-05,\n 1.1925e-04, -6.9678e-31, -2.9612e-05, 1.4610e-04, 1.6963e-04,\n -2.6295e-04, 5.6052e-45, 4.7689e-05, -1.7575e-04, 3.6474e-05,\n 5.3510e-05, -5.1009e-06, 9.4082e-05, 1.4956e-04, 6.5515e-05,\n -4.2547e-05, -1.0017e-04, 3.7764e-05, 1.7355e-28, 5.6052e-45,\n -2.5037e-05, 3.7096e-05, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.5992e-05, 2.3257e-06, 3.9828e-05, -4.1754e-05, 3.3585e-05,\n 5.6052e-45, 8.1778e-06, -3.5955e-06, 2.8306e-05, 5.6052e-45,\n -8.1166e-05, 1.1623e-04, -4.8540e-05, 5.6052e-45, -2.2035e-05,\n -3.7871e-05, 1.2515e-04, 9.4850e-05, -5.8377e-05, -5.0248e-05,\n -7.7020e-04, -4.0442e-05, 3.1163e-04, 5.8285e-05, 3.5595e-05,\n -1.1196e-04, 8.8795e-09, -2.3106e-04, 4.9203e-19, 7.4549e-06,\n -2.9382e-05, 1.2463e-04, 5.6052e-45, 6.8643e-05, 5.7257e-06,\n 2.4418e-04, 4.5223e-05, 4.7212e-05, -1.0576e-04, 9.7705e-05,\n -1.0929e-04, 5.6052e-45, 6.1244e-05, 2.1550e-05, -7.7412e-05,\n 5.6052e-45, -1.7830e-04, 1.4335e-04, -1.1816e-04, 1.2455e-04,\n 3.5963e-04, -5.2330e-05, 2.0618e-04, -2.8834e-04, 1.4231e-05,\n 9.5560e-05, -2.3463e-04, 1.3726e-04, -1.9363e-04, -1.1542e-04,\n 5.9719e-05, 1.0063e-04, -1.3939e-05, -5.4109e-05, 5.1683e-05,\n 1.1323e-04, 1.1217e-04, 8.4128e-05, -1.4960e-04, 1.0453e-04,\n 5.6052e-45, 5.4710e-05, -1.4260e-05, 4.4808e-05, 4.0193e-06,\n 6.7333e-05, -1.0266e-04, 2.2544e-04, -2.5572e-04, 1.5214e-04,\n -1.0401e-04, -9.3180e-05, -1.8644e-04, -1.3265e-05, 5.6218e-05,\n -6.8698e-05, -6.4009e-05, -7.5259e-05, 5.6052e-45, 5.5681e-05,\n -6.3020e-05, 7.3126e-05, -2.7347e-05, -2.0996e-04, 7.8351e-05,\n -7.5844e-05, 5.6052e-45, -2.9850e-05, -4.5902e-05, -9.4897e-05,\n 1.4031e-04, -1.8612e-04, 1.1664e-04, -1.2641e-05, -4.4997e-06,\n 1.2655e-04, 1.8093e-05, 5.6052e-45, 2.4961e-05, -4.5604e-04,\n -1.1351e-04, -2.6668e-04, 4.3833e-05, -2.9797e-05, -1.0482e-04,\n -1.7394e-18, -1.0907e-04, -4.1047e-05, 2.0687e-04, 6.5075e-05,\n -1.1219e-04, -6.0677e-05, -6.6832e-05, -9.9832e-05, 1.9359e-04,\n 4.2535e-05, -1.1121e-04, 1.2293e-04, 6.2709e-05, 1.7709e-05,\n 1.2128e-11, -3.2536e-05, 3.2604e-05, -3.6219e-05, 1.8014e-04,\n 1.2130e-04, 7.4951e-05, 5.6052e-45, 1.2233e-04, 5.6052e-45,\n 9.9826e-06, 3.7471e-05, 5.0985e-05, 9.5499e-31, 1.5749e-04,\n -1.2691e-04, -2.4912e-05, 1.2256e-05, -2.9885e-04, -5.6501e-05,\n 1.6152e-04, 5.6052e-45, 4.7886e-05, -3.1392e-05, 5.6052e-45,\n -3.0492e-04, 1.2449e-05, 1.7495e-04, -1.2557e-04, -6.6562e-05,\n -1.2466e-04, 5.3562e-41, 7.5010e-05, 1.0619e-04, -6.9164e-06,\n 7.9408e-05, 2.1803e-05, 2.1552e-05, -5.3166e-05, -1.1018e-05,\n -1.0263e-04, 6.3220e-05, -4.5029e-04, -1.7432e-05, -3.5489e-04,\n 1.9478e-04, -3.9741e-05, -3.7136e-06, 2.1685e-04, 2.1252e-05,\n 3.0014e-04, 1.7813e-04, 1.5587e-04, 2.4100e-19, -1.4856e-05,\n 5.6052e-45, 2.1316e-04, -2.5933e-35, -6.4867e-05, 8.8672e-05,\n 2.7863e-05, -1.9007e-04, -1.1425e-05, 1.5760e-04, 3.1228e-04,\n -2.7241e-04, 5.6052e-45, 6.6249e-05, 1.4112e-05, 1.8199e-05,\n 9.0675e-06, 5.6052e-45, 5.6052e-45, 5.6052e-45, 1.0783e-04,\n -2.0323e-05, 5.6052e-45, 1.3441e-04, 2.0043e-04, 6.8588e-05,\n 7.1324e-05, 5.6052e-45, -2.0448e-04, -6.6880e-05, 1.1991e-05,\n -2.2671e-04, -2.2429e-04, -4.4375e-04, -2.0082e-06, -1.5322e-05,\n 8.8474e-05, -1.5098e-05, -9.4342e-05, -9.2000e-05, 1.1421e-04,\n 8.1657e-05, 9.4784e-05, -2.4266e-05, -1.2188e-05, -7.4300e-06,\n -4.2374e-04, -3.5490e-05, -2.5278e-05, 2.5053e-05, -3.2132e-04,\n 3.1103e-04, -4.0482e-05, -1.7681e-05, 1.2131e-05, -1.0499e-17,\n 1.4120e-04, 5.6052e-45, -2.2712e-04, -2.8662e-04, -2.7610e-04,\n 1.4849e-05, 1.9036e-04, -5.6577e-08, -5.6052e-45, 5.6052e-45,\n 3.5628e-05, -2.1885e-04, -1.8366e-04, 5.6052e-45, -2.8512e-06,\n 8.1273e-27, 1.3676e-04, 1.9340e-04, -6.2535e-05, -2.0510e-05,\n -3.6166e-05, -1.4578e-04, 5.6052e-45, -8.7900e-05, -4.6426e-05,\n -1.0444e-04, -5.7292e-05, -1.2850e-05, 4.7803e-37, -2.4854e-05,\n -1.0326e-04, -1.1522e-04, -1.5457e-06, 1.5733e-04, -3.5736e-05,\n 4.3578e-06, 5.6052e-45, 3.1331e-04, -1.4163e-05, 5.6052e-45,\n 1.8118e-05, -1.6546e-04, -1.0439e-04, 6.6036e-05, 9.8913e-05,\n 5.7397e-05, 8.3186e-05, 6.0364e-05, -1.2176e-04, 1.7343e-04,\n 5.6052e-45, -2.2644e-04, -9.2222e-05, 8.9380e-05, -1.3487e-04,\n 5.0878e-05, -7.2112e-05, 7.9752e-05, 3.3456e-05, -3.7188e-05,\n 2.7403e-05, 2.0137e-04, 1.7045e-04, 1.6360e-04, 5.6052e-45,\n 4.5360e-05, 5.2170e-04, -2.3817e-04, -6.5981e-05, -4.7639e-05,\n 8.2872e-05, 2.8165e-05, -6.6102e-06, -3.9644e-05, 4.8232e-05,\n 6.3479e-05, -8.7849e-05, 6.2002e-13, 8.0774e-05, -5.6052e-45,\n -1.1994e-04, -5.1305e-05, 7.6211e-05, 2.3491e-05, 2.3746e-04,\n 4.0031e-06, -3.8844e-05, 2.6814e-04, -9.4837e-06, -2.1366e-04,\n -3.0659e-05, -9.7009e-05, 1.2654e-04, -9.0477e-05, 5.6052e-45,\n 5.6052e-45, -2.6142e-04, 8.2354e-05], device='cuda:0')", - "exp_avg_sq": "tensor([1.9787e-07, 5.4960e-09, 1.1300e-07, 1.6804e-07, 4.0614e-07, 5.1216e-08,\n 4.6980e-07, 3.4866e-07, 1.7439e-07, 4.0015e-07, 2.8661e-07, 2.8287e-08,\n 2.2284e-07, 1.1822e-07, 2.2500e-07, 2.4259e-07, 4.1483e-07, 6.1545e-08,\n 9.5374e-08, 2.7792e-07, 6.4184e-10, 1.1385e-07, 1.5269e-07, 2.8641e-07,\n 4.0428e-07, 3.1632e-07, 1.2460e-07, 1.1631e-10, 2.8061e-07, 2.8552e-07,\n 1.0643e-07, 2.1704e-07, 2.0715e-07, 2.9785e-07, 1.7895e-07, 1.7890e-07,\n 6.2461e-08, 2.7321e-07, 2.5917e-07, 2.4279e-07, 4.0728e-07, 1.9662e-07,\n 3.7759e-07, 3.2161e-07, 3.0701e-07, 1.6605e-07, 2.5025e-07, 2.1285e-07,\n 6.8579e-07, 2.5682e-07, 9.9758e-08, 1.1159e-06, 2.2410e-07, 1.8671e-07,\n 2.5009e-07, 1.5559e-07, 2.9956e-07, 1.1061e-07, 1.0357e-07, 2.4951e-07,\n 7.8470e-08, 1.8998e-07, 4.8784e-07, 2.3819e-07, 8.7685e-08, 1.6258e-07,\n 3.9505e-07, 2.9135e-07, 6.3024e-08, 1.7952e-08, 2.5887e-07, 4.5384e-08,\n 1.6177e-07, 6.0441e-07, 2.4969e-07, 2.8148e-07, 3.1500e-10, 4.1651e-07,\n 1.4168e-07, 1.9136e-07, 1.7970e-07, 2.5519e-07, 3.7545e-07, 8.6969e-10,\n 4.1920e-07, 3.2175e-07, 2.9942e-07, 4.1721e-07, 2.1455e-07, 4.0023e-07,\n 2.3185e-07, 1.6560e-07, 1.7416e-07, 2.8483e-07, 1.4649e-07, 4.3482e-07,\n 3.2429e-07, 4.0879e-07, 2.6879e-07, 3.0304e-07, 2.8274e-07, 3.0158e-07,\n 6.6636e-07, 4.3836e-07, 2.4077e-07, 3.1853e-07, 1.7897e-07, 2.2143e-07,\n 1.9495e-07, 1.1852e-07, 2.2477e-07, 4.4604e-07, 2.4669e-07, 7.5318e-08,\n 4.2339e-07, 2.7609e-07, 2.2352e-07, 1.4964e-07, 1.1498e-07, 2.3810e-08,\n 2.5876e-07, 4.5794e-07, 1.5655e-07, 5.9664e-08, 2.2381e-07, 3.1641e-07,\n 2.5708e-07, 1.9063e-09, 3.5785e-07, 2.9290e-07, 2.3824e-07, 1.9994e-07,\n 2.6063e-09, 2.8204e-07, 2.8045e-08, 2.2149e-07, 1.0561e-11, 2.9762e-07,\n 2.1885e-07, 8.3143e-08, 2.6277e-08, 1.4450e-08, 1.6451e-07, 4.1534e-12,\n 5.5984e-11, 7.4466e-08, 3.3294e-07, 3.0760e-07, 2.0007e-07, 1.2725e-07,\n 1.8955e-07, 3.1000e-07, 1.6888e-07, 2.2580e-07, 2.4992e-08, 2.7353e-07,\n 1.2937e-07, 2.6569e-07, 8.0852e-08, 2.2294e-07, 2.7319e-07, 1.6697e-07,\n 3.2299e-07, 4.1591e-07, 1.7182e-07, 2.3193e-07, 3.1699e-08, 1.0526e-13,\n 2.4305e-07, 1.8340e-07, 3.5148e-07, 3.2496e-07, 2.3349e-07, 2.9714e-07,\n 2.8921e-07, 2.4326e-07, 6.3664e-08, 2.9252e-07, 3.6861e-07, 4.8608e-07,\n 4.1012e-07, 2.6704e-07, 2.9031e-07, 8.4731e-08, 4.4213e-12, 6.1502e-17,\n 4.0849e-07, 2.4099e-07, 8.6675e-08, 8.9670e-08, 1.8568e-07, 1.7040e-07,\n 3.1860e-07, 4.0422e-07, 3.1660e-07, 1.5315e-07, 3.8357e-07, 3.3134e-07,\n 2.2169e-07, 6.7367e-19, 1.1606e-07, 1.4955e-07, 3.5342e-07, 3.3610e-07,\n 4.2231e-07, 1.9406e-07, 3.4893e-07, 2.1355e-07, 3.3644e-08, 5.0495e-10,\n 4.1043e-07, 7.0846e-10, 1.5706e-07, 4.4913e-07, 2.6942e-07, 2.5480e-07,\n 2.8717e-07, 1.4769e-07, 2.6766e-07, 2.1021e-07, 1.0581e-07, 3.8328e-07,\n 3.3786e-07, 2.0272e-07, 1.9837e-07, 9.9742e-08, 3.9717e-07, 4.3938e-07,\n 3.9772e-07, 3.0652e-08, 2.3608e-07, 2.4613e-07, 2.4020e-07, 4.3484e-07,\n 6.3114e-08, 3.4000e-07, 4.3882e-07, 1.1968e-11, 1.8443e-07, 2.7445e-07,\n 7.3912e-08, 3.2830e-08, 3.6227e-07, 2.3162e-07, 1.9343e-07, 2.6808e-07,\n 2.6612e-07, 3.7563e-07, 2.0068e-07, 4.1878e-08, 4.2593e-07, 5.0508e-07,\n 2.9274e-07, 3.9036e-07, 2.0392e-07, 3.9488e-07, 1.2804e-11, 1.4010e-07,\n 3.2694e-08, 5.5286e-07, 7.5396e-17, 3.0049e-07, 2.5423e-07, 1.8087e-07,\n 4.6354e-07, 2.0251e-07, 1.9328e-07, 1.8850e-07, 3.8694e-07, 2.1267e-07,\n 3.3415e-07, 2.4537e-07, 1.8378e-07, 3.1171e-07, 4.4294e-07, 2.1939e-07,\n 3.5608e-07, 6.5279e-11, 1.0816e-07, 4.7690e-07, 1.8614e-07, 4.9006e-07,\n 1.6984e-07, 2.8415e-07, 3.9908e-07, 1.7342e-07, 3.7931e-07, 3.0388e-07,\n 1.1356e-07, 4.8912e-07, 1.6712e-07, 2.7119e-07, 2.7078e-07, 1.1328e-07,\n 3.2003e-07, 3.2696e-07, 2.2132e-07, 4.2784e-07, 2.1075e-07, 9.7866e-08,\n 3.1647e-07, 4.4262e-07, 1.1758e-07, 1.2251e-07, 8.0221e-08, 1.5031e-10,\n 1.8498e-09, 1.2419e-07, 2.7385e-07, 2.9803e-07, 2.8875e-07, 9.7804e-12,\n 2.4710e-07, 1.7429e-07, 2.8125e-07, 2.1467e-11, 4.8201e-08, 6.4668e-07,\n 1.2869e-07, 1.4276e-07, 2.8334e-07, 6.6329e-08, 1.0736e-07, 2.8329e-07,\n 1.4484e-07, 7.7883e-08, 3.6745e-07, 2.8226e-07, 2.5990e-07, 2.6925e-07,\n 3.1118e-07, 1.3543e-07, 2.8249e-07, 1.7772e-07, 1.0112e-07, 4.9530e-07,\n 2.1232e-07, 1.0651e-07, 5.8496e-17, 3.0583e-10, 5.2431e-07, 2.2759e-07,\n 4.9460e-07, 1.0667e-07, 4.1261e-07, 2.1008e-07, 2.6537e-07, 6.6856e-07,\n 2.7858e-07, 3.2998e-07, 3.2988e-07, 2.8305e-07, 1.0520e-07, 1.5925e-08,\n 2.0489e-07, 3.9704e-07, 7.3340e-08, 4.6286e-08, 2.1087e-07, 4.2721e-07,\n 2.6884e-07, 2.6139e-07, 6.6044e-08, 4.0949e-07, 1.9711e-07, 7.6486e-08,\n 1.4768e-07, 3.6390e-07, 1.7253e-07, 9.1067e-08, 1.5973e-07, 5.7074e-08,\n 1.6770e-07, 1.8363e-07, 2.1989e-07, 2.0705e-07, 1.1498e-08, 6.0450e-08,\n 3.1357e-08, 1.4296e-07, 4.5719e-07, 3.2743e-07, 1.3067e-07, 2.8867e-07,\n 6.5530e-08, 3.4019e-07, 2.5523e-08, 2.9716e-07, 5.2069e-07, 1.5840e-07,\n 2.9442e-07, 3.4228e-07, 1.7393e-07, 8.7288e-08, 3.4933e-07, 4.2848e-07,\n 4.7862e-08, 2.7427e-08, 2.2537e-07, 7.1154e-08, 2.7043e-07, 3.2720e-07,\n 3.6741e-07, 5.1447e-07, 2.8477e-07, 3.1218e-07, 1.9029e-07, 3.8031e-07,\n 2.9843e-07, 3.6367e-07, 3.2553e-07, 4.0569e-13, 2.0285e-07, 2.8167e-07,\n 9.3794e-08, 2.2244e-07, 9.6574e-11, 3.5288e-07, 2.2628e-07, 1.8566e-07,\n 2.5817e-07, 3.7823e-07, 3.7120e-07, 2.7439e-11, 1.7998e-18, 1.3752e-07,\n 1.3430e-07, 1.5136e-08, 4.1611e-11, 8.7024e-11, 3.0947e-07, 2.2442e-07,\n 2.5766e-07, 2.1140e-07, 4.1888e-07, 9.6453e-09, 3.1538e-07, 2.1064e-07,\n 3.0039e-07, 2.2691e-10, 1.1761e-07, 2.0779e-07, 1.7273e-07, 2.7516e-10,\n 1.4389e-07, 4.6454e-07, 1.8539e-07, 2.1644e-07, 9.7542e-08, 9.3139e-08,\n 4.5987e-07, 1.4690e-07, 2.8715e-07, 2.8575e-07, 3.3087e-07, 2.8683e-07,\n 1.5031e-08, 5.1000e-07, 8.5275e-09, 6.4609e-08, 1.4915e-07, 1.7750e-07,\n 2.0776e-12, 2.5730e-07, 1.8606e-08, 2.5418e-07, 1.4350e-07, 1.4683e-07,\n 3.2795e-07, 1.3224e-07, 1.3860e-07, 8.4104e-10, 4.8989e-08, 3.5556e-07,\n 9.8230e-08, 3.7350e-11, 2.1002e-07, 2.1275e-07, 2.4015e-07, 4.3612e-07,\n 2.7251e-07, 3.6574e-07, 2.8081e-07, 4.2960e-07, 2.4329e-07, 2.5733e-07,\n 4.6028e-07, 3.6361e-07, 2.7757e-07, 2.4876e-07, 2.6837e-07, 1.1946e-07,\n 3.8260e-07, 1.8528e-07, 2.7602e-07, 1.5977e-07, 2.6357e-07, 3.0740e-07,\n 3.3309e-07, 1.2267e-07, 4.5022e-07, 5.1684e-07, 2.0089e-07, 1.6249e-07,\n 3.2634e-07, 1.4522e-07, 2.8576e-07, 3.2658e-07, 2.3547e-07, 3.4754e-07,\n 2.0625e-07, 3.2378e-07, 4.1414e-07, 3.8249e-07, 1.2767e-07, 5.3386e-08,\n 1.8244e-07, 1.3534e-07, 6.2360e-08, 3.2263e-07, 2.4196e-07, 1.1394e-07,\n 2.4923e-07, 2.3806e-07, 1.9702e-07, 2.6968e-07, 1.7869e-12, 3.7659e-07,\n 1.8706e-07, 2.2615e-07, 3.1584e-07, 4.0406e-07, 2.8103e-07, 3.2139e-07,\n 2.0145e-07, 5.7128e-08, 2.7118e-07, 2.0823e-07, 1.6825e-07, 2.7683e-07,\n 3.9637e-07, 3.0334e-07, 4.4156e-07, 5.2473e-07, 5.2753e-07, 7.4925e-10,\n 2.3900e-07, 2.9518e-07, 3.5120e-07, 3.4552e-07, 4.1989e-07, 3.6406e-07,\n 2.5464e-07, 2.6091e-07, 1.9751e-07, 4.8832e-07, 2.3650e-07, 3.2886e-07,\n 1.7686e-07, 2.6034e-07, 1.4182e-11, 4.1228e-07, 6.4464e-08, 1.8793e-07,\n 2.7504e-07, 2.5095e-07, 1.6190e-07, 2.9690e-10, 3.5443e-07, 1.4481e-08,\n 2.5138e-07, 1.3949e-07, 2.7057e-07, 1.2396e-08, 1.9369e-07, 3.1879e-07,\n 2.9602e-07, 8.4515e-08, 1.9143e-07, 2.0396e-07, 3.3760e-07, 3.7658e-11,\n 1.1979e-07, 9.7749e-08, 4.9991e-15, 3.7056e-07, 2.7733e-07, 2.6132e-07,\n 1.9219e-07, 1.8769e-07, 2.0875e-07, 5.8114e-12, 2.0218e-07, 2.8349e-07,\n 1.5433e-07, 1.7063e-07, 1.3770e-07, 2.0156e-07, 1.6915e-07, 1.5519e-07,\n 2.6530e-07, 1.9047e-07, 3.4263e-07, 1.8360e-07, 2.3873e-07, 3.8015e-07,\n 3.5195e-07, 4.6418e-08, 2.7560e-07, 2.2705e-07, 3.1265e-07, 2.1213e-07,\n 2.2522e-07, 1.0381e-07, 3.1899e-07, 2.6277e-11, 2.9310e-07, 3.0489e-08,\n 2.5398e-07, 4.7390e-07, 3.9462e-07, 3.5780e-07, 4.4137e-07, 4.4500e-07,\n 2.2057e-07, 2.8115e-07, 2.6071e-12, 2.9250e-07, 2.7853e-07, 5.4884e-08,\n 8.6325e-08, 9.8180e-09, 6.8446e-10, 2.1333e-11, 2.2386e-07, 4.4458e-07,\n 8.0758e-14, 1.9700e-07, 2.9701e-07, 3.4827e-07, 4.6211e-07, 8.6561e-12,\n 3.3383e-07, 2.4823e-07, 8.7855e-08, 2.2781e-07, 2.4956e-07, 2.6614e-07,\n 4.8229e-07, 1.1767e-07, 2.5167e-07, 2.5303e-07, 3.5467e-07, 3.1926e-07,\n 1.8272e-07, 2.1954e-07, 4.5158e-07, 1.1523e-07, 3.1622e-07, 9.1926e-09,\n 6.0841e-07, 2.4912e-07, 3.9583e-07, 2.9193e-07, 1.7358e-07, 5.4395e-07,\n 3.4455e-07, 1.0588e-07, 3.2158e-07, 3.1075e-09, 2.2421e-07, 7.6082e-13,\n 1.4163e-07, 1.9026e-07, 2.2687e-07, 2.6986e-07, 4.8505e-07, 3.8383e-09,\n 3.4211e-08, 8.1914e-11, 4.3863e-07, 2.9797e-07, 2.9739e-07, 1.9379e-07,\n 2.5495e-07, 3.5784e-08, 3.6520e-07, 1.6319e-07, 2.2808e-07, 3.7866e-07,\n 3.4859e-08, 1.1638e-07, 2.9372e-07, 1.0222e-07, 2.6341e-08, 1.7009e-07,\n 3.9296e-07, 2.4202e-07, 2.3250e-11, 1.9694e-07, 2.9199e-07, 2.6650e-07,\n 3.2000e-07, 3.9770e-07, 3.7168e-07, 2.1164e-07, 1.8510e-10, 8.0523e-07,\n 3.5398e-07, 1.4589e-12, 3.5360e-07, 1.7886e-07, 1.3013e-07, 3.3589e-07,\n 3.0434e-07, 1.0719e-07, 2.5082e-07, 2.9619e-07, 1.5261e-07, 3.7090e-07,\n 1.3019e-07, 5.0792e-07, 2.3075e-07, 3.0127e-07, 2.9691e-07, 1.0198e-07,\n 2.2873e-07, 4.0267e-07, 5.7454e-07, 4.4278e-07, 1.1760e-07, 4.0341e-07,\n 1.6072e-07, 3.4070e-07, 1.7479e-07, 1.0532e-07, 3.6764e-07, 2.6853e-07,\n 1.9878e-07, 1.5118e-07, 1.9043e-07, 4.3375e-07, 4.2423e-07, 2.2043e-07,\n 1.2792e-07, 2.2036e-07, 3.2919e-07, 1.1646e-07, 2.3857e-07, 3.4992e-07,\n 2.0766e-07, 9.7829e-08, 1.3398e-07, 1.9801e-07, 2.5945e-07, 3.9670e-07,\n 9.5537e-08, 1.7171e-07, 2.4698e-07, 2.7714e-07, 1.1461e-07, 1.6431e-07,\n 3.7028e-07, 1.9785e-07, 1.8528e-10, 2.8006e-17, 3.5061e-07, 1.6847e-07],\n device='cuda:0')" + "step": "tensor(7512.)", + "exp_avg": "tensor([ 7.2504e-05, -1.4659e-09, -1.7810e-05, 9.6097e-05, -9.7703e-05,\n 5.6052e-45, 8.1218e-05, 1.1349e-04, 1.9342e-05, -1.1096e-04,\n -6.2938e-04, 5.2981e-11, 5.6245e-05, 7.3831e-06, -3.8516e-05,\n -1.3947e-05, -1.3790e-04, 3.3722e-05, -1.7507e-04, -5.4347e-05,\n -1.2297e-10, 3.4194e-05, -8.8130e-05, -1.8006e-04, 2.7653e-05,\n 1.1498e-04, -1.1373e-04, 5.6052e-45, 3.4173e-05, 2.5284e-04,\n -8.0575e-05, -5.5898e-05, -1.9806e-06, -1.2373e-04, -8.0797e-05,\n 8.7798e-06, -6.5299e-07, 2.2180e-04, -9.0341e-05, -2.7487e-06,\n -9.5056e-05, 1.1102e-04, 4.4392e-05, -1.0761e-04, -2.9451e-05,\n 1.7221e-04, 6.5313e-05, 8.6605e-05, -5.0989e-05, 2.1521e-04,\n 5.6052e-45, 1.2304e-04, 4.3234e-05, -6.2325e-05, 2.2039e-04,\n -2.5784e-05, 8.5614e-06, -1.7957e-04, 1.5460e-06, -1.2214e-04,\n 5.8034e-05, 2.4906e-05, -2.2255e-04, -8.5075e-05, 7.4354e-05,\n 2.1618e-05, -3.9654e-05, 8.5306e-05, 9.3775e-07, -1.5786e-05,\n 9.4578e-07, -5.6052e-45, 1.0970e-04, 1.3020e-04, 8.5234e-05,\n 3.6678e-05, 5.6052e-45, -4.8925e-04, -3.4260e-05, -4.5233e-05,\n 7.3616e-05, 3.2690e-06, 8.8311e-05, -4.9290e-27, 6.3833e-05,\n 7.2079e-05, 4.7931e-05, 8.2575e-06, 7.5915e-05, 2.1784e-06,\n 7.2646e-05, 3.2162e-04, 1.3645e-05, -1.6555e-04, 1.0907e-04,\n -5.9572e-06, -8.4043e-05, 2.1170e-05, -8.7177e-05, 5.3531e-05,\n -5.7581e-05, -3.2924e-05, 6.3297e-05, 1.0614e-05, 1.8811e-04,\n 1.3824e-04, -1.1251e-05, 8.2832e-05, -3.5859e-05, 5.5885e-05,\n 1.0591e-04, -1.5845e-04, 2.0451e-04, -9.2315e-05, 5.3274e-05,\n -6.6810e-05, 2.6101e-05, 1.7253e-05, -3.2015e-06, 5.6052e-45,\n -1.0955e-04, 1.4040e-05, -3.6299e-05, -5.6052e-45, -8.7477e-05,\n 9.8645e-05, -2.5936e-05, -7.0608e-07, -3.7877e-05, -3.0499e-05,\n 9.7376e-06, 3.9085e-05, 5.6052e-45, 1.7604e-05, 9.5052e-06,\n -5.4087e-05, 3.7523e-33, 1.5277e-04, 1.6184e-04, 1.1537e-04,\n 5.6052e-45, 5.6052e-45, 8.4177e-05, -9.5347e-06, 5.6052e-45,\n 5.6052e-45, -6.1912e-06, 1.4257e-05, 5.6052e-45, 4.2527e-05,\n -2.0343e-04, -1.5120e-07, 1.1250e-04, -1.3756e-05, 3.4964e-05,\n -2.0044e-04, -3.0080e-05, 1.7539e-04, 2.3714e-18, -6.5276e-05,\n 7.2776e-05, -5.2775e-05, -1.1506e-04, 1.4212e-05, -9.2428e-05,\n 2.2842e-05, -3.0201e-05, 5.6052e-45, 4.6581e-05, 1.5650e-04,\n 2.3522e-04, 1.1380e-04, 4.1862e-05, -4.1595e-05, 1.5935e-05,\n -7.8212e-05, -1.5936e-06, 7.8816e-05, 1.1358e-04, -4.2100e-04,\n -1.8590e-05, 1.9947e-05, 7.5865e-07, -7.2621e-06, 5.6052e-45,\n 5.6052e-45, 1.5186e-04, -1.0582e-04, -3.3053e-05, 7.6032e-05,\n -9.8805e-05, -6.7210e-05, 2.5041e-05, 1.1384e-04, 1.1778e-04,\n -4.4278e-05, -6.6418e-05, -1.5948e-04, 3.6756e-05, 5.6052e-45,\n -1.3317e-05, -7.7978e-06, -1.7101e-05, -3.5774e-05, -5.1203e-05,\n 5.7918e-05, -2.7764e-05, -1.1698e-06, -8.3568e-06, 5.6052e-45,\n 5.6662e-05, -8.8273e-19, -7.9053e-05, 5.0268e-06, 1.0223e-05,\n -2.3522e-05, 1.3796e-04, -2.9585e-05, -1.2474e-04, -7.8081e-05,\n -3.8026e-06, 8.9563e-05, -1.8933e-04, -9.8188e-05, -1.1596e-04,\n -9.2040e-21, -1.8825e-05, -3.7102e-05, -4.0608e-05, 2.0935e-05,\n 1.0906e-04, 1.2532e-04, -5.3258e-05, 1.4111e-04, 5.6052e-45,\n -2.8270e-05, 6.7886e-05, 8.8102e-41, 2.5412e-06, -1.5233e-04,\n -5.1607e-05, 5.6052e-45, 4.7349e-05, -1.3799e-04, 9.1050e-05,\n 3.1474e-05, -1.1404e-04, -7.3875e-06, -2.5591e-05, 7.8228e-22,\n 1.2637e-05, 3.1959e-04, 1.9625e-04, 7.6764e-05, 8.5653e-05,\n 9.1424e-05, 5.6052e-45, 6.0713e-05, 3.0042e-05, -2.4996e-04,\n 5.6052e-45, 4.7818e-05, -1.4494e-04, -1.9802e-04, 9.2980e-05,\n -2.3367e-05, 4.9134e-05, 2.0355e-05, 5.3966e-05, -1.0593e-04,\n -8.7785e-05, 1.0933e-04, 1.8915e-04, -7.6685e-05, 2.2661e-05,\n 5.0705e-05, -3.2966e-05, 5.6052e-45, -2.0269e-05, 5.1595e-05,\n 5.6052e-45, 9.2374e-05, 2.2114e-05, 9.3308e-05, 1.8631e-04,\n -6.1662e-05, 6.2471e-05, -9.6496e-06, -4.5235e-05, -5.1556e-06,\n 7.4458e-06, -5.1508e-05, 7.2722e-05, -9.0189e-05, -1.3905e-04,\n -6.3468e-05, -7.2071e-05, -6.8184e-05, 7.0358e-05, 9.6912e-05,\n -3.2079e-04, -1.5744e-04, -1.8711e-04, 1.4191e-04, 1.7244e-04,\n -1.0289e-40, 1.0392e-05, -3.2135e-04, -6.0714e-06, 9.7351e-05,\n 2.7960e-05, 6.6562e-43, -3.5020e-05, 2.4200e-05, -4.5082e-06,\n 5.6052e-45, -9.3722e-05, 8.7119e-05, 6.7881e-05, -1.2211e-04,\n -6.1199e-05, 3.5478e-05, 4.5832e-05, 8.6315e-05, 1.9957e-04,\n 5.6052e-45, -2.9254e-04, -9.1530e-05, 6.7211e-08, 2.7611e-05,\n 8.8047e-05, 2.9869e-04, 4.1329e-05, -2.9901e-05, -5.5099e-05,\n 8.2196e-05, 4.1527e-05, -1.2886e-05, 5.6052e-45, 2.0472e-09,\n -3.2817e-04, 7.9753e-05, 2.9995e-05, 1.9559e-05, -3.1491e-04,\n -2.7844e-05, -3.5648e-05, -2.0479e-06, -9.4420e-05, 5.0236e-05,\n 1.5690e-04, -4.7097e-06, -4.9982e-05, 5.6052e-45, 2.5386e-04,\n 8.0873e-05, 2.6523e-05, -6.5275e-06, -1.3935e-07, 5.6346e-05,\n 2.1440e-04, 1.3592e-04, 5.6052e-45, 1.5982e-04, -1.9409e-04,\n 2.3901e-05, -7.1005e-05, -2.4043e-05, 8.5486e-05, 2.8906e-05,\n -1.5627e-05, 2.1743e-05, -1.4481e-05, 8.7188e-05, 2.1711e-05,\n 5.8985e-05, -1.1786e-09, 9.1741e-05, 9.8984e-17, -1.2256e-04,\n 5.0423e-05, 3.4439e-06, 3.4502e-06, 7.9687e-05, 5.6052e-45,\n 3.1082e-05, 9.8109e-05, 3.5252e-05, 3.5391e-05, 2.9228e-05,\n -3.7737e-05, 4.8151e-05, -5.6082e-05, 8.6735e-05, -9.4945e-05,\n 1.1524e-04, 1.5575e-04, 1.4224e-09, 5.6052e-45, 7.1216e-05,\n -1.3898e-04, 1.6342e-04, -4.7850e-05, -1.9263e-04, -9.0581e-05,\n 7.8404e-07, 5.6052e-45, 6.7763e-05, -9.1885e-05, -5.2287e-05,\n 8.1814e-06, 5.6052e-45, 2.9590e-05, -3.3670e-05, -2.4820e-06,\n -8.4733e-06, -1.7815e-05, -8.4414e-05, 6.3727e-05, 1.5902e-04,\n -1.9234e-05, 1.8212e-04, -4.6149e-04, 5.6052e-45, 5.6052e-45,\n -5.5280e-06, -2.0074e-05, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 2.0560e-04, 2.8282e-04, -2.8423e-04, -1.9699e-04, -1.3420e-04,\n 5.6052e-45, 9.9683e-05, -9.7093e-06, 3.8937e-05, -5.6052e-45,\n 1.3408e-04, 1.3425e-04, -3.2269e-05, 5.6052e-45, 1.2146e-05,\n -1.8063e-05, -6.1819e-05, 1.1259e-04, -1.8847e-05, -1.3320e-04,\n -1.3045e-04, 2.1847e-05, 6.1593e-05, -2.4373e-04, -3.5689e-05,\n -1.1485e-04, -1.6694e-04, 1.4238e-04, 5.6052e-45, -1.1939e-04,\n -6.5924e-05, -2.5922e-05, 5.6052e-45, 8.3399e-06, -2.7070e-05,\n 4.7282e-05, -2.9808e-05, -5.3766e-05, 1.1361e-04, 6.9751e-05,\n 9.2246e-05, 5.6052e-45, 5.2986e-05, 3.3308e-05, 4.3766e-05,\n -8.9322e-25, 1.0807e-04, -5.3058e-05, -5.7525e-05, -1.6923e-04,\n 1.2355e-05, 2.9891e-05, -1.4530e-04, -2.8705e-04, -7.0266e-05,\n -1.9626e-04, -3.0528e-05, 8.4310e-05, -1.2885e-04, 1.1774e-04,\n 1.8734e-04, -8.6356e-05, 1.6604e-04, 8.5562e-05, 1.3948e-04,\n -8.2810e-05, -7.1048e-05, 8.5683e-05, -3.4693e-05, 3.7757e-05,\n 5.6052e-45, -4.5616e-05, 2.3700e-05, 1.2524e-04, 2.5553e-04,\n 3.7218e-05, -6.5472e-05, 1.9509e-05, -1.3505e-04, -3.9350e-07,\n 2.4253e-04, 3.2880e-04, 1.2700e-04, -4.7309e-05, -5.1802e-05,\n -1.0119e-05, -1.2961e-04, -1.0202e-04, 5.6052e-45, 2.0273e-04,\n 2.8094e-04, 2.9570e-05, 2.5983e-04, 8.4209e-05, 1.4227e-04,\n -2.3152e-04, 5.6052e-45, 2.1566e-05, -1.2568e-04, -9.3252e-05,\n 1.5337e-05, 1.1635e-04, -4.0088e-05, 5.7605e-06, 3.5551e-05,\n -6.6025e-05, 6.7826e-05, 5.1559e-30, 9.6018e-06, -2.0611e-04,\n -3.0275e-05, -6.5166e-05, -2.2564e-05, 7.9556e-05, 1.3527e-04,\n -9.0547e-23, -4.2233e-05, -1.1186e-06, -1.0066e-04, -7.5431e-07,\n -1.7136e-04, 1.9814e-04, -1.2388e-04, -8.8481e-05, 4.7274e-05,\n -9.3003e-05, 3.4063e-05, 1.0773e-04, 3.6232e-05, 3.6882e-05,\n 5.6052e-45, -6.2560e-05, -6.0052e-05, -7.7398e-06, -1.1210e-04,\n 1.8685e-04, -2.9741e-05, 5.6052e-45, -1.1121e-04, 5.6052e-45,\n 2.9972e-05, -2.3542e-05, -2.0240e-05, 1.8160e-05, 4.1373e-05,\n 6.8472e-06, 9.3313e-05, 2.0514e-04, 1.6238e-04, 1.6676e-04,\n 1.2839e-05, 5.6052e-45, 2.3667e-05, 4.5064e-05, 5.6052e-45,\n -6.4317e-05, -1.8396e-04, 4.1933e-05, 6.9794e-05, -2.2815e-05,\n -2.8276e-05, 5.6052e-45, -1.1671e-04, 8.0423e-05, 1.6450e-04,\n 4.1985e-05, 5.7781e-05, 3.1496e-04, -6.0510e-05, -6.9774e-05,\n -7.6618e-05, 1.0367e-04, -2.9031e-05, -8.6432e-05, -8.8045e-06,\n 4.5230e-05, 4.3611e-05, 2.1199e-05, -1.4580e-04, 7.4114e-05,\n 1.0696e-04, -9.4603e-07, -8.2906e-04, -1.6243e-06, 5.5275e-05,\n 5.6052e-45, -2.0836e-04, -5.6052e-45, -4.9292e-05, -7.5164e-05,\n 1.6135e-04, 1.6148e-05, -6.1277e-05, 1.7202e-04, 5.6376e-06,\n 1.2328e-04, 5.6052e-45, 3.2617e-05, -1.1423e-05, 1.0516e-04,\n -2.4728e-05, 1.5747e-05, 5.6052e-45, 5.6052e-45, 1.0020e-04,\n 7.8660e-05, 5.6052e-45, 7.6908e-05, 1.0147e-04, 1.2600e-04,\n -3.1102e-06, 5.6052e-45, -1.2268e-04, 1.3220e-06, 6.0714e-05,\n -5.4230e-05, 5.4082e-05, -2.1044e-04, 1.0250e-04, -3.6403e-04,\n -2.3053e-04, 8.8411e-05, 1.3557e-04, -8.4307e-05, -5.2099e-05,\n 1.6079e-04, 1.7187e-04, -1.2405e-04, -1.3410e-04, -3.9664e-05,\n -1.6998e-04, -8.7308e-05, -1.6432e-04, 1.0634e-05, -1.2763e-04,\n -8.7049e-05, 1.5949e-04, 1.0473e-06, 9.6299e-05, 1.1191e-04,\n -2.6033e-05, 5.6052e-45, 2.1600e-04, 1.2307e-04, 6.2408e-05,\n 3.7390e-05, 2.2834e-04, 2.4632e-05, -5.6052e-45, 5.6052e-45,\n -2.5445e-05, -5.6750e-06, 4.3092e-05, 5.6052e-45, -1.5013e-04,\n -5.6052e-45, -1.8448e-05, -3.3228e-05, 1.0883e-04, 6.9427e-05,\n 5.3485e-05, -4.1406e-05, 5.6052e-45, -4.7909e-05, -3.4696e-05,\n -8.4972e-05, -2.7988e-05, 4.2301e-05, 5.6052e-45, -1.0807e-04,\n -1.4661e-04, 1.9384e-04, 1.0810e-04, -4.6562e-05, -4.0637e-05,\n 8.5437e-06, 5.6052e-45, -1.9545e-05, 2.3907e-04, 5.6052e-45,\n -9.5901e-06, 1.5144e-04, 4.4265e-05, 1.0717e-04, -7.8895e-05,\n -4.2341e-05, -3.6583e-05, -5.5301e-05, 3.9327e-05, 6.5386e-05,\n -3.7895e-25, 2.6720e-04, -8.4478e-06, -1.0924e-04, 5.7079e-05,\n 2.5695e-05, -1.6511e-04, -7.8789e-05, 4.5142e-05, 5.8352e-05,\n 3.6076e-05, 1.1824e-04, -5.3717e-05, -6.0429e-05, 5.6052e-45,\n -5.2280e-05, -6.7813e-05, -3.2084e-05, 1.3898e-04, 3.3738e-05,\n 1.1029e-04, -4.8019e-05, -2.6317e-04, -1.7929e-05, 1.1724e-04,\n 7.9672e-05, -3.3108e-05, -7.3609e-17, -2.8110e-05, -5.6052e-45,\n 2.2300e-04, 3.2946e-06, -7.8978e-05, 1.5620e-04, -1.9948e-05,\n -3.6455e-05, -2.2080e-05, -1.1481e-04, 5.3561e-05, 5.8068e-05,\n -1.0439e-05, -3.9750e-05, 1.6809e-04, 3.3410e-07, 5.6052e-45,\n 5.6052e-45, -2.5868e-05, -6.6843e-05], device='cuda:0')", + "exp_avg_sq": "tensor([1.6782e-07, 1.5712e-09, 1.1440e-07, 1.5294e-07, 3.3171e-07, 1.4635e-08,\n 3.3780e-07, 2.0531e-07, 1.2101e-07, 2.8587e-07, 2.4388e-07, 8.0832e-09,\n 1.8522e-07, 8.1257e-08, 1.9005e-07, 2.0007e-07, 2.7794e-07, 7.3459e-08,\n 9.4259e-08, 1.7963e-07, 1.8341e-10, 8.3141e-08, 1.1527e-07, 2.2083e-07,\n 3.0563e-07, 2.6791e-07, 1.2185e-07, 3.3235e-11, 2.2065e-07, 2.2777e-07,\n 7.8721e-08, 1.9370e-07, 1.7006e-07, 2.0230e-07, 1.5052e-07, 1.5626e-07,\n 1.7849e-08, 1.9220e-07, 1.9788e-07, 1.8929e-07, 2.7505e-07, 1.7092e-07,\n 2.8925e-07, 2.6349e-07, 2.3728e-07, 1.2915e-07, 1.7584e-07, 1.9158e-07,\n 4.4261e-07, 2.1718e-07, 2.8507e-08, 5.9058e-07, 1.7344e-07, 1.6047e-07,\n 1.8421e-07, 1.0751e-07, 2.1508e-07, 1.0717e-07, 7.8621e-08, 1.9718e-07,\n 7.3722e-08, 1.6085e-07, 3.6859e-07, 1.7583e-07, 7.5797e-08, 1.4106e-07,\n 3.0449e-07, 1.9353e-07, 5.1210e-08, 4.1670e-08, 2.2452e-07, 1.2969e-08,\n 1.3880e-07, 4.9444e-07, 2.6314e-07, 2.4458e-07, 9.0013e-11, 3.5644e-07,\n 1.2487e-07, 1.8523e-07, 1.0250e-07, 2.1040e-07, 2.9723e-07, 2.4852e-10,\n 2.3692e-07, 2.5070e-07, 2.2284e-07, 3.5875e-07, 1.9305e-07, 3.1198e-07,\n 2.1174e-07, 1.3317e-07, 1.6902e-07, 2.3644e-07, 1.2645e-07, 3.1108e-07,\n 2.3025e-07, 2.1466e-07, 1.9366e-07, 2.4501e-07, 2.1161e-07, 2.1507e-07,\n 4.2763e-07, 2.2186e-07, 1.8867e-07, 2.4118e-07, 1.5447e-07, 1.7496e-07,\n 1.5769e-07, 1.2207e-07, 1.8668e-07, 3.0449e-07, 1.8493e-07, 6.8369e-08,\n 2.8947e-07, 2.0021e-07, 1.8883e-07, 1.3003e-07, 8.5690e-08, 6.8038e-09,\n 1.8276e-07, 3.2675e-07, 1.3168e-07, 1.7050e-08, 1.8817e-07, 2.4485e-07,\n 1.7870e-07, 2.2026e-08, 2.8677e-07, 1.4359e-07, 1.7985e-07, 1.4624e-07,\n 7.4477e-10, 2.5652e-07, 4.0988e-08, 2.0158e-07, 3.0242e-12, 2.1044e-07,\n 1.7142e-07, 7.4657e-08, 7.5089e-09, 4.1292e-09, 1.5926e-07, 1.2542e-08,\n 1.5998e-11, 2.1279e-08, 2.6956e-07, 2.3958e-07, 5.7172e-08, 1.1799e-07,\n 1.5839e-07, 2.6730e-07, 1.6705e-07, 2.0318e-07, 5.4549e-08, 2.2866e-07,\n 1.0275e-07, 2.0186e-07, 2.3104e-08, 1.3938e-07, 2.3232e-07, 1.8334e-07,\n 2.7735e-07, 3.6935e-07, 1.3051e-07, 1.8350e-07, 1.5266e-08, 3.0080e-14,\n 2.0521e-07, 1.6644e-07, 2.8499e-07, 2.4615e-07, 2.0242e-07, 2.2911e-07,\n 2.2413e-07, 2.1223e-07, 9.5916e-08, 2.6615e-07, 2.5855e-07, 3.2116e-07,\n 3.0072e-07, 1.9275e-07, 2.1719e-07, 1.0832e-07, 1.2634e-12, 1.7575e-17,\n 3.0852e-07, 1.9539e-07, 8.2354e-08, 1.0496e-07, 1.4071e-07, 1.5075e-07,\n 2.5517e-07, 2.9074e-07, 2.4912e-07, 1.2457e-07, 3.1345e-07, 2.5940e-07,\n 1.8596e-07, 1.9251e-19, 6.8617e-08, 1.2609e-07, 2.8191e-07, 2.4431e-07,\n 3.0097e-07, 1.7462e-07, 2.5011e-07, 1.5695e-07, 2.6592e-08, 1.4429e-10,\n 3.1250e-07, 2.0245e-10, 1.3463e-07, 3.4215e-07, 2.2240e-07, 2.1980e-07,\n 2.2806e-07, 1.3061e-07, 2.1808e-07, 1.7601e-07, 7.9838e-08, 3.0541e-07,\n 2.9950e-07, 1.6916e-07, 1.0332e-07, 2.8502e-08, 2.2968e-07, 3.4490e-07,\n 3.1910e-07, 3.3036e-08, 1.8727e-07, 1.9399e-07, 2.0358e-07, 4.3955e-07,\n 1.8035e-08, 3.0309e-07, 2.9184e-07, 3.4224e-12, 1.4334e-07, 2.1380e-07,\n 4.9813e-08, 9.3816e-09, 2.8105e-07, 2.0178e-07, 1.7017e-07, 2.4754e-07,\n 2.0921e-07, 2.6232e-07, 2.0265e-07, 1.1967e-08, 2.6756e-07, 4.1044e-07,\n 1.9953e-07, 2.9086e-07, 1.8892e-07, 3.1307e-07, 3.6588e-12, 1.2687e-07,\n 6.6317e-08, 3.8187e-07, 2.1545e-17, 2.1796e-07, 2.0421e-07, 1.3393e-07,\n 3.0447e-07, 1.2996e-07, 1.5886e-07, 1.4212e-07, 2.8227e-07, 1.5110e-07,\n 2.4833e-07, 2.2931e-07, 1.4629e-07, 2.2944e-07, 3.4359e-07, 1.8080e-07,\n 2.9570e-07, 1.8654e-11, 9.2193e-08, 3.4186e-07, 5.3191e-08, 3.7907e-07,\n 9.9207e-08, 2.4011e-07, 3.0741e-07, 1.5317e-07, 3.1751e-07, 2.4348e-07,\n 1.2256e-07, 2.8936e-07, 1.4207e-07, 1.7559e-07, 1.1848e-07, 1.1442e-07,\n 2.5037e-07, 1.6528e-07, 1.6400e-07, 2.7770e-07, 1.5800e-07, 8.3767e-08,\n 3.0290e-07, 2.5625e-07, 1.0649e-07, 1.2185e-07, 1.2418e-07, 4.2953e-11,\n 8.2983e-09, 1.1091e-07, 2.3549e-07, 2.1468e-07, 2.0980e-07, 2.7948e-12,\n 2.1286e-07, 1.5166e-07, 2.4531e-07, 6.1345e-12, 5.0870e-08, 3.7808e-07,\n 9.6208e-08, 1.2539e-07, 2.3092e-07, 5.9797e-08, 1.0388e-07, 2.3216e-07,\n 1.2232e-07, 2.2256e-08, 2.4502e-07, 2.2763e-07, 1.9696e-07, 2.3664e-07,\n 2.3207e-07, 1.0991e-07, 2.1477e-07, 1.5921e-07, 9.0116e-08, 3.4881e-07,\n 1.8533e-07, 1.2697e-07, 1.6716e-17, 1.5038e-10, 3.6695e-07, 1.8758e-07,\n 2.9730e-07, 9.5674e-08, 3.0464e-07, 1.8715e-07, 2.2283e-07, 2.4406e-07,\n 2.6235e-07, 2.4534e-07, 2.5358e-07, 2.1792e-07, 9.2398e-08, 4.5508e-09,\n 1.8565e-07, 3.4562e-07, 7.6175e-08, 5.4580e-08, 1.8787e-07, 1.9090e-07,\n 2.3671e-07, 2.2287e-07, 1.8872e-08, 2.7881e-07, 1.6526e-07, 7.7623e-08,\n 1.4655e-07, 2.7833e-07, 1.5481e-07, 9.2955e-08, 1.2825e-07, 4.9332e-08,\n 1.3736e-07, 1.5457e-07, 1.9069e-07, 1.2117e-07, 3.2856e-09, 8.1677e-08,\n 8.9604e-09, 1.2906e-07, 4.1294e-07, 2.4337e-07, 1.1861e-07, 2.5325e-07,\n 1.8726e-08, 2.8194e-07, 6.5420e-08, 2.5388e-07, 3.6923e-07, 1.1901e-07,\n 2.1627e-07, 2.9603e-07, 1.9070e-07, 6.6897e-08, 2.7166e-07, 3.1219e-07,\n 6.2712e-08, 7.8381e-09, 6.4402e-08, 7.5119e-08, 1.6575e-07, 3.0388e-07,\n 2.6395e-07, 4.9351e-07, 2.1735e-07, 2.7155e-07, 5.4376e-08, 2.9716e-07,\n 2.3690e-07, 2.8869e-07, 2.6847e-07, 1.1593e-13, 1.3203e-07, 2.3479e-07,\n 9.3105e-08, 1.9501e-07, 1.4662e-08, 2.7102e-07, 1.8273e-07, 1.6082e-07,\n 1.8682e-07, 2.8035e-07, 3.2684e-07, 7.8409e-12, 5.1432e-19, 1.2607e-07,\n 8.3323e-08, 4.3253e-09, 1.1891e-11, 2.4868e-11, 2.4861e-07, 1.8990e-07,\n 1.9206e-07, 1.8210e-07, 3.1198e-07, 2.7562e-09, 2.6386e-07, 2.1307e-07,\n 2.3894e-07, 6.4840e-11, 1.0152e-07, 1.6460e-07, 1.5369e-07, 7.8628e-11,\n 1.4023e-07, 3.5226e-07, 1.4603e-07, 1.9087e-07, 1.0492e-07, 9.2753e-08,\n 4.0964e-07, 1.1630e-07, 2.2699e-07, 2.3372e-07, 2.6831e-07, 2.2701e-07,\n 2.3087e-08, 4.2725e-07, 2.4368e-09, 9.0936e-08, 1.3423e-07, 8.7220e-08,\n 5.9368e-13, 2.0422e-07, 4.4188e-08, 2.2109e-07, 1.1716e-07, 1.2799e-07,\n 2.5050e-07, 1.1476e-07, 1.1019e-07, 2.4033e-10, 6.4492e-08, 3.0113e-07,\n 8.5271e-08, 1.0673e-11, 1.8405e-07, 1.7215e-07, 1.0929e-07, 3.2189e-07,\n 2.1905e-07, 2.5613e-07, 2.3301e-07, 3.1303e-07, 2.0440e-07, 2.1593e-07,\n 3.6053e-07, 2.7533e-07, 2.0812e-07, 2.0968e-07, 1.9726e-07, 1.1806e-07,\n 2.9354e-07, 1.2057e-07, 2.2780e-07, 9.6785e-08, 2.1341e-07, 2.2465e-07,\n 2.6136e-07, 9.3345e-08, 1.2865e-07, 4.1300e-07, 1.6994e-07, 1.5051e-07,\n 2.5112e-07, 1.2003e-07, 2.5972e-07, 2.7941e-07, 1.9773e-07, 2.7367e-07,\n 1.7267e-07, 2.6160e-07, 3.9413e-07, 2.6591e-07, 1.1925e-07, 7.6332e-08,\n 1.5071e-07, 1.1002e-07, 1.7820e-08, 2.3857e-07, 1.9946e-07, 1.0184e-07,\n 1.9307e-07, 1.7157e-07, 1.6998e-07, 2.5773e-07, 5.1062e-13, 3.3501e-07,\n 1.5657e-07, 1.8735e-07, 2.4792e-07, 3.0163e-07, 2.2505e-07, 2.4916e-07,\n 1.7222e-07, 8.9450e-08, 1.9339e-07, 5.9503e-08, 1.2117e-07, 2.1916e-07,\n 3.3425e-07, 2.4054e-07, 3.3521e-07, 4.2125e-07, 3.1970e-07, 2.1410e-10,\n 2.0536e-07, 2.5145e-07, 2.7160e-07, 1.4841e-07, 2.8910e-07, 3.2759e-07,\n 1.9109e-07, 2.0538e-07, 1.7869e-07, 1.9607e-07, 2.1042e-07, 2.6446e-07,\n 1.7470e-07, 1.8024e-07, 4.0525e-12, 2.7167e-07, 1.6111e-07, 1.5787e-07,\n 2.2623e-07, 2.1324e-07, 1.4193e-07, 8.4841e-11, 2.3238e-07, 4.1381e-09,\n 2.1456e-07, 1.4496e-07, 2.1824e-07, 6.1526e-09, 1.2378e-07, 1.9359e-07,\n 2.4298e-07, 7.7605e-08, 1.4352e-07, 1.6166e-07, 2.4555e-07, 1.0761e-11,\n 1.0393e-07, 6.1830e-08, 1.4285e-15, 2.8325e-07, 2.2351e-07, 1.9808e-07,\n 1.5918e-07, 1.3498e-07, 1.8504e-07, 1.6607e-12, 1.5671e-07, 2.0291e-07,\n 1.2224e-07, 1.5209e-07, 1.0692e-07, 1.5860e-07, 1.3796e-07, 9.7461e-08,\n 2.4094e-07, 1.2925e-07, 2.4814e-07, 1.7335e-07, 2.1764e-07, 2.7917e-07,\n 3.1365e-07, 5.5106e-08, 2.3546e-07, 1.9184e-07, 2.2833e-07, 1.6643e-07,\n 2.3816e-07, 2.9703e-08, 2.3266e-07, 7.5090e-12, 2.3774e-07, 8.7125e-09,\n 2.2119e-07, 3.6057e-07, 3.1177e-07, 2.8470e-07, 2.9767e-07, 3.4004e-07,\n 1.3459e-07, 2.2874e-07, 7.4500e-13, 2.2037e-07, 2.3623e-07, 6.5812e-08,\n 5.5472e-08, 5.6689e-09, 1.9559e-10, 6.0961e-12, 2.0206e-07, 2.8071e-07,\n 2.3077e-14, 1.6071e-07, 2.6408e-07, 3.0076e-07, 3.7316e-07, 2.4735e-12,\n 2.5234e-07, 1.2501e-07, 9.3302e-08, 1.7419e-07, 2.1189e-07, 2.1962e-07,\n 3.2550e-07, 1.0065e-07, 2.0961e-07, 2.2622e-07, 3.1691e-07, 2.5840e-07,\n 1.5324e-07, 1.8625e-07, 3.5859e-07, 1.0292e-07, 2.0263e-07, 4.5551e-08,\n 4.6410e-07, 1.4435e-07, 3.2958e-07, 1.6194e-07, 1.3850e-07, 3.9176e-07,\n 3.1984e-07, 8.4640e-08, 2.0096e-07, 1.2768e-08, 1.8246e-07, 2.1741e-13,\n 1.1764e-07, 1.6521e-07, 1.8883e-07, 2.5208e-07, 3.7997e-07, 1.6822e-08,\n 9.7761e-09, 2.3407e-11, 3.4500e-07, 2.0695e-07, 2.0818e-07, 5.5376e-08,\n 1.9934e-07, 1.0226e-08, 2.6422e-07, 1.3403e-07, 1.7702e-07, 2.6924e-07,\n 5.0509e-08, 1.2703e-07, 8.3934e-08, 8.5006e-08, 6.9310e-08, 1.3270e-07,\n 3.0818e-07, 1.8432e-07, 6.6440e-12, 1.2731e-07, 2.3792e-07, 2.2518e-07,\n 2.4111e-07, 2.9570e-07, 3.0133e-07, 1.6171e-07, 5.2893e-11, 6.5503e-07,\n 3.4978e-07, 4.1688e-13, 2.9518e-07, 1.5138e-07, 9.6773e-08, 2.3331e-07,\n 2.2395e-07, 8.3985e-08, 1.9322e-07, 2.5786e-07, 1.3443e-07, 2.8039e-07,\n 3.7203e-08, 4.4083e-07, 1.9069e-07, 2.4926e-07, 1.8231e-07, 9.8289e-08,\n 2.0054e-07, 3.2654e-07, 5.2550e-07, 3.1957e-07, 9.6452e-08, 2.2921e-07,\n 1.7586e-07, 3.0296e-07, 4.9948e-08, 9.8532e-08, 2.7414e-07, 2.3151e-07,\n 1.6738e-07, 1.3774e-07, 1.6095e-07, 2.7772e-07, 1.8053e-07, 1.6434e-07,\n 1.3469e-07, 1.8872e-07, 2.1409e-07, 3.3280e-08, 1.9042e-07, 9.9992e-08,\n 1.7720e-07, 9.3871e-08, 1.2755e-07, 1.7656e-07, 2.2643e-07, 3.2375e-07,\n 7.9610e-08, 1.3374e-07, 2.1042e-07, 2.3386e-07, 8.6866e-08, 1.5206e-07,\n 2.7671e-07, 9.2331e-08, 5.2946e-11, 8.0030e-18, 2.5203e-07, 1.7546e-07],\n device='cuda:0')" }, "4": { - "step": "tensor(6260.)", - "exp_avg": "tensor([[-5.8735e-07, 4.8569e-25, 2.0672e-05, ..., 5.6052e-45,\n -4.0880e-05, 7.9777e-06],\n [ 9.2971e-06, 1.5598e-25, 1.5837e-06, ..., -5.6052e-45,\n -2.5973e-06, -5.9872e-06],\n [-1.1278e-05, 1.6732e-25, 2.5069e-05, ..., -5.6052e-45,\n 5.0401e-05, 2.4138e-05],\n ...,\n [ 1.6577e-06, 1.8463e-25, -2.1244e-05, ..., -5.6052e-45,\n 5.0692e-05, -7.5169e-06],\n [-8.4020e-06, 4.3994e-25, 1.0613e-05, ..., -5.6052e-45,\n -8.8597e-06, 1.5104e-05],\n [ 2.0186e-06, 3.7173e-25, -2.4462e-05, ..., -5.6052e-45,\n 5.2665e-05, 1.3639e-05]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.5131e-09, 6.8519e-14, 2.4536e-10, ..., 7.8498e-20, 1.7754e-09,\n 3.8327e-10],\n [2.9172e-09, 1.2635e-12, 5.1498e-10, ..., 4.0118e-18, 2.8007e-09,\n 1.8120e-09],\n [4.2391e-09, 7.6922e-15, 9.4877e-10, ..., 1.7497e-18, 3.0958e-09,\n 1.9701e-09],\n ...,\n [3.9652e-09, 1.2506e-12, 7.4609e-10, ..., 1.5308e-18, 1.7783e-09,\n 1.1084e-09],\n [4.7277e-09, 1.0482e-13, 6.8378e-10, ..., 2.2813e-20, 2.5905e-09,\n 2.7353e-09],\n [3.8885e-09, 1.6098e-13, 1.2032e-09, ..., 1.4620e-18, 2.5534e-09,\n 2.0446e-09]], device='cuda:0')" + "step": "tensor(7512.)", + "exp_avg": "tensor([[ 4.5632e-06, -3.5970e-11, -7.2489e-06, ..., 5.6052e-45,\n -5.7196e-07, -2.3426e-06],\n [-1.7750e-05, 5.6012e-12, 1.8249e-06, ..., -5.6052e-45,\n -1.1027e-05, 2.6085e-06],\n [-6.4006e-07, -5.0587e-11, -1.3868e-05, ..., -5.6052e-45,\n 2.1973e-05, -1.4698e-05],\n ...,\n [ 6.5974e-06, -4.5897e-11, -3.0239e-06, ..., -5.6052e-45,\n 8.7567e-06, 1.0037e-05],\n [ 2.1113e-06, 1.4446e-10, 9.4193e-07, ..., -5.6052e-45,\n -3.4310e-08, -5.3713e-06],\n [-5.3003e-06, -3.3248e-10, -7.7910e-06, ..., -5.6052e-45,\n -9.3499e-06, -3.6540e-06]], device='cuda:0')", + "exp_avg_sq": "tensor([[1.0607e-09, 2.0584e-14, 2.4135e-10, ..., 2.2431e-20, 1.2262e-09,\n 3.5673e-10],\n [2.1969e-09, 3.6143e-13, 5.2545e-10, ..., 1.1464e-18, 2.4705e-09,\n 1.5893e-09],\n [3.2540e-09, 9.0489e-15, 1.3605e-09, ..., 4.9998e-19, 2.5280e-09,\n 1.7764e-09],\n ...,\n [3.0365e-09, 3.5860e-13, 7.3736e-10, ..., 4.3743e-19, 1.3323e-09,\n 9.9376e-10],\n [3.5453e-09, 3.7564e-14, 8.5811e-10, ..., 6.5190e-21, 2.0806e-09,\n 2.4768e-09],\n [2.9416e-09, 4.9855e-14, 1.3465e-09, ..., 4.1779e-19, 1.7811e-09,\n 1.7255e-09]], device='cuda:0')" }, "5": { - "step": "tensor(5008.)", - "exp_avg": "tensor([[ 4.1077e-08, -1.3984e-08, 0.0000e+00, ..., 1.0988e-07,\n 4.0628e-09, 9.9676e-13],\n [-5.4944e-06, -6.3427e-06, -5.6052e-45, ..., 1.8983e-06,\n 2.2179e-07, 6.9046e-09],\n [ 1.2553e-07, -8.2448e-07, -2.8306e-43, ..., -8.5227e-08,\n 2.0630e-06, -1.0564e-08],\n ...,\n [ 7.8341e-07, 6.8223e-07, 5.6052e-45, ..., -4.3376e-07,\n 9.5179e-07, -3.3700e-06],\n [ 1.0092e-07, -8.9138e-07, -5.6052e-45, ..., 9.4612e-08,\n 1.6499e-06, 3.1060e-08],\n [ 1.8421e-06, 1.3908e-06, 5.6052e-45, ..., -4.0586e-06,\n -2.3236e-06, -2.7607e-06]], device='cuda:0')", - "exp_avg_sq": "tensor([[7.5349e-13, 7.0549e-13, 0.0000e+00, ..., 9.0973e-13, 1.8718e-12,\n 4.9006e-14],\n [2.7152e-10, 7.4804e-11, 8.4360e-14, ..., 3.1960e-10, 1.2772e-10,\n 1.9254e-11],\n [6.6905e-12, 8.3255e-12, 5.5893e-15, ..., 1.2933e-11, 7.6728e-11,\n 1.2273e-11],\n ...,\n [4.4578e-10, 1.3013e-10, 1.2235e-13, ..., 2.7417e-11, 8.1413e-11,\n 2.1616e-09],\n [1.7065e-11, 1.6678e-10, 1.1750e-15, ..., 1.4255e-10, 1.2765e-10,\n 7.0486e-12],\n [7.8672e-11, 3.3206e-11, 2.1205e-14, ..., 2.0737e-10, 4.8531e-11,\n 6.5964e-10]], device='cuda:0')" + "step": "tensor(6260.)", + "exp_avg": "tensor([[ 5.4320e-07, 2.5112e-07, 5.6052e-45, ..., 1.5735e-06,\n -6.8148e-08, -9.1376e-09],\n [ 5.6463e-07, 2.3119e-06, -5.6052e-45, ..., 1.3595e-06,\n 5.5618e-06, 2.2909e-07],\n [ 5.4435e-08, 3.4342e-07, -5.3641e-37, ..., 2.8727e-06,\n 4.8562e-08, 3.9716e-07],\n ...,\n [ 1.6038e-06, -8.4202e-07, -1.6381e-15, ..., -2.0679e-07,\n 6.5808e-07, 8.7196e-06],\n [ 5.7384e-07, -3.6656e-06, -5.6052e-45, ..., -2.7952e-07,\n 7.0624e-07, 2.4183e-07],\n [-8.9130e-07, -5.5013e-07, -1.9926e-42, ..., -1.7597e-06,\n 1.4912e-06, -1.7082e-07]], device='cuda:0')", + "exp_avg_sq": "tensor([[1.3985e-12, 3.4512e-12, 5.5557e-18, ..., 4.6889e-12, 1.8818e-12,\n 1.6998e-13],\n [1.9429e-10, 8.4029e-11, 2.4107e-14, ..., 2.9990e-10, 1.2616e-10,\n 2.2124e-11],\n [3.2450e-12, 5.3456e-12, 1.5972e-15, ..., 6.7679e-12, 6.1489e-11,\n 7.8617e-12],\n ...,\n [3.0389e-10, 6.8025e-11, 3.4964e-14, ..., 1.8137e-11, 6.6784e-11,\n 1.6892e-09],\n [8.8512e-12, 1.0373e-10, 3.3578e-16, ..., 1.4462e-10, 1.2270e-10,\n 2.4440e-12],\n [5.1302e-11, 3.2694e-11, 6.0604e-15, ..., 1.4584e-10, 9.2899e-11,\n 5.9803e-10]], device='cuda:0')" }, "6": { - "step": "tensor(5008.)", - "exp_avg": "tensor([-7.2008e-07, -3.0324e-05, -1.2121e-05, ..., 4.6873e-06,\n 5.6040e-06, 1.8653e-05], device='cuda:0')", - "exp_avg_sq": "tensor([6.4102e-11, 1.6039e-08, 4.9925e-09, ..., 1.4547e-08, 6.3299e-09,\n 9.7342e-09], device='cuda:0')" + "step": "tensor(6260.)", + "exp_avg": "tensor([ 1.4682e-05, -9.0957e-06, 9.8845e-06, ..., -1.0262e-05,\n 1.4127e-05, -3.0702e-06], device='cuda:0')", + "exp_avg_sq": "tensor([2.0110e-10, 1.4904e-08, 3.5646e-09, ..., 9.6344e-09, 5.2860e-09,\n 8.0342e-09], device='cuda:0')" }, "7": { - "step": "tensor(5008.)", - "exp_avg": "tensor([[-4.4023e-08, 7.2592e-08, 4.2675e-07, ..., -1.4111e-07,\n 3.2388e-08, -3.2449e-07],\n [ 3.0744e-08, -9.8427e-07, 2.1796e-07, ..., -2.7479e-07,\n 1.4039e-06, -6.8012e-08],\n [-2.4845e-08, 6.9495e-07, 3.8541e-07, ..., 2.7620e-07,\n 2.3047e-07, 1.5663e-06],\n ...,\n [ 1.5190e-10, -6.3821e-07, -6.1442e-07, ..., 7.7054e-07,\n -1.7749e-07, 1.1047e-07],\n [-1.8115e-08, 1.3022e-06, -3.2499e-07, ..., -1.0946e-07,\n 1.2228e-06, 6.6345e-07],\n [-1.7501e-07, -5.5627e-07, -2.9534e-08, ..., -1.1809e-07,\n -1.7901e-07, 4.4734e-07]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.6072e-12, 9.8176e-12, 2.4876e-11, ..., 1.0221e-11, 8.3117e-12,\n 9.9785e-12],\n [7.3777e-12, 2.5521e-11, 2.1757e-11, ..., 1.8811e-11, 1.0008e-11,\n 1.2684e-11],\n [9.2885e-13, 2.0288e-11, 1.8688e-11, ..., 1.8455e-11, 8.5347e-12,\n 2.0158e-11],\n ...,\n [1.8921e-12, 2.1504e-11, 1.4157e-11, ..., 2.1163e-11, 1.1271e-11,\n 2.5519e-11],\n [2.2871e-12, 2.9846e-11, 4.8118e-11, ..., 2.3456e-11, 9.3664e-12,\n 2.0419e-11],\n [1.6263e-12, 1.9157e-11, 1.0400e-10, ..., 1.7849e-11, 1.0994e-11,\n 1.5045e-11]], device='cuda:0')" + "step": "tensor(6260.)", + "exp_avg": "tensor([[ 7.1781e-07, -9.8490e-08, 1.0681e-06, ..., -4.6073e-07,\n -6.0009e-06, -1.0362e-06],\n [ 5.8528e-08, 9.1701e-07, 2.8097e-07, ..., -3.4345e-07,\n -1.1492e-06, -7.9879e-07],\n [ 2.7797e-07, 1.1085e-06, -3.7216e-07, ..., 6.6875e-08,\n -1.0726e-05, -1.3775e-06],\n ...,\n [ 1.7893e-08, -1.0760e-07, 4.9530e-07, ..., 7.6758e-07,\n 6.2710e-06, 1.8813e-06],\n [ 2.1236e-07, -3.1910e-07, -1.6600e-06, ..., -1.9081e-07,\n -3.8838e-06, -9.7871e-07],\n [-2.2817e-08, -6.5586e-08, -1.8786e-06, ..., 6.2706e-07,\n 8.7271e-07, 3.6174e-07]], device='cuda:0')", + "exp_avg_sq": "tensor([[9.2644e-13, 6.8421e-12, 2.0159e-11, ..., 6.6076e-12, 8.8399e-12,\n 6.6937e-12],\n [2.9660e-12, 1.7647e-11, 1.6013e-11, ..., 1.1292e-11, 7.6500e-12,\n 7.7266e-12],\n [1.0769e-12, 1.4492e-11, 1.0373e-11, ..., 1.1357e-11, 1.6475e-11,\n 1.3423e-11],\n ...,\n [9.6912e-13, 1.4544e-11, 1.0186e-11, ..., 1.4553e-11, 1.0822e-11,\n 1.5340e-11],\n [1.4052e-12, 2.1550e-11, 3.7486e-11, ..., 1.4449e-11, 6.6697e-12,\n 1.3723e-11],\n [1.4611e-12, 1.2709e-11, 8.2574e-11, ..., 1.1423e-11, 8.2069e-12,\n 1.0319e-11]], device='cuda:0')" }, "14": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "exp_avg": "tensor([5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([1.7587e-07], device='cuda:0')" + "exp_avg_sq": "tensor([5.0255e-08], device='cuda:0')" }, "15": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "exp_avg": "tensor([ 5.6052e-45, -5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([2.2688e-10, 1.2843e-08, 9.6561e-09], device='cuda:0')" + "exp_avg_sq": "tensor([6.4833e-11, 3.6701e-09, 2.7593e-09], device='cuda:0')" }, "16": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([1.3662e-05, 1.4077e-06, 1.5232e-06, 1.6271e-06], device='cuda:0')" + "exp_avg_sq": "tensor([3.9041e-06, 4.0227e-07, 4.3527e-07, 4.6496e-07], device='cuda:0')" }, "18": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([[1.4566e-11, 9.0929e-12, 0.0000e+00, ..., 5.9999e-11, 2.1578e-11,\n 6.9262e-13],\n [3.1077e-12, 7.9886e-12, 0.0000e+00, ..., 7.3445e-12, 2.5826e-11,\n 7.2784e-12],\n [9.3398e-13, 3.2961e-12, 0.0000e+00, ..., 1.1015e-12, 7.1412e-12,\n 2.3388e-12],\n ...,\n [1.1834e-13, 6.3651e-13, 0.0000e+00, ..., 9.0771e-13, 1.2570e-11,\n 7.0816e-14],\n [2.2796e-11, 2.3965e-11, 0.0000e+00, ..., 2.9006e-11, 1.2158e-10,\n 7.2961e-12],\n [7.1204e-14, 5.5338e-14, 0.0000e+00, ..., 1.0953e-13, 1.0864e-12,\n 2.5732e-13]], device='cuda:0')" + "exp_avg_sq": "tensor([[4.1624e-12, 2.5984e-12, 0.0000e+00, ..., 1.7145e-11, 6.1662e-12,\n 1.9792e-13],\n [8.8804e-13, 2.2828e-12, 0.0000e+00, ..., 2.0988e-12, 7.3799e-12,\n 2.0799e-12],\n [2.6689e-13, 9.4189e-13, 0.0000e+00, ..., 3.1477e-13, 2.0407e-12,\n 6.6833e-13],\n ...,\n [3.3818e-14, 1.8189e-13, 0.0000e+00, ..., 2.5939e-13, 3.5921e-12,\n 2.0236e-14],\n [6.5141e-12, 6.8481e-12, 0.0000e+00, ..., 8.2888e-12, 3.4742e-11,\n 2.0849e-12],\n [2.0347e-14, 1.5813e-14, 0.0000e+00, ..., 3.1298e-14, 3.1045e-13,\n 7.3530e-14]], device='cuda:0')" }, "19": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([2.0606e-08, 9.3954e-09, 2.1467e-09, 3.8972e-09, 1.4179e-09, 1.3740e-11,\n 2.4395e-10, 3.8693e-10, 9.8885e-10, 5.5979e-11, 3.1361e-09, 1.5948e-08,\n 6.5719e-09, 7.9609e-09, 1.2019e-09, 7.4954e-09, 1.7865e-08, 1.4808e-09,\n 9.9049e-09, 1.6225e-08, 1.9368e-10, 6.8170e-09, 1.4737e-09, 1.4571e-08,\n 2.9480e-09, 5.9659e-09, 2.6589e-09, 1.6540e-09, 4.5040e-10, 2.7903e-09,\n 8.3472e-09, 4.1833e-09, 1.5200e-10, 3.8372e-08, 1.9896e-08, 9.6730e-09,\n 1.0030e-09, 1.1061e-10, 7.6025e-09, 7.4092e-09, 3.6351e-08, 8.9915e-10,\n 1.6045e-10, 5.0813e-10, 1.2828e-09, 9.3198e-09, 4.0073e-09, 2.6122e-09,\n 2.5418e-09, 2.9797e-09, 1.2189e-10, 4.2907e-09, 1.0898e-08, 6.6486e-09,\n 1.4099e-09, 7.3521e-10, 5.0104e-09, 6.3822e-11, 1.9272e-09, 6.9291e-10,\n 2.1583e-09, 4.7684e-09, 3.0751e-09, 1.7823e-09, 1.7468e-10, 1.1375e-09,\n 4.1364e-09, 2.4332e-10, 5.2138e-09, 4.9074e-09, 4.6936e-11, 1.5843e-09,\n 4.1910e-09, 1.9052e-10, 1.6814e-09, 1.4406e-11, 1.1148e-09, 2.8467e-08,\n 5.6600e-10, 5.2312e-11, 8.7266e-09, 1.1680e-08, 1.9142e-12, 6.5134e-11,\n 2.0808e-09, 7.2274e-09, 2.1746e-10, 3.6989e-09, 4.7418e-10, 8.2244e-09,\n 1.3060e-09, 1.1929e-08, 1.5205e-08, 1.0759e-09, 1.5064e-09, 5.9965e-09,\n 3.9531e-10, 2.3574e-08, 7.8634e-09, 2.8165e-08, 8.0385e-09, 5.5699e-10,\n 1.1115e-10, 3.4015e-09, 2.8269e-10, 2.8079e-08, 7.0484e-10, 1.9970e-09,\n 4.4999e-09, 3.4717e-11, 3.6227e-08, 6.7111e-10, 1.0774e-08, 4.5758e-09,\n 1.5522e-09, 2.1353e-08, 5.3199e-10, 2.5884e-09, 7.1466e-09, 3.4193e-09,\n 3.7299e-09, 6.9018e-09, 6.8537e-10, 1.1521e-10, 4.4927e-11, 2.5097e-09,\n 5.9853e-08, 4.6670e-10, 3.9286e-09, 7.7916e-10, 1.0485e-10, 1.9248e-08,\n 1.6177e-09, 2.7093e-09, 1.0965e-08, 7.8298e-09, 6.6174e-09, 2.0289e-08,\n 8.7483e-09, 1.9246e-12, 7.9800e-11, 1.2948e-10, 7.0455e-09, 2.8913e-08,\n 1.6194e-08, 1.0639e-09, 1.0882e-09, 1.4047e-09, 8.6660e-09, 1.9625e-11,\n 4.4909e-10, 9.0218e-09, 3.1145e-09, 3.4686e-09, 7.7597e-09, 2.6695e-10,\n 2.4374e-09, 6.4040e-09, 1.2441e-09, 2.0743e-08, 9.8351e-09, 1.6139e-09,\n 3.8283e-09, 1.0394e-08, 6.5708e-10, 4.7012e-09, 4.6280e-08, 7.0317e-09,\n 4.8587e-09, 2.2404e-09, 3.2405e-10, 7.7298e-09, 3.1598e-08, 7.9476e-09,\n 1.0445e-08, 7.6081e-11, 1.0531e-09, 3.6363e-09, 4.1439e-09, 4.4426e-09,\n 5.4522e-09, 2.5457e-09, 5.1116e-10, 3.0694e-09, 5.8425e-09, 1.6212e-11,\n 1.5302e-09, 6.2847e-09, 1.5941e-09, 2.5334e-08, 1.8937e-08, 1.9640e-09,\n 1.3001e-09, 5.6315e-10, 3.7583e-12, 9.0749e-08, 1.0635e-09, 1.8311e-09,\n 2.0217e-09, 4.4324e-08, 4.4160e-13, 4.1393e-08, 7.7860e-09, 1.1846e-09,\n 7.5630e-09, 9.2929e-08, 8.5440e-09, 2.6665e-09, 3.8699e-11, 1.1994e-08,\n 9.6838e-10, 1.7504e-08, 2.7255e-09, 5.5532e-09, 4.2509e-11, 7.0713e-09,\n 5.4377e-08, 3.0782e-11, 5.9934e-10, 1.7578e-09, 1.2941e-09, 3.6175e-08,\n 4.5558e-09, 2.6011e-08, 3.6060e-09, 4.3788e-12, 9.2904e-11, 5.3223e-10,\n 1.9726e-10, 7.5946e-09, 6.4931e-11, 6.9400e-11, 1.3183e-08, 5.1555e-09,\n 4.0208e-09, 8.9585e-10, 1.0723e-09, 8.1170e-09, 1.0163e-09, 8.9298e-10,\n 1.0359e-08, 4.2160e-11, 2.2807e-09, 1.3985e-08, 2.4378e-09, 6.9836e-11,\n 1.3051e-10, 5.8687e-09, 6.6259e-10, 2.1107e-08, 1.1004e-08, 1.1057e-09,\n 9.5054e-09, 1.1223e-09, 3.1745e-08, 5.6132e-10], device='cuda:0')" + "exp_avg_sq": "tensor([5.8884e-09, 2.6848e-09, 6.1344e-10, 1.1137e-09, 4.0516e-10, 3.9262e-12,\n 6.9710e-11, 1.1057e-10, 2.8257e-10, 1.5996e-11, 8.9616e-10, 4.5574e-09,\n 1.8780e-09, 2.2749e-09, 3.4345e-10, 2.1419e-09, 5.1050e-09, 4.2316e-10,\n 2.8304e-09, 4.6365e-09, 5.5345e-11, 1.9480e-09, 4.2113e-10, 4.1639e-09,\n 8.4240e-10, 1.7048e-09, 7.5981e-10, 4.7265e-10, 1.2871e-10, 7.9734e-10,\n 2.3853e-09, 1.1954e-09, 4.3436e-11, 1.0965e-08, 5.6853e-09, 2.7641e-09,\n 2.8661e-10, 3.1608e-11, 2.1725e-09, 2.1172e-09, 1.0388e-08, 2.5694e-10,\n 4.5851e-11, 1.4520e-10, 3.6657e-10, 2.6632e-09, 1.1451e-09, 7.4645e-10,\n 7.2635e-10, 8.5148e-10, 3.4832e-11, 1.2261e-09, 3.1142e-09, 1.8999e-09,\n 4.0288e-10, 2.1009e-10, 1.4317e-09, 1.8238e-11, 5.5073e-10, 1.9800e-10,\n 6.1676e-10, 1.3626e-09, 8.7875e-10, 5.0930e-10, 4.9916e-11, 3.2506e-10,\n 1.1820e-09, 6.9532e-11, 1.4899e-09, 1.4023e-09, 1.3412e-11, 4.5272e-10,\n 1.1976e-09, 5.4443e-11, 4.8047e-10, 4.1165e-12, 3.1856e-10, 8.1348e-09,\n 1.6174e-10, 1.4949e-11, 2.4937e-09, 3.3378e-09, 5.4701e-13, 1.8613e-11,\n 5.9460e-10, 2.0653e-09, 6.2141e-11, 1.0570e-09, 1.3550e-10, 2.3502e-09,\n 3.7320e-10, 3.4087e-09, 4.3449e-09, 3.0745e-10, 4.3045e-10, 1.7135e-09,\n 1.1296e-10, 6.7363e-09, 2.2470e-09, 8.0484e-09, 2.2971e-09, 1.5916e-10,\n 3.1761e-11, 9.7199e-10, 8.0781e-11, 8.0239e-09, 2.0141e-10, 5.7067e-10,\n 1.2859e-09, 9.9208e-12, 1.0352e-08, 1.9177e-10, 3.0788e-09, 1.3076e-09,\n 4.4356e-10, 6.1017e-09, 1.5202e-10, 7.3964e-10, 2.0422e-09, 9.7711e-10,\n 1.0659e-09, 1.9722e-09, 1.9585e-10, 3.2921e-11, 1.2838e-11, 7.1717e-10,\n 1.7104e-08, 1.3336e-10, 1.1226e-09, 2.2265e-10, 2.9961e-11, 5.5002e-09,\n 4.6226e-10, 7.7421e-10, 3.1334e-09, 2.2374e-09, 1.8910e-09, 5.7976e-09,\n 2.4999e-09, 5.4997e-13, 2.2804e-11, 3.6999e-11, 2.0133e-09, 8.2621e-09,\n 4.6274e-09, 3.0402e-10, 3.1095e-10, 4.0142e-10, 2.4764e-09, 5.6079e-12,\n 1.2833e-10, 2.5781e-09, 8.8999e-10, 9.9118e-10, 2.2174e-09, 7.6284e-11,\n 6.9651e-10, 1.8300e-09, 3.5552e-10, 5.9274e-09, 2.8105e-09, 4.6119e-10,\n 1.0940e-09, 2.9702e-09, 1.8776e-10, 1.3434e-09, 1.3225e-08, 2.0094e-09,\n 1.3884e-09, 6.4021e-10, 9.2599e-11, 2.2089e-09, 9.0295e-09, 2.2711e-09,\n 2.9846e-09, 2.1741e-11, 3.0093e-10, 1.0391e-09, 1.1841e-09, 1.2695e-09,\n 1.5580e-09, 7.2746e-10, 1.4607e-10, 8.7711e-10, 1.6695e-09, 4.6328e-12,\n 4.3728e-10, 1.7959e-09, 4.5554e-10, 7.2394e-09, 5.4115e-09, 5.6123e-10,\n 3.7152e-10, 1.6093e-10, 1.0740e-12, 2.5932e-08, 3.0391e-10, 5.2324e-10,\n 5.7770e-10, 1.2666e-08, 1.2619e-13, 1.1828e-08, 2.2249e-09, 3.3851e-10,\n 2.1612e-09, 2.6555e-08, 2.4415e-09, 7.6196e-10, 1.1058e-11, 3.4275e-09,\n 2.7672e-10, 5.0019e-09, 7.7882e-10, 1.5869e-09, 1.2147e-11, 2.0207e-09,\n 1.5539e-08, 8.7963e-12, 1.7127e-10, 5.0232e-10, 3.6981e-10, 1.0337e-08,\n 1.3019e-09, 7.4329e-09, 1.0304e-09, 1.2513e-12, 2.6548e-11, 1.5209e-10,\n 5.6369e-11, 2.1702e-09, 1.8555e-11, 1.9832e-11, 3.7671e-09, 1.4732e-09,\n 1.1490e-09, 2.5600e-10, 3.0643e-10, 2.3195e-09, 2.9040e-10, 2.5518e-10,\n 2.9603e-09, 1.2048e-11, 6.5174e-10, 3.9963e-09, 6.9663e-10, 1.9956e-11,\n 3.7295e-11, 1.6770e-09, 1.8934e-10, 6.0316e-09, 3.1446e-09, 3.1596e-10,\n 2.7163e-09, 3.2072e-10, 9.0715e-09, 1.6040e-10], device='cuda:0')" }, "20": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([5.0837e-11, 4.6489e-11, 3.1264e-12, 2.0402e-11, 2.3935e-12, 1.9740e-13,\n 4.3405e-13, 2.0156e-12, 1.1969e-12, 7.0310e-14, 5.7767e-12, 2.8928e-11,\n 1.6655e-11, 3.7466e-11, 3.8567e-12, 1.7655e-11, 5.7906e-11, 2.5684e-12,\n 3.2040e-11, 4.6163e-11, 9.5665e-14, 2.8674e-11, 3.2421e-12, 5.9469e-11,\n 1.1559e-11, 1.1579e-11, 6.7244e-12, 5.4956e-12, 1.6345e-12, 4.9386e-12,\n 2.5492e-11, 1.2411e-11, 9.6807e-13, 1.0331e-10, 7.7087e-11, 2.4123e-11,\n 1.9913e-12, 2.0152e-15, 1.5798e-11, 1.3394e-11, 1.5376e-10, 2.1449e-12,\n 6.8263e-13, 2.2293e-12, 2.5628e-12, 2.0965e-11, 8.7288e-12, 5.5107e-12,\n 3.8874e-12, 4.6039e-12, 9.5507e-14, 1.1724e-11, 2.2626e-11, 1.4493e-11,\n 1.9313e-12, 3.4671e-12, 1.6209e-11, 1.3354e-13, 6.3606e-12, 1.9626e-12,\n 3.8969e-12, 3.5572e-11, 8.6256e-12, 4.2808e-12, 2.6434e-15, 1.8173e-12,\n 9.0315e-12, 1.3211e-13, 9.6269e-12, 1.5888e-11, 1.1045e-12, 3.6402e-12,\n 4.6533e-12, 6.3967e-13, 3.0752e-12, 9.3100e-13, 6.5003e-12, 1.4454e-10,\n 4.9540e-13, 1.4783e-16, 3.6552e-11, 3.0283e-11, 2.7830e-13, 3.2177e-13,\n 4.0930e-12, 1.6148e-11, 8.6660e-13, 1.7316e-11, 1.3897e-12, 1.5336e-11,\n 3.9559e-12, 2.7258e-11, 5.5273e-11, 1.3642e-12, 4.1675e-12, 2.1135e-11,\n 8.5597e-13, 5.0211e-11, 3.0602e-11, 7.9197e-11, 2.4092e-11, 2.5930e-12,\n 1.3592e-12, 6.0288e-12, 1.1792e-13, 6.5984e-11, 7.3680e-13, 4.7081e-12,\n 9.3694e-12, 4.9065e-14, 9.2464e-11, 2.0920e-12, 3.0535e-11, 9.6079e-12,\n 2.8324e-12, 8.6184e-11, 2.0960e-12, 5.5189e-12, 2.1548e-11, 8.2947e-12,\n 6.8629e-12, 1.4374e-11, 5.7599e-13, 1.1117e-14, 4.5848e-13, 3.8622e-12,\n 2.7065e-10, 1.7699e-12, 1.2548e-11, 1.1884e-12, 3.8420e-14, 5.7402e-11,\n 3.9586e-12, 8.0607e-12, 2.1048e-11, 1.7682e-11, 1.8587e-11, 5.2364e-11,\n 2.3933e-11, 3.2469e-13, 2.0464e-12, 7.4192e-13, 1.5481e-11, 8.4985e-11,\n 1.0176e-10, 4.2116e-12, 4.3713e-12, 3.2883e-12, 1.7952e-11, 4.9772e-15,\n 4.4954e-12, 1.9783e-11, 1.0056e-11, 3.9488e-12, 1.8204e-11, 1.6173e-12,\n 3.5971e-12, 1.0911e-11, 2.0467e-12, 8.2042e-11, 1.9179e-11, 1.8265e-12,\n 8.9190e-12, 3.4649e-11, 1.9094e-12, 6.1096e-12, 1.5909e-10, 1.3336e-11,\n 1.4868e-11, 5.7095e-12, 1.6607e-13, 3.0832e-11, 7.6113e-11, 2.9662e-11,\n 1.9661e-11, 2.9267e-13, 1.6457e-12, 1.6265e-11, 1.2693e-11, 1.1603e-11,\n 1.4080e-11, 4.9266e-12, 6.2074e-13, 4.6085e-12, 1.4626e-11, 8.7429e-18,\n 2.3574e-12, 1.3981e-11, 1.4039e-12, 6.3436e-11, 4.9428e-11, 2.4737e-12,\n 3.8664e-12, 3.0811e-12, 3.5427e-13, 3.5179e-10, 1.3539e-12, 2.9321e-12,\n 6.1185e-12, 1.2182e-10, 1.0495e-13, 1.1122e-10, 1.4050e-11, 1.6182e-12,\n 2.5157e-11, 2.3684e-10, 1.8507e-11, 6.7287e-12, 1.2685e-14, 1.8743e-11,\n 1.3265e-12, 4.3632e-11, 1.8506e-11, 1.6947e-11, 2.5405e-14, 1.2280e-11,\n 2.5870e-10, 1.0554e-14, 2.2501e-12, 2.1057e-12, 1.6530e-12, 1.5912e-10,\n 2.5351e-11, 6.3299e-11, 1.0661e-11, 9.3165e-14, 4.5158e-15, 1.5781e-12,\n 1.9538e-13, 2.1354e-11, 5.2482e-13, 4.8876e-14, 7.7762e-11, 2.2765e-11,\n 1.2169e-11, 2.0868e-12, 1.3716e-12, 2.3608e-11, 2.7892e-12, 8.9159e-13,\n 2.5651e-11, 9.8212e-15, 6.8431e-12, 3.6019e-11, 3.6967e-12, 1.1972e-14,\n 3.0740e-14, 1.1226e-11, 1.0451e-12, 5.6963e-11, 2.2360e-11, 1.5541e-12,\n 2.1291e-11, 2.2211e-12, 1.0166e-10, 2.9559e-12], device='cuda:0')" + "exp_avg_sq": "tensor([1.4527e-11, 1.3285e-11, 8.9339e-13, 5.8300e-12, 6.8396e-13, 5.6409e-14,\n 1.2403e-13, 5.7598e-13, 3.4202e-13, 2.0092e-14, 1.6507e-12, 8.2663e-12,\n 4.7593e-12, 1.0706e-11, 1.1021e-12, 5.0450e-12, 1.6547e-11, 7.3393e-13,\n 9.1555e-12, 1.3191e-11, 2.7337e-14, 8.1938e-12, 9.2646e-13, 1.6994e-11,\n 3.3031e-12, 3.3089e-12, 1.9216e-12, 1.5704e-12, 4.6707e-13, 1.4112e-12,\n 7.2846e-12, 3.5466e-12, 2.7663e-13, 2.9521e-11, 2.2028e-11, 6.8934e-12,\n 5.6902e-13, 5.7587e-16, 4.5143e-12, 3.8274e-12, 4.3938e-11, 6.1294e-13,\n 1.9507e-13, 6.3703e-13, 7.3233e-13, 5.9909e-12, 2.4943e-12, 1.5747e-12,\n 1.1109e-12, 1.3156e-12, 2.7292e-14, 3.3501e-12, 6.4656e-12, 4.1416e-12,\n 5.5188e-13, 9.9076e-13, 4.6320e-12, 3.8159e-14, 1.8176e-12, 5.6083e-13,\n 1.1136e-12, 1.0165e-11, 2.4648e-12, 1.2233e-12, 7.5537e-16, 5.1930e-13,\n 2.5808e-12, 3.7751e-14, 2.7510e-12, 4.5402e-12, 3.1561e-13, 1.0402e-12,\n 1.3297e-12, 1.8279e-13, 8.7875e-13, 2.6604e-13, 1.8575e-12, 4.1304e-11,\n 1.4156e-13, 4.2244e-17, 1.0445e-11, 8.6535e-12, 7.9526e-14, 9.1947e-14,\n 1.1696e-12, 4.6143e-12, 2.4764e-13, 4.9481e-12, 3.9712e-13, 4.3823e-12,\n 1.1304e-12, 7.7891e-12, 1.5795e-11, 3.8982e-13, 1.1909e-12, 6.0395e-12,\n 2.4460e-13, 1.4348e-11, 8.7449e-12, 2.2631e-11, 6.8845e-12, 7.4097e-13,\n 3.8840e-13, 1.7228e-12, 3.3697e-14, 1.8856e-11, 2.1055e-13, 1.3454e-12,\n 2.6774e-12, 1.4021e-14, 2.6422e-11, 5.9779e-13, 8.7255e-12, 2.7455e-12,\n 8.0938e-13, 2.4628e-11, 5.9894e-13, 1.5771e-12, 6.1575e-12, 2.3703e-12,\n 1.9611e-12, 4.1075e-12, 1.6460e-13, 3.1767e-15, 1.3101e-13, 1.1037e-12,\n 7.7339e-11, 5.0577e-13, 3.5856e-12, 3.3958e-13, 1.0979e-14, 1.6403e-11,\n 1.1312e-12, 2.3034e-12, 6.0147e-12, 5.0528e-12, 5.3113e-12, 1.4963e-11,\n 6.8389e-12, 9.2784e-14, 5.8477e-13, 2.1201e-13, 4.4238e-12, 2.4285e-11,\n 2.9077e-11, 1.2035e-12, 1.2491e-12, 9.3964e-13, 5.1298e-12, 1.4223e-15,\n 1.2846e-12, 5.6530e-12, 2.8735e-12, 1.1284e-12, 5.2019e-12, 4.6215e-13,\n 1.0279e-12, 3.1178e-12, 5.8485e-13, 2.3444e-11, 5.4806e-12, 5.2193e-13,\n 2.5487e-12, 9.9013e-12, 5.4564e-13, 1.7459e-12, 4.5460e-11, 3.8110e-12,\n 4.2485e-12, 1.6315e-12, 4.7454e-14, 8.8104e-12, 2.1750e-11, 8.4762e-12,\n 5.6181e-12, 8.3633e-14, 4.7028e-13, 4.6479e-12, 3.6271e-12, 3.3156e-12,\n 4.0235e-12, 1.4078e-12, 1.7738e-13, 1.3169e-12, 4.1794e-12, 2.4984e-18,\n 6.7364e-13, 3.9953e-12, 4.0118e-13, 1.8127e-11, 1.4124e-11, 7.0688e-13,\n 1.1049e-12, 8.8046e-13, 1.0124e-13, 1.0053e-10, 3.8690e-13, 8.3786e-13,\n 1.7484e-12, 3.4811e-11, 2.9989e-14, 3.1781e-11, 4.0149e-12, 4.6241e-13,\n 7.1888e-12, 6.7679e-11, 5.2885e-12, 1.9228e-12, 3.6249e-15, 5.3560e-12,\n 3.7905e-13, 1.2468e-11, 5.2883e-12, 4.8427e-12, 7.2597e-15, 3.5091e-12,\n 7.3927e-11, 3.0160e-15, 6.4300e-13, 6.0172e-13, 4.7235e-13, 4.5471e-11,\n 7.2442e-12, 1.8088e-11, 3.0464e-12, 2.6623e-14, 1.2904e-15, 4.5095e-13,\n 5.5831e-14, 6.1022e-12, 1.4997e-13, 1.3967e-14, 2.2221e-11, 6.5054e-12,\n 3.4774e-12, 5.9633e-13, 3.9194e-13, 6.7463e-12, 7.9703e-13, 2.5478e-13,\n 7.3299e-12, 2.8065e-15, 1.9555e-12, 1.0293e-11, 1.0564e-12, 3.4212e-15,\n 8.7842e-15, 3.2079e-12, 2.9863e-13, 1.6278e-11, 6.3897e-12, 4.4410e-13,\n 6.0841e-12, 6.3471e-13, 2.9051e-11, 8.4468e-13], device='cuda:0')" }, "21": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([8.7193e-11, 4.4508e-11, 6.1277e-12, 2.0029e-11, 3.1496e-12, 1.3380e-13,\n 6.0683e-13, 3.2961e-12, 2.0118e-12, 3.9426e-14, 9.5755e-12, 5.4047e-11,\n 1.9061e-11, 3.7143e-11, 6.4909e-12, 3.4239e-11, 7.7432e-11, 4.0941e-12,\n 2.8456e-11, 6.7186e-11, 1.0816e-13, 3.2105e-11, 7.4311e-12, 6.2731e-11,\n 1.5443e-11, 1.7882e-11, 1.2383e-11, 8.6888e-12, 3.3658e-12, 7.1992e-12,\n 3.7434e-11, 1.9476e-11, 1.5138e-12, 1.3206e-10, 8.5087e-11, 4.1555e-11,\n 1.4390e-12, 4.1827e-16, 3.3554e-11, 2.5484e-11, 1.1825e-10, 4.9894e-12,\n 1.6464e-12, 3.9380e-12, 4.3938e-12, 2.9960e-11, 1.8505e-11, 1.2898e-11,\n 6.3870e-12, 9.6134e-12, 1.5608e-13, 1.8970e-11, 3.5524e-11, 2.0838e-11,\n 3.5157e-12, 4.7561e-12, 2.3158e-11, 1.2040e-13, 1.0739e-11, 4.1243e-12,\n 5.1648e-12, 2.5052e-11, 1.4068e-11, 8.1531e-12, 4.3431e-14, 2.8430e-12,\n 1.9556e-11, 3.5698e-13, 2.3180e-11, 2.3782e-11, 1.2142e-12, 8.0418e-12,\n 1.4568e-11, 1.3804e-12, 3.9982e-12, 8.5327e-13, 6.9230e-12, 1.2580e-10,\n 7.4656e-13, 1.1864e-16, 4.2155e-11, 5.2598e-11, 3.0659e-13, 6.9611e-13,\n 1.0607e-11, 3.3656e-11, 1.5304e-12, 1.9584e-11, 1.5377e-12, 3.6193e-11,\n 6.8366e-12, 5.1342e-11, 4.8966e-11, 2.8534e-12, 6.7788e-12, 1.6196e-11,\n 1.1227e-12, 8.0495e-11, 3.6871e-11, 1.1796e-10, 3.5435e-11, 3.8705e-12,\n 1.5501e-12, 9.2719e-12, 1.7419e-13, 9.2677e-11, 1.2553e-12, 1.0451e-11,\n 1.2663e-11, 3.2559e-14, 1.2166e-10, 4.0837e-12, 4.8465e-11, 1.6859e-11,\n 3.7283e-12, 6.7451e-11, 2.8814e-12, 6.9796e-12, 3.2689e-11, 1.6531e-11,\n 1.0569e-11, 2.0359e-11, 1.0597e-12, 3.0656e-14, 8.5377e-13, 7.1995e-12,\n 2.4062e-10, 2.6192e-12, 1.9836e-11, 2.1940e-12, 4.4356e-14, 8.3359e-11,\n 8.2576e-12, 1.2836e-11, 3.4564e-11, 2.5553e-11, 2.9313e-11, 6.8423e-11,\n 3.8947e-11, 3.9599e-13, 1.5474e-12, 1.2136e-12, 2.2550e-11, 9.8913e-11,\n 7.5255e-11, 6.3054e-12, 6.2416e-12, 7.3417e-12, 2.8414e-11, 1.1112e-14,\n 4.1223e-12, 2.7142e-11, 1.5787e-11, 9.9493e-12, 2.1545e-11, 2.4589e-12,\n 6.1008e-12, 2.3217e-11, 3.2995e-12, 8.7918e-11, 3.2296e-11, 4.4906e-12,\n 9.0052e-12, 4.6745e-11, 3.5309e-12, 1.3427e-11, 1.6525e-10, 2.2414e-11,\n 2.2273e-11, 1.0944e-11, 4.5643e-13, 2.0129e-11, 1.0838e-10, 3.7671e-11,\n 3.1408e-11, 5.9795e-13, 2.4481e-12, 1.9439e-11, 9.4806e-12, 2.1240e-11,\n 1.4400e-11, 6.3610e-12, 1.0352e-12, 8.3857e-12, 2.4494e-11, 2.5134e-15,\n 4.6966e-12, 1.7934e-11, 4.2851e-12, 1.0400e-10, 6.3202e-11, 5.1856e-12,\n 7.0746e-12, 3.7169e-12, 4.6532e-13, 3.7237e-10, 2.1514e-12, 3.9536e-12,\n 9.8160e-12, 1.5619e-10, 2.6248e-13, 1.7233e-10, 2.5053e-11, 2.6300e-12,\n 3.5473e-11, 3.2130e-10, 2.5606e-11, 6.2672e-12, 3.2042e-14, 4.1699e-11,\n 1.6047e-12, 7.5488e-11, 1.5535e-11, 2.6827e-11, 3.0372e-14, 2.3587e-11,\n 1.8260e-10, 7.1149e-15, 3.7872e-12, 3.9228e-12, 2.4038e-12, 1.5170e-10,\n 2.3632e-11, 1.1067e-10, 1.8530e-11, 1.7153e-13, 3.0955e-15, 3.5857e-12,\n 4.9844e-13, 3.2643e-11, 1.0499e-12, 1.0557e-13, 5.8959e-11, 2.4372e-11,\n 1.8550e-11, 4.4858e-12, 2.7769e-12, 3.6414e-11, 5.7864e-12, 2.1648e-12,\n 4.6752e-11, 9.2814e-15, 1.1101e-11, 4.3758e-11, 6.9385e-12, 2.3144e-14,\n 2.8395e-14, 1.8693e-11, 1.0436e-12, 9.0881e-11, 3.5112e-11, 2.5757e-12,\n 2.8538e-11, 3.7685e-12, 1.0527e-10, 4.1190e-12], device='cuda:0')" + "exp_avg_sq": "tensor([2.4916e-11, 1.2718e-11, 1.7510e-12, 5.7235e-12, 9.0003e-13, 3.8233e-14,\n 1.7341e-13, 9.4189e-13, 5.7490e-13, 1.1266e-14, 2.7363e-12, 1.5444e-11,\n 5.4467e-12, 1.0614e-11, 1.8548e-12, 9.7841e-12, 2.2127e-11, 1.1699e-12,\n 8.1315e-12, 1.9199e-11, 3.0907e-14, 9.1743e-12, 2.1235e-12, 1.7926e-11,\n 4.4130e-12, 5.1100e-12, 3.5385e-12, 2.4829e-12, 9.6182e-13, 2.0572e-12,\n 1.0697e-11, 5.5655e-12, 4.3257e-13, 3.7738e-11, 2.4314e-11, 1.1875e-11,\n 4.1120e-13, 1.1952e-16, 9.5883e-12, 7.2822e-12, 3.3792e-11, 1.4258e-12,\n 4.7048e-13, 1.1253e-12, 1.2556e-12, 8.5612e-12, 5.2879e-12, 3.6857e-12,\n 1.8251e-12, 2.7471e-12, 4.4602e-14, 5.4208e-12, 1.0151e-11, 5.9546e-12,\n 1.0046e-12, 1.3591e-12, 6.6175e-12, 3.4407e-14, 3.0687e-12, 1.1785e-12,\n 1.4759e-12, 7.1589e-12, 4.0202e-12, 2.3298e-12, 1.2411e-14, 8.1240e-13,\n 5.5883e-12, 1.0201e-13, 6.6240e-12, 6.7960e-12, 3.4696e-13, 2.2980e-12,\n 4.1628e-12, 3.9446e-13, 1.1425e-12, 2.4383e-13, 1.9783e-12, 3.5948e-11,\n 2.1333e-13, 3.3902e-17, 1.2046e-11, 1.5030e-11, 8.7610e-14, 1.9892e-13,\n 3.0311e-12, 9.6174e-12, 4.3731e-13, 5.5962e-12, 4.3941e-13, 1.0343e-11,\n 1.9536e-12, 1.4671e-11, 1.3992e-11, 8.1539e-13, 1.9371e-12, 4.6281e-12,\n 3.2081e-13, 2.3002e-11, 1.0536e-11, 3.3709e-11, 1.0126e-11, 1.1060e-12,\n 4.4294e-13, 2.6495e-12, 4.9776e-14, 2.6483e-11, 3.5870e-13, 2.9864e-12,\n 3.6185e-12, 9.3040e-15, 3.4764e-11, 1.1670e-12, 1.3849e-11, 4.8175e-12,\n 1.0654e-12, 1.9275e-11, 8.2337e-13, 1.9945e-12, 9.3412e-12, 4.7238e-12,\n 3.0203e-12, 5.8176e-12, 3.0281e-13, 8.7602e-15, 2.4397e-13, 2.0573e-12,\n 6.8758e-11, 7.4846e-13, 5.6681e-12, 6.2695e-13, 1.2675e-14, 2.3820e-11,\n 2.3597e-12, 3.6678e-12, 9.8771e-12, 7.3019e-12, 8.3765e-12, 1.9552e-11,\n 1.1129e-11, 1.1316e-13, 4.4217e-13, 3.4678e-13, 6.4438e-12, 2.8265e-11,\n 2.1505e-11, 1.8018e-12, 1.7836e-12, 2.0980e-12, 8.1194e-12, 3.1753e-15,\n 1.1780e-12, 7.7562e-12, 4.5114e-12, 2.8431e-12, 6.1566e-12, 7.0266e-13,\n 1.7433e-12, 6.6344e-12, 9.4285e-13, 2.5123e-11, 9.2290e-12, 1.2832e-12,\n 2.5733e-12, 1.3358e-11, 1.0090e-12, 3.8369e-12, 4.7221e-11, 6.4049e-12,\n 6.3647e-12, 3.1275e-12, 1.3043e-13, 5.7521e-12, 3.0972e-11, 1.0765e-11,\n 8.9751e-12, 1.7087e-13, 6.9957e-13, 5.5549e-12, 2.7092e-12, 6.0695e-12,\n 4.1149e-12, 1.8177e-12, 2.9580e-13, 2.3963e-12, 6.9995e-12, 7.1822e-16,\n 1.3421e-12, 5.1248e-12, 1.2245e-12, 2.9718e-11, 1.8061e-11, 1.4818e-12,\n 2.0216e-12, 1.0621e-12, 1.3297e-13, 1.0641e-10, 6.1478e-13, 1.1298e-12,\n 2.8050e-12, 4.4633e-11, 7.5005e-14, 4.9246e-11, 7.1592e-12, 7.5153e-13,\n 1.0137e-11, 9.1815e-11, 7.3171e-12, 1.7909e-12, 9.1561e-15, 1.1916e-11,\n 4.5854e-13, 2.1571e-11, 4.4393e-12, 7.6660e-12, 8.6792e-15, 6.7402e-12,\n 5.2179e-11, 2.0332e-15, 1.0822e-12, 1.1210e-12, 6.8690e-13, 4.3349e-11,\n 6.7531e-12, 3.1624e-11, 5.2950e-12, 4.9017e-14, 8.8456e-16, 1.0246e-12,\n 1.4243e-13, 9.3281e-12, 3.0003e-13, 3.0167e-14, 1.6848e-11, 6.9646e-12,\n 5.3008e-12, 1.2818e-12, 7.9352e-13, 1.0406e-11, 1.6535e-12, 6.1861e-13,\n 1.3360e-11, 2.6522e-15, 3.1722e-12, 1.2504e-11, 1.9827e-12, 6.6136e-15,\n 8.1141e-15, 5.3418e-12, 2.9822e-13, 2.5970e-11, 1.0034e-11, 7.3603e-13,\n 8.1550e-12, 1.0769e-12, 3.0082e-11, 1.1770e-12], device='cuda:0')" }, "22": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([[6.1609e-12, 2.3143e-11, 0.0000e+00, ..., 3.3698e-11, 5.5105e-11,\n 8.4312e-12],\n [2.8259e-12, 3.1539e-12, 0.0000e+00, ..., 5.2046e-12, 1.6054e-11,\n 1.0205e-14],\n [3.7940e-12, 2.0610e-12, 0.0000e+00, ..., 3.4659e-12, 7.4052e-12,\n 5.9638e-13],\n ...,\n [6.4804e-12, 7.9899e-12, 0.0000e+00, ..., 9.8687e-12, 1.8674e-11,\n 5.9402e-12],\n [1.9731e-11, 6.2648e-12, 0.0000e+00, ..., 6.2276e-12, 3.9516e-11,\n 3.7619e-12],\n [9.6579e-14, 1.3525e-13, 0.0000e+00, ..., 1.4584e-12, 2.3245e-12,\n 4.7443e-13]], device='cuda:0')" + "exp_avg_sq": "tensor([[1.7605e-12, 6.6133e-12, 0.0000e+00, ..., 9.6295e-12, 1.5747e-11,\n 2.4093e-12],\n [8.0753e-13, 9.0125e-13, 0.0000e+00, ..., 1.4873e-12, 4.5876e-12,\n 2.9161e-15],\n [1.0842e-12, 5.8895e-13, 0.0000e+00, ..., 9.9041e-13, 2.1161e-12,\n 1.7042e-13],\n ...,\n [1.8518e-12, 2.2832e-12, 0.0000e+00, ..., 2.8201e-12, 5.3362e-12,\n 1.6975e-12],\n [5.6383e-12, 1.7902e-12, 0.0000e+00, ..., 1.7796e-12, 1.1292e-11,\n 1.0750e-12],\n [2.7598e-14, 3.8650e-14, 0.0000e+00, ..., 4.1676e-13, 6.6423e-13,\n 1.3557e-13]], device='cuda:0')" }, "23": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([2.4068e-08, 1.8525e-09, 4.0916e-09, 3.0792e-09, 1.0500e-09, 5.6321e-10,\n 2.4226e-09, 1.4414e-12, 9.6605e-10, 1.6385e-10, 1.8190e-08, 3.6018e-09,\n 4.8459e-09, 3.1442e-09, 9.2011e-10, 2.2707e-09, 9.5160e-09, 6.2247e-10,\n 1.8016e-09, 1.9190e-08, 7.5750e-11, 7.2855e-09, 3.1714e-09, 1.8322e-09,\n 2.4519e-10, 7.6415e-09, 2.4138e-09, 3.8378e-10, 2.6835e-10, 2.1374e-09,\n 1.0558e-08, 2.1983e-09, 2.6967e-10, 7.7247e-08, 7.6302e-09, 5.3465e-09,\n 9.1561e-10, 6.6355e-14, 7.1345e-09, 3.5982e-08, 4.5253e-09, 2.7623e-09,\n 4.8831e-10, 2.8092e-09, 1.6403e-08, 1.2171e-08, 6.8825e-09, 4.0271e-09,\n 1.0146e-09, 1.7478e-08, 1.9599e-10, 3.1591e-08, 2.1897e-08, 4.9939e-09,\n 5.0674e-10, 9.1554e-11, 1.2013e-08, 3.8195e-10, 5.4853e-09, 1.1060e-09,\n 2.5813e-09, 2.9600e-09, 1.5748e-08, 8.3587e-09, 3.8140e-10, 9.8680e-10,\n 5.8901e-09, 1.0504e-09, 8.2478e-09, 1.9830e-10, 2.9131e-13, 3.3840e-09,\n 6.2236e-09, 1.0933e-10, 2.6395e-09, 3.2257e-10, 9.1474e-10, 1.7556e-08,\n 2.5811e-10, 1.1033e-10, 1.4105e-09, 1.2370e-08, 3.5278e-13, 9.1676e-10,\n 2.5687e-10, 9.3183e-09, 3.7498e-09, 2.1012e-09, 3.7071e-09, 5.4300e-08,\n 7.2739e-10, 1.2243e-08, 1.5614e-08, 2.3614e-09, 2.9628e-08, 2.2137e-09,\n 2.6213e-09, 1.6643e-08, 5.5929e-09, 6.5217e-09, 1.4488e-08, 8.1564e-10,\n 1.5711e-10, 5.8532e-09, 2.1691e-10, 3.1242e-08, 1.7600e-09, 3.5624e-09,\n 1.0882e-09, 5.0319e-12, 2.5549e-08, 1.8281e-09, 2.5374e-08, 7.7458e-09,\n 8.4746e-10, 1.4096e-09, 1.6795e-09, 1.4273e-09, 1.0769e-09, 2.6173e-09,\n 7.8344e-09, 4.5816e-09, 7.7843e-10, 6.2094e-13, 4.6075e-12, 2.8193e-09,\n 1.5686e-08, 1.0171e-08, 8.2302e-10, 1.3666e-08, 1.0723e-11, 6.7534e-08,\n 3.0725e-09, 5.7113e-09, 9.7981e-09, 2.6307e-08, 8.3463e-09, 2.6521e-08,\n 7.2343e-09, 3.3002e-12, 9.4631e-11, 4.9411e-10, 3.1760e-09, 2.8597e-08,\n 4.9864e-09, 1.3503e-09, 1.1949e-09, 4.6312e-10, 2.8427e-08, 6.1040e-10,\n 4.9852e-10, 2.3746e-09, 2.5947e-09, 1.0327e-08, 3.5206e-09, 4.7152e-10,\n 2.4857e-09, 2.6399e-08, 7.4635e-10, 3.6347e-08, 9.2380e-09, 1.3430e-08,\n 1.4019e-09, 1.2221e-08, 9.3478e-09, 7.7021e-09, 7.4211e-08, 1.0440e-08,\n 1.1482e-08, 1.4849e-09, 4.7364e-10, 1.9972e-09, 1.3764e-08, 4.2342e-09,\n 7.8969e-09, 8.4580e-11, 6.9569e-10, 4.6405e-10, 9.4203e-10, 3.3164e-09,\n 4.8717e-10, 2.6873e-09, 1.7735e-09, 2.3770e-09, 1.8339e-08, 6.2330e-11,\n 1.8575e-09, 5.6874e-09, 1.0923e-09, 2.9482e-08, 3.1086e-09, 4.4414e-09,\n 4.7966e-09, 3.7903e-10, 1.6125e-12, 8.4350e-08, 1.0028e-09, 1.0395e-09,\n 6.2056e-09, 1.6158e-08, 6.5511e-13, 4.5870e-09, 1.7383e-08, 2.0567e-10,\n 1.0448e-08, 3.4773e-08, 5.4796e-09, 1.1255e-09, 4.2040e-12, 4.7524e-08,\n 9.7694e-10, 1.6298e-08, 5.0081e-10, 2.2393e-09, 3.7892e-10, 4.5201e-09,\n 5.9757e-09, 2.0549e-11, 4.5472e-10, 5.8861e-10, 1.9657e-09, 7.6375e-09,\n 1.4150e-09, 2.7077e-08, 4.7726e-09, 1.9497e-10, 6.0251e-11, 2.9102e-10,\n 1.0816e-09, 2.4785e-09, 7.5208e-11, 2.6411e-10, 4.3182e-09, 3.4969e-09,\n 1.4047e-08, 4.4011e-09, 1.0287e-09, 7.8449e-09, 1.1734e-09, 2.0582e-09,\n 7.5240e-09, 1.5629e-11, 9.8400e-09, 1.2209e-08, 4.2397e-09, 9.6427e-12,\n 3.3672e-13, 4.2070e-08, 2.0311e-10, 1.5460e-08, 5.3040e-09, 9.1214e-10,\n 6.9643e-09, 7.2863e-09, 1.6912e-08, 4.7694e-10], device='cuda:0')" + "exp_avg_sq": "tensor([6.8776e-09, 5.2936e-10, 1.1692e-09, 8.7991e-10, 3.0004e-10, 1.6094e-10,\n 6.9228e-10, 4.1189e-13, 2.7605e-10, 4.6823e-11, 5.1980e-09, 1.0292e-09,\n 1.3848e-09, 8.9847e-10, 2.6293e-10, 6.4888e-10, 2.7193e-09, 1.7788e-10,\n 5.1481e-10, 5.4836e-09, 2.1646e-11, 2.0819e-09, 9.0626e-10, 5.2356e-10,\n 7.0064e-11, 2.1836e-09, 6.8977e-10, 1.0967e-10, 7.6684e-11, 6.1078e-10,\n 3.0170e-09, 6.2819e-10, 7.7061e-11, 2.2074e-08, 2.1804e-09, 1.5278e-09,\n 2.6164e-10, 1.8962e-14, 2.0387e-09, 1.0282e-08, 1.2931e-09, 7.8935e-10,\n 1.3954e-10, 8.0274e-10, 4.6873e-09, 3.4779e-09, 1.9667e-09, 1.1508e-09,\n 2.8993e-10, 4.9945e-09, 5.6004e-11, 9.0275e-09, 6.2572e-09, 1.4270e-09,\n 1.4481e-10, 2.6162e-11, 3.4329e-09, 1.0915e-10, 1.5675e-09, 3.1605e-10,\n 7.3762e-10, 8.4586e-10, 4.5001e-09, 2.3886e-09, 1.0899e-10, 2.8199e-10,\n 1.6831e-09, 3.0017e-10, 2.3569e-09, 5.6665e-11, 8.3243e-14, 9.6702e-10,\n 1.7784e-09, 3.1241e-11, 7.5426e-10, 9.2178e-11, 2.6140e-10, 5.0168e-09,\n 7.3756e-11, 3.1528e-11, 4.0306e-10, 3.5348e-09, 1.0081e-13, 2.6197e-10,\n 7.3403e-11, 2.6628e-09, 1.0715e-09, 6.0043e-10, 1.0593e-09, 1.5517e-08,\n 2.0786e-10, 3.4986e-09, 4.4618e-09, 6.7479e-10, 8.4663e-09, 6.3259e-10,\n 7.4906e-10, 4.7558e-09, 1.5982e-09, 1.8636e-09, 4.1400e-09, 2.3308e-10,\n 4.4895e-11, 1.6726e-09, 6.1984e-11, 8.9277e-09, 5.0292e-10, 1.0180e-09,\n 3.1095e-10, 1.4379e-12, 7.3009e-09, 5.2240e-10, 7.2509e-09, 2.2134e-09,\n 2.4217e-10, 4.0281e-10, 4.7992e-10, 4.0785e-10, 3.0774e-10, 7.4792e-10,\n 2.2387e-09, 1.3092e-09, 2.2244e-10, 1.7744e-13, 1.3166e-12, 8.0565e-10,\n 4.4824e-09, 2.9065e-09, 2.3518e-10, 3.9050e-09, 3.0641e-12, 1.9298e-08,\n 8.7798e-10, 1.6321e-09, 2.7999e-09, 7.5174e-09, 2.3850e-09, 7.5785e-09,\n 2.0673e-09, 9.4305e-13, 2.7042e-11, 1.4120e-10, 9.0755e-10, 8.1719e-09,\n 1.4249e-09, 3.8585e-10, 3.4145e-10, 1.3234e-10, 8.1232e-09, 1.7443e-10,\n 1.4245e-10, 6.7857e-10, 7.4146e-10, 2.9511e-09, 1.0060e-09, 1.3474e-10,\n 7.1030e-10, 7.5437e-09, 2.1328e-10, 1.0386e-08, 2.6398e-09, 3.8376e-09,\n 4.0062e-10, 3.4921e-09, 2.6712e-09, 2.2009e-09, 2.1206e-08, 2.9832e-09,\n 3.2810e-09, 4.2431e-10, 1.3535e-10, 5.7071e-10, 3.9331e-09, 1.2100e-09,\n 2.2566e-09, 2.4169e-11, 1.9880e-10, 1.3260e-10, 2.6919e-10, 9.4768e-10,\n 1.3921e-10, 7.6793e-10, 5.0679e-10, 6.7926e-10, 5.2405e-09, 1.7811e-11,\n 5.3080e-10, 1.6252e-09, 3.1212e-10, 8.4248e-09, 8.8830e-10, 1.2692e-09,\n 1.3707e-09, 1.0831e-10, 4.6079e-13, 2.4104e-08, 2.8655e-10, 2.9705e-10,\n 1.7733e-09, 4.6173e-09, 1.8720e-13, 1.3108e-09, 4.9672e-09, 5.8771e-11,\n 2.9855e-09, 9.9365e-09, 1.5658e-09, 3.2163e-10, 1.2013e-12, 1.3580e-08,\n 2.7917e-10, 4.6574e-09, 1.4311e-10, 6.3991e-10, 1.0828e-10, 1.2917e-09,\n 1.7076e-09, 5.8720e-12, 1.2994e-10, 1.6820e-10, 5.6173e-10, 2.1825e-09,\n 4.0434e-10, 7.7374e-09, 1.3638e-09, 5.5715e-11, 1.7217e-11, 8.3161e-11,\n 3.0908e-10, 7.0824e-10, 2.1491e-11, 7.5473e-11, 1.2339e-09, 9.9926e-10,\n 4.0141e-09, 1.2577e-09, 2.9397e-10, 2.2417e-09, 3.3532e-10, 5.8815e-10,\n 2.1500e-09, 4.4662e-12, 2.8119e-09, 3.4888e-09, 1.2115e-09, 2.7555e-12,\n 9.6221e-14, 1.2022e-08, 5.8041e-11, 4.4178e-09, 1.5157e-09, 2.6065e-10,\n 1.9901e-09, 2.0821e-09, 4.8328e-09, 1.3629e-10], device='cuda:0')" }, "24": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([6.7459e-11, 5.6995e-12, 6.4632e-12, 1.3986e-11, 1.6358e-12, 2.2088e-12,\n 5.6793e-12, 1.9008e-14, 1.1622e-12, 1.1149e-13, 6.4869e-11, 8.8477e-12,\n 1.0079e-11, 7.5010e-12, 2.5774e-12, 6.1835e-12, 2.2919e-11, 1.2666e-12,\n 3.9531e-12, 4.2664e-11, 9.6346e-16, 2.4999e-11, 9.7106e-12, 5.0563e-12,\n 1.0334e-12, 1.4048e-11, 5.3874e-12, 1.3409e-12, 1.0528e-12, 5.2779e-12,\n 4.0411e-11, 4.0571e-12, 6.6688e-13, 3.0584e-10, 1.7572e-11, 1.1262e-11,\n 2.2260e-12, 1.9514e-14, 1.9645e-11, 7.9935e-11, 8.7740e-12, 9.2396e-12,\n 1.1149e-12, 2.3251e-11, 4.8961e-11, 2.3934e-11, 3.1321e-11, 1.0294e-11,\n 1.6190e-12, 6.6987e-11, 1.7227e-13, 9.5156e-11, 5.4492e-11, 8.4844e-12,\n 4.3288e-13, 3.0550e-13, 3.9433e-11, 2.4703e-13, 1.5303e-11, 2.7578e-12,\n 5.7921e-12, 8.5241e-12, 6.7881e-11, 2.2954e-11, 2.6983e-13, 1.3843e-12,\n 1.5709e-11, 1.1586e-12, 1.6990e-11, 9.8805e-13, 3.4525e-13, 6.8430e-12,\n 1.2691e-11, 6.5315e-13, 4.7326e-12, 2.2113e-12, 2.9261e-12, 4.5239e-11,\n 4.1373e-13, 1.1756e-13, 3.3534e-12, 3.2552e-11, 3.1341e-14, 7.3233e-12,\n 1.1854e-12, 2.2646e-11, 1.6852e-11, 4.8089e-12, 6.7183e-12, 1.9280e-10,\n 1.9936e-12, 2.9210e-11, 5.4635e-11, 3.4997e-12, 1.9085e-10, 4.2505e-12,\n 3.3823e-12, 3.3439e-11, 1.4548e-11, 1.6775e-11, 6.1373e-11, 6.1503e-12,\n 7.8793e-13, 1.5211e-11, 9.0753e-14, 5.8048e-11, 3.4404e-12, 8.9438e-12,\n 1.7058e-12, 2.1672e-13, 5.5600e-11, 6.9412e-12, 1.5868e-10, 1.5820e-11,\n 7.3562e-13, 2.4806e-12, 4.6563e-12, 2.1678e-12, 3.8832e-12, 8.7645e-12,\n 2.2499e-11, 1.1659e-11, 8.6824e-13, 2.7668e-15, 1.0854e-13, 4.7279e-12,\n 3.4516e-11, 6.0411e-11, 1.8367e-12, 4.5985e-11, 1.3664e-14, 2.1881e-10,\n 7.6672e-12, 1.3320e-11, 1.4535e-11, 9.5560e-11, 1.8944e-11, 7.6995e-11,\n 1.7230e-11, 2.4545e-14, 9.7151e-13, 1.8169e-12, 5.3888e-12, 8.9275e-11,\n 1.0517e-11, 4.5763e-12, 3.7119e-12, 1.3196e-12, 9.9073e-11, 8.7216e-13,\n 2.1549e-12, 3.9805e-12, 7.5624e-12, 3.5470e-11, 5.8198e-12, 2.3056e-12,\n 3.6871e-12, 7.8449e-11, 1.6748e-12, 1.1149e-10, 1.9408e-11, 5.1643e-11,\n 1.9684e-12, 4.1386e-11, 5.6843e-11, 1.7232e-11, 2.5442e-10, 1.9788e-11,\n 4.9682e-11, 5.5388e-12, 3.5136e-13, 2.8608e-12, 2.5939e-11, 1.2916e-11,\n 1.4569e-11, 2.7848e-13, 7.9999e-13, 1.1027e-12, 1.6791e-12, 5.7407e-12,\n 8.8488e-13, 6.2834e-12, 3.0028e-12, 4.0829e-12, 3.9279e-11, 4.0042e-16,\n 3.3644e-12, 1.3836e-11, 1.4413e-12, 8.0094e-11, 3.8851e-12, 1.1468e-11,\n 2.2002e-11, 1.3705e-12, 3.9858e-14, 3.3310e-10, 1.8349e-12, 1.6081e-12,\n 2.4695e-11, 2.3218e-11, 1.0947e-13, 7.7812e-12, 5.9381e-11, 2.4409e-13,\n 3.2020e-11, 6.5276e-11, 9.1309e-12, 1.7710e-12, 1.3692e-15, 1.5129e-10,\n 2.0401e-12, 3.6533e-11, 2.3768e-12, 4.9061e-12, 4.4097e-13, 8.6896e-12,\n 1.1195e-11, 8.3558e-15, 1.3971e-12, 7.1971e-13, 4.7857e-12, 2.2039e-11,\n 6.4540e-12, 8.7061e-11, 1.6575e-11, 7.6258e-13, 2.7416e-14, 3.0512e-12,\n 1.7089e-12, 7.1250e-12, 3.9067e-13, 4.2965e-13, 7.4007e-12, 1.4059e-11,\n 3.5419e-11, 1.4363e-11, 1.6120e-12, 1.7217e-11, 3.6114e-12, 3.0862e-12,\n 1.9110e-11, 6.6285e-15, 8.3193e-11, 2.9799e-11, 6.9415e-12, 6.5653e-15,\n 1.1957e-13, 1.9847e-10, 2.1455e-13, 4.7901e-11, 1.0120e-11, 1.3666e-12,\n 1.5950e-11, 1.5106e-11, 3.6879e-11, 1.7887e-12], device='cuda:0')" + "exp_avg_sq": "tensor([1.9277e-11, 1.6287e-12, 1.8469e-12, 3.9966e-12, 4.6745e-13, 6.3118e-13,\n 1.6229e-12, 5.4316e-15, 3.3209e-13, 3.1859e-14, 1.8537e-11, 2.5283e-12,\n 2.8802e-12, 2.1435e-12, 7.3652e-13, 1.7670e-12, 6.5492e-12, 3.6195e-13,\n 1.1296e-12, 1.2192e-11, 2.7532e-16, 7.1437e-12, 2.7749e-12, 1.4449e-12,\n 2.9530e-13, 4.0144e-12, 1.5395e-12, 3.8317e-13, 3.0084e-13, 1.5082e-12,\n 1.1548e-11, 1.1593e-12, 1.9057e-13, 8.7397e-11, 5.0214e-12, 3.2183e-12,\n 6.3609e-13, 5.5764e-15, 5.6138e-12, 2.2842e-11, 2.5072e-12, 2.6403e-12,\n 3.1860e-13, 6.6441e-12, 1.3991e-11, 6.8392e-12, 8.9502e-12, 2.9417e-12,\n 4.6265e-13, 1.9142e-11, 4.9228e-14, 2.7192e-11, 1.5572e-11, 2.4245e-12,\n 1.2370e-13, 8.7300e-14, 1.1268e-11, 7.0591e-14, 4.3729e-12, 7.8807e-13,\n 1.6551e-12, 2.4358e-12, 1.9397e-11, 6.5593e-12, 7.7105e-14, 3.9557e-13,\n 4.4889e-12, 3.3107e-13, 4.8550e-12, 2.8234e-13, 9.8658e-14, 1.9555e-12,\n 3.6266e-12, 1.8664e-13, 1.3524e-12, 6.3189e-13, 8.3617e-13, 1.2927e-11,\n 1.1823e-13, 3.3594e-14, 9.5826e-13, 9.3021e-12, 8.9558e-15, 2.0927e-12,\n 3.3873e-13, 6.4714e-12, 4.8156e-12, 1.3742e-12, 1.9198e-12, 5.5094e-11,\n 5.6969e-13, 8.3471e-12, 1.5612e-11, 1.0001e-12, 5.4537e-11, 1.2146e-12,\n 9.6652e-13, 9.5556e-12, 4.1572e-12, 4.7936e-12, 1.7538e-11, 1.7575e-12,\n 2.2516e-13, 4.3466e-12, 2.5933e-14, 1.6588e-11, 9.8312e-13, 2.5557e-12,\n 4.8744e-13, 6.1930e-14, 1.5888e-11, 1.9835e-12, 4.5344e-11, 4.5207e-12,\n 2.1021e-13, 7.0886e-13, 1.3306e-12, 6.1947e-13, 1.1097e-12, 2.5045e-12,\n 6.4292e-12, 3.3316e-12, 2.4811e-13, 7.9064e-16, 3.1015e-14, 1.3510e-12,\n 9.8631e-12, 1.7263e-11, 5.2486e-13, 1.3141e-11, 3.9046e-15, 6.2527e-11,\n 2.1910e-12, 3.8062e-12, 4.1534e-12, 2.7307e-11, 5.4133e-12, 2.2002e-11,\n 4.9236e-12, 7.0141e-15, 2.7762e-13, 5.1920e-13, 1.5399e-12, 2.5511e-11,\n 3.0052e-12, 1.3077e-12, 1.0607e-12, 3.7707e-13, 2.8311e-11, 2.4923e-13,\n 6.1578e-13, 1.1375e-12, 2.1610e-12, 1.0136e-11, 1.6630e-12, 6.5885e-13,\n 1.0536e-12, 2.2418e-11, 4.7858e-13, 3.1858e-11, 5.5459e-12, 1.4757e-11,\n 5.6250e-13, 1.1826e-11, 1.6243e-11, 4.9243e-12, 7.2702e-11, 5.6545e-12,\n 1.4197e-11, 1.5828e-12, 1.0040e-13, 8.1751e-13, 7.4124e-12, 3.6908e-12,\n 4.1631e-12, 7.9577e-14, 2.2860e-13, 3.1511e-13, 4.7981e-13, 1.6405e-12,\n 2.5286e-13, 1.7955e-12, 8.5807e-13, 1.1667e-12, 1.1224e-11, 1.1442e-16,\n 9.6140e-13, 3.9538e-12, 4.1187e-13, 2.2888e-11, 1.1102e-12, 3.2772e-12,\n 6.2873e-12, 3.9163e-13, 1.1390e-14, 9.5185e-11, 5.2432e-13, 4.5951e-13,\n 7.0568e-12, 6.6348e-12, 3.1282e-14, 2.2235e-12, 1.6969e-11, 6.9750e-14,\n 9.1500e-12, 1.8653e-11, 2.6092e-12, 5.0609e-13, 3.9127e-16, 4.3232e-11,\n 5.8296e-13, 1.0439e-11, 6.7919e-13, 1.4020e-12, 1.2601e-13, 2.4831e-12,\n 3.1990e-12, 2.3877e-15, 3.9924e-13, 2.0566e-13, 1.3675e-12, 6.2977e-12,\n 1.8443e-12, 2.4878e-11, 4.7364e-12, 2.1791e-13, 7.8343e-15, 8.7190e-13,\n 4.8833e-13, 2.0360e-12, 1.1164e-13, 1.2278e-13, 2.1148e-12, 4.0174e-12,\n 1.0121e-11, 4.1043e-12, 4.6065e-13, 4.9199e-12, 1.0320e-12, 8.8191e-13,\n 5.4609e-12, 1.8942e-15, 2.3773e-11, 8.5153e-12, 1.9836e-12, 1.8761e-15,\n 3.4169e-14, 5.6714e-11, 6.1309e-14, 1.3688e-11, 2.8920e-12, 3.9052e-13,\n 4.5577e-12, 4.3167e-12, 1.0539e-11, 5.1115e-13], device='cuda:0')" }, "25": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([9.7916e-11, 8.6332e-12, 1.1649e-11, 1.5799e-11, 2.4511e-12, 3.5090e-12,\n 7.1045e-12, 2.2312e-14, 2.1021e-12, 1.9432e-13, 5.8027e-11, 1.3053e-11,\n 1.4581e-11, 1.4568e-11, 4.6632e-12, 1.1356e-11, 4.1490e-11, 1.7635e-12,\n 5.7668e-12, 8.2942e-11, 1.5129e-14, 3.3209e-11, 1.5813e-11, 8.5709e-12,\n 1.3011e-12, 2.4950e-11, 1.1449e-11, 2.3554e-12, 1.8845e-12, 6.3287e-12,\n 4.7840e-11, 9.7838e-12, 1.6700e-12, 2.7230e-10, 3.3379e-11, 2.3270e-11,\n 1.7315e-12, 1.4756e-14, 3.2026e-11, 1.2650e-10, 1.5848e-11, 1.3349e-11,\n 2.7796e-12, 1.5458e-11, 5.3639e-11, 4.2551e-11, 3.1631e-11, 1.9443e-11,\n 2.6285e-12, 5.6316e-11, 1.7442e-13, 1.2900e-10, 7.1136e-11, 1.5890e-11,\n 9.1393e-13, 7.4466e-13, 5.3577e-11, 6.8353e-13, 2.3567e-11, 6.1642e-12,\n 6.6510e-12, 1.3886e-11, 6.7668e-11, 3.6309e-11, 4.6829e-13, 2.6266e-12,\n 2.6184e-11, 2.4435e-12, 3.5054e-11, 1.2204e-12, 2.8528e-13, 1.6146e-11,\n 2.0014e-11, 6.8967e-13, 7.5523e-12, 2.4939e-12, 5.3743e-12, 7.2772e-11,\n 6.7860e-13, 1.0420e-13, 7.4685e-12, 5.5422e-11, 8.3653e-14, 6.2007e-12,\n 1.5548e-12, 3.9635e-11, 1.8831e-11, 9.5001e-12, 1.1548e-11, 2.1822e-10,\n 3.5930e-12, 5.3056e-11, 4.9723e-11, 7.3413e-12, 1.2697e-10, 6.8243e-12,\n 8.2792e-12, 5.8824e-11, 2.4928e-11, 3.0490e-11, 6.2359e-11, 5.4812e-12,\n 1.4934e-12, 1.7017e-11, 2.6296e-13, 1.0673e-10, 4.3026e-12, 1.7245e-11,\n 3.1220e-12, 3.8673e-13, 8.7722e-11, 9.3942e-12, 1.0829e-10, 2.6683e-11,\n 2.2385e-12, 4.5805e-12, 8.6547e-12, 4.3216e-12, 5.2588e-12, 1.3534e-11,\n 2.3168e-11, 1.3958e-11, 1.7474e-12, 1.8117e-15, 1.1486e-13, 8.6173e-12,\n 6.6872e-11, 4.6279e-11, 4.1619e-12, 4.1620e-11, 1.8135e-14, 2.7384e-10,\n 1.3925e-11, 2.4534e-11, 3.1241e-11, 8.8130e-11, 3.7173e-11, 9.0881e-11,\n 3.1144e-11, 5.3417e-14, 1.2322e-12, 3.3120e-12, 9.8857e-12, 9.6233e-11,\n 2.1159e-11, 7.2945e-12, 6.0104e-12, 2.4256e-12, 9.6361e-11, 9.8618e-13,\n 3.5047e-12, 7.8655e-12, 1.2564e-11, 3.0585e-11, 1.0150e-11, 3.3570e-12,\n 6.6837e-12, 8.9433e-11, 2.2222e-12, 1.4818e-10, 3.0704e-11, 4.0448e-11,\n 3.3205e-12, 5.4601e-11, 4.3388e-11, 2.4199e-11, 2.5903e-10, 3.5448e-11,\n 4.8836e-11, 7.2593e-12, 9.0626e-13, 5.6698e-12, 4.7673e-11, 1.9874e-11,\n 2.5656e-11, 6.5372e-13, 1.5458e-12, 2.2567e-12, 2.1735e-12, 1.5212e-11,\n 1.2980e-12, 6.6707e-12, 4.8594e-12, 7.3577e-12, 7.6629e-11, 6.7447e-15,\n 6.0334e-12, 1.6858e-11, 2.8684e-12, 1.2252e-10, 1.0440e-11, 1.2914e-11,\n 2.2778e-11, 2.1776e-12, 1.0499e-13, 3.3840e-10, 2.3918e-12, 2.7548e-12,\n 2.8851e-11, 5.5438e-11, 2.6486e-13, 2.0373e-11, 5.4687e-11, 4.6136e-13,\n 4.3260e-11, 1.2208e-10, 1.8025e-11, 2.9023e-12, 2.4287e-15, 1.7131e-10,\n 1.9015e-12, 6.8448e-11, 3.5380e-12, 1.0915e-11, 5.9572e-13, 1.5157e-11,\n 1.9931e-11, 8.2035e-14, 2.7689e-12, 1.3343e-12, 4.0005e-12, 3.3400e-11,\n 7.5538e-12, 1.1342e-10, 2.2576e-11, 1.5670e-12, 2.1748e-14, 2.6924e-12,\n 2.2690e-12, 1.1621e-11, 8.3502e-13, 4.6279e-13, 1.9418e-11, 1.7817e-11,\n 6.0354e-11, 2.0746e-11, 2.3900e-12, 3.3176e-11, 6.8275e-12, 5.8514e-12,\n 3.3369e-11, 1.4545e-14, 4.6317e-11, 4.0607e-11, 1.2405e-11, 4.9396e-14,\n 2.4730e-13, 1.4315e-10, 3.4487e-13, 6.2741e-11, 1.6204e-11, 2.1040e-12,\n 2.1538e-11, 2.1766e-11, 5.4999e-11, 3.2305e-12], device='cuda:0')" + "exp_avg_sq": "tensor([2.7980e-11, 2.4670e-12, 3.3287e-12, 4.5147e-12, 7.0042e-13, 1.0027e-12,\n 2.0302e-12, 6.3758e-15, 6.0070e-13, 5.5529e-14, 1.6582e-11, 3.7299e-12,\n 4.1667e-12, 4.1629e-12, 1.3325e-12, 3.2451e-12, 1.1856e-11, 5.0393e-13,\n 1.6479e-12, 2.3701e-11, 4.3231e-15, 9.4896e-12, 4.5187e-12, 2.4492e-12,\n 3.7180e-13, 7.1295e-12, 3.2716e-12, 6.7307e-13, 5.3850e-13, 1.8085e-12,\n 1.3671e-11, 2.7958e-12, 4.7722e-13, 7.7813e-11, 9.5382e-12, 6.6496e-12,\n 4.9478e-13, 4.2167e-15, 9.1517e-12, 3.6147e-11, 4.5286e-12, 3.8145e-12,\n 7.9429e-13, 4.4172e-12, 1.5328e-11, 1.2159e-11, 9.0389e-12, 5.5561e-12,\n 7.5112e-13, 1.6093e-11, 4.9843e-14, 3.6863e-11, 2.0328e-11, 4.5406e-12,\n 2.6116e-13, 2.1279e-13, 1.5310e-11, 1.9532e-13, 6.7344e-12, 1.7615e-12,\n 1.9006e-12, 3.9681e-12, 1.9337e-11, 1.0376e-11, 1.3382e-13, 7.5057e-13,\n 7.4822e-12, 6.9826e-13, 1.0017e-11, 3.4873e-13, 8.1521e-14, 4.6138e-12,\n 5.7191e-12, 1.9708e-13, 2.1581e-12, 7.1264e-13, 1.5358e-12, 2.0795e-11,\n 1.9392e-13, 2.9777e-14, 2.1342e-12, 1.5837e-11, 2.3905e-14, 1.7719e-12,\n 4.4431e-13, 1.1326e-11, 5.3810e-12, 2.7147e-12, 3.3001e-12, 6.2357e-11,\n 1.0267e-12, 1.5161e-11, 1.4209e-11, 2.0978e-12, 3.6282e-11, 1.9501e-12,\n 2.3658e-12, 1.6809e-11, 7.1234e-12, 8.7127e-12, 1.7820e-11, 1.5663e-12,\n 4.2675e-13, 4.8628e-12, 7.5144e-14, 3.0500e-11, 1.2295e-12, 4.9278e-12,\n 8.9214e-13, 1.1051e-13, 2.5067e-11, 2.6845e-12, 3.0944e-11, 7.6249e-12,\n 6.3968e-13, 1.3089e-12, 2.4731e-12, 1.2349e-12, 1.5027e-12, 3.8674e-12,\n 6.6205e-12, 3.9887e-12, 4.9932e-13, 5.1771e-16, 3.2822e-14, 2.4625e-12,\n 1.9109e-11, 1.3225e-11, 1.1893e-12, 1.1893e-11, 5.1822e-15, 7.8253e-11,\n 3.9791e-12, 7.0108e-12, 8.9272e-12, 2.5184e-11, 1.0622e-11, 2.5970e-11,\n 8.8997e-12, 1.5264e-14, 3.5212e-13, 9.4643e-13, 2.8249e-12, 2.7499e-11,\n 6.0463e-12, 2.0845e-12, 1.7175e-12, 6.9314e-13, 2.7536e-11, 2.8181e-13,\n 1.0015e-12, 2.2476e-12, 3.5901e-12, 8.7398e-12, 2.9004e-12, 9.5930e-13,\n 1.9099e-12, 2.5556e-11, 6.3502e-13, 4.2345e-11, 8.7738e-12, 1.1558e-11,\n 9.4887e-13, 1.5603e-11, 1.2399e-11, 6.9150e-12, 7.4019e-11, 1.0130e-11,\n 1.3955e-11, 2.0744e-12, 2.5897e-13, 1.6202e-12, 1.3623e-11, 5.6790e-12,\n 7.3314e-12, 1.8681e-13, 4.4172e-13, 6.4486e-13, 6.2111e-13, 4.3469e-12,\n 3.7091e-13, 1.9062e-12, 1.3886e-12, 2.1025e-12, 2.1897e-11, 1.9274e-15,\n 1.7241e-12, 4.8172e-12, 8.1966e-13, 3.5010e-11, 2.9834e-12, 3.6903e-12,\n 6.5089e-12, 6.2226e-13, 3.0001e-14, 9.6701e-11, 6.8349e-13, 7.8721e-13,\n 8.2443e-12, 1.5842e-11, 7.5687e-14, 5.8218e-12, 1.5627e-11, 1.3184e-13,\n 1.2362e-11, 3.4886e-11, 5.1508e-12, 8.2934e-13, 6.9401e-16, 4.8954e-11,\n 5.4336e-13, 1.9560e-11, 1.0110e-12, 3.1191e-12, 1.7023e-13, 4.3312e-12,\n 5.6954e-12, 2.3442e-14, 7.9124e-13, 3.8129e-13, 1.1432e-12, 9.5442e-12,\n 2.1585e-12, 3.2410e-11, 6.4512e-12, 4.4779e-13, 6.2148e-15, 7.6938e-13,\n 6.4838e-13, 3.3207e-12, 2.3861e-13, 1.3225e-13, 5.5488e-12, 5.0915e-12,\n 1.7247e-11, 5.9283e-12, 6.8297e-13, 9.4802e-12, 1.9510e-12, 1.6721e-12,\n 9.5356e-12, 4.1563e-15, 1.3236e-11, 1.1604e-11, 3.5448e-12, 1.4115e-14,\n 7.0669e-14, 4.0907e-11, 9.8550e-14, 1.7929e-11, 4.6305e-12, 6.0123e-13,\n 6.1547e-12, 6.2198e-12, 1.5716e-11, 9.2314e-13], device='cuda:0')" }, "26": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([[4.7214e-12, 6.5265e-12, 0.0000e+00, ..., 1.3814e-11, 2.2276e-11,\n 2.7211e-12],\n [4.3650e-13, 6.0684e-13, 0.0000e+00, ..., 1.6960e-14, 5.4672e-13,\n 1.7459e-14],\n [3.5281e-12, 2.7571e-12, 0.0000e+00, ..., 2.8378e-12, 3.0650e-12,\n 1.9935e-12],\n ...,\n [1.7568e-12, 1.4438e-12, 0.0000e+00, ..., 1.0185e-12, 9.9470e-12,\n 2.8708e-13],\n [1.0310e-11, 3.1905e-12, 0.0000e+00, ..., 2.7496e-12, 2.2534e-11,\n 4.9905e-12],\n [2.0849e-12, 3.6655e-13, 0.0000e+00, ..., 5.0572e-13, 4.3443e-12,\n 4.5673e-14]], device='cuda:0')" + "exp_avg_sq": "tensor([[1.3492e-12, 1.8650e-12, 0.0000e+00, ..., 3.9476e-12, 6.3655e-12,\n 7.7757e-13],\n [1.2473e-13, 1.7341e-13, 0.0000e+00, ..., 4.8463e-15, 1.5623e-13,\n 4.9891e-15],\n [1.0082e-12, 7.8785e-13, 0.0000e+00, ..., 8.1094e-13, 8.7584e-13,\n 5.6967e-13],\n ...,\n [5.0203e-13, 4.1257e-13, 0.0000e+00, ..., 2.9104e-13, 2.8424e-12,\n 8.2034e-14],\n [2.9461e-12, 9.1172e-13, 0.0000e+00, ..., 7.8571e-13, 6.4394e-12,\n 1.4261e-12],\n [5.9578e-13, 1.0474e-13, 0.0000e+00, ..., 1.4451e-13, 1.2414e-12,\n 1.3051e-14]], device='cuda:0')" }, "27": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([8.0017e-09, 3.5886e-10, 2.8484e-09, 3.7011e-09, 1.1270e-09, 2.8408e-10,\n 3.7384e-09, 9.0982e-11, 1.6777e-09, 2.5605e-10, 8.5059e-09, 2.9439e-08,\n 5.0118e-09, 5.8939e-09, 6.1493e-09, 2.5454e-08, 2.8527e-09, 2.8281e-09,\n 9.5461e-09, 1.1099e-08, 3.5591e-11, 1.1925e-09, 2.4051e-09, 2.9819e-09,\n 2.7617e-09, 5.1198e-09, 4.0701e-09, 7.9944e-10, 1.1682e-10, 1.3306e-09,\n 7.0711e-09, 4.6717e-09, 7.3253e-10, 3.5931e-08, 1.3014e-08, 7.9766e-09,\n 2.3729e-10, 4.4974e-12, 3.3460e-09, 1.0805e-07, 1.6118e-08, 2.2661e-09,\n 7.1033e-10, 4.4458e-10, 1.8635e-08, 3.2817e-08, 5.1118e-09, 1.1324e-08,\n 1.4108e-09, 9.0657e-09, 1.7483e-10, 1.9301e-08, 4.9210e-09, 4.4334e-09,\n 7.0456e-10, 6.1458e-11, 8.4043e-09, 3.1188e-10, 5.5857e-09, 4.8882e-10,\n 2.8705e-09, 5.4953e-09, 1.3853e-08, 7.3829e-09, 2.0823e-10, 5.1522e-10,\n 5.5390e-09, 3.6060e-10, 3.5124e-09, 6.6737e-10, 1.2184e-11, 3.4039e-09,\n 1.4386e-08, 1.0566e-09, 2.9127e-09, 1.9471e-11, 1.8915e-10, 1.1751e-08,\n 7.5831e-10, 2.5040e-13, 6.9552e-09, 8.3705e-09, 1.1987e-11, 7.1052e-10,\n 4.9591e-09, 2.1182e-08, 1.7027e-09, 7.3138e-09, 1.4488e-09, 1.2081e-08,\n 8.5557e-09, 5.2625e-09, 1.3376e-08, 2.4406e-09, 4.3098e-09, 6.6058e-10,\n 2.2470e-09, 1.5383e-08, 3.7325e-09, 3.1067e-08, 2.7521e-08, 6.7337e-10,\n 1.1969e-11, 3.6962e-09, 2.0773e-10, 2.4372e-08, 1.9533e-09, 1.7839e-09,\n 5.3154e-09, 3.3147e-11, 2.6347e-09, 1.7193e-11, 1.7172e-08, 1.6821e-08,\n 1.3259e-09, 1.4704e-08, 1.2556e-09, 2.3982e-09, 3.5531e-09, 3.6494e-09,\n 1.2719e-09, 2.8937e-10, 5.2182e-10, 6.9224e-13, 6.7809e-11, 3.2660e-09,\n 2.7756e-08, 4.5843e-09, 1.1592e-08, 5.2709e-09, 2.9109e-11, 3.3448e-08,\n 5.3927e-09, 4.9845e-09, 1.2172e-08, 7.1024e-09, 7.9153e-09, 1.0621e-08,\n 7.2145e-09, 2.7796e-11, 2.1054e-10, 3.8792e-10, 2.8563e-09, 3.6347e-08,\n 6.9445e-09, 4.2423e-10, 6.5492e-09, 7.9116e-10, 6.6866e-09, 1.6541e-10,\n 3.2224e-10, 1.1331e-08, 8.2662e-09, 1.0345e-08, 2.1966e-09, 3.7033e-10,\n 6.5868e-09, 1.9500e-08, 3.4209e-09, 4.3105e-09, 5.3255e-09, 6.1143e-09,\n 3.3194e-09, 4.4322e-09, 4.1498e-09, 3.9059e-09, 2.8038e-08, 2.1595e-08,\n 3.1523e-09, 1.1894e-08, 1.1842e-09, 8.5449e-10, 1.7647e-08, 1.6790e-09,\n 1.2603e-08, 9.3739e-10, 3.6067e-10, 1.1379e-09, 1.2170e-09, 1.3532e-09,\n 1.5950e-09, 1.3729e-09, 5.0009e-10, 3.5625e-09, 5.4818e-08, 8.3738e-12,\n 2.7177e-09, 2.2789e-09, 1.6088e-09, 1.6442e-08, 4.8335e-09, 5.9276e-10,\n 2.4445e-09, 3.3702e-09, 2.5598e-11, 2.9862e-08, 1.5713e-09, 2.3929e-10,\n 7.4631e-09, 3.3116e-08, 2.7731e-14, 3.2384e-09, 1.6074e-08, 3.5772e-09,\n 2.9871e-09, 2.3320e-08, 5.7675e-09, 1.2426e-09, 1.2386e-12, 2.7968e-08,\n 5.6812e-10, 2.8025e-08, 1.7218e-11, 6.2144e-09, 8.9191e-11, 7.4668e-09,\n 1.0510e-08, 1.7867e-13, 8.8848e-10, 1.2354e-09, 1.0516e-09, 3.4202e-08,\n 1.8081e-09, 7.1740e-09, 1.6152e-09, 6.6459e-11, 3.6999e-11, 1.3181e-10,\n 1.2085e-09, 2.4079e-09, 2.0187e-10, 2.9620e-11, 1.2599e-08, 8.9925e-10,\n 6.6103e-10, 4.6097e-09, 3.9212e-09, 5.7696e-09, 2.3862e-09, 1.8515e-09,\n 6.2569e-10, 1.4174e-12, 4.7505e-09, 3.0804e-09, 1.3781e-09, 4.4022e-11,\n 2.8443e-11, 2.3533e-08, 1.0934e-09, 2.7132e-08, 8.6251e-09, 9.4150e-10,\n 1.2950e-08, 3.6602e-09, 7.6460e-09, 1.4532e-09], device='cuda:0')" + "exp_avg_sq": "tensor([2.2866e-09, 1.0255e-10, 8.1395e-10, 1.0576e-09, 3.2206e-10, 8.1179e-11,\n 1.0683e-09, 2.5999e-11, 4.7942e-10, 7.3168e-11, 2.4306e-09, 8.4125e-09,\n 1.4322e-09, 1.6842e-09, 1.7572e-09, 7.2737e-09, 8.1518e-10, 8.0816e-10,\n 2.7279e-09, 3.1715e-09, 1.0170e-11, 3.4075e-10, 6.8727e-10, 8.5211e-10,\n 7.8919e-10, 1.4630e-09, 1.1631e-09, 2.2845e-10, 3.3382e-11, 3.8023e-10,\n 2.0206e-09, 1.3350e-09, 2.0933e-10, 1.0268e-08, 3.7190e-09, 2.2794e-09,\n 6.7808e-11, 1.2852e-12, 9.5615e-10, 3.0875e-08, 4.6060e-09, 6.4755e-10,\n 2.0298e-10, 1.2704e-10, 5.3252e-09, 9.3777e-09, 1.4607e-09, 3.2361e-09,\n 4.0314e-10, 2.5906e-09, 4.9958e-11, 5.5153e-09, 1.4062e-09, 1.2669e-09,\n 2.0133e-10, 1.7562e-11, 2.4016e-09, 8.9124e-11, 1.5962e-09, 1.3968e-10,\n 8.2025e-10, 1.5703e-09, 3.9586e-09, 2.1097e-09, 5.9504e-11, 1.4723e-10,\n 1.5828e-09, 1.0304e-10, 1.0037e-09, 1.9071e-10, 3.4816e-12, 9.7268e-10,\n 4.1110e-09, 3.0194e-10, 8.3232e-10, 5.5639e-12, 5.4052e-11, 3.3579e-09,\n 2.1669e-10, 7.1554e-14, 1.9875e-09, 2.3919e-09, 3.4253e-12, 2.0304e-10,\n 1.4171e-09, 6.0530e-09, 4.8657e-10, 2.0900e-09, 4.1401e-10, 3.4524e-09,\n 2.4449e-09, 1.5038e-09, 3.8223e-09, 6.9742e-10, 1.2316e-09, 1.8877e-10,\n 6.4209e-10, 4.3959e-09, 1.0666e-09, 8.8777e-09, 7.8643e-09, 1.9242e-10,\n 3.4203e-12, 1.0562e-09, 5.9360e-11, 6.9644e-09, 5.5817e-10, 5.0975e-10,\n 1.5189e-09, 9.4719e-12, 7.5289e-10, 4.9130e-12, 4.9069e-09, 4.8066e-09,\n 3.7890e-10, 4.2019e-09, 3.5880e-10, 6.8530e-10, 1.0153e-09, 1.0429e-09,\n 3.6344e-10, 8.2691e-11, 1.4911e-10, 1.9781e-13, 1.9377e-11, 9.3330e-10,\n 7.9316e-09, 1.3100e-09, 3.3125e-09, 1.5062e-09, 8.3181e-12, 9.5582e-09,\n 1.5410e-09, 1.4244e-09, 3.4784e-09, 2.0296e-09, 2.2618e-09, 3.0351e-09,\n 2.0616e-09, 7.9430e-12, 6.0163e-11, 1.1085e-10, 8.1620e-10, 1.0386e-08,\n 1.9844e-09, 1.2123e-10, 1.8715e-09, 2.2608e-10, 1.9108e-09, 4.7267e-11,\n 9.2083e-11, 3.2379e-09, 2.3621e-09, 2.9560e-09, 6.2770e-10, 1.0582e-10,\n 1.8822e-09, 5.5722e-09, 9.7756e-10, 1.2318e-09, 1.5218e-09, 1.7472e-09,\n 9.4855e-10, 1.2665e-09, 1.1858e-09, 1.1162e-09, 8.0121e-09, 6.1711e-09,\n 9.0080e-10, 3.3987e-09, 3.3838e-10, 2.4418e-10, 5.0427e-09, 4.7979e-10,\n 3.6013e-09, 2.6787e-10, 1.0307e-10, 3.2518e-10, 3.4776e-10, 3.8668e-10,\n 4.5579e-10, 3.9231e-10, 1.4290e-10, 1.0180e-09, 1.5665e-08, 2.3929e-12,\n 7.7661e-10, 6.5121e-10, 4.5974e-10, 4.6984e-09, 1.3812e-09, 1.6939e-10,\n 6.9852e-10, 9.6307e-10, 7.3149e-12, 8.5332e-09, 4.4900e-10, 6.8379e-11,\n 2.1326e-09, 9.4632e-09, 7.9244e-15, 9.2539e-10, 4.5931e-09, 1.0222e-09,\n 8.5358e-10, 6.6640e-09, 1.6481e-09, 3.5508e-10, 3.5394e-13, 7.9921e-09,\n 1.6235e-10, 8.0082e-09, 4.9203e-12, 1.7758e-09, 2.5487e-11, 2.1337e-09,\n 3.0033e-09, 5.1056e-14, 2.5389e-10, 3.5303e-10, 3.0051e-10, 9.7734e-09,\n 5.1669e-10, 2.0500e-09, 4.6155e-10, 1.8991e-11, 1.0573e-11, 3.7665e-11,\n 3.4532e-10, 6.8807e-10, 5.7685e-11, 8.4641e-12, 3.6002e-09, 2.5697e-10,\n 1.8890e-10, 1.3173e-09, 1.1205e-09, 1.6487e-09, 6.8187e-10, 5.2908e-10,\n 1.7879e-10, 4.0502e-13, 1.3575e-09, 8.8025e-10, 3.9381e-10, 1.2580e-11,\n 8.1277e-12, 6.7248e-09, 3.1246e-10, 7.7532e-09, 2.4647e-09, 2.6904e-10,\n 3.7007e-09, 1.0459e-09, 2.1849e-09, 4.1527e-10], device='cuda:0')" }, "28": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([1.4317e-11, 1.7761e-12, 5.2149e-12, 1.2093e-11, 2.1572e-12, 1.3330e-12,\n 7.6957e-12, 5.0994e-13, 2.6088e-12, 3.1743e-13, 1.5953e-11, 7.7825e-11,\n 1.0940e-11, 1.4743e-11, 2.3648e-11, 1.0553e-10, 8.6074e-12, 6.9530e-12,\n 3.2184e-11, 2.2007e-11, 3.8922e-15, 3.1577e-12, 8.9512e-12, 6.0688e-12,\n 9.8478e-12, 1.2349e-11, 8.4726e-12, 2.5220e-12, 3.0015e-13, 2.3293e-12,\n 1.9141e-11, 1.1116e-11, 1.7240e-12, 9.0160e-11, 3.4044e-11, 2.4080e-11,\n 5.0835e-13, 1.9206e-15, 8.1180e-12, 6.6052e-10, 3.6168e-11, 7.2491e-12,\n 2.3722e-12, 1.3266e-12, 6.3084e-11, 1.4481e-10, 1.5782e-11, 5.1584e-11,\n 1.5321e-12, 1.8803e-11, 1.7420e-13, 4.8747e-11, 9.4453e-12, 9.7962e-12,\n 9.9713e-13, 2.3293e-13, 2.6933e-11, 2.6510e-13, 2.2820e-11, 1.6978e-12,\n 6.6490e-12, 3.9456e-11, 6.6051e-11, 2.5858e-11, 9.7426e-14, 5.9612e-13,\n 1.4997e-11, 3.1839e-13, 5.7506e-12, 2.1653e-12, 4.7749e-13, 1.0266e-11,\n 3.3451e-11, 2.4932e-12, 6.9827e-12, 2.0434e-13, 6.7983e-13, 3.3701e-11,\n 1.0754e-12, 1.9151e-14, 2.5846e-11, 1.8012e-11, 2.0921e-14, 3.4518e-12,\n 1.9651e-11, 6.1637e-11, 4.9304e-12, 3.1211e-11, 3.5457e-12, 2.7683e-11,\n 4.8484e-11, 1.3181e-11, 3.3047e-11, 3.5740e-12, 9.3222e-12, 6.4545e-13,\n 4.5360e-12, 2.7324e-11, 8.7845e-12, 8.6939e-11, 1.8019e-10, 3.0724e-12,\n 6.8726e-14, 8.4265e-12, 1.1010e-13, 5.3897e-11, 5.9558e-12, 4.8591e-12,\n 1.5905e-11, 1.3743e-16, 6.9257e-12, 3.4760e-13, 5.2540e-11, 3.3151e-11,\n 1.8521e-12, 3.8715e-11, 4.2007e-12, 4.6004e-12, 8.9126e-12, 9.6011e-12,\n 1.6685e-12, 1.1877e-12, 9.6600e-13, 4.0961e-15, 2.7005e-13, 6.1579e-12,\n 7.0688e-11, 1.4237e-11, 5.8992e-11, 1.1948e-11, 6.2804e-15, 8.7344e-11,\n 2.5026e-11, 9.6870e-12, 3.7946e-11, 1.2891e-11, 1.3546e-11, 1.8043e-11,\n 1.5547e-11, 1.5278e-15, 1.0496e-12, 2.3625e-12, 3.3976e-12, 1.0380e-10,\n 1.7829e-11, 1.2544e-12, 2.9706e-11, 2.0812e-12, 1.1826e-11, 3.3501e-14,\n 8.8193e-13, 2.9442e-11, 5.0972e-11, 4.1918e-11, 2.2123e-12, 3.1777e-12,\n 3.1197e-11, 5.1991e-11, 7.5208e-12, 9.2699e-12, 7.2519e-12, 1.0222e-11,\n 1.6279e-11, 8.1721e-12, 1.2910e-11, 9.1411e-12, 5.0787e-11, 7.1228e-11,\n 6.4857e-12, 2.8009e-11, 2.1829e-12, 1.2437e-12, 3.8725e-11, 3.3033e-12,\n 3.6856e-11, 5.6416e-12, 7.8195e-13, 2.6440e-12, 2.3735e-12, 4.7552e-12,\n 2.1488e-12, 2.9943e-12, 5.6379e-13, 6.6208e-12, 1.6857e-10, 5.6987e-14,\n 4.3166e-12, 4.1422e-12, 1.7919e-12, 3.6067e-11, 7.1404e-12, 9.8535e-13,\n 8.5896e-12, 1.5502e-11, 1.0713e-13, 6.6319e-11, 5.3133e-12, 3.6748e-13,\n 3.4643e-11, 8.2040e-11, 1.0951e-14, 7.3772e-12, 5.8623e-11, 1.1771e-11,\n 1.1046e-11, 4.6375e-11, 1.1226e-11, 3.2267e-12, 4.5276e-15, 5.9797e-11,\n 8.9542e-13, 7.6046e-11, 1.9838e-13, 1.6623e-11, 9.3342e-14, 1.6848e-11,\n 1.4810e-11, 8.0388e-14, 2.9784e-12, 2.1745e-12, 2.6592e-12, 1.5490e-10,\n 6.5571e-12, 1.8776e-11, 4.4865e-12, 7.1965e-13, 1.0225e-15, 7.5602e-13,\n 1.6640e-12, 5.4954e-12, 7.7738e-13, 5.6442e-16, 4.6946e-11, 3.0372e-12,\n 2.4021e-12, 1.6655e-11, 1.4250e-11, 1.1929e-11, 1.0754e-11, 4.1497e-12,\n 3.2295e-12, 2.8935e-14, 1.5945e-11, 5.1569e-12, 2.5135e-12, 4.4248e-14,\n 5.1250e-14, 7.8238e-11, 2.5039e-12, 8.8587e-11, 1.4683e-11, 1.6232e-12,\n 5.2703e-11, 6.6369e-12, 1.3793e-11, 9.3942e-12], device='cuda:0')" + "exp_avg_sq": "tensor([4.0913e-12, 5.0754e-13, 1.4902e-12, 3.4557e-12, 6.1644e-13, 3.8090e-13,\n 2.1991e-12, 1.4572e-13, 7.4548e-13, 9.0709e-14, 4.5588e-12, 2.2239e-11,\n 3.1263e-12, 4.2130e-12, 6.7576e-12, 3.0156e-11, 2.4596e-12, 1.9869e-12,\n 9.1968e-12, 6.2887e-12, 1.1122e-15, 9.0235e-13, 2.5579e-12, 1.7342e-12,\n 2.8141e-12, 3.5287e-12, 2.4211e-12, 7.2069e-13, 8.5772e-14, 6.6561e-13,\n 5.4696e-12, 3.1764e-12, 4.9265e-13, 2.5764e-11, 9.7285e-12, 6.8810e-12,\n 1.4527e-13, 5.4882e-16, 2.3198e-12, 1.8875e-10, 1.0335e-11, 2.0715e-12,\n 6.7786e-13, 3.7908e-13, 1.8027e-11, 4.1381e-11, 4.5098e-12, 1.4741e-11,\n 4.3781e-13, 5.3730e-12, 4.9778e-14, 1.3930e-11, 2.6991e-12, 2.7993e-12,\n 2.8494e-13, 6.6561e-14, 7.6964e-12, 7.5754e-14, 6.5209e-12, 4.8516e-13,\n 1.9000e-12, 1.1275e-11, 1.8875e-11, 7.3891e-12, 2.7840e-14, 1.7035e-13,\n 4.2856e-12, 9.0981e-14, 1.6433e-12, 6.1876e-13, 1.3645e-13, 2.9336e-12,\n 9.5590e-12, 7.1246e-13, 1.9954e-12, 5.8392e-14, 1.9427e-13, 9.6305e-12,\n 3.0731e-13, 5.4725e-15, 7.3858e-12, 5.1470e-12, 5.9783e-15, 9.8637e-13,\n 5.6155e-12, 1.7613e-11, 1.4089e-12, 8.9187e-12, 1.0132e-12, 7.9108e-12,\n 1.3855e-11, 3.7665e-12, 9.4435e-12, 1.0213e-12, 2.6639e-12, 1.8444e-13,\n 1.2962e-12, 7.8080e-12, 2.5102e-12, 2.4844e-11, 5.1489e-11, 8.7797e-13,\n 1.9639e-14, 2.4079e-12, 3.1462e-14, 1.5401e-11, 1.7019e-12, 1.3885e-12,\n 4.5449e-12, 3.9273e-17, 1.9791e-12, 9.9329e-14, 1.5014e-11, 9.4731e-12,\n 5.2924e-13, 1.1063e-11, 1.2004e-12, 1.3146e-12, 2.5468e-12, 2.7436e-12,\n 4.7680e-13, 3.3940e-13, 2.7604e-13, 1.1705e-15, 7.7169e-14, 1.7597e-12,\n 2.0200e-11, 4.0683e-12, 1.6857e-11, 3.4143e-12, 1.7947e-15, 2.4959e-11,\n 7.1513e-12, 2.7681e-12, 1.0843e-11, 3.6837e-12, 3.8708e-12, 5.1559e-12,\n 4.4426e-12, 4.3659e-16, 2.9993e-13, 6.7510e-13, 9.7089e-13, 2.9662e-11,\n 5.0948e-12, 3.5847e-13, 8.4887e-12, 5.9472e-13, 3.3795e-12, 9.5731e-15,\n 2.5202e-13, 8.4133e-12, 1.4566e-11, 1.1978e-11, 6.3217e-13, 9.0805e-13,\n 8.9147e-12, 1.4857e-11, 2.1491e-12, 2.6489e-12, 2.0723e-12, 2.9211e-12,\n 4.6520e-12, 2.3352e-12, 3.6890e-12, 2.6121e-12, 1.4513e-11, 2.0354e-11,\n 1.8533e-12, 8.0038e-12, 6.2377e-13, 3.5539e-13, 1.1066e-11, 9.4394e-13,\n 1.0532e-11, 1.6121e-12, 2.2345e-13, 7.5553e-13, 6.7825e-13, 1.3588e-12,\n 6.1402e-13, 8.5565e-13, 1.6111e-13, 1.8919e-12, 4.8170e-11, 1.6284e-14,\n 1.2335e-12, 1.1837e-12, 5.1205e-13, 1.0307e-11, 2.0404e-12, 2.8157e-13,\n 2.4545e-12, 4.4299e-12, 3.0614e-14, 1.8951e-11, 1.5183e-12, 1.0501e-13,\n 9.8994e-12, 2.3444e-11, 3.1294e-15, 2.1081e-12, 1.6752e-11, 3.3638e-12,\n 3.1565e-12, 1.3252e-11, 3.2079e-12, 9.2206e-13, 1.2938e-15, 1.7088e-11,\n 2.5587e-13, 2.1731e-11, 5.6689e-14, 4.7501e-12, 2.6673e-14, 4.8143e-12,\n 4.2322e-12, 2.2971e-14, 8.5110e-13, 6.2139e-13, 7.5989e-13, 4.4264e-11,\n 1.8737e-12, 5.3652e-12, 1.2821e-12, 2.0564e-13, 2.9218e-16, 2.1604e-13,\n 4.7550e-13, 1.5703e-12, 2.2214e-13, 1.6129e-16, 1.3415e-11, 8.6789e-13,\n 6.8641e-13, 4.7593e-12, 4.0721e-12, 3.4087e-12, 3.0731e-12, 1.1858e-12,\n 9.2285e-13, 8.2684e-15, 4.5565e-12, 1.4736e-12, 7.1826e-13, 1.2644e-14,\n 1.4645e-14, 2.2357e-11, 7.1549e-13, 2.5315e-11, 4.1957e-12, 4.6384e-13,\n 1.5060e-11, 1.8965e-12, 3.9413e-12, 2.6844e-12], device='cuda:0')" }, "29": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([3.1934e-11, 1.7548e-12, 9.5380e-12, 1.6244e-11, 2.6124e-12, 1.8145e-12,\n 1.1524e-11, 8.2556e-13, 4.7574e-12, 5.2007e-13, 2.8074e-11, 1.0744e-10,\n 1.6735e-11, 2.3987e-11, 2.7972e-11, 1.0933e-10, 1.4694e-11, 9.5436e-12,\n 3.1386e-11, 4.7888e-11, 5.3961e-16, 5.6248e-12, 1.1653e-11, 1.3780e-11,\n 1.3448e-11, 1.6997e-11, 1.8173e-11, 4.2824e-12, 7.4018e-13, 4.1504e-12,\n 3.1968e-11, 2.0094e-11, 3.8576e-12, 1.3043e-10, 5.6288e-11, 3.6099e-11,\n 4.0786e-13, 8.8243e-15, 1.4885e-11, 3.9315e-10, 5.6244e-11, 1.1427e-11,\n 3.8567e-12, 2.6185e-12, 6.4825e-11, 1.1508e-10, 2.3489e-11, 5.1193e-11,\n 3.8265e-12, 3.1635e-11, 3.0413e-13, 8.2052e-11, 1.6192e-11, 1.5211e-11,\n 1.8151e-12, 5.9917e-13, 3.6588e-11, 8.1168e-13, 2.6313e-11, 2.7560e-12,\n 8.4016e-12, 2.6976e-11, 6.0028e-11, 3.2869e-11, 3.4338e-13, 1.4800e-12,\n 2.5024e-11, 7.8122e-13, 1.5474e-11, 3.2953e-12, 3.0835e-13, 1.6307e-11,\n 4.8055e-11, 5.7258e-12, 8.9462e-12, 1.7169e-13, 1.0987e-12, 5.1285e-11,\n 1.5885e-12, 2.2281e-14, 3.0793e-11, 3.7463e-11, 8.0329e-14, 4.0207e-12,\n 2.3664e-11, 9.1668e-11, 7.7300e-12, 3.1877e-11, 4.9830e-12, 5.3273e-11,\n 3.9486e-11, 2.3459e-11, 4.6866e-11, 7.9087e-12, 1.8906e-11, 1.9321e-12,\n 7.3549e-12, 5.6698e-11, 1.7801e-11, 1.2517e-10, 1.1883e-10, 4.0422e-12,\n 1.1804e-13, 1.1385e-11, 1.6452e-13, 8.9039e-11, 5.0101e-12, 8.5903e-12,\n 1.6950e-11, 3.3300e-15, 9.3849e-12, 9.4640e-14, 7.1665e-11, 5.9992e-11,\n 4.2314e-12, 5.1302e-11, 6.1076e-12, 7.5868e-12, 1.6143e-11, 1.5409e-11,\n 4.0441e-12, 9.6075e-13, 1.1130e-12, 1.0681e-14, 6.3718e-13, 1.0326e-11,\n 1.1777e-10, 2.1180e-11, 5.1027e-11, 1.7080e-11, 3.8194e-16, 1.3466e-10,\n 2.5564e-11, 2.2433e-11, 4.1464e-11, 2.4354e-11, 3.4466e-11, 3.9113e-11,\n 3.1651e-11, 3.2870e-15, 1.5643e-12, 2.7625e-12, 9.8289e-12, 1.2729e-10,\n 2.9236e-11, 2.3724e-12, 2.9924e-11, 4.1682e-12, 2.2539e-11, 2.7323e-13,\n 1.9310e-12, 3.8285e-11, 3.7551e-11, 3.4339e-11, 6.7048e-12, 2.6920e-12,\n 2.0163e-11, 7.1214e-11, 1.0577e-11, 1.9883e-11, 1.8895e-11, 2.0743e-11,\n 8.9897e-12, 1.8959e-11, 1.8681e-11, 1.2735e-11, 1.0302e-10, 7.4763e-11,\n 1.5245e-11, 5.3110e-11, 2.9510e-12, 2.9100e-12, 6.3050e-11, 7.7669e-12,\n 4.2246e-11, 5.7697e-12, 1.0398e-12, 5.7344e-12, 3.1274e-12, 6.8295e-12,\n 4.7389e-12, 3.9979e-12, 1.3354e-12, 1.1079e-11, 2.1501e-10, 1.0683e-13,\n 9.0852e-12, 7.1491e-12, 4.6496e-12, 6.7857e-11, 1.7713e-11, 1.8368e-12,\n 1.1523e-11, 1.6939e-11, 3.9290e-13, 1.2930e-10, 4.2499e-12, 6.2984e-13,\n 3.4761e-11, 1.2378e-10, 1.0052e-13, 1.4839e-11, 5.5452e-11, 9.8390e-12,\n 1.4643e-11, 9.0929e-11, 1.9394e-11, 3.7931e-12, 3.8584e-14, 1.0189e-10,\n 1.2424e-12, 1.1283e-10, 1.8708e-13, 2.7328e-11, 7.3957e-14, 2.7662e-11,\n 3.6318e-11, 9.4482e-14, 4.9319e-12, 3.9444e-12, 2.2155e-12, 1.4192e-10,\n 8.9853e-12, 3.3201e-11, 8.4421e-12, 9.4607e-13, 1.5929e-15, 1.2281e-12,\n 3.1780e-12, 1.0719e-11, 1.3879e-12, 1.6804e-14, 5.5566e-11, 4.4800e-12,\n 3.2184e-12, 2.1342e-11, 1.1416e-11, 2.6129e-11, 1.1947e-11, 5.4501e-12,\n 2.5427e-12, 6.7851e-14, 2.1833e-11, 1.0716e-11, 4.6031e-12, 5.3829e-14,\n 2.2359e-14, 8.1531e-11, 2.5996e-12, 1.1203e-10, 2.9068e-11, 2.5553e-12,\n 4.2582e-11, 1.1780e-11, 2.9119e-11, 8.0715e-12], device='cuda:0')" + "exp_avg_sq": "tensor([9.1253e-12, 5.0143e-13, 2.7256e-12, 4.6418e-12, 7.4652e-13, 5.1852e-13,\n 3.2930e-12, 2.3591e-13, 1.3595e-12, 1.4861e-13, 8.0222e-12, 3.0701e-11,\n 4.7823e-12, 6.8544e-12, 7.9932e-12, 3.1243e-11, 4.1991e-12, 2.7272e-12,\n 8.9687e-12, 1.3684e-11, 1.5420e-16, 1.6073e-12, 3.3301e-12, 3.9376e-12,\n 3.8428e-12, 4.8570e-12, 5.1932e-12, 1.2237e-12, 2.1151e-13, 1.1860e-12,\n 9.1352e-12, 5.7421e-12, 1.1023e-12, 3.7270e-11, 1.6085e-11, 1.0316e-11,\n 1.1655e-13, 2.5216e-15, 4.2534e-12, 1.1235e-10, 1.6072e-11, 3.2654e-12,\n 1.1021e-12, 7.4826e-13, 1.8524e-11, 3.2886e-11, 6.7123e-12, 1.4629e-11,\n 1.0934e-12, 9.0400e-12, 8.6908e-14, 2.3447e-11, 4.6269e-12, 4.3467e-12,\n 5.1868e-13, 1.7122e-13, 1.0455e-11, 2.3194e-13, 7.5190e-12, 7.8756e-13,\n 2.4008e-12, 7.7087e-12, 1.7153e-11, 9.3927e-12, 9.8123e-14, 4.2292e-13,\n 7.1508e-12, 2.2324e-13, 4.4217e-12, 9.4167e-13, 8.8113e-14, 4.6598e-12,\n 1.3732e-11, 1.6362e-12, 2.5564e-12, 4.9063e-14, 3.1397e-13, 1.4655e-11,\n 4.5391e-13, 6.3671e-15, 8.7994e-12, 1.0705e-11, 2.2955e-14, 1.1490e-12,\n 6.7623e-12, 2.6195e-11, 2.2089e-12, 9.1092e-12, 1.4239e-12, 1.5223e-11,\n 1.1283e-11, 6.7036e-12, 1.3392e-11, 2.2600e-12, 5.4024e-12, 5.5211e-13,\n 2.1017e-12, 1.6202e-11, 5.0868e-12, 3.5767e-11, 3.3955e-11, 1.1551e-12,\n 3.3732e-14, 3.2532e-12, 4.7012e-14, 2.5444e-11, 1.4317e-12, 2.4547e-12,\n 4.8436e-12, 9.5156e-16, 2.6818e-12, 2.7044e-14, 2.0479e-11, 1.7143e-11,\n 1.2092e-12, 1.4660e-11, 1.7453e-12, 2.1680e-12, 4.6129e-12, 4.4032e-12,\n 1.1556e-12, 2.7454e-13, 3.1806e-13, 3.0523e-15, 1.8208e-13, 2.9506e-12,\n 3.3652e-11, 6.0522e-12, 1.4581e-11, 4.8807e-12, 1.0914e-16, 3.8480e-11,\n 7.3051e-12, 6.4104e-12, 1.1849e-11, 6.9593e-12, 9.8489e-12, 1.1177e-11,\n 9.0444e-12, 9.3928e-16, 4.4700e-13, 7.8940e-13, 2.8087e-12, 3.6374e-11,\n 8.3543e-12, 6.7793e-13, 8.5509e-12, 1.1911e-12, 6.4408e-12, 7.8079e-14,\n 5.5181e-13, 1.0940e-11, 1.0730e-11, 9.8125e-12, 1.9160e-12, 7.6926e-13,\n 5.7617e-12, 2.0350e-11, 3.0224e-12, 5.6817e-12, 5.3995e-12, 5.9275e-12,\n 2.5689e-12, 5.4177e-12, 5.3381e-12, 3.6390e-12, 2.9438e-11, 2.1364e-11,\n 4.3564e-12, 1.5176e-11, 8.4327e-13, 8.3155e-13, 1.8017e-11, 2.2195e-12,\n 1.2072e-11, 1.6487e-12, 2.9714e-13, 1.6386e-12, 8.9367e-13, 1.9516e-12,\n 1.3542e-12, 1.1424e-12, 3.8159e-13, 3.1660e-12, 6.1442e-11, 3.0527e-14,\n 2.5962e-12, 2.0429e-12, 1.3287e-12, 1.9391e-11, 5.0617e-12, 5.2488e-13,\n 3.2928e-12, 4.8404e-12, 1.1228e-13, 3.6949e-11, 1.2144e-12, 1.7998e-13,\n 9.9332e-12, 3.5370e-11, 2.8723e-14, 4.2402e-12, 1.5846e-11, 2.8116e-12,\n 4.1845e-12, 2.5984e-11, 5.5419e-12, 1.0839e-12, 1.1026e-14, 2.9115e-11,\n 3.5501e-13, 3.2241e-11, 5.3460e-14, 7.8092e-12, 2.1134e-14, 7.9047e-12,\n 1.0378e-11, 2.6999e-14, 1.4093e-12, 1.1271e-12, 6.3309e-13, 4.0554e-11,\n 2.5676e-12, 9.4875e-12, 2.4124e-12, 2.7035e-13, 4.5520e-16, 3.5094e-13,\n 9.0812e-13, 3.0630e-12, 3.9661e-13, 4.8020e-15, 1.5878e-11, 1.2802e-12,\n 9.1968e-13, 6.0986e-12, 3.2622e-12, 7.4665e-12, 3.4139e-12, 1.5574e-12,\n 7.2660e-13, 1.9389e-14, 6.2390e-12, 3.0622e-12, 1.3154e-12, 1.5382e-14,\n 6.3892e-15, 2.3298e-11, 7.4286e-13, 3.2014e-11, 8.3064e-12, 7.3020e-13,\n 1.2168e-11, 3.3661e-12, 8.3210e-12, 2.3065e-12], device='cuda:0')" }, "30": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([[2.9404e-12, 9.9306e-12, 0.0000e+00, ..., 3.4161e-11, 1.9439e-11,\n 2.0781e-11],\n [4.6973e-12, 1.7545e-12, 0.0000e+00, ..., 1.4674e-12, 1.6997e-11,\n 7.6646e-13],\n [3.8258e-12, 6.7511e-12, 0.0000e+00, ..., 4.6961e-12, 3.1858e-11,\n 1.8835e-12],\n ...,\n [9.2494e-13, 2.5181e-12, 0.0000e+00, ..., 3.2028e-12, 4.7602e-12,\n 2.9387e-14],\n [2.1119e-11, 1.8774e-11, 0.0000e+00, ..., 2.7300e-11, 1.1645e-10,\n 2.3740e-11],\n [1.7049e-13, 2.2796e-13, 0.0000e+00, ..., 8.7930e-13, 2.0564e-12,\n 2.1398e-14]], device='cuda:0')" + "exp_avg_sq": "tensor([[8.4025e-13, 2.8377e-12, 0.0000e+00, ..., 9.7617e-12, 5.5548e-12,\n 5.9383e-12],\n [1.3423e-12, 5.0136e-13, 0.0000e+00, ..., 4.1932e-13, 4.8570e-12,\n 2.1902e-13],\n [1.0933e-12, 1.9292e-12, 0.0000e+00, ..., 1.3420e-12, 9.1036e-12,\n 5.3823e-13],\n ...,\n [2.6431e-13, 7.1958e-13, 0.0000e+00, ..., 9.1522e-13, 1.3603e-12,\n 8.3976e-15],\n [6.0349e-12, 5.3648e-12, 0.0000e+00, ..., 7.8012e-12, 3.3277e-11,\n 6.7838e-12],\n [4.8720e-14, 6.5142e-14, 0.0000e+00, ..., 2.5127e-13, 5.8762e-13,\n 6.1147e-15]], device='cuda:0')" }, "31": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([1.8221e-08, 5.6640e-09, 7.2018e-09, 2.2413e-10, 2.0364e-09, 2.1541e-10,\n 2.2094e-09, 1.8894e-10, 1.2910e-09, 8.2745e-10, 1.6998e-08, 1.6755e-08,\n 2.2996e-09, 1.6601e-09, 3.2482e-09, 4.3943e-10, 8.2496e-09, 2.4540e-09,\n 8.2403e-09, 2.2943e-08, 3.7738e-11, 6.5768e-09, 1.4492e-09, 1.3464e-08,\n 9.9305e-10, 2.1807e-09, 1.4419e-08, 2.3310e-10, 8.1870e-10, 2.4998e-09,\n 4.3365e-09, 1.4275e-08, 1.8865e-11, 2.5691e-08, 1.4782e-08, 1.7548e-08,\n 6.6557e-10, 2.3723e-11, 9.5354e-09, 6.2961e-08, 2.6461e-08, 7.4517e-10,\n 6.7723e-10, 6.8263e-10, 4.8119e-09, 1.0984e-08, 6.6944e-09, 3.0939e-09,\n 6.7389e-10, 9.3074e-09, 2.4947e-10, 2.3308e-08, 1.1682e-08, 7.7831e-09,\n 3.4453e-10, 7.1556e-10, 5.2383e-09, 6.1283e-10, 3.8836e-09, 1.2279e-09,\n 1.5005e-09, 3.8834e-09, 6.3922e-09, 1.5723e-08, 1.6472e-10, 3.9191e-09,\n 3.7746e-09, 9.1257e-10, 7.5192e-09, 1.0917e-09, 7.6560e-12, 3.0421e-09,\n 1.2485e-08, 1.2168e-09, 7.4507e-10, 9.6445e-12, 1.9585e-09, 2.1912e-08,\n 8.5736e-10, 6.1528e-12, 8.9283e-09, 2.8513e-08, 1.2744e-11, 2.0972e-10,\n 3.5412e-09, 2.1109e-08, 9.5305e-10, 4.3294e-09, 3.1672e-09, 8.1074e-10,\n 4.4851e-09, 1.6172e-08, 4.9262e-09, 5.7154e-09, 1.8329e-08, 4.5684e-09,\n 8.1005e-09, 2.7484e-08, 4.2840e-09, 3.3246e-08, 3.0969e-09, 6.3425e-10,\n 1.2356e-12, 3.2793e-09, 3.5997e-11, 3.9629e-08, 1.1344e-09, 4.4831e-09,\n 3.4088e-09, 2.3696e-11, 3.0048e-08, 1.3292e-09, 9.3338e-09, 1.4057e-08,\n 2.4560e-09, 1.9571e-08, 4.2185e-10, 1.5540e-09, 2.4313e-09, 1.3049e-08,\n 1.3608e-09, 1.0419e-09, 5.8009e-10, 9.7413e-12, 9.6428e-11, 7.0110e-09,\n 1.1732e-08, 1.7285e-09, 5.8141e-09, 1.1915e-08, 1.3010e-10, 2.6716e-08,\n 6.5243e-09, 9.7674e-10, 1.1309e-08, 4.2786e-09, 2.1409e-08, 4.1486e-09,\n 3.2693e-09, 2.3521e-12, 8.3370e-12, 8.0587e-11, 1.4662e-08, 5.4359e-08,\n 9.5185e-09, 1.8740e-09, 2.2996e-09, 3.6381e-09, 3.2128e-09, 2.7454e-10,\n 8.4052e-10, 7.4460e-09, 5.3998e-09, 1.3216e-09, 1.6174e-09, 9.5671e-10,\n 2.5728e-09, 2.8957e-08, 1.2813e-09, 1.3880e-09, 4.5871e-09, 1.4360e-09,\n 1.3674e-09, 4.0966e-09, 1.3002e-09, 6.1012e-09, 2.7976e-08, 2.0240e-09,\n 9.5547e-09, 1.7324e-09, 8.7816e-11, 3.1728e-09, 5.8639e-09, 9.1727e-10,\n 1.1869e-08, 2.0177e-09, 1.4010e-09, 1.3382e-09, 1.4917e-09, 4.0330e-09,\n 4.8699e-09, 1.4530e-09, 2.1739e-09, 3.6102e-09, 3.1899e-08, 1.8642e-11,\n 2.5011e-09, 9.4017e-09, 1.6695e-09, 3.0636e-08, 5.3221e-09, 3.1337e-09,\n 1.8017e-09, 6.2825e-10, 2.7564e-12, 8.4957e-08, 1.1871e-09, 2.5340e-10,\n 2.4901e-09, 1.1650e-08, 1.3371e-11, 3.4560e-08, 2.3823e-09, 2.9728e-09,\n 2.9207e-09, 8.0255e-09, 1.4570e-08, 2.9670e-10, 6.7857e-12, 5.5618e-09,\n 1.6067e-09, 4.0845e-09, 1.5222e-10, 1.0165e-08, 2.0130e-10, 2.2135e-08,\n 2.5684e-08, 1.0110e-10, 2.5293e-10, 2.5731e-09, 1.0481e-09, 1.4341e-08,\n 1.3336e-09, 1.7363e-08, 2.3419e-09, 3.2148e-13, 2.9780e-12, 2.2961e-10,\n 3.5637e-10, 3.7527e-09, 9.8776e-11, 3.3320e-10, 3.8770e-09, 2.2491e-09,\n 1.5349e-08, 1.1814e-10, 1.0787e-09, 7.1858e-09, 2.0075e-09, 2.3517e-09,\n 4.2644e-09, 1.7390e-11, 1.6886e-09, 6.0244e-09, 2.4756e-09, 5.9730e-11,\n 1.3920e-10, 1.1506e-08, 7.9750e-10, 2.5751e-09, 2.0709e-08, 3.1722e-10,\n 1.2916e-08, 1.1420e-09, 3.0749e-08, 7.3548e-10], device='cuda:0')" + "exp_avg_sq": "tensor([5.2068e-09, 1.6185e-09, 2.0580e-09, 6.4047e-11, 5.8193e-10, 6.1554e-11,\n 6.3134e-10, 5.3993e-11, 3.6892e-10, 2.3645e-10, 4.8575e-09, 4.7879e-09,\n 6.5711e-10, 4.7438e-10, 9.2820e-10, 1.2557e-10, 2.3574e-09, 7.0124e-10,\n 2.3547e-09, 6.5561e-09, 1.0784e-11, 1.8794e-09, 4.1412e-10, 3.8473e-09,\n 2.8377e-10, 6.2316e-10, 4.1204e-09, 6.6611e-11, 2.3395e-10, 7.1434e-10,\n 1.2392e-09, 4.0791e-09, 5.3907e-12, 7.3413e-09, 4.2239e-09, 5.0144e-09,\n 1.9019e-10, 6.7790e-12, 2.7248e-09, 1.7992e-08, 7.5615e-09, 2.1294e-10,\n 1.9352e-10, 1.9507e-10, 1.3750e-09, 3.1388e-09, 1.9130e-09, 8.8411e-10,\n 1.9257e-10, 2.6597e-09, 7.1287e-11, 6.6604e-09, 3.3381e-09, 2.2241e-09,\n 9.8453e-11, 2.0448e-10, 1.4969e-09, 1.7512e-10, 1.1098e-09, 3.5088e-10,\n 4.2877e-10, 1.1097e-09, 1.8266e-09, 4.4931e-09, 4.7071e-11, 1.1199e-09,\n 1.0786e-09, 2.6077e-10, 2.1487e-09, 3.1195e-10, 2.1878e-12, 8.6930e-10,\n 3.5677e-09, 3.4770e-10, 2.1291e-10, 2.7560e-12, 5.5967e-10, 6.2616e-09,\n 2.4500e-10, 1.7582e-12, 2.5513e-09, 8.1479e-09, 3.6416e-12, 5.9929e-11,\n 1.0119e-09, 6.0320e-09, 2.7234e-10, 1.2372e-09, 9.0506e-10, 2.3168e-10,\n 1.2816e-09, 4.6212e-09, 1.4077e-09, 1.6332e-09, 5.2376e-09, 1.3055e-09,\n 2.3148e-09, 7.8537e-09, 1.2242e-09, 9.5002e-09, 8.8498e-10, 1.8124e-10,\n 3.5309e-13, 9.3709e-10, 1.0287e-11, 1.1324e-08, 3.2415e-10, 1.2811e-09,\n 9.7409e-10, 6.7712e-12, 8.5865e-09, 3.7982e-10, 2.6672e-09, 4.0168e-09,\n 7.0182e-10, 5.5925e-09, 1.2055e-10, 4.4407e-10, 6.9477e-10, 3.7287e-09,\n 3.8885e-10, 2.9774e-10, 1.6577e-10, 2.7837e-12, 2.7555e-11, 2.0034e-09,\n 3.3527e-09, 4.9393e-10, 1.6614e-09, 3.4049e-09, 3.7176e-11, 7.6343e-09,\n 1.8644e-09, 2.7911e-10, 3.2317e-09, 1.2226e-09, 6.1177e-09, 1.1855e-09,\n 9.3422e-10, 6.7212e-13, 2.3824e-12, 2.3028e-11, 4.1898e-09, 1.5534e-08,\n 2.7200e-09, 5.3550e-10, 6.5713e-10, 1.0396e-09, 9.1808e-10, 7.8452e-11,\n 2.4019e-10, 2.1278e-09, 1.5430e-09, 3.7765e-10, 4.6219e-10, 2.7339e-10,\n 7.3520e-10, 8.2746e-09, 3.6614e-10, 3.9664e-10, 1.3108e-09, 4.1035e-10,\n 3.9074e-10, 1.1706e-09, 3.7154e-10, 1.7435e-09, 7.9944e-09, 5.7837e-10,\n 2.7303e-09, 4.9506e-10, 2.5094e-11, 9.0665e-10, 1.6757e-09, 2.6212e-10,\n 3.3916e-09, 5.7656e-10, 4.0035e-10, 3.8239e-10, 4.2625e-10, 1.1525e-09,\n 1.3916e-09, 4.1520e-10, 6.2120e-10, 1.0316e-09, 9.1154e-09, 5.3270e-12,\n 7.1470e-10, 2.6866e-09, 4.7708e-10, 8.7544e-09, 1.5208e-09, 8.9548e-10,\n 5.1486e-10, 1.7953e-10, 7.8766e-13, 2.4277e-08, 3.3921e-10, 7.2412e-11,\n 7.1157e-10, 3.3290e-09, 3.8208e-12, 9.8759e-09, 6.8075e-10, 8.4951e-10,\n 8.3460e-10, 2.2933e-09, 4.1635e-09, 8.4784e-11, 1.9391e-12, 1.5893e-09,\n 4.5912e-10, 1.1672e-09, 4.3497e-11, 2.9047e-09, 5.7523e-11, 6.3254e-09,\n 7.3393e-09, 2.8889e-11, 7.2278e-11, 7.3528e-10, 2.9949e-10, 4.0979e-09,\n 3.8109e-10, 4.9617e-09, 6.6922e-10, 9.1865e-14, 8.5100e-13, 6.5612e-11,\n 1.0184e-10, 1.0724e-09, 2.8226e-11, 9.5216e-11, 1.1079e-09, 6.4271e-10,\n 4.3860e-09, 3.3758e-11, 3.0824e-10, 2.0534e-09, 5.7367e-10, 6.7201e-10,\n 1.2186e-09, 4.9694e-12, 4.8252e-10, 1.7215e-09, 7.0742e-10, 1.7068e-11,\n 3.9779e-11, 3.2881e-09, 2.2789e-10, 7.3585e-10, 5.9178e-09, 9.0649e-11,\n 3.6907e-09, 3.2634e-10, 8.7868e-09, 2.1017e-10], device='cuda:0')" }, "32": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([3.9274e-11, 1.8154e-11, 2.0117e-11, 1.1906e-12, 4.5109e-12, 6.3944e-13,\n 3.2225e-12, 1.5377e-12, 1.7615e-12, 1.9861e-12, 3.1294e-11, 2.7548e-11,\n 4.7595e-12, 4.2067e-12, 1.0725e-11, 3.1908e-12, 1.9686e-11, 4.2706e-12,\n 2.0162e-11, 5.4554e-11, 1.8157e-14, 2.3088e-11, 6.9070e-12, 2.8359e-11,\n 3.3333e-12, 2.5237e-12, 6.8857e-11, 7.1771e-13, 2.8473e-12, 4.7980e-12,\n 6.5610e-12, 1.0397e-10, 1.7740e-13, 4.3926e-11, 4.3960e-11, 6.6829e-11,\n 8.3350e-13, 2.6876e-14, 2.7819e-11, 1.5937e-10, 7.8260e-11, 1.2393e-12,\n 1.8045e-12, 3.1076e-12, 6.2382e-12, 1.9701e-11, 2.7615e-11, 7.4501e-12,\n 8.6012e-13, 2.1615e-11, 2.7246e-13, 7.1405e-11, 2.4278e-11, 1.6954e-11,\n 4.0315e-13, 5.1025e-12, 1.1286e-11, 6.7818e-13, 1.9661e-11, 4.2338e-12,\n 3.4507e-12, 1.5927e-11, 2.0294e-11, 7.0608e-11, 1.9859e-13, 1.1405e-11,\n 1.0285e-11, 1.4633e-12, 1.7465e-11, 2.3667e-12, 2.2176e-15, 1.1095e-11,\n 3.1915e-11, 4.0233e-12, 1.1264e-12, 3.6471e-14, 1.3923e-11, 6.3695e-11,\n 1.7853e-12, 3.2955e-16, 5.2120e-11, 1.5693e-10, 7.7389e-14, 1.1779e-12,\n 1.5935e-11, 6.1642e-11, 3.6494e-12, 1.6894e-11, 6.7462e-12, 4.2576e-12,\n 1.2206e-11, 3.8752e-11, 7.1334e-12, 1.5094e-11, 6.2655e-11, 9.1153e-12,\n 2.1666e-11, 6.2905e-11, 1.0920e-11, 9.6296e-11, 6.8475e-12, 3.8434e-12,\n 1.7631e-13, 6.2818e-12, 1.8272e-14, 1.2201e-10, 2.2538e-12, 2.2718e-11,\n 5.4785e-12, 8.7161e-18, 9.2027e-11, 3.9136e-12, 2.3521e-11, 2.4464e-11,\n 4.3730e-12, 6.1481e-11, 1.3202e-12, 2.5427e-12, 4.8434e-12, 5.6399e-11,\n 1.6583e-12, 1.4929e-12, 6.1549e-13, 4.5575e-13, 8.9504e-13, 2.3210e-11,\n 2.1008e-11, 4.3488e-12, 1.7894e-11, 4.4923e-11, 6.7795e-15, 5.0140e-11,\n 2.5175e-11, 2.9899e-12, 1.8745e-11, 6.1572e-12, 7.3065e-11, 9.3592e-12,\n 8.2805e-12, 3.6218e-16, 3.4257e-15, 3.9662e-13, 3.5235e-11, 2.2497e-10,\n 3.2989e-11, 7.1245e-12, 5.6435e-12, 2.1725e-11, 3.8026e-12, 2.7362e-13,\n 5.5107e-12, 1.5203e-11, 2.1368e-11, 2.1559e-12, 2.1327e-12, 7.6453e-12,\n 4.5530e-12, 7.3567e-11, 1.9358e-12, 6.0424e-12, 1.0166e-11, 2.1350e-12,\n 1.6548e-12, 1.3961e-11, 3.1183e-12, 1.3247e-11, 5.2430e-11, 3.8783e-12,\n 4.6006e-11, 5.7063e-12, 2.7667e-14, 5.9548e-12, 9.5274e-12, 2.7998e-12,\n 3.3688e-11, 1.3987e-11, 2.0638e-12, 4.3191e-12, 3.1162e-12, 8.6227e-12,\n 1.1031e-11, 2.2319e-12, 4.6375e-12, 5.9159e-12, 7.4060e-11, 1.2028e-15,\n 4.2156e-12, 3.4294e-11, 1.9098e-12, 6.6839e-11, 8.9582e-12, 7.7147e-12,\n 6.9713e-12, 1.9747e-12, 7.1110e-15, 3.4806e-10, 1.2752e-12, 1.8777e-13,\n 5.1962e-12, 2.2328e-11, 4.6073e-13, 1.5176e-10, 4.1798e-12, 8.2054e-12,\n 5.8246e-12, 1.7914e-11, 3.7585e-11, 5.2139e-13, 4.8957e-16, 9.6792e-12,\n 3.3138e-12, 9.2538e-12, 4.8039e-13, 3.5599e-11, 1.4536e-13, 6.4158e-11,\n 6.0519e-11, 1.2249e-14, 7.0403e-13, 5.5154e-12, 2.0422e-12, 3.4862e-11,\n 4.5081e-12, 3.6924e-11, 6.4172e-12, 2.9153e-14, 2.1617e-14, 8.4001e-13,\n 5.0630e-13, 1.1330e-11, 2.9576e-13, 1.9148e-13, 1.3417e-11, 7.8683e-12,\n 4.8381e-11, 8.0127e-13, 1.5938e-12, 1.5348e-11, 8.5778e-12, 4.3071e-12,\n 8.3768e-12, 9.4670e-15, 4.7940e-12, 1.7174e-11, 4.0482e-12, 4.7172e-15,\n 5.9826e-14, 1.8826e-11, 1.2602e-12, 8.7980e-12, 6.0900e-11, 2.9653e-13,\n 4.6311e-11, 2.1521e-12, 8.2001e-11, 2.3181e-12], device='cuda:0')" + "exp_avg_sq": "tensor([1.1223e-11, 5.1877e-12, 5.7486e-12, 3.4023e-13, 1.2890e-12, 1.8272e-13,\n 9.2084e-13, 4.3941e-13, 5.0337e-13, 5.6755e-13, 8.9426e-12, 7.8721e-12,\n 1.3601e-12, 1.2021e-12, 3.0649e-12, 9.1179e-13, 5.6256e-12, 1.2204e-12,\n 5.7616e-12, 1.5589e-11, 5.1886e-15, 6.5977e-12, 1.9737e-12, 8.1039e-12,\n 9.5251e-13, 7.2115e-13, 1.9676e-11, 2.0509e-13, 8.1365e-13, 1.3711e-12,\n 1.8749e-12, 2.9711e-11, 5.0693e-14, 1.2552e-11, 1.2562e-11, 1.9097e-11,\n 2.3818e-13, 7.6800e-15, 7.9494e-12, 4.5541e-11, 2.2363e-11, 3.5414e-13,\n 5.1565e-13, 8.8801e-13, 1.7826e-12, 5.6296e-12, 7.8913e-12, 2.1289e-12,\n 2.4579e-13, 6.1765e-12, 7.7856e-14, 2.0405e-11, 6.9376e-12, 4.8447e-12,\n 1.1520e-13, 1.4581e-12, 3.2249e-12, 1.9379e-13, 5.6184e-12, 1.2099e-12,\n 9.8605e-13, 4.5513e-12, 5.7993e-12, 2.0177e-11, 5.6749e-14, 3.2590e-12,\n 2.9389e-12, 4.1815e-13, 4.9907e-12, 6.7631e-13, 6.3370e-16, 3.1704e-12,\n 9.1198e-12, 1.1497e-12, 3.2188e-13, 1.0422e-14, 3.9785e-12, 1.8201e-11,\n 5.1016e-13, 9.4172e-17, 1.4894e-11, 4.4845e-11, 2.2114e-14, 3.3660e-13,\n 4.5534e-12, 1.7615e-11, 1.0428e-12, 4.8276e-12, 1.9278e-12, 1.2166e-12,\n 3.4879e-12, 1.1074e-11, 2.0384e-12, 4.3132e-12, 1.7904e-11, 2.6048e-12,\n 6.1912e-12, 1.7976e-11, 3.1205e-12, 2.7517e-11, 1.9567e-12, 1.0983e-12,\n 5.0381e-14, 1.7951e-12, 5.2214e-15, 3.4866e-11, 6.4404e-13, 6.4917e-12,\n 1.5655e-12, 2.4907e-18, 2.6297e-11, 1.1184e-12, 6.7212e-12, 6.9907e-12,\n 1.2496e-12, 1.7569e-11, 3.7727e-13, 7.2660e-13, 1.3840e-12, 1.6116e-11,\n 4.7386e-13, 4.2659e-13, 1.7588e-13, 1.3023e-13, 2.5577e-13, 6.6325e-12,\n 6.0032e-12, 1.2427e-12, 5.1132e-12, 1.2837e-11, 1.9373e-15, 1.4328e-11,\n 7.1940e-12, 8.5438e-13, 5.3565e-12, 1.7595e-12, 2.0879e-11, 2.6745e-12,\n 2.3662e-12, 1.0350e-16, 9.7892e-16, 1.1334e-13, 1.0069e-11, 6.4287e-11,\n 9.4269e-12, 2.0359e-12, 1.6127e-12, 6.2082e-12, 1.0866e-12, 7.8188e-14,\n 1.5747e-12, 4.3445e-12, 6.1061e-12, 6.1606e-13, 6.0943e-13, 2.1847e-12,\n 1.3011e-12, 2.1022e-11, 5.5317e-13, 1.7267e-12, 2.9049e-12, 6.1010e-13,\n 4.7287e-13, 3.9895e-12, 8.9109e-13, 3.7856e-12, 1.4982e-11, 1.1083e-12,\n 1.3147e-11, 1.6306e-12, 7.9061e-15, 1.7016e-12, 2.7225e-12, 8.0007e-13,\n 9.6267e-12, 3.9970e-12, 5.8975e-13, 1.2342e-12, 8.9047e-13, 2.4640e-12,\n 3.1522e-12, 6.3778e-13, 1.3252e-12, 1.6905e-12, 2.1163e-11, 3.4372e-16,\n 1.2046e-12, 9.7997e-12, 5.4575e-13, 1.9100e-11, 2.5599e-12, 2.2045e-12,\n 1.9921e-12, 5.6427e-13, 2.0320e-15, 9.9460e-11, 3.6440e-13, 5.3658e-14,\n 1.4849e-12, 6.3803e-12, 1.3166e-13, 4.3368e-11, 1.1944e-12, 2.3448e-12,\n 1.6644e-12, 5.1190e-12, 1.0740e-11, 1.4899e-13, 1.3990e-16, 2.7659e-12,\n 9.4694e-13, 2.6443e-12, 1.3728e-13, 1.0173e-11, 4.1537e-14, 1.8334e-11,\n 1.7294e-11, 3.5004e-15, 2.0118e-13, 1.5761e-12, 5.8357e-13, 9.9620e-12,\n 1.2882e-12, 1.0551e-11, 1.8338e-12, 8.3307e-15, 6.1773e-15, 2.4004e-13,\n 1.4468e-13, 3.2376e-12, 8.4516e-14, 5.4717e-14, 3.8340e-12, 2.2484e-12,\n 1.3825e-11, 2.2897e-13, 4.5545e-13, 4.3857e-12, 2.4512e-12, 1.2308e-12,\n 2.3937e-12, 2.7053e-15, 1.3699e-12, 4.9077e-12, 1.1568e-12, 1.3480e-15,\n 1.7096e-14, 5.3797e-12, 3.6011e-13, 2.5141e-12, 1.7403e-11, 8.4737e-14,\n 1.3234e-11, 6.1499e-13, 2.3433e-11, 6.6242e-13], device='cuda:0')" }, "33": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([7.5326e-11, 2.6558e-11, 2.1104e-11, 1.3712e-12, 4.4045e-12, 1.7125e-12,\n 6.0641e-12, 1.9858e-12, 2.9704e-12, 1.3816e-12, 5.5497e-11, 5.9627e-11,\n 7.1590e-12, 7.9347e-12, 1.6872e-11, 1.8263e-12, 3.6970e-11, 6.2897e-12,\n 2.6500e-11, 9.5658e-11, 1.4967e-14, 3.0500e-11, 8.5262e-12, 5.6574e-11,\n 5.5242e-12, 6.8938e-12, 6.4719e-11, 1.3667e-12, 4.9579e-12, 6.5542e-12,\n 2.0590e-11, 6.6831e-11, 1.9951e-13, 8.7980e-11, 6.2681e-11, 7.7593e-11,\n 1.4008e-12, 4.8241e-14, 4.2497e-11, 2.2049e-10, 8.8591e-11, 4.3884e-12,\n 4.0150e-12, 4.2567e-12, 1.6227e-11, 3.6663e-11, 3.2323e-11, 1.5100e-11,\n 1.7875e-12, 3.1510e-11, 2.3460e-13, 9.9106e-11, 3.9705e-11, 2.3611e-11,\n 6.8542e-13, 5.4702e-12, 2.3259e-11, 1.5061e-12, 2.0060e-11, 6.7288e-12,\n 3.3398e-12, 1.9836e-11, 3.0468e-11, 6.9986e-11, 2.7397e-13, 1.0254e-11,\n 1.8771e-11, 1.9558e-12, 3.2517e-11, 5.5613e-12, 5.5478e-17, 1.5243e-11,\n 4.0280e-11, 6.9120e-12, 2.0678e-12, 1.6763e-13, 1.1348e-11, 9.0567e-11,\n 1.3077e-12, 8.9810e-15, 4.2286e-11, 1.2390e-10, 1.0567e-13, 1.8902e-12,\n 1.7979e-11, 8.8212e-11, 5.4504e-12, 2.1437e-11, 1.0320e-11, 3.5736e-12,\n 2.0634e-11, 7.1451e-11, 1.6036e-11, 1.8224e-11, 7.9955e-11, 1.3154e-11,\n 2.4764e-11, 9.3898e-11, 2.1400e-11, 1.3967e-10, 1.4933e-11, 4.1574e-12,\n 3.7364e-13, 9.1650e-12, 1.1073e-14, 1.3372e-10, 2.3251e-12, 2.2319e-11,\n 9.9705e-12, 1.1546e-14, 1.0115e-10, 7.2264e-12, 4.2047e-11, 4.8330e-11,\n 6.1062e-12, 6.4947e-11, 2.2551e-12, 4.3913e-12, 1.0979e-11, 5.9999e-11,\n 4.1362e-12, 2.9805e-12, 1.0928e-12, 2.4122e-13, 1.1122e-12, 2.0612e-11,\n 4.9365e-11, 8.7100e-12, 2.6921e-11, 3.5909e-11, 3.6592e-14, 1.0946e-10,\n 3.0648e-11, 5.1202e-12, 3.7125e-11, 1.3828e-11, 9.0570e-11, 1.4312e-11,\n 1.5366e-11, 1.4749e-14, 3.1936e-14, 7.0042e-13, 4.4997e-11, 1.8321e-10,\n 4.1819e-11, 9.9099e-12, 1.0960e-11, 1.9476e-11, 1.0670e-11, 2.9111e-13,\n 5.6703e-12, 2.3676e-11, 2.6269e-11, 3.9955e-12, 4.9595e-12, 6.9205e-12,\n 7.2297e-12, 9.8903e-11, 3.6554e-12, 6.2277e-12, 1.5326e-11, 4.2122e-12,\n 3.3620e-12, 1.8826e-11, 6.6846e-12, 1.9113e-11, 9.8514e-11, 6.6932e-12,\n 4.5396e-11, 8.3662e-12, 1.5270e-13, 9.0478e-12, 2.0685e-11, 5.0089e-12,\n 3.8412e-11, 1.2330e-11, 3.5258e-12, 7.6270e-12, 3.1362e-12, 1.8324e-11,\n 1.3342e-11, 3.8250e-12, 5.2101e-12, 1.0373e-11, 1.3472e-10, 9.9532e-15,\n 7.8698e-12, 2.7098e-11, 4.5912e-12, 1.2424e-10, 1.7942e-11, 9.0964e-12,\n 9.2508e-12, 4.0216e-12, 6.7478e-15, 3.4320e-10, 2.7920e-12, 4.6587e-13,\n 1.2341e-11, 4.1783e-11, 8.0916e-13, 1.4878e-10, 8.4440e-12, 7.4997e-12,\n 1.3880e-11, 3.0270e-11, 4.4786e-11, 8.4800e-13, 2.3207e-14, 1.9939e-11,\n 3.4819e-12, 1.7109e-11, 1.1537e-12, 4.4752e-11, 1.6319e-13, 7.3044e-11,\n 8.6934e-11, 1.4988e-14, 1.4351e-12, 6.4310e-12, 1.6047e-12, 6.2543e-11,\n 7.5057e-12, 7.4372e-11, 1.1226e-11, 7.7261e-14, 1.7699e-14, 1.7735e-12,\n 8.4685e-13, 1.8159e-11, 6.5684e-13, 3.8980e-13, 1.7606e-11, 1.1617e-11,\n 6.6657e-11, 6.0551e-13, 2.5955e-12, 3.2245e-11, 1.0979e-11, 6.2260e-12,\n 1.9164e-11, 9.8857e-15, 9.0175e-12, 2.1268e-11, 7.3676e-12, 2.1421e-16,\n 8.3170e-14, 3.7367e-11, 1.7784e-12, 1.1841e-11, 6.9613e-11, 6.4526e-13,\n 3.8576e-11, 3.5449e-12, 1.0689e-10, 4.5275e-12], device='cuda:0')" + "exp_avg_sq": "tensor([2.1525e-11, 7.5892e-12, 6.0307e-12, 3.9183e-13, 1.2586e-12, 4.8935e-13,\n 1.7329e-12, 5.6747e-13, 8.4883e-13, 3.9481e-13, 1.5859e-11, 1.7039e-11,\n 2.0457e-12, 2.2674e-12, 4.8212e-12, 5.2189e-13, 1.0565e-11, 1.7973e-12,\n 7.5725e-12, 2.7335e-11, 4.2768e-15, 8.7157e-12, 2.4364e-12, 1.6166e-11,\n 1.5786e-12, 1.9700e-12, 1.8494e-11, 3.9054e-13, 1.4168e-12, 1.8729e-12,\n 5.8836e-12, 1.9098e-11, 5.7012e-14, 2.5141e-11, 1.7912e-11, 2.2173e-11,\n 4.0030e-13, 1.3785e-14, 1.2144e-11, 6.3006e-11, 2.5315e-11, 1.2540e-12,\n 1.1473e-12, 1.2164e-12, 4.6370e-12, 1.0477e-11, 9.2365e-12, 4.3148e-12,\n 5.1079e-13, 9.0043e-12, 6.7038e-14, 2.8320e-11, 1.1346e-11, 6.7471e-12,\n 1.9587e-13, 1.5631e-12, 6.6465e-12, 4.3037e-13, 5.7322e-12, 1.9228e-12,\n 9.5438e-13, 5.6683e-12, 8.7064e-12, 1.9999e-11, 7.8290e-14, 2.9300e-12,\n 5.3639e-12, 5.5888e-13, 9.2919e-12, 1.5892e-12, 1.5853e-17, 4.3558e-12,\n 1.1510e-11, 1.9752e-12, 5.9088e-13, 4.7900e-14, 3.2426e-12, 2.5880e-11,\n 3.7367e-13, 2.5664e-15, 1.2083e-11, 3.5405e-11, 3.0197e-14, 5.4013e-13,\n 5.1377e-12, 2.5207e-11, 1.5575e-12, 6.1258e-12, 2.9490e-12, 1.0212e-12,\n 5.8963e-12, 2.0418e-11, 4.5824e-12, 5.2075e-12, 2.2848e-11, 3.7587e-12,\n 7.0766e-12, 2.6832e-11, 6.1152e-12, 3.9913e-11, 4.2672e-12, 1.1880e-12,\n 1.0677e-13, 2.6190e-12, 3.1641e-15, 3.8213e-11, 6.6441e-13, 6.3779e-12,\n 2.8491e-12, 3.2993e-15, 2.8905e-11, 2.0650e-12, 1.2015e-11, 1.3811e-11,\n 1.7449e-12, 1.8559e-11, 6.4440e-13, 1.2549e-12, 3.1374e-12, 1.7145e-11,\n 1.1820e-12, 8.5170e-13, 3.1226e-13, 6.8929e-14, 3.1783e-13, 5.8900e-12,\n 1.4106e-11, 2.4889e-12, 7.6928e-12, 1.0261e-11, 1.0456e-14, 3.1279e-11,\n 8.7578e-12, 1.4631e-12, 1.0609e-11, 3.9516e-12, 2.5881e-11, 4.0897e-12,\n 4.3909e-12, 4.2147e-15, 9.1258e-15, 2.0015e-13, 1.2858e-11, 5.2353e-11,\n 1.1950e-11, 2.8318e-12, 3.1320e-12, 5.5654e-12, 3.0491e-12, 8.3188e-14,\n 1.6203e-12, 6.7656e-12, 7.5065e-12, 1.1417e-12, 1.4172e-12, 1.9776e-12,\n 2.0660e-12, 2.8262e-11, 1.0445e-12, 1.7796e-12, 4.3796e-12, 1.2037e-12,\n 9.6073e-13, 5.3797e-12, 1.9102e-12, 5.4617e-12, 2.8151e-11, 1.9126e-12,\n 1.2972e-11, 2.3907e-12, 4.3636e-14, 2.5855e-12, 5.9108e-12, 1.4313e-12,\n 1.0977e-11, 3.5235e-12, 1.0075e-12, 2.1795e-12, 8.9620e-13, 5.2363e-12,\n 3.8126e-12, 1.0930e-12, 1.4888e-12, 2.9642e-12, 3.8499e-11, 2.8442e-15,\n 2.2489e-12, 7.7434e-12, 1.3120e-12, 3.5502e-11, 5.1271e-12, 2.5994e-12,\n 2.6435e-12, 1.1492e-12, 1.9282e-15, 9.8071e-11, 7.9784e-13, 1.3312e-13,\n 3.5266e-12, 1.1940e-11, 2.3122e-13, 4.2516e-11, 2.4130e-12, 2.1431e-12,\n 3.9663e-12, 8.6499e-12, 1.2798e-11, 2.4232e-13, 6.6316e-15, 5.6978e-12,\n 9.9498e-13, 4.8891e-12, 3.2969e-13, 1.2788e-11, 4.6632e-14, 2.0873e-11,\n 2.4842e-11, 4.2828e-15, 4.1010e-13, 1.8377e-12, 4.5857e-13, 1.7872e-11,\n 2.1448e-12, 2.1252e-11, 3.2080e-12, 2.2078e-14, 5.0578e-15, 5.0679e-13,\n 2.4199e-13, 5.1891e-12, 1.8770e-13, 1.1139e-13, 5.0310e-12, 3.3197e-12,\n 1.9048e-11, 1.7303e-13, 7.4169e-13, 9.2144e-12, 3.1373e-12, 1.7791e-12,\n 5.4762e-12, 2.8249e-15, 2.5768e-12, 6.0776e-12, 2.1053e-12, 6.1213e-17,\n 2.3766e-14, 1.0678e-11, 5.0819e-13, 3.3836e-12, 1.9893e-11, 1.8439e-13,\n 1.1024e-11, 1.0130e-12, 3.0546e-11, 1.2938e-12], device='cuda:0')" }, "34": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([[2.6849e-13, 9.6948e-13, 5.9143e-13, ..., 1.3386e-13, 7.3290e-13,\n 4.4041e-13],\n [1.9109e-14, 2.8588e-15, 3.1159e-14, ..., 8.3152e-14, 9.6665e-14,\n 3.4452e-15],\n [9.3147e-14, 4.0966e-13, 1.2917e-13, ..., 6.6098e-14, 3.9804e-13,\n 7.6162e-14],\n ...,\n [2.3308e-12, 5.5287e-12, 5.7296e-12, ..., 1.0181e-12, 8.5963e-12,\n 1.2441e-11],\n [1.5564e-11, 2.4617e-11, 3.7995e-11, ..., 8.4373e-12, 3.7176e-11,\n 6.8257e-11],\n [2.1354e-10, 4.7365e-10, 5.6600e-10, ..., 1.2049e-10, 7.1290e-10,\n 1.1185e-09]], device='cuda:0')" + "exp_avg_sq": "tensor([[7.6722e-14, 2.7704e-13, 1.6901e-13, ..., 3.8251e-14, 2.0943e-13,\n 1.2585e-13],\n [5.4607e-15, 8.1694e-16, 8.9038e-15, ..., 2.3761e-14, 2.7623e-14,\n 9.8450e-16],\n [2.6617e-14, 1.1706e-13, 3.6912e-14, ..., 1.8888e-14, 1.1374e-13,\n 2.1764e-14],\n ...,\n [6.6605e-13, 1.5799e-12, 1.6373e-12, ..., 2.9093e-13, 2.4565e-12,\n 3.5552e-12],\n [4.4476e-12, 7.0346e-12, 1.0857e-11, ..., 2.4110e-12, 1.0623e-11,\n 1.9505e-11],\n [6.1021e-11, 1.3535e-10, 1.6174e-10, ..., 3.4432e-11, 2.0372e-10,\n 3.1962e-10]], device='cuda:0')" }, "35": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([5.4132e-12, 1.7397e-13, 1.5531e-12, 1.0220e-13, 6.8432e-13, 1.0117e-12,\n 1.4939e-13, 1.2923e-13, 6.2956e-15, 1.0547e-12, 8.5550e-13, 1.4796e-12,\n 3.6750e-14, 3.1111e-14, 3.7853e-13, 2.4827e-13, 8.1897e-14, 3.2414e-12,\n 1.1062e-12, 1.4168e-14, 5.5779e-14, 4.5263e-13, 1.0451e-12, 4.0314e-14,\n 1.9525e-12, 2.2899e-13, 3.2423e-13, 3.3060e-14, 7.5353e-14, 1.5414e-13,\n 1.5384e-13, 2.5588e-12, 3.7081e-13, 1.9849e-14, 7.4876e-16, 7.6500e-13,\n 2.7223e-13, 1.0967e-13, 3.2793e-15, 1.4622e-16, 4.4427e-13, 5.5595e-14,\n 2.7147e-13, 1.2818e-13, 2.1928e-14, 2.4963e-13, 3.1609e-13, 3.2070e-13,\n 1.6933e-13, 7.1953e-13, 2.7923e-13, 7.9419e-14, 1.0363e-13, 6.1026e-14,\n 2.2892e-15, 3.1126e-15, 7.1239e-14, 5.7180e-15, 2.2921e-13, 9.7482e-15,\n 8.0954e-13, 2.6999e-13, 6.3713e-16, 4.9117e-13, 9.4990e-13, 3.2353e-13,\n 4.8509e-13, 9.6050e-14, 1.2301e-14, 7.5826e-15, 4.9701e-13, 5.2501e-14,\n 4.1775e-15, 2.0525e-13, 1.1959e-13, 3.2236e-13, 2.3710e-13, 6.1403e-14,\n 1.0169e-18, 2.2267e-13, 1.6958e-13, 1.4926e-12, 2.2841e-14, 2.1102e-13,\n 2.8707e-14, 1.4484e-13, 8.4433e-14, 2.3364e-16, 1.4367e-13, 1.8646e-14,\n 6.6693e-16, 1.0821e-13, 9.6760e-14, 2.7575e-12, 9.0854e-13, 6.3181e-13,\n 6.8834e-14, 3.2481e-12, 2.2436e-13, 4.1608e-13, 3.8672e-14, 4.4486e-14,\n 2.1910e-14, 2.3796e-13, 2.8925e-15, 8.2985e-13, 1.3627e-12, 1.5147e-13,\n 1.7199e-13, 2.5857e-14, 2.6156e-13, 2.9702e-12, 2.2469e-14, 3.4028e-12,\n 4.7200e-13, 2.1027e-12, 2.0634e-19, 1.9288e-12, 1.4050e-12, 6.9537e-13,\n 2.6612e-12, 1.1906e-14, 1.3893e-12, 5.2388e-15, 7.5825e-13, 1.2630e-12,\n 1.1446e-12, 8.2902e-13, 1.5303e-12, 2.3157e-12, 2.6632e-13, 5.1942e-13,\n 1.5643e-14, 8.9259e-13, 2.5644e-13, 2.3456e-13, 5.4234e-13, 7.3434e-13,\n 6.1700e-16, 1.4327e-13, 1.8013e-12, 6.6288e-15, 3.7886e-13, 1.3978e-13,\n 8.7503e-13, 3.6400e-12, 7.2377e-13, 9.2955e-15, 2.2956e-12, 3.4813e-14,\n 7.4039e-14, 6.0147e-13, 2.4974e-12, 2.0527e-13, 8.8194e-13, 1.3231e-12,\n 9.7432e-15, 2.4570e-15, 9.0065e-14, 4.4279e-13, 4.2243e-13, 1.1716e-13,\n 1.7636e-12, 9.1224e-13, 1.0158e-13, 9.8018e-16, 1.9574e-14, 2.7731e-15,\n 5.5374e-14, 2.4936e-13, 5.0422e-14, 3.0588e-14, 2.1117e-13, 2.5581e-13,\n 2.3285e-14, 6.5329e-15, 2.4479e-13, 5.6516e-14, 3.5099e-14, 5.0023e-15,\n 4.0785e-13, 1.1863e-15, 8.8212e-14, 3.8081e-13, 1.0841e-13, 5.0552e-14,\n 3.7777e-13, 1.2682e-13, 1.2726e-12, 4.1548e-14, 3.5486e-13, 2.9250e-14,\n 3.2142e-13, 3.1917e-12, 4.5048e-13, 9.5994e-13, 4.9102e-13, 1.4353e-12,\n 1.6416e-16, 1.7430e-14, 8.3488e-14, 1.2396e-13, 2.1370e-13, 1.7604e-12,\n 4.4196e-14, 1.3233e-13, 9.8956e-15, 1.4904e-13, 1.7168e-14, 3.9867e-13,\n 6.8222e-17, 2.6808e-13, 1.8113e-14, 1.9746e-13, 1.8168e-12, 9.2099e-13,\n 1.3530e-12, 3.6590e-13, 4.6303e-13, 1.4405e-12, 3.6581e-13, 6.9853e-15,\n 1.8520e-13, 1.6474e-12, 1.0538e-14, 2.1749e-13, 1.3170e-12, 4.6824e-13,\n 4.1306e-13, 3.9962e-13, 3.4587e-13, 1.0298e-14, 1.5190e-13, 8.0241e-14,\n 3.7225e-13, 1.0217e-13, 2.5727e-15, 2.3601e-14, 9.5305e-14, 1.0037e-14,\n 9.2326e-13, 2.5465e-13, 3.8985e-14, 3.0861e-13, 9.4522e-15, 3.5294e-13,\n 9.4798e-13, 1.2164e-13, 4.0304e-13, 1.2564e-13, 9.5848e-14, 9.8571e-13,\n 3.8284e-13, 1.0907e-12, 1.0203e-12, 2.2117e-13, 3.5089e-29, 1.5511e-30,\n 6.3283e-28, 4.4494e-29, 7.0852e-28, 2.7846e-29, 1.8266e-29, 3.8694e-28,\n 2.5357e-28, 3.6908e-28, 1.1142e-29, 1.3801e-29, 3.8705e-28, 3.4785e-28,\n 1.9176e-29, 4.2256e-29, 1.5567e-28, 9.9175e-32, 9.8407e-28, 1.5576e-28,\n 5.1256e-29, 3.5011e-28, 1.5571e-28, 1.5985e-28, 1.1688e-28, 3.7331e-29,\n 9.1718e-29, 1.0497e-29, 7.9312e-28, 1.6301e-28, 1.4160e-28, 1.2612e-28,\n 6.8264e-28, 2.0679e-28, 2.2306e-29, 7.9042e-29, 1.2561e-28, 1.7923e-30,\n 1.7253e-29, 5.4119e-30, 3.7259e-31, 1.6652e-29, 8.2462e-29, 1.1838e-30,\n 1.3524e-28, 3.9761e-29, 2.1011e-28, 1.0856e-30, 1.7184e-28, 1.4803e-30,\n 4.5301e-29, 5.1744e-33, 6.3617e-29, 3.0284e-28, 8.7874e-29, 1.6866e-29,\n 6.5793e-28, 8.4642e-29, 1.2226e-27, 9.8374e-29, 2.8262e-30, 1.5720e-29,\n 1.4271e-28, 8.4614e-29, 2.1801e-29, 1.1362e-28, 4.1582e-28, 5.4487e-28,\n 2.9017e-28, 1.1924e-29, 1.0152e-27, 7.4004e-28, 7.7771e-28, 1.2980e-27,\n 2.6854e-27, 8.4441e-29, 1.6735e-27, 3.5590e-29, 1.2326e-28, 2.4580e-29,\n 3.6223e-28, 1.4348e-27, 8.5728e-28, 5.5858e-28, 1.4862e-28, 2.6880e-28,\n 1.3081e-27, 2.9420e-27, 1.6476e-28, 4.5126e-28, 2.5976e-28, 1.3602e-27,\n 1.4272e-28, 8.3397e-30, 9.4642e-29, 9.0357e-29, 8.4031e-28, 2.8568e-30,\n 1.7133e-28, 1.2389e-30, 9.7410e-29, 4.5995e-28, 6.7622e-28, 5.7953e-29,\n 1.5288e-29, 6.6781e-29, 1.1555e-28, 6.6269e-30, 4.0705e-30, 1.9027e-29,\n 2.1995e-29, 3.7612e-30, 8.7054e-29, 1.0881e-28, 1.5392e-28, 1.3960e-28,\n 9.9277e-29, 1.3110e-30, 4.4281e-32, 2.5511e-30, 1.8069e-30, 1.8770e-28,\n 2.6024e-28, 1.9048e-28, 8.9937e-29, 2.4402e-28, 1.2095e-30, 4.0893e-28,\n 4.3537e-30, 5.0007e-29, 2.0369e-29, 6.4908e-28, 2.5944e-28, 4.0565e-28,\n 3.3748e-28, 7.1481e-30, 6.3580e-30, 7.5053e-29, 1.5026e-28, 4.6507e-28,\n 4.9160e-29, 2.3322e-29, 1.3382e-29, 2.0269e-30, 7.6294e-28, 7.1760e-29,\n 2.1567e-28, 1.6250e-29, 2.5551e-29, 1.6527e-27, 2.8094e-28, 3.1944e-30,\n 2.2801e-28, 2.7066e-28, 7.8557e-28, 4.6698e-28, 5.9814e-28, 2.8614e-28,\n 4.3182e-30, 3.3673e-27, 7.6292e-28, 9.3566e-30, 9.2283e-28, 2.3444e-28,\n 2.0404e-28, 1.2015e-29, 1.3852e-28, 6.6030e-29, 6.8780e-29, 2.7034e-28,\n 2.1621e-28, 2.5729e-32, 2.0060e-28, 2.0276e-29, 3.1217e-30, 4.7710e-29,\n 4.8852e-28, 3.4571e-28, 1.7875e-30, 2.7229e-28, 7.9044e-29, 5.6772e-29,\n 9.5136e-28, 9.5287e-28, 1.0085e-28, 5.1830e-30, 9.7366e-28, 9.0213e-28,\n 4.2066e-28, 5.3503e-29, 5.4597e-29, 2.1683e-28, 2.5428e-28, 2.8477e-28,\n 1.4013e-27, 9.8635e-29, 3.0897e-29, 4.8783e-28, 9.4163e-29, 4.7062e-28,\n 7.4272e-29, 8.0507e-29, 7.6351e-29, 3.6369e-28, 7.0961e-28, 5.1172e-28,\n 3.4087e-30, 1.2925e-29, 1.6609e-28, 6.8624e-32, 5.0372e-29, 3.1605e-28,\n 3.3733e-29, 9.7882e-29, 1.4441e-28, 2.8824e-28, 9.0057e-29, 1.4329e-28,\n 2.9657e-28, 1.6230e-28, 1.4014e-29, 3.2494e-29, 1.0351e-29, 1.8812e-29,\n 1.2786e-28, 8.4791e-29, 2.3916e-28, 8.2138e-28, 5.1789e-28, 3.5540e-28,\n 2.5436e-28, 3.0378e-28, 1.0972e-29, 6.2930e-28, 1.9045e-29, 2.8699e-29,\n 1.3538e-28, 7.8502e-29, 1.4303e-28, 2.8894e-29, 2.2167e-29, 4.4495e-33,\n 1.3428e-29, 1.1111e-29, 2.0669e-28, 3.6744e-28, 2.8585e-29, 2.4696e-28,\n 1.0866e-28, 3.2475e-28, 2.6835e-30, 9.1194e-29, 2.5991e-29, 3.8474e-28,\n 5.9666e-29, 3.4405e-29, 7.1399e-10, 1.1177e-08, 1.9675e-09, 1.8763e-09,\n 1.4845e-08, 1.3190e-08, 1.7993e-10, 9.5904e-11, 2.4043e-10, 2.4064e-09,\n 3.1874e-12, 6.4155e-10, 1.1646e-08, 3.9173e-10, 6.4458e-10, 2.9514e-09,\n 4.3329e-09, 8.9002e-10, 7.9333e-09, 2.9932e-09, 3.3198e-12, 9.3232e-12,\n 1.6101e-09, 1.0607e-08, 8.9361e-09, 3.1132e-11, 4.0251e-09, 2.1649e-10,\n 1.1326e-09, 2.2977e-10, 5.0089e-09, 4.1825e-09, 1.6211e-08, 1.7401e-10,\n 3.3165e-09, 4.7609e-09, 2.8085e-09, 2.8425e-09, 2.5069e-09, 9.9086e-11,\n 1.5353e-08, 5.0276e-11, 2.4836e-10, 4.3728e-10, 7.4349e-09, 6.3579e-10,\n 1.3904e-09, 5.8019e-10, 1.1160e-10, 1.0626e-09, 2.1301e-09, 3.8048e-10,\n 6.2130e-09, 5.8638e-09, 6.3195e-10, 4.4286e-10, 6.8526e-10, 4.9667e-11,\n 4.0125e-09, 1.5221e-09, 3.2444e-09, 3.2109e-09, 3.3807e-10, 8.7544e-11,\n 1.8074e-13, 8.8117e-12, 2.7217e-11, 8.1398e-11, 1.1739e-09, 8.3657e-11,\n 1.0454e-09, 2.7261e-09, 3.9234e-09, 6.2538e-10, 1.4172e-09, 1.7894e-11,\n 7.9923e-10, 2.3667e-08, 1.2418e-10, 1.1931e-09, 1.6018e-09, 1.0955e-08,\n 4.0205e-12, 3.1071e-09, 8.9526e-10, 4.3239e-09, 2.7606e-09, 3.2057e-09,\n 2.6194e-13, 1.0225e-08, 6.8735e-09, 6.2574e-12, 3.0784e-10, 7.4047e-09,\n 3.4392e-10, 1.4365e-09, 2.0302e-11, 5.1021e-09, 9.8236e-10, 4.7747e-09,\n 2.4886e-09, 8.6299e-09, 1.0869e-10, 2.0867e-09, 4.2783e-09, 1.5590e-08,\n 6.8941e-09, 1.2168e-11, 2.6139e-10, 2.6176e-09, 1.6361e-09, 3.0214e-10,\n 4.6096e-10, 5.4494e-10, 1.6161e-09, 2.4036e-10, 2.0364e-09, 4.0415e-11,\n 6.2549e-09, 7.9055e-11, 6.1285e-09, 7.0209e-09, 2.1632e-09, 6.3750e-11,\n 7.3932e-10, 8.6549e-11, 2.0239e-09, 4.7969e-09, 4.7898e-10, 1.2426e-10,\n 4.9874e-09, 2.4747e-10, 2.9128e-09, 1.6413e-09, 8.4555e-10, 5.7589e-12,\n 1.1750e-10, 1.2445e-08, 1.6037e-10, 8.6340e-09, 1.4908e-12, 4.2736e-09,\n 1.6193e-09, 1.6175e-09, 4.7741e-10, 5.8166e-09, 2.1881e-09, 8.4980e-10,\n 3.5274e-09, 1.1009e-11, 8.0170e-09, 5.0497e-10, 6.2172e-11, 8.3364e-10,\n 1.6482e-09, 6.4645e-09, 1.2458e-09, 2.9611e-09, 4.5930e-11, 7.4478e-09,\n 2.1775e-09, 2.2254e-09, 1.7463e-09, 1.3658e-08, 1.2439e-11, 3.8641e-10,\n 1.1851e-08, 9.2497e-10, 1.0927e-09, 5.6846e-11, 1.8935e-09, 4.8989e-09,\n 4.2568e-09, 6.4737e-09, 1.5709e-10, 1.9267e-09, 1.1244e-09, 1.0062e-09,\n 1.0283e-09, 6.4778e-10, 1.6391e-09, 4.1121e-09, 3.1583e-10, 7.0122e-09,\n 7.8309e-09, 1.9028e-09, 1.8688e-09, 1.7787e-09, 1.2865e-11, 3.8062e-10,\n 3.8764e-10, 8.3601e-10, 3.7819e-09, 4.7902e-10, 2.0843e-10, 2.8830e-09,\n 7.9559e-13, 7.0089e-09, 5.9267e-11, 7.6795e-09, 3.2672e-09, 9.1839e-09,\n 2.3706e-10, 1.5709e-09, 6.1686e-09, 2.3211e-09, 2.2535e-10, 5.9539e-09,\n 4.5362e-10, 2.8859e-09, 1.4660e-11, 4.4464e-10, 1.2987e-08, 2.8333e-09,\n 1.3035e-08, 1.4666e-09, 2.0940e-09, 5.1867e-12, 7.2429e-10, 1.8265e-09,\n 1.6795e-11, 1.2295e-10, 1.5778e-09, 5.0519e-10, 7.2586e-10, 4.1345e-10,\n 4.6845e-09, 4.6924e-10, 1.1850e-10, 3.9115e-11, 1.2952e-11, 2.2557e-11,\n 2.8264e-10, 3.8740e-09, 2.7027e-10, 5.5263e-10, 6.9069e-11, 6.5955e-09,\n 3.2343e-09, 6.2715e-09, 1.4270e-08, 2.9555e-10, 7.7238e-09, 3.1521e-09,\n 6.9997e-10, 2.6217e-08, 5.6577e-10, 2.6979e-09, 9.0457e-10, 6.4911e-10,\n 3.3768e-10, 2.9007e-09, 1.5626e-09, 7.5473e-11, 3.9088e-10, 6.7166e-09],\n device='cuda:0')" + "exp_avg_sq": "tensor([1.5469e-12, 4.9713e-14, 4.4380e-13, 2.9205e-14, 1.9555e-13, 2.8910e-13,\n 4.2689e-14, 3.6929e-14, 1.7990e-15, 3.0140e-13, 2.4447e-13, 4.2282e-13,\n 1.0502e-14, 8.8903e-15, 1.0817e-13, 7.0944e-14, 2.3403e-14, 9.2625e-13,\n 3.1610e-13, 4.0486e-15, 1.5939e-14, 1.2934e-13, 2.9865e-13, 1.1520e-14,\n 5.5795e-13, 6.5436e-14, 9.2651e-14, 9.4472e-15, 2.1533e-14, 4.4046e-14,\n 4.3962e-14, 7.3118e-13, 1.0596e-13, 5.6721e-15, 2.1396e-16, 2.1860e-13,\n 7.7791e-14, 3.1340e-14, 9.3709e-16, 4.1784e-17, 1.2695e-13, 1.5887e-14,\n 7.7575e-14, 3.6629e-14, 6.2660e-15, 7.1333e-14, 9.0325e-14, 9.1642e-14,\n 4.8387e-14, 2.0561e-13, 7.9792e-14, 2.2695e-14, 2.9613e-14, 1.7439e-14,\n 6.5417e-16, 8.8944e-16, 2.0357e-14, 1.6340e-15, 6.5499e-14, 2.7856e-15,\n 2.3133e-13, 7.7151e-14, 1.8206e-16, 1.4036e-13, 2.7144e-13, 9.2452e-14,\n 1.3862e-13, 2.7447e-14, 3.5152e-15, 2.1668e-15, 1.4202e-13, 1.5003e-14,\n 1.1938e-15, 5.8651e-14, 3.4174e-14, 9.2118e-14, 6.7754e-14, 1.7546e-14,\n 2.9059e-19, 6.3631e-14, 4.8459e-14, 4.2653e-13, 6.5269e-15, 6.0301e-14,\n 8.2032e-15, 4.1390e-14, 2.4127e-14, 6.6763e-17, 4.1054e-14, 5.3283e-15,\n 1.9058e-16, 3.0921e-14, 2.7650e-14, 7.8799e-13, 2.5962e-13, 1.8055e-13,\n 1.9670e-14, 9.2816e-13, 6.4113e-14, 1.1890e-13, 1.1051e-14, 1.2712e-14,\n 6.2610e-15, 6.8000e-14, 8.2655e-16, 2.3714e-13, 3.8941e-13, 4.3284e-14,\n 4.9149e-14, 7.3888e-15, 7.4743e-14, 8.4876e-13, 6.4206e-15, 9.7238e-13,\n 1.3488e-13, 6.0086e-13, 5.8964e-20, 5.5118e-13, 4.0150e-13, 1.9871e-13,\n 7.6045e-13, 3.4021e-15, 3.9702e-13, 1.4970e-15, 2.1668e-13, 3.6090e-13,\n 3.2707e-13, 2.3690e-13, 4.3729e-13, 6.6174e-13, 7.6104e-14, 1.4843e-13,\n 4.4701e-15, 2.5507e-13, 7.3281e-14, 6.7026e-14, 1.5498e-13, 2.0984e-13,\n 1.7631e-16, 4.0941e-14, 5.1472e-13, 1.8942e-15, 1.0826e-13, 3.9944e-14,\n 2.5005e-13, 1.0402e-12, 2.0682e-13, 2.6563e-15, 6.5600e-13, 9.9481e-15,\n 2.1157e-14, 1.7188e-13, 7.1367e-13, 5.8658e-14, 2.5202e-13, 3.7810e-13,\n 2.7842e-15, 7.0212e-16, 2.5737e-14, 1.2653e-13, 1.2071e-13, 3.3478e-14,\n 5.0397e-13, 2.6068e-13, 2.9028e-14, 2.8009e-16, 5.5934e-15, 7.9245e-16,\n 1.5823e-14, 7.1256e-14, 1.4408e-14, 8.7409e-15, 6.0343e-14, 7.3099e-14,\n 6.6538e-15, 1.8668e-15, 6.9949e-14, 1.6150e-14, 1.0030e-14, 1.4295e-15,\n 1.1655e-13, 3.3901e-16, 2.5207e-14, 1.0882e-13, 3.0980e-14, 1.4446e-14,\n 1.0795e-13, 3.6241e-14, 3.6365e-13, 1.1873e-14, 1.0141e-13, 8.3584e-15,\n 9.1848e-14, 9.1205e-13, 1.2873e-13, 2.7431e-13, 1.4031e-13, 4.1014e-13,\n 4.6910e-17, 4.9807e-15, 2.3857e-14, 3.5422e-14, 6.1067e-14, 5.0306e-13,\n 1.2629e-14, 3.7814e-14, 2.8277e-15, 4.2590e-14, 4.9058e-15, 1.1392e-13,\n 1.9495e-17, 7.6606e-14, 5.1758e-15, 5.6424e-14, 5.1916e-13, 2.6318e-13,\n 3.8662e-13, 1.0456e-13, 1.3231e-13, 4.1162e-13, 1.0453e-13, 1.9961e-15,\n 5.2924e-14, 4.7076e-13, 3.0112e-15, 6.2149e-14, 3.7634e-13, 1.3380e-13,\n 1.1803e-13, 1.1420e-13, 9.8835e-14, 2.9427e-15, 4.3407e-14, 2.2930e-14,\n 1.0637e-13, 2.9195e-14, 7.3517e-16, 6.7442e-15, 2.7234e-14, 2.8681e-15,\n 2.6383e-13, 7.2768e-14, 1.1140e-14, 8.8188e-14, 2.7011e-15, 1.0086e-13,\n 2.7089e-13, 3.4759e-14, 1.1517e-13, 3.5903e-14, 2.7389e-14, 2.8168e-13,\n 1.0940e-13, 3.1169e-13, 2.9157e-13, 6.3202e-14, 1.0027e-29, 4.4325e-31,\n 1.8084e-28, 1.2715e-29, 2.0246e-28, 7.9573e-30, 5.2196e-30, 1.1057e-28,\n 7.2459e-29, 1.0547e-28, 3.1840e-30, 3.9437e-30, 1.1060e-28, 9.9401e-29,\n 5.4798e-30, 1.2075e-29, 4.4484e-29, 2.8340e-32, 2.8120e-28, 4.4510e-29,\n 1.4647e-29, 1.0005e-28, 4.4496e-29, 4.5678e-29, 3.3399e-29, 1.0668e-29,\n 2.6209e-29, 2.9995e-30, 2.2664e-28, 4.6582e-29, 4.0464e-29, 3.6039e-29,\n 1.9507e-28, 5.9091e-29, 6.3742e-30, 2.2587e-29, 3.5893e-29, 5.1216e-31,\n 4.9301e-30, 1.5465e-30, 1.0647e-31, 4.7585e-30, 2.3564e-29, 3.3827e-31,\n 3.8647e-29, 1.1362e-29, 6.0040e-29, 3.1022e-31, 4.9106e-29, 4.2302e-31,\n 1.2945e-29, 1.4786e-33, 1.8179e-29, 8.6540e-29, 2.5111e-29, 4.8195e-30,\n 1.8801e-28, 2.4187e-29, 3.4936e-28, 2.8111e-29, 8.0761e-31, 4.4922e-30,\n 4.0782e-29, 2.4179e-29, 6.2297e-30, 3.2467e-29, 1.1882e-28, 1.5570e-28,\n 8.2919e-29, 3.4074e-30, 2.9011e-28, 2.1147e-28, 2.2224e-28, 3.7093e-28,\n 7.6739e-28, 2.4130e-29, 4.7822e-28, 1.0170e-29, 3.5222e-29, 7.0239e-30,\n 1.0351e-28, 4.1001e-28, 2.4497e-28, 1.5962e-28, 4.2469e-29, 7.6811e-29,\n 3.7380e-28, 8.4070e-28, 4.7081e-29, 1.2895e-28, 7.4228e-29, 3.8870e-28,\n 4.0784e-29, 2.3831e-30, 2.7045e-29, 2.5820e-29, 2.4012e-28, 8.1636e-31,\n 4.8958e-29, 3.5403e-31, 2.7836e-29, 1.3144e-28, 1.9324e-28, 1.6561e-29,\n 4.3687e-30, 1.9083e-29, 3.3020e-29, 1.8937e-30, 1.1632e-30, 5.4372e-30,\n 6.2853e-30, 1.0748e-30, 2.4877e-29, 3.1093e-29, 4.3983e-29, 3.9890e-29,\n 2.8369e-29, 3.7462e-31, 1.2654e-32, 7.2900e-31, 5.1633e-31, 5.3636e-29,\n 7.4366e-29, 5.4431e-29, 2.5700e-29, 6.9731e-29, 3.4562e-31, 1.1686e-28,\n 1.2441e-30, 1.4290e-29, 5.8207e-30, 1.8548e-28, 7.4138e-29, 1.1592e-28,\n 9.6439e-29, 2.0426e-30, 1.8169e-30, 2.1447e-29, 4.2938e-29, 1.3290e-28,\n 1.4048e-29, 6.6645e-30, 3.8240e-30, 5.7921e-31, 2.1802e-28, 2.0506e-29,\n 6.1630e-29, 4.6436e-30, 7.3014e-30, 4.7226e-28, 8.0281e-29, 9.1282e-31,\n 6.5155e-29, 7.7342e-29, 2.2448e-28, 1.3344e-28, 1.7092e-28, 8.1767e-29,\n 1.2340e-30, 9.6223e-28, 2.1801e-28, 2.6737e-30, 2.6371e-28, 6.6992e-29,\n 5.8305e-29, 3.4334e-30, 3.9584e-29, 1.8868e-29, 1.9654e-29, 7.7250e-29,\n 6.1783e-29, 7.3521e-33, 5.7323e-29, 5.7940e-30, 8.9205e-31, 1.3634e-29,\n 1.3960e-28, 9.8789e-29, 5.1078e-31, 7.7810e-29, 2.2588e-29, 1.6223e-29,\n 2.7186e-28, 2.7229e-28, 2.8820e-29, 1.4811e-30, 2.7823e-28, 2.5779e-28,\n 1.2021e-28, 1.5289e-29, 1.5602e-29, 6.1961e-29, 7.2661e-29, 8.1374e-29,\n 4.0043e-28, 2.8186e-29, 8.8291e-30, 1.3940e-28, 2.6908e-29, 1.3448e-28,\n 2.1224e-29, 2.3006e-29, 2.1818e-29, 1.0393e-28, 2.0278e-28, 1.4623e-28,\n 9.7407e-31, 3.6935e-30, 4.7463e-29, 1.9610e-32, 1.4394e-29, 9.0314e-29,\n 9.6395e-30, 2.7970e-29, 4.1268e-29, 8.2367e-29, 2.5734e-29, 4.0946e-29,\n 8.4746e-29, 4.6378e-29, 4.0047e-30, 9.2854e-30, 2.9580e-30, 5.3757e-30,\n 3.6537e-29, 2.4230e-29, 6.8342e-29, 2.3472e-28, 1.4799e-28, 1.0156e-28,\n 7.2685e-29, 8.6808e-29, 3.1353e-30, 1.7983e-28, 5.4423e-30, 8.2011e-30,\n 3.8686e-29, 2.2432e-29, 4.0872e-29, 8.2567e-30, 6.3345e-30, 1.2715e-33,\n 3.8371e-30, 3.1752e-30, 5.9063e-29, 1.0500e-28, 8.1683e-30, 7.0572e-29,\n 3.1050e-29, 9.2801e-29, 7.6682e-31, 2.6059e-29, 7.4272e-30, 1.0994e-28,\n 1.7050e-29, 9.8315e-30, 2.0403e-10, 3.1940e-09, 5.6222e-10, 5.3615e-10,\n 4.2422e-09, 3.7693e-09, 5.1416e-11, 2.7405e-11, 6.8706e-11, 6.8766e-10,\n 9.1082e-13, 1.8333e-10, 3.3279e-09, 1.1194e-10, 1.8419e-10, 8.4339e-10,\n 1.2382e-09, 2.5433e-10, 2.2670e-09, 8.5533e-10, 9.4866e-13, 2.6642e-12,\n 4.6009e-10, 3.0311e-09, 2.5535e-09, 8.8961e-12, 1.1502e-09, 6.1863e-11,\n 3.2364e-10, 6.5659e-11, 1.4313e-09, 1.1952e-09, 4.6324e-09, 4.9725e-11,\n 9.4772e-10, 1.3605e-09, 8.0255e-10, 8.1227e-10, 7.1638e-10, 2.8314e-11,\n 4.3871e-09, 1.4367e-11, 7.0971e-11, 1.2496e-10, 2.1246e-09, 1.8168e-10,\n 3.9731e-10, 1.6579e-10, 3.1890e-11, 3.0365e-10, 6.0868e-10, 1.0873e-10,\n 1.7754e-09, 1.6756e-09, 1.8059e-10, 1.2655e-10, 1.9582e-10, 1.4193e-11,\n 1.1466e-09, 4.3495e-10, 9.2711e-10, 9.1755e-10, 9.6606e-11, 2.5016e-11,\n 5.1649e-14, 2.5180e-12, 7.7775e-12, 2.3260e-11, 3.3545e-10, 2.3906e-11,\n 2.9872e-10, 7.7901e-10, 1.1211e-09, 1.7871e-10, 4.0498e-10, 5.1135e-12,\n 2.2839e-10, 6.7630e-09, 3.5485e-11, 3.4093e-10, 4.5772e-10, 3.1304e-09,\n 1.1489e-12, 8.8787e-10, 2.5583e-10, 1.2356e-09, 7.8886e-10, 9.1605e-10,\n 7.4850e-14, 2.9219e-09, 1.9642e-09, 1.7881e-12, 8.7967e-11, 2.1159e-09,\n 9.8279e-11, 4.1048e-10, 5.8016e-12, 1.4580e-09, 2.8072e-10, 1.3644e-09,\n 7.1115e-10, 2.4661e-09, 3.1059e-11, 5.9629e-10, 1.2226e-09, 4.4551e-09,\n 1.9701e-09, 3.4771e-12, 7.4696e-11, 7.4800e-10, 4.6753e-10, 8.6338e-11,\n 1.3172e-10, 1.5572e-10, 4.6182e-10, 6.8685e-11, 5.8192e-10, 1.1549e-11,\n 1.7874e-09, 2.2591e-11, 1.7513e-09, 2.0063e-09, 6.1816e-10, 1.8217e-11,\n 2.1126e-10, 2.4732e-11, 5.7836e-10, 1.3708e-09, 1.3687e-10, 3.5507e-11,\n 1.4252e-09, 7.0717e-11, 8.3236e-10, 4.6901e-10, 2.4162e-10, 1.6456e-12,\n 3.3576e-11, 3.5562e-09, 4.5828e-11, 2.4672e-09, 4.2601e-13, 1.2212e-09,\n 4.6272e-10, 4.6223e-10, 1.3642e-10, 1.6621e-09, 6.2528e-10, 2.4284e-10,\n 1.0080e-09, 3.1459e-12, 2.2909e-09, 1.4430e-10, 1.7766e-11, 2.3822e-10,\n 4.7099e-10, 1.8473e-09, 3.5598e-10, 8.4615e-10, 1.3125e-11, 2.1283e-09,\n 6.2225e-10, 6.3592e-10, 4.9903e-10, 3.9029e-09, 3.5545e-12, 1.1042e-10,\n 3.3866e-09, 2.6432e-10, 3.1225e-10, 1.6244e-11, 5.4109e-10, 1.3999e-09,\n 1.2164e-09, 1.8499e-09, 4.4888e-11, 5.5058e-10, 3.2130e-10, 2.8752e-10,\n 2.9385e-10, 1.8511e-10, 4.6838e-10, 1.1751e-09, 9.0250e-11, 2.0038e-09,\n 2.2377e-09, 5.4374e-10, 5.3403e-10, 5.0828e-10, 3.6762e-12, 1.0877e-10,\n 1.1077e-10, 2.3890e-10, 1.0807e-09, 1.3688e-10, 5.9560e-11, 8.2385e-10,\n 2.2735e-13, 2.0029e-09, 1.6936e-11, 2.1945e-09, 9.3362e-10, 2.6244e-09,\n 6.7742e-11, 4.4889e-10, 1.7627e-09, 6.6328e-10, 6.4397e-11, 1.7014e-09,\n 1.2963e-10, 8.2468e-10, 4.1892e-12, 1.2706e-10, 3.7112e-09, 8.0963e-10,\n 3.7247e-09, 4.1909e-10, 5.9839e-10, 1.4821e-12, 2.0697e-10, 5.2193e-10,\n 4.7994e-12, 3.5133e-11, 4.5087e-10, 1.4436e-10, 2.0742e-10, 1.1815e-10,\n 1.3386e-09, 1.3409e-10, 3.3861e-11, 1.1177e-11, 3.7012e-12, 6.4459e-12,\n 8.0767e-11, 1.1070e-09, 7.7232e-11, 1.5792e-10, 1.9737e-11, 1.8847e-09,\n 9.2421e-10, 1.7921e-09, 4.0777e-09, 8.4456e-11, 2.2071e-09, 9.0075e-10,\n 2.0002e-10, 7.4917e-09, 1.6167e-10, 7.7095e-10, 2.5849e-10, 1.8549e-10,\n 9.6494e-11, 8.2889e-10, 4.4652e-10, 2.1567e-11, 1.1170e-10, 1.9193e-09],\n device='cuda:0')" }, "36": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([[2.3916e-10, 1.3355e-13, 1.3355e-09, ..., 3.4586e-12, 3.4818e-11,\n 9.6246e-11],\n [3.2417e-10, 3.3785e-14, 1.8943e-09, ..., 5.4215e-12, 5.1450e-11,\n 1.3588e-10],\n [1.8306e-10, 2.3285e-14, 1.0706e-09, ..., 2.8288e-12, 3.0470e-11,\n 6.9231e-11],\n ...,\n [1.3852e-10, 1.9771e-14, 8.4007e-10, ..., 1.2993e-12, 2.0959e-11,\n 5.5722e-11],\n [1.8197e-11, 4.8669e-14, 1.0024e-10, ..., 3.4863e-13, 1.9703e-12,\n 7.5544e-12],\n [9.6284e-11, 1.1602e-13, 5.6499e-10, ..., 1.6443e-12, 1.3158e-11,\n 3.8664e-11]], device='cuda:0')" + "exp_avg_sq": "tensor([[6.8342e-11, 3.8163e-14, 3.8163e-10, ..., 9.8833e-13, 9.9496e-12,\n 2.7503e-11],\n [9.2635e-11, 9.6544e-15, 5.4131e-10, ..., 1.5492e-12, 1.4702e-11,\n 3.8828e-11],\n [5.2309e-11, 6.6539e-15, 3.0594e-10, ..., 8.0835e-13, 8.7071e-12,\n 1.9783e-11],\n ...,\n [3.9582e-11, 5.6497e-15, 2.4006e-10, ..., 3.7129e-13, 5.9892e-12,\n 1.5923e-11],\n [5.2000e-12, 1.3907e-14, 2.8644e-11, ..., 9.9624e-14, 5.6303e-13,\n 2.1587e-12],\n [2.7514e-11, 3.3154e-14, 1.6145e-10, ..., 4.6987e-13, 3.7600e-12,\n 1.1049e-11]], device='cuda:0')" }, "37": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([8.6521e-09, 1.2037e-08, 6.4371e-09, 3.4176e-08, 1.0372e-10, 9.9126e-10,\n 1.8153e-08, 3.4086e-08, 2.7570e-08, 1.3729e-08, 5.9507e-10, 2.8667e-08,\n 1.4162e-09, 7.0624e-11, 3.7278e-09, 3.5617e-11, 1.5570e-08, 2.9486e-08,\n 8.2463e-09, 1.7617e-09, 3.2107e-09, 2.3631e-09, 1.1249e-08, 1.4420e-10,\n 3.7484e-10, 3.1807e-09, 3.2796e-10, 4.7950e-09, 2.2723e-09, 7.1341e-09,\n 1.5600e-11, 3.8282e-09, 3.8737e-09, 1.3491e-09, 1.7369e-10, 7.5948e-09,\n 2.6005e-10, 4.3670e-08, 7.3232e-09, 5.2118e-10, 1.6452e-08, 1.0717e-09,\n 3.1490e-09, 2.8475e-10, 5.7460e-09, 9.8548e-09, 1.2374e-10, 1.2873e-11,\n 8.0839e-09, 7.3448e-09, 1.0492e-08, 4.3729e-10, 4.5600e-09, 1.7092e-08,\n 2.1834e-10, 1.9098e-09, 1.7276e-08, 6.8540e-11, 7.9279e-10, 1.4233e-09,\n 5.9747e-09, 4.6765e-09, 9.6207e-10, 9.1411e-09, 5.7467e-11, 1.4409e-08,\n 1.9625e-09, 3.7520e-09, 4.2249e-09, 1.5470e-09, 7.2679e-10, 1.4239e-08,\n 1.7446e-08, 3.6551e-08, 2.3587e-10, 2.8537e-11, 1.4393e-09, 1.2289e-08,\n 1.8622e-10, 3.5342e-11, 2.8888e-09, 1.8317e-10, 7.3241e-09, 1.2581e-09,\n 1.3844e-09, 7.1560e-10, 1.3983e-09, 2.0104e-09, 2.0724e-09, 2.4178e-08,\n 2.1071e-11, 1.4375e-08, 5.9609e-09, 1.6016e-08, 3.9033e-08, 2.4879e-08,\n 2.5264e-09, 3.4061e-08, 1.2244e-09, 4.3048e-10, 1.5161e-08, 5.2118e-12,\n 6.6090e-09, 3.9928e-09, 3.7640e-10, 1.0564e-08, 6.3490e-12, 6.2212e-08,\n 1.2889e-08, 1.3371e-08, 4.7254e-11, 3.8677e-10, 1.1978e-09, 1.1949e-08,\n 3.7665e-09, 1.0041e-08, 2.1233e-11, 1.2410e-08, 1.5605e-08, 6.2562e-09,\n 2.0864e-08, 8.2424e-09, 6.4423e-09, 9.2326e-10, 4.3649e-08, 2.2889e-08,\n 5.8177e-09, 2.4043e-09, 4.9843e-09, 1.7741e-11, 5.1904e-08, 5.0444e-09,\n 6.5528e-11, 1.9351e-11, 3.3714e-08, 1.3005e-08, 1.9582e-09, 8.6047e-10,\n 7.9572e-11, 2.6055e-11, 2.1493e-08, 4.4705e-11, 3.2250e-08, 6.2753e-10,\n 7.2805e-10, 1.5434e-09, 1.2842e-08, 1.0772e-08, 3.1283e-09, 3.7672e-11,\n 6.2025e-10, 1.9792e-10, 2.3486e-11, 2.1235e-08, 1.8789e-09, 2.7858e-08,\n 6.6839e-10, 1.5616e-08, 2.1129e-09, 2.1163e-08, 5.5321e-09, 9.0048e-10,\n 9.5002e-09, 8.2650e-09, 1.1467e-08, 2.5272e-08, 5.3098e-08, 2.5087e-11,\n 9.9628e-11, 4.8316e-09, 2.7939e-09, 4.2537e-11, 1.7134e-08, 2.2054e-09,\n 2.4771e-08, 3.0890e-11, 1.1576e-08, 1.2917e-09, 5.5733e-09, 5.5932e-09,\n 5.9631e-08, 5.9374e-09, 3.8339e-09, 5.6999e-12, 1.1171e-09, 1.7315e-08,\n 7.0726e-11, 6.3258e-09, 1.6116e-08, 1.4359e-09, 1.6466e-08, 1.4318e-09,\n 4.1509e-09, 3.7994e-09, 2.5106e-08, 1.3435e-08, 1.5807e-08, 1.1604e-10,\n 5.0356e-09, 3.2251e-08, 3.4481e-11, 2.9914e-09, 1.5420e-09, 2.2612e-08,\n 2.2092e-08, 3.1598e-09, 1.0452e-08, 2.1560e-09, 1.5444e-10, 5.8911e-11,\n 6.9963e-11, 3.5183e-09, 1.8128e-08, 1.6135e-08, 5.8224e-09, 2.0254e-08,\n 2.4601e-09, 1.2169e-08, 8.8650e-11, 3.1932e-10, 6.0976e-12, 1.7187e-09,\n 1.9065e-08, 1.4207e-09, 4.7914e-10, 9.8124e-09, 2.7622e-09, 8.6252e-09,\n 6.1337e-10, 1.2167e-08, 1.7027e-10, 6.6576e-09, 2.0694e-08, 8.7289e-11,\n 1.7231e-08, 3.8666e-10, 4.0916e-09, 1.5358e-11, 3.2489e-08, 6.5481e-10,\n 2.9780e-09, 6.7227e-10, 7.8653e-10, 4.6069e-09, 6.7799e-09, 4.6983e-09,\n 1.6984e-08, 6.4585e-09, 2.7494e-08, 2.6418e-09, 1.4300e-08, 3.0999e-13,\n 1.1271e-09, 5.3104e-09, 6.8644e-10, 3.6539e-09], device='cuda:0')" + "exp_avg_sq": "tensor([2.4724e-09, 3.4396e-09, 1.8394e-09, 9.7661e-09, 2.9639e-11, 2.8326e-10,\n 5.1874e-09, 9.7404e-09, 7.8784e-09, 3.9232e-09, 1.7005e-10, 8.1918e-09,\n 4.0469e-10, 2.0181e-11, 1.0653e-09, 1.0178e-11, 4.4494e-09, 8.4259e-09,\n 2.3564e-09, 5.0342e-10, 9.1748e-10, 6.7528e-10, 3.2145e-09, 4.1207e-11,\n 1.0711e-10, 9.0890e-10, 9.3718e-11, 1.3702e-09, 6.4934e-10, 2.0386e-09,\n 4.4577e-12, 1.0940e-09, 1.1069e-09, 3.8552e-10, 4.9634e-11, 2.1703e-09,\n 7.4312e-11, 1.2479e-08, 2.0927e-09, 1.4893e-10, 4.7012e-09, 3.0624e-10,\n 8.9984e-10, 8.1369e-11, 1.6420e-09, 2.8161e-09, 3.5360e-11, 3.6786e-12,\n 2.3100e-09, 2.0988e-09, 2.9983e-09, 1.2496e-10, 1.3031e-09, 4.8841e-09,\n 6.2393e-11, 5.4574e-10, 4.9367e-09, 1.9586e-11, 2.2655e-10, 4.0673e-10,\n 1.7073e-09, 1.3363e-09, 2.7492e-10, 2.6121e-09, 1.6422e-11, 4.1174e-09,\n 5.6081e-10, 1.0722e-09, 1.2073e-09, 4.4208e-10, 2.0769e-10, 4.0689e-09,\n 4.9853e-09, 1.0445e-08, 6.7402e-11, 8.1547e-12, 4.1130e-10, 3.5117e-09,\n 5.3213e-11, 1.0099e-11, 8.2549e-10, 5.2343e-11, 2.0929e-09, 3.5951e-10,\n 3.9560e-10, 2.0449e-10, 3.9957e-10, 5.7448e-10, 5.9222e-10, 6.9091e-09,\n 6.0211e-12, 4.1076e-09, 1.7034e-09, 4.5767e-09, 1.1154e-08, 7.1092e-09,\n 7.2194e-10, 9.7331e-09, 3.4987e-10, 1.2301e-10, 4.3323e-09, 1.4893e-12,\n 1.8886e-09, 1.1410e-09, 1.0756e-10, 3.0187e-09, 1.8143e-12, 1.7778e-08,\n 3.6831e-09, 3.8209e-09, 1.3503e-11, 1.1052e-10, 3.4227e-10, 3.4144e-09,\n 1.0763e-09, 2.8692e-09, 6.0674e-12, 3.5464e-09, 4.4594e-09, 1.7878e-09,\n 5.9621e-09, 2.3553e-09, 1.8410e-09, 2.6383e-10, 1.2473e-08, 6.5408e-09,\n 1.6625e-09, 6.8704e-10, 1.4243e-09, 5.0695e-12, 1.4832e-08, 1.4415e-09,\n 1.8725e-11, 5.5298e-12, 9.6340e-09, 3.7162e-09, 5.5957e-10, 2.4589e-10,\n 2.2738e-11, 7.4453e-12, 6.1418e-09, 1.2775e-11, 9.2156e-09, 1.7932e-10,\n 2.0804e-10, 4.4103e-10, 3.6697e-09, 3.0781e-09, 8.9394e-10, 1.0765e-11,\n 1.7724e-10, 5.6556e-11, 6.7113e-12, 6.0680e-09, 5.3691e-10, 7.9606e-09,\n 1.9100e-10, 4.4625e-09, 6.0378e-10, 6.0474e-09, 1.5809e-09, 2.5732e-10,\n 2.7148e-09, 2.3618e-09, 3.2767e-09, 7.2216e-09, 1.5173e-08, 7.1688e-12,\n 2.8470e-11, 1.3807e-09, 7.9839e-10, 1.2155e-11, 4.8961e-09, 6.3020e-10,\n 7.0785e-09, 8.8272e-12, 3.3079e-09, 3.6912e-10, 1.5926e-09, 1.5983e-09,\n 1.7040e-08, 1.6967e-09, 1.0956e-09, 1.6288e-12, 3.1922e-10, 4.9479e-09,\n 2.0211e-11, 1.8077e-09, 4.6053e-09, 4.1033e-10, 4.7053e-09, 4.0914e-10,\n 1.1862e-09, 1.0857e-09, 7.1742e-09, 3.8391e-09, 4.5170e-09, 3.3159e-11,\n 1.4390e-09, 9.2158e-09, 9.8534e-12, 8.5480e-10, 4.4063e-10, 6.4615e-09,\n 6.3128e-09, 9.0294e-10, 2.9866e-09, 6.1610e-10, 4.4132e-11, 1.6834e-11,\n 1.9992e-11, 1.0054e-09, 5.1802e-09, 4.6107e-09, 1.6638e-09, 5.7878e-09,\n 7.0300e-10, 3.4775e-09, 2.5332e-11, 9.1249e-11, 1.7425e-12, 4.9114e-10,\n 5.4478e-09, 4.0599e-10, 1.3692e-10, 2.8040e-09, 7.8933e-10, 2.4647e-09,\n 1.7527e-10, 3.4767e-09, 4.8656e-11, 1.9025e-09, 5.9134e-09, 2.4943e-11,\n 4.9240e-09, 1.1049e-10, 1.1692e-09, 4.3887e-12, 9.2839e-09, 1.8712e-10,\n 8.5099e-10, 1.9211e-10, 2.2476e-10, 1.3165e-09, 1.9374e-09, 1.3426e-09,\n 4.8532e-09, 1.8456e-09, 7.8567e-09, 7.5492e-10, 4.0862e-09, 8.8583e-14,\n 3.2207e-10, 1.5175e-09, 1.9616e-10, 1.0441e-09], device='cuda:0')" }, "38": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([[6.0536e-08, 1.6024e-08, 2.2937e-08, ..., 2.9215e-09, 1.7832e-08,\n 3.2929e-08],\n [6.2506e-09, 1.6247e-09, 2.4324e-09, ..., 3.2392e-10, 1.8005e-09,\n 3.2645e-09],\n [6.8091e-09, 1.7555e-09, 2.5344e-09, ..., 3.1189e-10, 2.0211e-09,\n 3.7756e-09],\n [7.1340e-09, 1.9696e-09, 2.6820e-09, ..., 3.3829e-10, 2.1298e-09,\n 3.9542e-09]], device='cuda:0')" + "exp_avg_sq": "tensor([[1.7299e-08, 4.5790e-09, 6.5544e-09, ..., 8.3484e-10, 5.0958e-09,\n 9.4098e-09],\n [1.7862e-09, 4.6427e-10, 6.9507e-10, ..., 9.2562e-11, 5.1449e-10,\n 9.3285e-10],\n [1.9458e-09, 5.0165e-10, 7.2422e-10, ..., 8.9125e-11, 5.7755e-10,\n 1.0789e-09],\n [2.0386e-09, 5.6283e-10, 7.6639e-10, ..., 9.6670e-11, 6.0860e-10,\n 1.1300e-09]], device='cuda:0')" }, "39": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([1.5180e-06, 1.5641e-07, 1.6925e-07, 1.8079e-07], device='cuda:0')" + "exp_avg_sq": "tensor([4.3379e-07, 4.4697e-08, 4.8363e-08, 5.1662e-08], device='cuda:0')" }, "40": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([[6.0536e-08, 1.6024e-08, 2.2937e-08, ..., 2.9215e-09, 1.7832e-08,\n 3.2929e-08],\n [6.2506e-09, 1.6247e-09, 2.4324e-09, ..., 3.2392e-10, 1.8005e-09,\n 3.2645e-09],\n [6.8091e-09, 1.7555e-09, 2.5344e-09, ..., 3.1189e-10, 2.0211e-09,\n 3.7756e-09],\n [7.1340e-09, 1.9696e-09, 2.6820e-09, ..., 3.3829e-10, 2.1298e-09,\n 3.9542e-09]], device='cuda:0')" + "exp_avg_sq": "tensor([[1.7299e-08, 4.5790e-09, 6.5544e-09, ..., 8.3484e-10, 5.0958e-09,\n 9.4098e-09],\n [1.7862e-09, 4.6427e-10, 6.9507e-10, ..., 9.2562e-11, 5.1449e-10,\n 9.3285e-10],\n [1.9458e-09, 5.0165e-10, 7.2422e-10, ..., 8.9125e-11, 5.7755e-10,\n 1.0789e-09],\n [2.0386e-09, 5.6283e-10, 7.6639e-10, ..., 9.6670e-11, 6.0860e-10,\n 1.1300e-09]], device='cuda:0')" }, "41": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([1.5180e-06, 1.5641e-07, 1.6925e-07, 1.8079e-07], device='cuda:0')" + "exp_avg_sq": "tensor([4.3379e-07, 4.4697e-08, 4.8363e-08, 5.1662e-08], device='cuda:0')" }, "42": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "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([[6.0536e-08, 1.6024e-08, 2.2937e-08, ..., 2.9215e-09, 1.7832e-08,\n 3.2929e-08],\n [6.2506e-09, 1.6247e-09, 2.4324e-09, ..., 3.2392e-10, 1.8005e-09,\n 3.2645e-09],\n [6.8091e-09, 1.7555e-09, 2.5344e-09, ..., 3.1189e-10, 2.0211e-09,\n 3.7756e-09],\n [7.1340e-09, 1.9696e-09, 2.6820e-09, ..., 3.3829e-10, 2.1298e-09,\n 3.9542e-09]], device='cuda:0')" + "exp_avg_sq": "tensor([[1.7299e-08, 4.5790e-09, 6.5544e-09, ..., 8.3484e-10, 5.0958e-09,\n 9.4098e-09],\n [1.7862e-09, 4.6427e-10, 6.9507e-10, ..., 9.2562e-11, 5.1449e-10,\n 9.3285e-10],\n [1.9458e-09, 5.0165e-10, 7.2422e-10, ..., 8.9125e-11, 5.7755e-10,\n 1.0789e-09],\n [2.0386e-09, 5.6283e-10, 7.6639e-10, ..., 9.6670e-11, 6.0860e-10,\n 1.1300e-09]], device='cuda:0')" }, "43": { - "step": "tensor(5008.)", + "step": "tensor(6260.)", "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([1.5180e-06, 1.5641e-07, 1.6925e-07, 1.8079e-07], device='cuda:0')" + "exp_avg_sq": "tensor([4.3379e-07, 4.4697e-08, 4.8363e-08, 5.1662e-08], device='cuda:0')" }, "8": { - "step": "tensor(3756.)", - "exp_avg": "tensor([[-4.8903e-07, -1.4045e-07, 0.0000e+00, ..., -7.3340e-07,\n 1.4998e-05, 2.1285e-07],\n [-1.9784e-08, 1.2568e-06, 0.0000e+00, ..., -1.4949e-07,\n -3.2219e-06, -4.5879e-07],\n [ 1.6308e-07, 5.0047e-07, 5.6052e-45, ..., -1.2668e-06,\n -2.1327e-06, -3.3276e-07],\n ...,\n [ 3.7087e-07, 2.1190e-06, 0.0000e+00, ..., 2.7206e-06,\n 3.8326e-06, -1.6441e-08],\n [-1.9052e-07, 3.0003e-07, 4.2088e-11, ..., 5.3748e-07,\n 9.0616e-06, 2.4403e-07],\n [ 1.0096e-07, 3.6192e-08, 0.0000e+00, ..., -4.0276e-08,\n -8.6625e-08, 8.4811e-09]], device='cuda:0')", - "exp_avg_sq": "tensor([[5.7180e-11, 9.6704e-12, 0.0000e+00, ..., 6.1756e-11, 8.3537e-11,\n 7.5953e-11],\n [3.4743e-11, 5.2152e-11, 0.0000e+00, ..., 3.9450e-11, 6.6994e-10,\n 2.2686e-11],\n [1.5295e-10, 1.7517e-11, 9.5494e-17, ..., 3.3178e-11, 2.2306e-10,\n 4.8251e-11],\n ...,\n [6.5790e-11, 6.3602e-11, 0.0000e+00, ..., 2.4873e-11, 2.1711e-10,\n 9.2870e-12],\n [7.1215e-11, 7.2712e-11, 3.3815e-18, ..., 1.1008e-10, 1.3688e-10,\n 1.2221e-11],\n [8.9506e-11, 3.5121e-11, 0.0000e+00, ..., 3.3409e-11, 1.1314e-11,\n 1.1286e-11]], device='cuda:0')" + "step": "tensor(5008.)", + "exp_avg": "tensor([[ 2.7322e-07, 1.1414e-07, 0.0000e+00, ..., 2.4205e-07,\n -5.1692e-07, -1.4170e-06],\n [ 2.2839e-06, 1.5650e-06, 5.6052e-45, ..., -9.9496e-07,\n 2.6799e-06, -3.6803e-07],\n [-2.3394e-06, 1.9509e-06, -5.6052e-45, ..., -1.0197e-06,\n 2.8309e-06, -8.4491e-08],\n ...,\n [-3.2771e-07, 4.9473e-07, 0.0000e+00, ..., 8.5927e-07,\n 4.6747e-06, 1.5459e-08],\n [ 3.1645e-07, 6.8645e-08, 5.6052e-45, ..., 2.6501e-07,\n -2.1393e-06, 4.6323e-07],\n [ 3.3234e-07, 4.0286e-07, 0.0000e+00, ..., -1.5698e-07,\n 5.0433e-07, -9.5544e-10]], device='cuda:0')", + "exp_avg_sq": "tensor([[3.8411e-11, 1.6491e-11, 0.0000e+00, ..., 4.1278e-11, 6.3133e-11,\n 4.8459e-11],\n [3.3939e-11, 2.8431e-11, 1.0593e-17, ..., 2.3176e-11, 5.6537e-10,\n 1.3718e-11],\n [9.9958e-11, 1.7606e-11, 3.4223e-17, ..., 2.5640e-11, 2.3179e-10,\n 5.3783e-11],\n ...,\n [4.5590e-11, 4.8537e-11, 0.0000e+00, ..., 1.1090e-11, 2.7077e-10,\n 3.4143e-12],\n [4.7733e-11, 4.1754e-11, 9.6628e-19, ..., 6.4338e-11, 8.9093e-11,\n 6.6330e-12],\n [1.0614e-10, 2.6774e-11, 0.0000e+00, ..., 4.4612e-11, 1.0074e-11,\n 3.2938e-12]], device='cuda:0')" }, "9": { - "step": "tensor(3756.)", - "exp_avg": "tensor([ 2.4792e-05, 1.3956e-05, -4.1845e-06, ..., 4.0973e-05,\n -1.0948e-05, 3.6630e-08], device='cuda:0')", - "exp_avg_sq": "tensor([6.9141e-09, 1.2114e-08, 8.5247e-09, ..., 9.6235e-09, 8.3769e-09,\n 5.0240e-09], device='cuda:0')" + "step": "tensor(5008.)", + "exp_avg": "tensor([-3.3946e-06, 7.3244e-07, 2.2150e-05, ..., 1.8893e-05,\n -4.6869e-06, -4.4450e-05], device='cuda:0')", + "exp_avg_sq": "tensor([6.4343e-09, 8.5541e-09, 7.6326e-09, ..., 7.9219e-09, 6.1019e-09,\n 5.4206e-09], device='cuda:0')" }, "10": { - "step": "tensor(3756.)", - "exp_avg": "tensor([[-7.2394e-07, 9.9176e-07, 3.3189e-07, ..., 1.7271e-07,\n -1.6048e-07, 1.9659e-07],\n [ 2.8910e-07, 2.1317e-06, 3.0884e-07, ..., -6.8876e-07,\n -9.0864e-08, 3.8183e-07],\n [ 5.6754e-07, -1.7926e-06, 1.3411e-06, ..., -3.3674e-07,\n 2.2415e-07, -2.7177e-07],\n ...,\n [-8.5051e-07, 3.0009e-06, -1.7829e-07, ..., -7.6412e-07,\n 5.3012e-07, -1.4112e-07],\n [ 5.1514e-07, 4.0788e-07, -6.8600e-07, ..., -5.6059e-07,\n -1.2135e-07, -6.5150e-08],\n [-1.0949e-06, 3.6378e-07, -7.0776e-08, ..., 6.4191e-08,\n 5.4157e-07, -3.2797e-07]], device='cuda:0')", - "exp_avg_sq": "tensor([[8.4067e-12, 1.2408e-11, 1.0855e-11, ..., 5.0959e-12, 8.8449e-12,\n 1.4629e-11],\n [6.8501e-12, 2.5491e-11, 1.0946e-11, ..., 1.3110e-11, 9.4138e-12,\n 8.2610e-12],\n [1.0422e-11, 2.5428e-11, 1.3102e-11, ..., 1.3734e-11, 1.3396e-11,\n 1.5140e-11],\n ...,\n [1.7972e-11, 3.0262e-11, 1.4742e-11, ..., 1.4682e-11, 1.1497e-11,\n 3.0129e-11],\n [1.2878e-11, 2.3289e-11, 1.1450e-11, ..., 1.6605e-11, 9.6001e-12,\n 9.2498e-12],\n [8.9619e-12, 2.5655e-11, 1.4756e-11, ..., 2.5977e-11, 1.3260e-11,\n 9.4459e-12]], device='cuda:0')" + "step": "tensor(5008.)", + "exp_avg": "tensor([[-6.1846e-07, -5.8479e-07, 2.5030e-07, ..., -6.3079e-07,\n 2.6942e-07, -1.4575e-06],\n [-6.7804e-07, -5.5368e-07, -2.4734e-07, ..., -1.2554e-07,\n -1.8391e-07, -7.5793e-07],\n [-2.0860e-07, 2.3801e-07, -7.6508e-07, ..., -8.8244e-07,\n -2.5882e-07, -7.3373e-07],\n ...,\n [-3.7170e-07, 5.6562e-07, -5.9484e-07, ..., 8.6176e-07,\n 4.4681e-08, 1.7646e-06],\n [ 3.6604e-08, 1.1395e-07, -4.7692e-07, ..., 5.9535e-07,\n -3.3610e-07, 7.8428e-07],\n [ 5.5183e-07, 2.0077e-07, 4.0415e-07, ..., -6.0778e-07,\n 5.5857e-08, 4.5510e-07]], device='cuda:0')", + "exp_avg_sq": "tensor([[3.6347e-12, 5.6132e-12, 8.3232e-12, ..., 3.1764e-12, 4.8772e-12,\n 7.4605e-12],\n [4.4338e-12, 1.3370e-11, 7.4670e-12, ..., 9.4064e-12, 5.8145e-12,\n 6.1145e-12],\n [6.6633e-12, 1.4536e-11, 8.9892e-12, ..., 8.4219e-12, 7.9061e-12,\n 1.0989e-11],\n ...,\n [9.4204e-12, 1.5763e-11, 9.7963e-12, ..., 9.4881e-12, 7.4499e-12,\n 1.6717e-11],\n [7.2053e-12, 1.2562e-11, 7.2204e-12, ..., 1.0506e-11, 5.4578e-12,\n 6.4776e-12],\n [5.1055e-12, 1.5171e-11, 1.0771e-11, ..., 1.8017e-11, 7.9056e-12,\n 6.2604e-12]], device='cuda:0')" }, "11": { - "step": "tensor(2504.)", - "exp_avg": "tensor([[ 1.2778e-06, 1.0016e-06, 1.4296e-09, ..., 1.3046e-06,\n 3.0553e-06, 5.1079e-06],\n [ 8.0536e-07, -5.7379e-08, 6.2426e-23, ..., 1.0864e-07,\n -5.8014e-07, 2.4104e-06],\n [ 4.7670e-07, 6.8543e-09, 0.0000e+00, ..., 2.0272e-07,\n -4.9184e-06, 3.7464e-08],\n ...,\n [ 3.7641e-07, 3.5784e-09, 5.6052e-45, ..., 3.5863e-09,\n 2.7128e-08, 3.5433e-09],\n [ 2.3868e-07, 7.5190e-08, 0.0000e+00, ..., 2.8456e-06,\n 7.6243e-06, -1.0378e-07],\n [-4.4136e-07, 2.0859e-07, 5.1894e-12, ..., -6.9046e-07,\n -4.3875e-07, 1.5809e-09]], device='cuda:0')", - "exp_avg_sq": "tensor([[6.1974e-11, 1.0273e-10, 3.3333e-15, ..., 7.2439e-10, 2.1659e-10,\n 6.2048e-11],\n [1.9480e-10, 1.6404e-11, 3.0942e-17, ..., 3.5723e-11, 1.8729e-10,\n 1.4771e-10],\n [1.1885e-10, 3.8970e-11, 0.0000e+00, ..., 1.9061e-11, 2.5919e-10,\n 1.5255e-11],\n ...,\n [1.1889e-10, 1.8331e-11, 1.4347e-17, ..., 6.5394e-12, 3.5567e-11,\n 8.2972e-12],\n [7.3297e-11, 3.0906e-11, 0.0000e+00, ..., 2.4757e-10, 1.0390e-09,\n 8.8339e-12],\n [3.3567e-11, 9.7932e-11, 2.4873e-15, ..., 6.1352e-10, 8.3351e-11,\n 1.6817e-11]], device='cuda:0')" + "step": "tensor(3756.)", + "exp_avg": "tensor([[-5.3524e-07, 9.6118e-07, 1.2363e-25, ..., 5.1984e-06,\n 6.2675e-08, -1.0419e-06],\n [ 2.9958e-06, 6.9983e-08, -7.8282e-29, ..., -1.2293e-08,\n 5.2561e-07, -5.0621e-07],\n [-9.9318e-07, -2.0644e-08, 0.0000e+00, ..., 1.9619e-08,\n 1.2122e-06, 4.6951e-08],\n ...,\n [-3.1998e-07, -1.1345e-08, 3.6434e-44, ..., -3.5600e-08,\n 1.2326e-06, -3.4479e-08],\n [ 1.1720e-07, 1.5895e-07, 0.0000e+00, ..., -1.9828e-07,\n 1.7987e-06, 1.1621e-08],\n [-2.7216e-07, -1.2180e-06, 4.5601e-10, ..., 3.7599e-06,\n -7.3007e-08, -2.6747e-08]], device='cuda:0')", + "exp_avg_sq": "tensor([[2.6335e-11, 5.1120e-11, 1.5334e-15, ..., 3.4880e-10, 1.1002e-10,\n 3.9218e-11],\n [1.0686e-10, 7.4373e-12, 1.1449e-17, ..., 1.4784e-11, 3.0855e-10,\n 1.3192e-10],\n [8.1004e-11, 2.8370e-11, 0.0000e+00, ..., 9.3603e-12, 1.9376e-10,\n 1.4890e-11],\n ...,\n [1.5055e-10, 7.8671e-12, 4.1943e-18, ..., 7.9320e-12, 2.6623e-11,\n 7.3317e-12],\n [4.4816e-11, 2.5245e-11, 0.0000e+00, ..., 2.7306e-10, 9.8597e-10,\n 2.8800e-12],\n [1.5426e-11, 5.0527e-11, 3.0999e-15, ..., 4.6796e-10, 6.6555e-11,\n 5.9601e-12]], device='cuda:0')" }, "12": { - "step": "tensor(2504.)", - "exp_avg": "tensor([ 2.4118e-05, 4.5042e-05, -3.2372e-06, ..., -8.2233e-07,\n 4.6033e-05, -2.9744e-06], device='cuda:0')", - "exp_avg_sq": "tensor([1.1918e-08, 1.0005e-08, 9.7556e-09, ..., 6.0090e-09, 1.3704e-08,\n 1.0873e-08], device='cuda:0')" + "step": "tensor(3756.)", + "exp_avg": "tensor([-7.4234e-06, -8.7620e-06, 2.2323e-05, ..., 3.0812e-06,\n 1.5739e-05, 3.5401e-05], device='cuda:0')", + "exp_avg_sq": "tensor([6.4784e-09, 6.2796e-09, 7.2937e-09, ..., 5.7122e-09, 1.1787e-08,\n 6.9539e-09], device='cuda:0')" }, "13": { - "step": "tensor(2504.)", - "exp_avg": "tensor([[-8.3479e-07, 5.4764e-07, -1.0775e-06, ..., 1.1893e-07,\n 4.6882e-07, -8.5413e-07],\n [-3.7459e-08, -3.5614e-07, 7.4919e-07, ..., -8.6935e-07,\n 3.3207e-07, -7.6068e-07],\n [-5.8790e-07, 1.2587e-07, -1.2202e-06, ..., -8.2444e-07,\n -2.7528e-07, -3.6729e-07],\n ...,\n [-1.3363e-06, -1.0929e-06, -4.5763e-07, ..., 4.2793e-07,\n -7.8387e-07, 6.9590e-08],\n [-4.0714e-07, -1.3150e-06, -1.2504e-07, ..., -4.0317e-07,\n 4.5231e-07, -2.0854e-07],\n [ 1.0503e-06, -5.4846e-07, -1.1344e-07, ..., -1.7724e-07,\n -2.3885e-07, -6.0194e-07]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.5478e-11, 7.0544e-12, 3.1112e-11, ..., 5.3519e-12, 5.3167e-12,\n 8.8194e-12],\n [1.5035e-11, 9.8603e-12, 4.0443e-11, ..., 1.0026e-11, 6.5719e-12,\n 1.8348e-11],\n [1.5861e-11, 1.0703e-11, 4.2777e-11, ..., 2.0316e-11, 1.0770e-11,\n 1.3061e-11],\n ...,\n [2.0773e-11, 2.2542e-11, 2.2820e-11, ..., 7.9171e-12, 1.1264e-11,\n 1.9694e-11],\n [2.3086e-11, 2.1862e-11, 1.4965e-11, ..., 1.5300e-11, 8.7356e-12,\n 1.6244e-11],\n [1.7134e-11, 1.4242e-11, 9.0803e-12, ..., 8.5805e-12, 7.6396e-12,\n 2.8204e-11]], device='cuda:0')" + "step": "tensor(3756.)", + "exp_avg": "tensor([[-1.9878e-08, -3.5001e-07, -2.2776e-07, ..., -4.6995e-07,\n -5.4793e-08, 3.7963e-07],\n [ 2.9615e-07, 1.2807e-07, 2.6242e-07, ..., -4.3392e-07,\n 1.5130e-07, 1.7820e-06],\n [-2.1330e-07, 1.3158e-07, 2.0135e-07, ..., 2.8594e-07,\n -3.6920e-07, -7.8347e-07],\n ...,\n [ 1.4045e-07, 1.7456e-07, -2.6393e-07, ..., 5.9195e-07,\n -2.3096e-07, -1.2911e-06],\n [ 9.6099e-08, 2.3452e-07, -5.0690e-08, ..., 1.9186e-06,\n 9.6991e-09, 1.1152e-06],\n [ 3.4397e-07, 7.6174e-08, 1.2815e-07, ..., 4.4035e-07,\n -3.4829e-07, 3.5899e-07]], device='cuda:0')", + "exp_avg_sq": "tensor([[6.0186e-12, 3.6141e-12, 2.1950e-11, ..., 3.2164e-12, 3.1609e-12,\n 4.3140e-12],\n [6.5656e-12, 4.7244e-12, 3.0155e-11, ..., 5.8952e-12, 3.6596e-12,\n 7.5404e-12],\n [7.9908e-12, 5.2944e-12, 3.1349e-11, ..., 1.5430e-11, 5.8450e-12,\n 6.3165e-12],\n ...,\n [9.9630e-12, 9.3546e-12, 1.2246e-11, ..., 4.8785e-12, 6.6605e-12,\n 8.7260e-12],\n [1.0101e-11, 1.0969e-11, 9.4197e-12, ..., 1.3302e-11, 4.7231e-12,\n 7.9104e-12],\n [8.0239e-12, 7.7243e-12, 5.3173e-12, ..., 6.9940e-12, 4.4882e-12,\n 1.1129e-11]], device='cuda:0')" } }, "param_groups": [ { - "lr": 0.005000500000000001, + "lr": 0.0034555695366224513, "name": "shared", "betas": [ 0.9, @@ -242,7 +242,7 @@ ] }, { - "lr": 0.005000500000000001, + "lr": 0.0034555695366224513, "name": "scale_384", "betas": [ 0.9, @@ -265,7 +265,7 @@ ] }, { - "lr": 0.005000500000000001, + "lr": 0.0034555695366224513, "name": "scale_768", "betas": [ 0.9, @@ -288,7 +288,7 @@ ] }, { - "lr": 0.005000500000000001, + "lr": 0.0034555695366224513, "name": "scale_1024", "betas": [ 0.9, @@ -311,7 +311,7 @@ ] }, { - "lr": 0.005000500000000001, + "lr": 0.0034555695366224513, "name": "scale_1280", "betas": [ 0.9, @@ -334,7 +334,7 @@ ] }, { - "lr": 0.0025005, + "lr": 0.001728112022559819, "name": "fusion", "betas": [ 0.9, @@ -390,7 +390,7 @@ "T_i": 10, "T_mult": 2, "eta_min": 1e-06, - "T_cur": 5, + "T_cur": 6, "base_lrs": [ 0.01, 0.01, @@ -399,27 +399,27 @@ 0.01, 0.005 ], - "last_epoch": 5, + "last_epoch": 6, "_step_count": 0, "_is_initial": false, "_get_lr_called_within_step": false, "_last_lr": [ - 0.005000500000000001, - 0.005000500000000001, - 0.005000500000000001, - 0.005000500000000001, - 0.005000500000000001, - 0.0025005 + 0.0034555695366224513, + 0.0034555695366224513, + 0.0034555695366224513, + 0.0034555695366224513, + 0.0034555695366224513, + 0.001728112022559819 ] }, "metrics": { - "best_val_acc": 82.08, - "best_epoch": 4, + "best_val_acc": 82.2, + "best_epoch": 5, "scale_accuracies": { - "384": 82.08, - "768": 82.112, - "1024": 82.248, - "1280": 81.954 + "384": 82.2, + "768": 82.33, + "1024": 82.326, + "1280": 82.178 } }, "train_config": {