diff --git "a/weights/final_model_metadata.json" "b/weights/final_model_metadata.json" --- "a/weights/final_model_metadata.json" +++ "b/weights/final_model_metadata.json" @@ -1,226 +1,316 @@ { - "epoch": 19, + "epoch": 9, "optimizer_state_dict": { "state": { "0": { - "step": "tensor(25040.)", - "exp_avg": "tensor([[ 4.7855e-07, -2.8362e-05, 6.1365e-06, ..., 2.3855e-06,\n -7.0232e-06, 6.6765e-07],\n [-7.8494e-06, 2.6277e-05, -2.8105e-06, ..., 1.4638e-05,\n -5.1096e-06, -1.3632e-06],\n [ 1.9106e-06, 2.9793e-06, 5.8921e-07, ..., -8.6641e-07,\n -3.0229e-06, -1.1548e-07],\n ...,\n [ 1.2039e-05, 2.1023e-05, 2.6477e-05, ..., -8.6529e-06,\n 9.0014e-06, 3.5765e-05],\n [-2.8159e-05, 2.4238e-05, 9.9622e-06, ..., 1.3619e-05,\n -6.8560e-06, 2.8892e-05],\n [ 3.1999e-06, 2.0921e-05, 5.1910e-06, ..., -3.8031e-06,\n 1.4771e-05, -1.1220e-05]], device='cuda:0')", - "exp_avg_sq": "tensor([[7.5278e-09, 8.6641e-09, 4.3816e-09, ..., 4.8477e-09, 4.5908e-09,\n 3.6245e-09],\n [6.9600e-09, 6.2931e-09, 7.3132e-09, ..., 5.2736e-09, 4.0526e-09,\n 3.1510e-09],\n [3.2098e-10, 3.6571e-10, 1.8603e-10, ..., 2.8248e-10, 2.7618e-10,\n 1.9356e-10],\n ...,\n [7.0773e-09, 5.6125e-09, 5.2420e-09, ..., 4.0586e-09, 4.1386e-09,\n 3.1464e-09],\n [8.4677e-09, 7.3359e-09, 6.3060e-09, ..., 6.2596e-09, 4.9606e-09,\n 3.8731e-09],\n [2.6034e-09, 3.8147e-09, 2.3548e-09, ..., 1.5739e-09, 1.7697e-09,\n 1.3828e-09]], device='cuda:0')" + "step": "tensor(12520.)", + "exp_avg": "tensor([[-6.3960e-05, -6.0342e-06, 2.5991e-05, ..., 7.9031e-06,\n -1.0817e-05, 1.0993e-06],\n [ 8.8662e-06, -1.3754e-05, -6.6039e-06, ..., -7.5104e-06,\n 6.7735e-06, 1.1299e-05],\n [ 2.0843e-05, -5.0735e-05, -6.4116e-05, ..., 4.3576e-05,\n 6.4431e-05, -3.8258e-05],\n ...,\n [-3.0679e-06, -6.5986e-05, -5.2217e-05, ..., 1.6762e-06,\n 6.4565e-05, -6.0786e-05],\n [-5.6052e-45, -5.6052e-45, 5.6052e-45, ..., -5.6052e-45,\n 5.6052e-45, -5.6052e-45],\n [ 1.9304e-05, 9.2353e-05, -7.0835e-05, ..., 7.7714e-05,\n 4.0903e-05, -2.3749e-05]], device='cuda:0')", + "exp_avg_sq": "tensor([[1.4227e-08, 9.8451e-09, 5.8651e-09, ..., 7.5263e-09, 6.9033e-09,\n 6.9766e-09],\n [5.8534e-09, 5.5103e-09, 3.1861e-09, ..., 2.9842e-09, 3.1072e-09,\n 2.1654e-09],\n [1.4612e-08, 1.2444e-08, 1.6592e-08, ..., 1.1548e-08, 9.1652e-09,\n 8.7804e-09],\n ...,\n [1.3218e-08, 1.5949e-08, 2.4706e-08, ..., 1.1718e-08, 9.9901e-09,\n 1.0700e-08],\n [5.9810e-16, 3.1201e-15, 9.4819e-16, ..., 5.5587e-18, 2.9919e-15,\n 1.7631e-17],\n [1.3356e-08, 1.3326e-08, 1.5134e-08, ..., 1.3426e-08, 7.8721e-09,\n 6.7177e-09]], device='cuda:0')" }, "1": { - "step": "tensor(25040.)", - "exp_avg": "tensor([-2.7998e-04, 6.3148e-04, -1.9552e-05, ..., -4.1836e-04,\n -4.2263e-04, 4.7905e-04], device='cuda:0')", - "exp_avg_sq": "tensor([1.0341e-05, 9.6406e-06, 6.0121e-07, ..., 8.6792e-06, 1.1257e-05,\n 3.9933e-06], device='cuda:0')" + "step": "tensor(12520.)", + "exp_avg": "tensor([-9.0691e-04, 7.2409e-04, 3.0193e-03, ..., 1.0748e-03,\n 5.6052e-45, 3.6514e-03], device='cuda:0')", + "exp_avg_sq": "tensor([1.4690e-05, 6.2727e-06, 2.1278e-05, ..., 2.5137e-05, 2.0251e-11,\n 2.1901e-05], device='cuda:0')" }, "2": { - "step": "tensor(25040.)", - "exp_avg": "tensor([[ 3.4862e-06, -1.0523e-07, -3.6105e-10, ..., -9.3377e-07,\n -4.4853e-06, 1.6478e-06],\n [ 1.0881e-05, 2.0274e-08, 1.5601e-06, ..., -1.2355e-06,\n -6.5253e-07, -6.3509e-09],\n [ 3.4015e-07, 3.3309e-06, -2.2469e-07, ..., -2.0356e-06,\n 5.7929e-07, -3.4297e-06],\n ...,\n [ 0.0000e+00, 5.6052e-45, 0.0000e+00, ..., 5.6052e-45,\n 5.6052e-45, 0.0000e+00],\n [-2.7422e-06, 8.3219e-07, -3.0563e-07, ..., 3.1836e-06,\n 1.8240e-06, -7.1395e-07],\n [ 2.5603e-07, 5.9040e-07, 5.9978e-08, ..., -1.7944e-08,\n -2.0139e-06, 2.9062e-08]], device='cuda:0')", - "exp_avg_sq": "tensor([[6.9834e-10, 8.9197e-11, 2.8956e-11, ..., 1.2364e-09, 2.9846e-10,\n 1.3164e-10],\n [3.2663e-09, 1.0254e-10, 1.2508e-09, ..., 9.5398e-10, 2.8868e-10,\n 1.1235e-11],\n [6.3681e-10, 8.5574e-10, 1.5433e-10, ..., 1.1856e-09, 2.1785e-09,\n 3.8453e-11],\n ...,\n [0.0000e+00, 7.2328e-27, 0.0000e+00, ..., 2.1189e-28, 2.5306e-28,\n 0.0000e+00],\n [3.3579e-09, 4.1970e-10, 1.2735e-10, ..., 2.1111e-09, 3.6014e-10,\n 8.0853e-10],\n [1.6551e-09, 7.1761e-10, 2.9226e-11, ..., 3.8150e-10, 3.1043e-09,\n 2.9589e-10]], device='cuda:0')" + "step": "tensor(12520.)", + "exp_avg": "tensor([[-2.7164e-05, -1.9900e-07, 9.5126e-07, ..., 9.4927e-06,\n 5.6052e-45, -9.6058e-07],\n [-2.7432e-05, 5.1023e-07, -2.3293e-06, ..., 2.8908e-07,\n -5.6052e-45, -2.5127e-06],\n [-3.1435e-07, -3.2044e-07, -1.2302e-07, ..., 2.4341e-06,\n -5.6052e-45, -1.9378e-07],\n ...,\n [ 5.8933e-07, 1.5870e-07, -7.7837e-07, ..., 9.6177e-07,\n 5.6052e-45, 2.0668e-06],\n [ 3.8947e-06, -6.5624e-09, -9.0244e-06, ..., -9.2989e-07,\n -5.6052e-45, 6.2904e-06],\n [-1.1632e-05, -7.0983e-06, 8.5217e-06, ..., 1.5592e-06,\n 5.6052e-45, 2.1561e-06]], device='cuda:0')", + "exp_avg_sq": "tensor([[1.9624e-09, 2.2615e-10, 1.6196e-09, ..., 6.6825e-09, 4.7686e-16,\n 6.7964e-10],\n [2.0070e-09, 7.9523e-10, 4.2056e-10, ..., 2.2196e-10, 2.6458e-14,\n 4.2702e-09],\n [2.3410e-10, 3.6115e-10, 3.2898e-10, ..., 6.0066e-10, 1.8233e-13,\n 1.4097e-09],\n ...,\n [4.7294e-11, 4.8290e-11, 5.5586e-10, ..., 7.2295e-11, 1.8310e-14,\n 1.7603e-11],\n [5.4413e-10, 2.2035e-11, 1.9991e-10, ..., 3.0986e-10, 3.6935e-17,\n 4.8808e-10],\n [3.3296e-09, 3.1063e-10, 6.0098e-09, ..., 1.7746e-09, 3.2723e-15,\n 2.6791e-09]], device='cuda:0')" }, "3": { - "step": "tensor(25040.)", - "exp_avg": "tensor([-2.3417e-05, 1.0869e-04, 5.0860e-05, -2.0679e-05, 1.0940e-05,\n -6.3716e-06, -1.1589e-04, 4.9230e-05, -2.2433e-05, 7.8860e-06,\n 3.5839e-05, 5.3028e-05, -5.7574e-05, 9.9217e-06, 2.6153e-04,\n -5.4614e-05, -7.9927e-05, 4.3496e-05, 1.9399e-04, -3.2363e-06,\n 2.5123e-05, 8.5945e-06, -5.2042e-05, -6.1650e-05, -2.3809e-05,\n 3.5911e-05, 2.8088e-06, 5.6052e-45, 7.9537e-05, -1.8822e-05,\n -3.1249e-05, -9.5124e-06, -3.1025e-05, 2.9351e-05, -2.8860e-05,\n -1.7788e-05, 5.5750e-05, -2.6437e-05, 6.3399e-05, 7.1006e-06,\n 5.1195e-05, -5.3072e-05, 7.3677e-05, 6.8925e-05, -5.8142e-06,\n 8.1950e-08, -1.1390e-04, -7.3443e-05, 1.4399e-05, 1.2382e-04,\n -2.0351e-05, -7.3158e-05, 4.7580e-05, 1.2104e-05, 3.4692e-05,\n -1.3044e-05, -5.2895e-05, -5.7057e-05, 3.5187e-06, -8.6241e-05,\n -5.0459e-06, 6.1589e-05, 1.2421e-04, 2.3834e-05, 2.4572e-05,\n -1.9903e-05, 1.3943e-05, -7.9278e-05, 2.1805e-05, 3.3724e-05,\n -9.4838e-06, -6.3505e-05, 1.5431e-05, -7.0602e-05, -4.0375e-05,\n 8.4890e-05, 3.1639e-05, -7.3266e-05, -8.2743e-06, -8.3554e-05,\n 1.6140e-05, 1.7322e-05, 6.9952e-05, -4.5774e-06, 5.5513e-05,\n -7.7110e-05, -3.2018e-05, 8.5333e-05, 1.2339e-05, 3.9323e-05,\n -3.3312e-05, 4.4363e-05, 1.2088e-04, -7.9918e-05, 2.4782e-05,\n -2.0157e-05, 2.7429e-05, -4.9655e-05, 1.1388e-05, -2.0318e-05,\n -3.2760e-05, 2.7578e-07, -1.5805e-04, 9.8155e-06, -1.5212e-05,\n 8.6702e-05, -3.7241e-05, -6.3969e-05, 3.9605e-05, -1.0433e-06,\n -2.7178e-05, -3.9835e-05, -7.3296e-06, -2.6904e-05, -2.3320e-05,\n -1.8482e-06, -3.4229e-06, -2.0749e-05, 6.8529e-05, 5.6052e-45,\n -3.8389e-06, -5.0350e-05, -2.2720e-05, 2.1243e-06, 1.3075e-04,\n 8.6677e-06, -5.1888e-05, 2.9661e-05, -5.3901e-05, 2.4973e-05,\n 9.8852e-06, -1.5108e-05, 5.6052e-45, -4.6532e-05, 8.1966e-05,\n 6.8804e-05, 4.4820e-06, -1.6495e-05, -2.6936e-05, 2.0875e-05,\n -3.4943e-05, 5.6052e-45, -2.5878e-05, 1.1841e-06, 5.6052e-45,\n 5.6052e-45, 4.3776e-05, 5.1759e-05, 5.6052e-45, 8.7514e-06,\n -7.4270e-05, -6.8552e-07, -5.7954e-05, 1.6541e-05, -5.5310e-05,\n -1.0141e-04, 3.0188e-05, 3.0022e-06, 5.6052e-45, 5.6146e-05,\n 1.0592e-05, -1.2533e-04, -1.6736e-04, 3.8657e-05, 1.0466e-04,\n -1.0396e-04, 6.0132e-06, 5.6052e-45, -7.5488e-06, 7.4029e-05,\n 1.1692e-07, -2.8829e-05, 4.4952e-05, -7.9139e-06, 1.1082e-04,\n -1.2981e-05, -1.0684e-04, 1.8862e-05, -5.5697e-05, -3.0518e-05,\n -3.2962e-05, -2.3387e-05, -2.3661e-04, 5.4104e-05, 2.0828e-05,\n 5.6052e-45, 1.0441e-04, 2.2550e-05, -7.4156e-05, 1.0652e-05,\n -1.6022e-06, -9.8403e-06, 1.1428e-04, 1.2335e-04, 2.5260e-05,\n 3.3257e-05, 6.8477e-05, 6.5280e-06, 8.9212e-05, 5.6052e-45,\n 2.5781e-05, -1.7584e-05, -3.1680e-05, -6.9079e-05, 3.3577e-05,\n 1.0441e-05, 7.9169e-06, 1.3519e-05, -5.4950e-06, 5.9122e-05,\n 5.8427e-05, -7.7070e-06, -2.9243e-05, 1.6918e-05, -9.3517e-05,\n -5.5299e-06, 2.7353e-05, 1.2932e-05, -1.1321e-04, -1.4087e-05,\n 4.5730e-05, -4.0543e-06, 2.0128e-06, -2.9839e-06, 3.0967e-05,\n 1.0607e-06, -2.7993e-05, 1.6548e-06, 6.4309e-06, 3.7802e-05,\n -3.1499e-07, -1.2325e-06, 2.8199e-05, -1.8746e-05, 5.6052e-45,\n -8.3653e-05, 4.5681e-05, 5.6052e-45, 3.6873e-05, 3.5082e-05,\n 1.4882e-04, -3.4230e-05, -4.8431e-05, -2.6842e-05, 2.5021e-05,\n 3.5697e-05, 8.6234e-06, -2.3245e-05, -3.4960e-05, 1.8614e-05,\n -6.4279e-05, 6.3831e-05, -2.4537e-05, -2.1178e-05, 3.5083e-05,\n -3.9646e-04, 5.6052e-45, 1.5414e-05, -1.5911e-06, 7.5868e-05,\n 5.6052e-45, -1.9916e-05, 1.5159e-05, -3.8783e-05, 1.2012e-05,\n -4.1169e-05, 7.3680e-06, -1.4271e-05, 7.8516e-05, 1.5750e-05,\n -4.1940e-06, -1.8588e-05, 4.3696e-05, -5.1770e-05, -1.3639e-05,\n 1.4840e-06, -2.9052e-05, -1.2809e-06, -2.4812e-05, 2.8432e-05,\n 4.9698e-08, 2.8866e-05, 3.7695e-05, -1.2246e-05, 6.9908e-06,\n -9.7580e-05, 3.3683e-07, -1.7445e-05, 1.4177e-05, -4.2066e-06,\n -3.5414e-05, -1.0041e-05, 5.9036e-06, 1.6073e-05, 6.7210e-05,\n 7.7319e-05, -1.0682e-04, -3.2485e-05, -3.5182e-05, -1.0756e-04,\n -1.2817e-05, 1.2925e-05, -1.4820e-05, -1.5358e-05, -4.1672e-05,\n -6.4453e-05, 1.9520e-05, 6.9213e-05, 6.7571e-05, -1.1525e-05,\n 1.5110e-05, 5.6052e-45, 4.2821e-05, 1.9012e-05, 8.2789e-05,\n 5.6052e-45, 1.7125e-05, -3.8602e-05, 2.0533e-05, 1.0396e-05,\n -2.5968e-05, 1.2791e-05, 2.9435e-05, 1.4803e-05, -8.5510e-05,\n 5.6052e-45, 5.0868e-05, 6.3225e-05, 4.0152e-05, -3.2611e-05,\n 1.0751e-04, -2.1428e-05, 5.1900e-05, -6.1643e-06, -3.4757e-05,\n -2.5864e-05, 3.8154e-05, -1.5429e-05, 5.6052e-45, -2.7859e-05,\n 5.9847e-05, 7.3820e-05, 5.0649e-05, -4.9991e-05, 2.3503e-05,\n -7.0443e-05, -4.6426e-05, 8.1839e-05, -7.2296e-06, 1.9290e-05,\n 3.0952e-05, -2.9861e-05, 1.4697e-05, 5.6052e-45, 1.1975e-04,\n 1.1474e-04, 8.2174e-05, 3.3650e-05, 9.8003e-05, -3.6013e-06,\n 3.0776e-05, 5.4368e-05, 5.6052e-45, -5.2574e-05, 1.9225e-06,\n -1.5645e-05, -8.7506e-06, 1.9282e-06, 6.3703e-07, -5.4891e-06,\n -3.7549e-05, 3.3294e-05, 1.2946e-05, 9.0820e-05, 3.8182e-05,\n 3.3370e-05, -2.5296e-07, 2.2593e-05, -9.7731e-06, 1.0657e-04,\n -1.5258e-05, 3.6311e-05, -2.5219e-06, -1.2653e-05, 5.6052e-45,\n -3.9828e-05, -5.2843e-05, -2.0941e-04, 6.2772e-05, -5.6157e-05,\n -9.3354e-06, 6.7556e-05, 5.2356e-05, 1.4235e-06, -3.0392e-05,\n 9.3062e-05, -1.6556e-05, -3.7741e-05, -3.3301e-06, -7.3157e-05,\n -1.8686e-05, -4.7856e-05, 5.2285e-05, 2.1052e-05, 6.1708e-05,\n 8.2419e-05, 4.1136e-06, -6.6875e-05, 2.1566e-05, -8.0902e-05,\n 6.0522e-05, -5.6052e-45, -4.0707e-05, -4.3265e-05, 2.4748e-05,\n 8.2249e-06, 1.1793e-05, 1.9633e-04, 2.1058e-06, 2.5392e-05,\n 4.8291e-05, 5.4061e-05, -1.5555e-05, -1.0215e-04, -2.8745e-06,\n -1.9555e-05, -2.3926e-05, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 2.8059e-05, 1.1178e-05, 3.6317e-05, -4.2448e-05, 4.8013e-05,\n -4.4770e-05, -1.8677e-05, 1.5416e-05, 6.8403e-05, 1.5013e-04,\n -1.3413e-05, 3.1211e-05, -2.0441e-05, 5.6052e-45, 3.8583e-05,\n -1.6123e-05, -2.5162e-05, -2.0423e-05, 1.0552e-04, -5.1513e-05,\n -6.3589e-05, -5.3522e-05, -3.8040e-05, -6.1606e-06, -1.5163e-04,\n -2.6281e-05, 4.1089e-05, 1.0197e-04, 5.6052e-45, 9.7272e-05,\n 5.8079e-05, 1.9560e-05, 2.6933e-05, -2.1879e-05, 3.6683e-05,\n -4.3979e-06, 3.8709e-05, -8.0339e-06, 1.4417e-05, -4.5544e-06,\n 1.2192e-05, 5.6052e-45, 4.0745e-05, -3.3265e-06, -5.4629e-05,\n -2.9782e-05, 4.6080e-05, 4.8436e-05, -2.6694e-06, 8.2390e-06,\n 2.5490e-05, 1.1124e-04, 4.6261e-06, 1.0959e-05, 3.0985e-05,\n -2.3084e-05, 9.5142e-05, 8.5315e-05, 6.9060e-05, -7.2174e-06,\n -5.2024e-05, 5.6447e-06, 4.9549e-05, 1.0093e-05, 2.9826e-05,\n 5.2224e-06, -1.2561e-04, -1.3390e-04, -1.1311e-05, 1.1167e-05,\n 8.6068e-07, 7.4644e-05, 1.4270e-04, 2.1062e-05, -2.3120e-05,\n 9.6362e-06, 5.2068e-05, -1.1231e-05, 1.0859e-04, -1.5335e-04,\n 1.6361e-04, -1.4166e-05, 3.6272e-04, 7.8985e-05, -8.9354e-06,\n -6.8768e-05, 5.0716e-05, 6.4886e-05, 5.5478e-10, 1.9653e-05,\n 9.4578e-06, -3.0212e-05, -5.1889e-05, -1.4030e-05, 4.8059e-05,\n -2.0071e-05, -1.4633e-05, -4.8723e-05, 5.1856e-06, -5.4642e-05,\n 4.3399e-06, -5.4512e-05, 8.6155e-05, -8.5661e-05, -6.7195e-05,\n -2.8196e-05, 1.8616e-05, 5.6052e-45, 4.7356e-05, 6.2364e-06,\n -5.3005e-05, 8.2914e-05, -5.3938e-05, 6.4984e-05, 6.0860e-05,\n 5.6052e-45, 8.6213e-06, -6.4817e-07, 5.1057e-05, 3.3903e-05,\n -1.9045e-05, 4.3795e-05, 2.8627e-05, 3.4347e-05, -1.6047e-05,\n 2.0271e-05, -2.8412e-05, -3.0607e-05, 5.1312e-05, -1.1938e-05,\n 5.6052e-45, -1.9952e-05, -9.2206e-06, 4.9283e-05, 2.6113e-06,\n 5.0904e-05, -6.6084e-06, 5.6052e-45, 4.2698e-05, 5.6052e-45,\n -1.7648e-06, 9.6147e-06, -3.2764e-06, -2.1465e-05, 1.5792e-05,\n -1.2625e-05, -2.4434e-05, 7.2588e-05, 2.6208e-05, 6.1977e-05,\n 2.0359e-05, 1.1401e-06, 1.2691e-05, -1.3343e-04, 5.6052e-45,\n -8.3869e-05, -5.1079e-05, 4.7920e-05, -3.7080e-05, -1.1720e-06,\n -2.0207e-05, 5.6052e-45, 2.8148e-06, -8.7754e-05, -4.2531e-05,\n -1.6659e-04, -1.4460e-05, 1.0587e-04, 3.3210e-05, 3.0487e-05,\n -7.0828e-05, 2.1012e-06, 4.4710e-05, 6.0191e-05, -1.1361e-05,\n -7.3688e-06, -4.1773e-05, 5.8160e-06, -5.0593e-05, 1.1708e-04,\n -5.0747e-05, 4.5735e-05, -2.9939e-06, -7.5677e-06, -4.3382e-05,\n 5.6052e-45, -1.3393e-04, 7.8651e-06, 2.0024e-05, 8.4476e-05,\n -1.4322e-05, 8.3338e-05, 5.6620e-06, -3.0921e-05, -7.0068e-05,\n -1.5279e-05, 5.6052e-45, 9.5309e-05, -7.7610e-05, -1.1064e-05,\n -3.0272e-05, -2.6524e-05, 5.6052e-45, 5.6052e-45, 1.7499e-05,\n 3.9179e-05, 5.6052e-45, -5.4034e-05, 1.5054e-05, -6.5313e-05,\n 6.6295e-05, 5.6052e-45, -1.1523e-05, 3.8356e-05, 9.7480e-06,\n -6.7581e-05, 4.9965e-05, 1.3604e-07, -1.0779e-05, 3.4718e-06,\n -1.0014e-04, 4.4166e-05, -1.0541e-04, 2.5247e-04, -3.5192e-05,\n 4.2830e-05, 7.1035e-05, -1.2968e-04, -4.7387e-05, -6.1922e-05,\n -1.2536e-06, 7.5690e-06, -7.5038e-05, 6.0050e-05, 2.1750e-05,\n -7.1275e-05, 8.5723e-05, -3.1638e-05, -3.2120e-05, 2.5243e-05,\n -2.8837e-05, 5.6052e-45, 1.9648e-05, -4.7032e-06, -7.3381e-05,\n -6.1251e-05, -9.2262e-05, 3.8867e-05, -3.2997e-05, 5.6052e-45,\n 6.0342e-05, -2.0883e-05, -5.5067e-05, 6.4870e-06, 3.5546e-05,\n 2.0735e-05, 3.1603e-05, 2.9079e-05, -3.9304e-05, 9.7804e-05,\n -3.8339e-05, 1.0194e-05, -8.0390e-06, -1.6234e-05, 4.5563e-05,\n 3.6863e-05, -1.9910e-05, -1.1792e-05, -6.4681e-05, -4.2362e-05,\n -3.6591e-05, 1.1972e-05, 1.3586e-06, 1.2773e-05, -1.5848e-05,\n 6.7115e-05, 5.6052e-45, -8.2598e-05, -7.0288e-05, 5.6052e-45,\n 5.4285e-05, -6.2144e-05, 5.2431e-05, -2.9231e-06, -1.4096e-04,\n 8.6763e-06, -8.8740e-05, -4.4875e-05, 1.2115e-06, 6.1375e-05,\n 5.6052e-45, 5.9038e-05, 8.9667e-05, 7.2413e-05, -3.3509e-05,\n -3.0299e-05, -4.7211e-05, -5.7025e-05, 8.2712e-05, -6.4883e-05,\n -1.3720e-05, 4.3999e-05, 1.8569e-05, -3.5992e-06, -2.7626e-05,\n 8.8106e-05, -1.5047e-04, -2.0950e-06, -1.4603e-05, -6.1229e-07,\n 1.9062e-05, -5.8297e-05, -9.5922e-06, 4.3905e-05, -5.7983e-05,\n 2.1920e-05, 5.9032e-05, 2.5369e-07, 1.6492e-06, 1.0936e-05,\n -7.8156e-06, -6.3830e-05, -2.9258e-05, 3.7923e-05, 6.5043e-05,\n 2.3446e-05, 4.3672e-05, -5.5858e-05, -6.4516e-05, -3.0839e-05,\n -2.9728e-05, 1.0922e-05, 6.8355e-05, 1.7797e-05, 5.6052e-45,\n 5.6052e-45, 1.5786e-05, -4.3993e-05], device='cuda:0')", - "exp_avg_sq": "tensor([3.5714e-08, 4.2195e-08, 4.2070e-08, 5.0195e-08, 5.6416e-08, 8.8783e-09,\n 6.2200e-08, 3.6730e-08, 3.2619e-08, 4.6578e-08, 3.4253e-08, 2.1778e-08,\n 4.8512e-08, 2.3433e-08, 4.7870e-08, 4.7230e-08, 3.5016e-08, 5.0400e-08,\n 5.1007e-08, 2.9132e-08, 2.9921e-08, 1.4612e-08, 3.5507e-08, 4.2934e-08,\n 5.2852e-08, 7.5930e-08, 4.0909e-08, 8.0459e-19, 4.3165e-08, 5.6683e-08,\n 1.8561e-08, 5.2949e-08, 3.0597e-08, 5.1428e-08, 2.9313e-08, 4.7267e-08,\n 2.3932e-08, 3.4459e-08, 3.8489e-08, 3.6140e-08, 5.4017e-08, 4.3897e-08,\n 5.7105e-08, 5.6889e-08, 4.2984e-08, 3.2924e-08, 4.5531e-08, 5.6885e-08,\n 9.7997e-08, 7.0402e-08, 2.4964e-08, 8.5104e-08, 3.3489e-08, 3.1780e-08,\n 4.0222e-08, 3.2004e-08, 5.3415e-08, 4.5543e-08, 2.5640e-08, 3.7763e-08,\n 1.9826e-08, 3.5679e-08, 6.7216e-08, 2.1787e-08, 2.0129e-08, 5.0714e-08,\n 5.2045e-08, 7.8444e-08, 1.8341e-08, 4.0035e-08, 3.9597e-08, 2.3180e-08,\n 3.6857e-08, 1.0950e-07, 5.0519e-08, 6.0441e-08, 5.8476e-08, 7.5716e-08,\n 2.7079e-08, 5.8034e-08, 4.2024e-08, 5.0756e-08, 4.6124e-08, 1.3330e-08,\n 3.8684e-08, 4.6217e-08, 2.8999e-08, 7.3370e-08, 4.3101e-08, 7.2999e-08,\n 6.8850e-08, 3.1691e-08, 4.6899e-08, 4.2469e-08, 3.1139e-08, 4.1863e-08,\n 5.0861e-08, 2.9664e-08, 3.8329e-08, 6.1649e-08, 5.4333e-08, 3.0538e-08,\n 1.0696e-07, 2.5761e-08, 4.8254e-08, 3.5802e-08, 3.3423e-08, 2.8055e-08,\n 2.8363e-08, 3.3830e-08, 4.6844e-08, 4.1828e-08, 2.8621e-08, 2.3101e-08,\n 4.4745e-08, 4.5296e-08, 4.8170e-08, 4.2291e-08, 3.7300e-08, 1.6471e-16,\n 3.8593e-08, 4.3045e-08, 3.5618e-08, 8.7155e-09, 3.2998e-08, 4.5164e-08,\n 4.2247e-08, 4.1546e-08, 5.4208e-08, 3.2164e-08, 3.1487e-08, 3.2802e-08,\n 1.8030e-17, 5.5712e-08, 9.2164e-08, 5.0581e-08, 9.1211e-09, 4.8965e-08,\n 2.7716e-08, 3.9257e-08, 2.0059e-08, 9.9964e-17, 3.3914e-08, 3.5449e-08,\n 3.8729e-19, 5.1515e-16, 6.4656e-08, 3.6277e-08, 1.3841e-15, 2.9372e-08,\n 4.4576e-08, 5.7282e-08, 3.3646e-08, 4.0951e-08, 3.3842e-08, 4.4192e-08,\n 3.8061e-08, 4.6242e-08, 5.7020e-16, 2.6984e-08, 7.2613e-08, 8.0873e-08,\n 6.5749e-08, 8.1786e-08, 3.4085e-08, 3.6370e-08, 1.4096e-08, 7.2821e-22,\n 4.1791e-08, 3.8411e-08, 6.4891e-08, 6.1009e-08, 3.6322e-08, 4.6572e-08,\n 3.5167e-08, 3.9745e-08, 4.8572e-08, 4.9176e-08, 2.8450e-08, 7.3245e-08,\n 5.8586e-08, 4.4754e-08, 6.4884e-08, 4.7663e-08, 1.8828e-08, 4.2546e-25,\n 4.2638e-08, 4.1293e-08, 2.9716e-08, 3.1730e-08, 2.0827e-08, 4.4682e-08,\n 5.6512e-08, 6.0316e-08, 4.5271e-08, 2.2155e-08, 6.4660e-08, 4.5080e-08,\n 4.8603e-08, 4.6603e-27, 2.0719e-08, 2.4742e-08, 6.8110e-08, 3.4900e-08,\n 6.6641e-08, 4.4008e-08, 3.5291e-08, 2.7545e-08, 3.1248e-08, 2.0106e-08,\n 4.6193e-08, 1.4143e-08, 2.7515e-08, 6.6355e-08, 5.7627e-08, 4.5365e-08,\n 5.2786e-08, 2.8011e-08, 3.9638e-08, 3.7790e-08, 1.6652e-08, 5.3574e-08,\n 5.1297e-07, 2.7227e-08, 2.1871e-08, 2.3754e-08, 6.0067e-08, 6.2595e-08,\n 6.0076e-08, 2.1235e-08, 3.3629e-08, 2.9627e-08, 4.7749e-08, 1.1186e-07,\n 4.3661e-16, 6.9190e-08, 2.9150e-08, 8.2852e-20, 3.0023e-08, 3.8205e-08,\n 1.7176e-08, 2.5839e-08, 5.4845e-08, 5.2216e-08, 4.2455e-08, 8.0395e-08,\n 4.2014e-08, 3.5090e-08, 5.7986e-08, 1.2522e-08, 5.6391e-08, 7.8770e-08,\n 2.9215e-08, 3.8634e-08, 3.6845e-08, 8.1741e-08, 8.8576e-20, 4.8989e-08,\n 5.5824e-08, 7.7287e-08, 5.2158e-25, 3.5903e-08, 4.8412e-08, 2.5388e-08,\n 4.4904e-08, 3.4901e-08, 2.8421e-08, 2.7027e-08, 4.4749e-08, 7.6977e-08,\n 3.5495e-08, 5.1287e-08, 3.3184e-08, 3.4538e-08, 7.4813e-08, 3.7536e-08,\n 6.1123e-08, 6.9678e-09, 2.7708e-08, 6.6054e-08, 1.3816e-08, 6.9213e-08,\n 3.1072e-08, 9.5217e-08, 4.0004e-08, 5.5677e-08, 8.4864e-08, 4.2772e-08,\n 4.1656e-08, 4.0245e-08, 4.8103e-08, 2.3324e-08, 2.0696e-08, 3.2530e-08,\n 5.1591e-08, 2.8560e-08, 3.8631e-08, 4.7127e-08, 2.7439e-08, 2.5715e-08,\n 6.8616e-08, 3.6207e-08, 2.7302e-08, 2.9296e-08, 6.1588e-08, 2.0041e-08,\n 2.2018e-08, 2.1408e-08, 5.0680e-08, 4.9138e-08, 3.8973e-08, 7.4765e-20,\n 3.3821e-08, 2.9734e-08, 6.3005e-08, 1.4851e-19, 2.6841e-08, 8.7396e-08,\n 2.5378e-08, 3.5604e-08, 4.3660e-08, 1.0179e-08, 3.5411e-08, 4.5809e-08,\n 3.2022e-08, 5.3880e-16, 3.7588e-08, 4.1293e-08, 3.1774e-08, 4.9296e-08,\n 4.8275e-08, 2.0507e-08, 3.3075e-08, 3.6011e-08, 2.3840e-08, 6.9646e-08,\n 4.2227e-08, 4.9394e-08, 4.0467e-25, 3.1616e-08, 5.8113e-08, 4.9651e-08,\n 5.9726e-08, 3.2507e-08, 6.0135e-08, 4.7975e-08, 5.9395e-08, 2.3924e-08,\n 5.4742e-08, 4.1563e-08, 5.2543e-08, 4.0124e-08, 3.3827e-08, 1.1017e-16,\n 4.6846e-08, 8.3508e-08, 2.6703e-08, 2.9675e-08, 4.9575e-08, 4.1811e-08,\n 3.8735e-08, 5.0736e-08, 4.5690e-16, 3.4054e-08, 2.6239e-08, 2.0735e-08,\n 3.8285e-08, 5.0903e-08, 5.0928e-08, 3.1937e-08, 2.2724e-08, 2.1889e-08,\n 2.6712e-08, 4.0483e-08, 3.8153e-08, 4.3586e-08, 3.5002e-08, 3.4114e-08,\n 2.3053e-08, 2.7313e-08, 9.7037e-08, 3.0764e-08, 2.3880e-08, 4.9223e-08,\n 4.5333e-16, 7.6357e-08, 3.6102e-08, 8.6626e-08, 4.9531e-08, 2.4246e-08,\n 4.0240e-08, 7.2893e-08, 6.4687e-08, 1.4387e-08, 4.6115e-08, 4.5692e-08,\n 1.9294e-08, 1.3621e-08, 8.9760e-09, 2.7775e-08, 5.4089e-08, 6.9034e-08,\n 5.3879e-08, 7.0058e-08, 4.0486e-08, 7.2680e-08, 2.3034e-08, 4.2008e-08,\n 4.3288e-08, 5.6290e-08, 9.1703e-08, 3.7532e-15, 2.4764e-08, 4.2337e-08,\n 3.1297e-08, 4.7579e-08, 1.6787e-08, 5.0558e-08, 3.9933e-08, 4.1196e-08,\n 4.9170e-08, 5.3746e-08, 7.6945e-08, 3.2570e-08, 1.2952e-08, 4.0888e-08,\n 1.6118e-08, 1.0471e-16, 1.1411e-18, 6.0202e-19, 3.8037e-08, 5.7808e-08,\n 3.6485e-08, 6.0038e-08, 4.7491e-08, 2.8265e-08, 6.0017e-08, 8.2710e-08,\n 3.9974e-08, 1.6397e-08, 2.7432e-08, 2.9842e-08, 3.5887e-08, 2.0581e-18,\n 5.2630e-08, 7.3204e-08, 3.5699e-08, 4.0776e-08, 3.5486e-08, 3.0236e-08,\n 8.0719e-08, 2.7351e-08, 3.4042e-08, 4.8540e-08, 4.5188e-08, 5.2141e-08,\n 3.9037e-08, 6.9510e-08, 5.8993e-17, 5.0802e-08, 3.1206e-08, 1.1944e-08,\n 5.6883e-08, 3.2814e-08, 8.1684e-08, 5.8604e-08, 3.2984e-08, 3.1812e-08,\n 5.1302e-08, 3.4367e-08, 3.4538e-08, 5.8182e-18, 3.3488e-08, 5.8658e-08,\n 3.0308e-08, 3.2044e-08, 3.9365e-08, 4.5578e-08, 1.8120e-08, 6.9898e-08,\n 4.3291e-08, 4.0489e-08, 4.0697e-08, 4.3128e-08, 5.6771e-08, 4.3411e-08,\n 5.0151e-08, 5.6984e-08, 4.3909e-08, 3.4014e-08, 3.8510e-08, 3.3093e-08,\n 5.2987e-08, 1.9173e-08, 6.2614e-08, 2.6881e-08, 4.0993e-08, 4.1699e-08,\n 5.3079e-08, 1.9877e-08, 8.0988e-09, 9.8599e-08, 3.5362e-08, 4.7492e-08,\n 3.7052e-08, 3.6309e-08, 6.6926e-08, 5.6371e-08, 5.7214e-08, 5.0092e-08,\n 4.6631e-08, 6.4789e-08, 1.3505e-07, 6.0959e-08, 4.4790e-08, 6.0400e-08,\n 4.2454e-08, 2.5907e-08, 1.5997e-08, 5.1180e-08, 3.3247e-08, 3.5483e-08,\n 6.6128e-08, 3.4300e-08, 5.4209e-08, 6.0474e-08, 2.7109e-08, 7.8824e-08,\n 2.4206e-08, 5.1711e-08, 4.1105e-08, 3.7830e-08, 4.1495e-08, 5.0710e-08,\n 6.0144e-08, 4.3236e-08, 5.0128e-08, 1.4444e-15, 3.1120e-08, 5.0861e-08,\n 7.2373e-08, 5.8782e-08, 6.2105e-08, 9.2809e-08, 4.1135e-08, 5.1920e-18,\n 5.7624e-08, 5.0727e-08, 5.9877e-08, 5.2890e-08, 4.9626e-08, 6.8141e-08,\n 4.5690e-08, 4.0694e-08, 4.3888e-08, 6.3899e-08, 4.9898e-08, 4.4317e-08,\n 5.1740e-08, 3.2706e-08, 9.8106e-20, 5.2944e-08, 5.1324e-08, 3.0141e-08,\n 6.0090e-08, 5.4050e-08, 4.2432e-08, 2.0539e-18, 3.8610e-08, 1.0018e-16,\n 6.2007e-08, 3.4811e-08, 4.8742e-08, 6.0448e-08, 2.0728e-08, 2.7884e-08,\n 4.1104e-08, 2.6871e-08, 3.1728e-08, 3.4164e-08, 4.3092e-08, 2.3688e-08,\n 2.7084e-08, 1.8654e-08, 3.4583e-23, 5.4448e-08, 4.0134e-08, 3.4790e-08,\n 3.4155e-08, 2.7166e-08, 4.6880e-08, 4.0203e-20, 2.9974e-08, 3.8917e-08,\n 3.8675e-08, 2.8791e-08, 2.0465e-08, 3.6510e-08, 4.4944e-08, 2.9061e-08,\n 7.3347e-08, 3.0575e-08, 3.5769e-08, 5.2383e-08, 6.0262e-08, 4.6953e-08,\n 6.8950e-08, 2.0393e-08, 5.4101e-08, 4.1387e-08, 4.4190e-08, 4.0968e-08,\n 5.3212e-08, 2.7240e-08, 2.9881e-08, 1.8178e-19, 7.3386e-08, 1.4659e-08,\n 5.2404e-08, 6.5940e-08, 7.2260e-08, 6.0334e-08, 4.0490e-08, 4.9560e-08,\n 4.2481e-08, 5.5129e-08, 1.8036e-20, 4.5183e-08, 6.1317e-08, 2.0626e-08,\n 1.4966e-08, 1.8529e-08, 4.7350e-18, 1.4758e-19, 4.8770e-08, 4.8936e-08,\n 5.6231e-22, 3.4811e-08, 6.2188e-08, 4.9916e-08, 6.7262e-08, 5.9892e-20,\n 3.7161e-08, 4.1464e-08, 5.5098e-08, 3.4389e-08, 7.0250e-08, 4.7463e-08,\n 3.8994e-08, 3.7693e-08, 4.6490e-08, 5.6724e-08, 6.9827e-08, 5.5025e-08,\n 3.3536e-08, 3.5866e-08, 5.6932e-08, 2.3012e-08, 4.7965e-08, 3.7320e-08,\n 9.4374e-08, 2.7562e-08, 6.2337e-08, 5.3580e-08, 2.7207e-08, 7.1244e-08,\n 7.7094e-08, 2.6056e-08, 3.5339e-08, 4.3253e-08, 3.9004e-08, 5.2633e-21,\n 1.9679e-08, 5.0144e-08, 3.8522e-08, 5.7417e-08, 6.0234e-08, 6.6882e-08,\n 4.1823e-08, 6.0517e-19, 6.1952e-08, 4.3980e-08, 3.1722e-08, 3.1005e-08,\n 4.6434e-08, 2.5988e-08, 5.6344e-08, 3.7220e-08, 3.1061e-08, 6.7188e-08,\n 4.0794e-08, 3.7701e-08, 1.1269e-08, 3.1198e-08, 4.8462e-08, 3.0247e-08,\n 5.8089e-08, 3.2215e-08, 5.9115e-08, 2.6653e-08, 3.9430e-08, 5.3427e-08,\n 3.7167e-08, 4.8069e-08, 4.2709e-08, 3.6091e-08, 1.2805e-18, 1.5561e-07,\n 7.9020e-08, 1.0092e-20, 6.0930e-08, 4.2642e-08, 2.2182e-08, 4.2900e-08,\n 4.3782e-08, 1.8594e-08, 5.0350e-08, 5.5179e-08, 3.7685e-08, 4.5081e-08,\n 9.0101e-16, 7.3360e-08, 4.3803e-08, 5.8831e-08, 4.5127e-08, 2.8550e-08,\n 6.0273e-08, 6.3813e-08, 1.2234e-07, 4.9970e-08, 2.4521e-08, 3.6487e-08,\n 6.2055e-08, 6.3225e-08, 2.9158e-08, 2.9581e-08, 3.6150e-08, 5.2161e-08,\n 3.8915e-08, 4.2503e-08, 4.0311e-08, 3.4229e-08, 1.8899e-08, 2.7328e-08,\n 4.2164e-08, 5.9004e-08, 3.7582e-08, 1.8496e-08, 3.4376e-08, 3.8007e-08,\n 3.7086e-08, 4.1891e-08, 4.1854e-08, 3.3834e-08, 6.4460e-08, 6.7874e-08,\n 2.5467e-08, 2.7834e-08, 4.4089e-08, 4.4331e-08, 2.4962e-08, 4.0705e-08,\n 4.0052e-08, 1.4557e-08, 1.2818e-18, 1.9374e-25, 5.3238e-08, 4.0484e-08],\n device='cuda:0')" + "step": "tensor(12520.)", + "exp_avg": "tensor([-2.3097e-05, 9.5465e-05, -8.4887e-05, 1.8510e-04, 1.3161e-04,\n -2.0138e-05, -1.4139e-04, 3.0738e-05, -1.0147e-04, 1.1279e-04,\n 5.3551e-04, -5.0594e-05, 2.7870e-05, -2.8186e-05, -1.8885e-04,\n 1.8148e-04, -1.8156e-05, -1.8771e-05, 1.0264e-04, 4.3141e-05,\n 1.0310e-05, -9.1166e-05, 7.8901e-05, -3.3479e-06, 5.7091e-05,\n 1.1384e-04, -7.9914e-05, 1.9652e-04, -2.5717e-05, 3.2720e-06,\n -6.0814e-06, 4.1105e-06, 4.4543e-05, -1.6229e-05, -3.3145e-05,\n 1.2382e-04, 3.2658e-05, -1.1043e-04, 1.5420e-04, -3.8682e-05,\n -1.1848e-05, 5.8904e-05, 1.2257e-05, -9.3945e-06, 1.8034e-04,\n 9.1726e-05, -5.6052e-45, -5.7588e-05, -4.4043e-05, 1.0230e-04,\n -1.8961e-05, 7.8207e-05, 1.1652e-04, -2.0387e-04, 1.1119e-04,\n -1.0507e-04, -9.7519e-05, -1.4766e-05, 6.2325e-05, 9.5417e-05,\n 1.9512e-04, 1.9354e-04, -7.6378e-05, 1.7317e-05, -7.5884e-05,\n -3.8450e-05, 4.8033e-05, 9.6934e-05, 1.7354e-04, 5.6695e-05,\n -1.0674e-04, -9.1333e-05, 1.1460e-04, -1.6287e-04, 1.4066e-04,\n -3.2969e-04, -1.7241e-04, -8.8311e-05, 7.6578e-05, 2.0743e-05,\n -3.3686e-04, -9.5934e-05, 9.1859e-05, 9.0062e-05, -4.4188e-05,\n 5.8586e-04, -2.1020e-06, -1.5128e-04, 6.2619e-05, 1.3123e-04,\n -1.0706e-04, 3.3948e-04, -9.1203e-05, 2.3025e-05, 6.6483e-05,\n -3.3932e-05, 2.9560e-04, 3.8678e-04, 8.5746e-05, -1.4657e-04,\n 3.5971e-04, 7.8842e-05, -3.0828e-05, 1.7587e-04, -2.1526e-05,\n 5.8349e-05, -8.9016e-05, -1.5025e-04, 6.3920e-06, -3.5173e-05,\n -1.8645e-05, 6.7288e-05, -8.4056e-05, 1.0504e-04, -4.8835e-04,\n 9.3352e-05, 2.5125e-04, -2.2435e-04, 2.8625e-04, 2.9043e-04,\n 1.4384e-04, -1.1013e-04, -7.8196e-04, -3.0329e-05, -1.2110e-04,\n -5.3806e-05, -5.4022e-05, 2.8966e-04, 2.5352e-04, -2.1548e-05,\n -1.6928e-04, 3.4110e-05, -8.5611e-05, -2.3438e-04, 1.0973e-04,\n 2.1551e-05, 3.3693e-04, -1.3017e-04, -1.4229e-04, -1.7197e-04,\n 1.2198e-04, 3.2057e-04, 3.5783e-05, -9.1998e-04, -9.7898e-05,\n -4.7958e-05, -2.3403e-04, 1.7019e-04, -1.2190e-04, -1.2435e-04,\n -2.1264e-04, -7.0273e-05, -4.7520e-05, -3.4855e-05, 1.7900e-04,\n 3.7459e-05, -1.8914e-05, 1.6363e-04, 2.8416e-04, -9.4851e-06,\n -6.3028e-05, -6.1105e-05, -7.1743e-05, 1.0626e-04, 4.5321e-05,\n 3.9030e-06, -1.0642e-07, 8.2713e-05, 1.6905e-04, 1.7992e-04,\n 1.3156e-04, 1.2906e-04, 8.0033e-05, 2.7912e-04, 1.1921e-05,\n 3.6470e-04, 2.8856e-05, -1.0755e-04, -2.6323e-04, 1.1200e-04,\n 1.0333e-05, -5.7287e-05, -1.6896e-04, 1.3798e-04, -3.5929e-04,\n 8.5352e-06, 9.7074e-05, -7.6636e-05, 7.1150e-05, -1.5857e-04,\n -4.2728e-05, 1.3065e-04, 2.5683e-04, 1.0638e-04, -1.5501e-04,\n 2.9864e-05, -7.6376e-04, -1.4754e-04, 1.8833e-04, -1.5678e-04,\n -2.7757e-04, -2.0727e-05, -8.6255e-05, -5.9114e-06, 1.3423e-04,\n -1.0213e-04, 3.3842e-05, 4.7509e-05, -6.0057e-05, 5.8698e-05,\n 1.2088e-04, 1.6870e-05, 5.6052e-45, 2.6568e-04, -2.7662e-05,\n 8.2736e-05, 1.2345e-05, 8.8218e-05, -8.5139e-06, -3.0057e-04,\n 1.6387e-04, -6.1195e-05, -3.4784e-04, 2.3584e-05, 6.3180e-05,\n -3.1708e-04, 3.9810e-04, -2.6866e-06, -4.5599e-05, 8.1166e-07,\n -2.4455e-04, -6.2089e-05, -9.8747e-05, 6.3951e-05, -5.1049e-04,\n -1.6604e-04, -2.6183e-04, 1.9005e-04, -5.2653e-05, 1.1844e-04,\n -1.5078e-04, 1.0557e-04, -5.6954e-05, 1.8391e-05, 1.1196e-05,\n -1.3285e-04, -1.6921e-04, -6.2150e-05, -6.3824e-05, 2.3192e-04,\n -1.4176e-05, -5.5748e-05, 3.4016e-05, -4.3338e-05, 2.5997e-04,\n -1.2537e-04, -5.1138e-05, -4.3932e-05, -1.2502e-04, 3.3399e-05,\n -4.4549e-04, 9.2137e-05, -7.2620e-05, 1.0123e-04, 2.1431e-05,\n 2.0540e-05, 1.9755e-04, -6.9849e-05, 7.0310e-05, -9.9446e-05,\n 1.3434e-05, -1.0451e-04, -7.7842e-05, 1.7967e-04, -6.7510e-05,\n 1.2798e-04, 2.3808e-04, 1.1131e-04, 8.8340e-05, -5.5274e-06,\n 3.2821e-05, -1.6958e-04, 3.7256e-05, 3.1093e-05, 1.2233e-04,\n -2.5047e-04, 3.2305e-05, 5.0378e-05, 1.0068e-04, -1.5496e-04,\n -3.9147e-05, -3.3405e-04, 4.4631e-04, -1.5522e-04, 1.4786e-05,\n -1.7125e-05, 1.8812e-04, 2.0777e-05, 8.9404e-05, -6.4116e-05,\n 6.1663e-06, 1.7845e-04, -2.6063e-04, -7.2249e-05, -9.1835e-05,\n -2.9728e-05, 1.1194e-06, -5.4563e-04, -1.3660e-04, 1.0344e-04,\n -2.5862e-04, 1.8876e-04, 2.7539e-04, -2.9561e-04, 1.7747e-04,\n 4.3308e-05, -5.8373e-05, 2.6540e-05, -1.6754e-04, -5.1464e-05,\n 5.2062e-05, -6.8142e-05, 1.2304e-05, -1.4105e-04, -1.4604e-05,\n 2.8377e-06, -3.8877e-05, -5.3468e-05, 1.5921e-04, -1.0798e-04,\n -1.3162e-04, -1.5323e-04, 8.4145e-05, 1.0660e-04, -5.6187e-04,\n -3.3303e-05, -6.9691e-05, -1.6655e-05, -4.8823e-05, -2.5034e-05,\n 5.9345e-05, -1.3957e-05, 6.7960e-05, 1.4905e-04, 2.0097e-04,\n 3.5805e-04, 1.8035e-04, -5.3022e-06, -3.9822e-05, 5.5282e-04,\n 1.3070e-04, 1.3386e-04, 3.2436e-04, 1.5350e-04, 2.6154e-05,\n 1.4979e-05, -3.8418e-05, -1.9032e-04, 7.4158e-05, 6.3868e-05,\n -6.4798e-05, 9.5889e-06, -7.5351e-05, 5.8865e-06, -1.4279e-05,\n 1.6290e-04, 1.6383e-04, 4.6368e-06, -1.1138e-04, 2.8730e-04,\n 9.0168e-05, 3.9317e-04, -7.7682e-05, -2.2131e-04, -5.3471e-05,\n 7.2889e-05, -4.3777e-05, 1.9659e-04, -2.1230e-04, -3.5902e-04,\n 9.0394e-06, 6.1946e-05, -2.6224e-04, -2.8862e-04, 4.7006e-04,\n 1.3675e-05, 9.7516e-05, -1.8979e-05, -5.3750e-05, 1.3230e-04,\n 2.3560e-06, 8.6359e-05, 2.6651e-05, -1.6681e-05, -1.2589e-05,\n -1.4659e-04, -5.5320e-05, 1.7960e-05, -6.4628e-05, -1.9698e-04,\n 2.0112e-04, 1.1762e-04, -4.8798e-05, 2.2323e-04, 2.1938e-04,\n -9.1546e-05, -1.2086e-04, -1.0658e-04, -1.1210e-04, 1.3857e-04,\n -2.4781e-04, 3.3030e-04, 3.5094e-05, 2.1677e-04, 8.2244e-05,\n -1.7420e-04, 2.6636e-04, 7.7121e-05, 3.5095e-04, -1.4248e-05,\n 2.7391e-06, 4.9044e-06, 6.8081e-05, 4.7585e-05, 3.2127e-04,\n -6.0759e-05, -7.9097e-05, 1.6906e-04, -2.3544e-05, -2.9171e-05,\n -1.6745e-04, 1.5603e-04, 4.6529e-05, -1.1040e-04, 3.6129e-05,\n -1.3145e-04, 5.1521e-05, -1.4818e-04, -4.8436e-05, 1.4861e-04,\n -2.1237e-05, 2.7486e-04, 1.0794e-04, -1.8279e-05, 4.2193e-05,\n -1.3471e-04, 7.8152e-05, 1.3767e-04, 6.3685e-05, -1.7406e-04,\n -1.4160e-04, 5.6353e-05, -1.4992e-05, -1.9234e-05, 2.8758e-05,\n -2.9798e-04, 2.7565e-05, 6.8054e-05, -1.4859e-05, -1.6320e-04,\n -2.6463e-05, 1.3570e-04, -1.1376e-04, 1.6880e-04, 5.2559e-04,\n -7.0014e-05, -1.8083e-04, -6.5876e-05, -4.1378e-05, 5.8489e-05,\n -1.6953e-04, 3.5184e-05, -1.0442e-04, -2.4437e-05, 2.0582e-04,\n 1.6997e-04, -5.2732e-05, -1.0792e-06, -5.5454e-05, -8.3733e-05,\n -1.9213e-05, -8.9905e-05, 7.2450e-05, 1.7878e-04, 1.3550e-04,\n -1.9496e-05, 1.1908e-04, -3.5069e-04, 2.0520e-04, -3.6586e-04,\n -1.0590e-04, 6.5640e-05, -7.6636e-05, -8.0065e-05, 1.2526e-04,\n -5.9940e-05, 2.7638e-04, -1.4376e-04, 1.3936e-05, -5.2534e-05,\n 4.7914e-04, 1.6016e-05, 1.3879e-04, 2.0653e-04, 5.4986e-05,\n -1.4737e-06, 7.5473e-05, 1.4020e-04, -9.5014e-05, 3.4207e-05,\n 9.7805e-05, 2.9363e-04], device='cuda:0')", + "exp_avg_sq": "tensor([1.8256e-07, 2.1425e-07, 1.8096e-07, 3.1086e-07, 1.5436e-07, 2.5972e-07,\n 1.4867e-07, 2.7652e-07, 1.7006e-07, 1.7839e-07, 3.9224e-07, 2.4319e-07,\n 1.0806e-07, 2.2200e-07, 1.0709e-07, 2.9199e-07, 2.3372e-07, 3.3276e-07,\n 2.1700e-07, 1.4956e-07, 1.4182e-07, 2.2349e-07, 3.9689e-07, 2.1595e-07,\n 3.0088e-07, 3.0177e-07, 3.0431e-07, 4.0482e-07, 1.2525e-07, 1.2114e-07,\n 3.0681e-07, 3.4064e-07, 3.0594e-07, 4.1693e-07, 1.4970e-07, 2.5318e-07,\n 2.2596e-07, 4.0247e-07, 3.4422e-07, 2.9641e-07, 4.3696e-07, 4.2409e-07,\n 3.1674e-07, 1.5198e-07, 3.0235e-07, 3.3255e-07, 8.9036e-10, 1.9861e-07,\n 3.8588e-07, 2.1598e-07, 2.0388e-07, 1.7505e-07, 3.5890e-07, 2.2420e-07,\n 2.3949e-07, 4.0970e-07, 3.1663e-07, 2.9832e-07, 3.6877e-07, 3.3916e-07,\n 2.4072e-07, 3.1500e-07, 2.7670e-07, 2.9994e-07, 2.0455e-07, 1.4815e-07,\n 2.6749e-07, 2.0407e-07, 2.9794e-07, 1.6954e-07, 2.6729e-07, 1.1179e-07,\n 2.4321e-07, 2.9745e-07, 2.0795e-07, 2.8625e-07, 3.4843e-07, 3.0315e-07,\n 3.7308e-07, 2.8408e-07, 2.0312e-07, 2.0092e-07, 2.3537e-07, 2.5766e-07,\n 2.0972e-07, 1.7113e-07, 2.6095e-07, 4.3421e-07, 2.0788e-07, 2.9926e-07,\n 2.1054e-07, 3.3935e-07, 1.9373e-07, 3.1144e-07, 1.0913e-07, 3.9414e-07,\n 3.6177e-07, 4.0089e-07, 4.9945e-07, 1.9958e-07, 3.4982e-07, 2.5925e-07,\n 3.0261e-07, 2.3794e-07, 1.6194e-07, 1.4839e-07, 3.0522e-07, 1.7743e-07,\n 2.0349e-07, 1.8151e-07, 7.4737e-08, 3.7157e-07, 1.5902e-07, 3.3497e-07,\n 2.8562e-07, 3.2404e-07, 2.5718e-07, 3.6781e-07, 2.0814e-07, 1.8925e-07,\n 3.3441e-07, 2.5547e-07, 5.9334e-07, 3.0809e-07, 3.7081e-07, 1.1751e-07,\n 4.0517e-07, 1.8862e-07, 2.1909e-07, 1.4162e-07, 1.1979e-07, 1.7339e-07,\n 2.1070e-07, 1.3662e-07, 1.5900e-07, 2.3966e-07, 3.8422e-07, 2.7259e-07,\n 2.4649e-07, 4.2156e-07, 3.7968e-07, 3.0587e-07, 1.6132e-07, 6.2563e-07,\n 2.5627e-07, 2.9693e-07, 3.4808e-07, 2.3952e-07, 2.3560e-07, 4.3328e-07,\n 3.5887e-07, 3.1825e-07, 1.2514e-07, 3.5611e-07, 1.9743e-07, 9.0684e-08,\n 3.1427e-07, 2.8987e-07, 1.5878e-07, 1.8582e-07, 2.4428e-07, 1.9156e-07,\n 2.1997e-07, 2.2396e-07, 2.8756e-07, 2.1900e-07, 4.5996e-07, 3.1181e-07,\n 3.7335e-07, 1.8851e-07, 3.9430e-07, 1.5814e-07, 7.1152e-08, 3.3474e-07,\n 2.2523e-07, 2.2151e-07, 3.3801e-07, 2.8776e-07, 2.3862e-07, 1.6238e-07,\n 1.5302e-07, 2.3336e-07, 1.9019e-07, 1.3591e-07, 3.5657e-07, 3.3209e-07,\n 3.1976e-07, 2.7448e-07, 2.9121e-07, 2.9496e-07, 3.1630e-07, 4.0501e-07,\n 3.6675e-07, 2.9018e-07, 4.0665e-07, 3.0313e-07, 4.4880e-07, 1.6705e-07,\n 1.4267e-07, 3.9530e-07, 4.0331e-07, 2.1055e-07, 3.1743e-07, 4.3339e-07,\n 1.8589e-07, 2.1722e-07, 1.3684e-07, 4.7503e-07, 1.2302e-07, 1.0313e-07,\n 2.4157e-07, 2.2642e-07, 6.4589e-10, 2.7511e-07, 1.5819e-07, 2.6258e-07,\n 2.6209e-07, 1.4568e-07, 2.8857e-07, 2.8944e-07, 2.2921e-07, 1.9293e-07,\n 1.8708e-07, 2.5509e-07, 3.2730e-07, 3.7669e-07, 4.1516e-07, 2.1599e-07,\n 3.2565e-07, 2.1794e-07, 2.0155e-07, 2.9262e-07, 3.0381e-07, 2.8204e-07,\n 2.9562e-07, 1.8101e-07, 3.2179e-07, 2.3236e-07, 2.7288e-07, 3.1693e-07,\n 1.4705e-07, 3.3113e-07, 2.3241e-07, 2.2905e-07, 1.1740e-07, 2.0800e-07,\n 3.1191e-07, 1.8555e-07, 1.7903e-07, 2.9601e-07, 2.2307e-07, 3.4995e-07,\n 3.1301e-07, 3.5587e-07, 3.8555e-07, 2.8922e-07, 2.2781e-07, 2.0783e-07,\n 3.7546e-07, 2.3805e-07, 3.5527e-07, 3.5037e-07, 2.8838e-07, 2.6903e-07,\n 9.8683e-08, 2.0528e-07, 2.8616e-07, 2.7834e-07, 4.4921e-07, 3.2371e-07,\n 2.9430e-07, 3.0961e-07, 1.9665e-07, 4.7703e-07, 1.3199e-07, 1.4511e-07,\n 1.6614e-07, 3.3142e-07, 2.3435e-07, 1.3907e-07, 3.2215e-07, 2.1930e-07,\n 2.4806e-07, 2.3886e-07, 2.7431e-07, 3.8069e-07, 1.5978e-07, 2.5274e-07,\n 2.3241e-07, 3.9467e-07, 3.3011e-07, 2.9075e-07, 3.8669e-07, 1.6693e-07,\n 1.0705e-07, 3.2581e-07, 2.2301e-07, 2.4748e-07, 3.3591e-07, 2.5318e-07,\n 1.7251e-07, 2.2753e-07, 3.6364e-07, 3.8079e-07, 2.1385e-07, 1.7824e-07,\n 2.1364e-07, 4.4789e-07, 1.8411e-07, 3.2213e-07, 2.4491e-07, 1.7791e-07,\n 1.3328e-07, 2.6314e-07, 3.6003e-07, 3.1611e-07, 1.6867e-07, 3.3040e-07,\n 1.0868e-07, 2.9242e-07, 2.3903e-07, 1.7826e-07, 4.1388e-07, 1.8012e-07,\n 1.6430e-07, 3.0997e-07, 2.7268e-07, 1.8449e-07, 4.0652e-07, 2.5353e-07,\n 3.5883e-07, 3.4857e-07, 2.4792e-07, 1.9784e-07, 3.7018e-07, 2.7227e-07,\n 3.7122e-07, 1.7250e-07, 9.2883e-08, 2.7772e-07, 2.9366e-07, 2.6413e-07,\n 2.6411e-07, 9.2334e-08, 1.7285e-07, 1.8836e-07, 1.3916e-07, 1.0605e-07,\n 1.9253e-07, 2.9649e-07, 2.1877e-07, 2.1372e-07, 2.4435e-07, 2.7261e-07,\n 3.4448e-07, 1.5889e-07, 2.6325e-07, 3.0022e-07, 2.6954e-07, 2.2011e-07,\n 2.7123e-07, 1.1561e-07, 2.6991e-07, 3.4243e-07, 3.4209e-07, 2.0396e-07,\n 2.4188e-08, 3.6758e-07, 1.5635e-07, 3.6905e-07, 2.6259e-07, 3.8687e-07,\n 4.7996e-07, 1.8078e-07, 2.1399e-07, 2.5266e-07, 1.4999e-07, 2.1441e-07,\n 1.4788e-07, 2.1741e-07, 9.1990e-08, 2.6614e-07, 3.7700e-07, 3.6182e-07,\n 3.6236e-07, 2.6526e-07, 3.3645e-07, 1.1933e-07, 2.2411e-07, 3.3755e-07,\n 2.1433e-07, 2.4179e-07, 3.1439e-07, 3.4441e-07, 1.0257e-07, 1.6522e-07,\n 1.9009e-07, 2.2387e-07, 2.0718e-07, 2.9718e-07, 4.3104e-07, 1.5074e-07,\n 3.9462e-07, 1.9716e-07, 4.2145e-07, 2.9546e-07, 2.5878e-07, 4.3282e-07,\n 2.6977e-07, 3.2924e-07, 3.1842e-07, 2.2467e-07, 2.6467e-07, 3.6254e-07,\n 3.0816e-07, 1.7705e-07, 3.1104e-07, 2.2346e-07, 2.1714e-07, 1.9056e-07,\n 1.3208e-07, 2.5582e-07, 2.1585e-07, 1.4291e-07, 4.6411e-07, 2.5803e-07,\n 1.5700e-07, 2.7100e-07, 1.7117e-07, 3.7821e-07, 3.3190e-07, 2.8111e-07,\n 1.8160e-07, 2.7962e-07, 1.7832e-07, 1.5373e-07, 2.8135e-07, 1.4392e-07,\n 1.2374e-07, 1.5946e-07, 3.9163e-07, 2.9864e-07, 1.5609e-07, 1.6107e-07,\n 3.6032e-07, 2.3995e-07, 1.0812e-07, 3.3769e-07, 2.4657e-07, 3.6158e-07,\n 3.8523e-07, 1.2528e-07, 2.2048e-07, 3.0105e-07, 4.2159e-07, 3.1306e-07,\n 1.5309e-07, 2.6877e-07, 1.0997e-07, 2.1401e-07, 2.8960e-07, 2.2950e-07,\n 2.6709e-07, 3.3779e-07, 4.1425e-07, 9.7348e-08, 1.9150e-07, 3.3333e-07,\n 1.7449e-07, 2.9185e-07, 4.2516e-07, 1.4202e-07, 2.8163e-07, 2.7556e-07,\n 2.6623e-07, 1.8766e-07, 2.1143e-07, 3.4654e-07, 1.7227e-07, 2.8115e-07,\n 4.5815e-07, 2.0989e-07, 1.9896e-07, 2.8303e-07, 3.1401e-07, 3.3674e-07,\n 9.0994e-08, 3.1324e-07, 3.4415e-07, 3.7619e-07, 3.6785e-07, 3.0748e-07,\n 2.2138e-07, 3.8593e-07, 2.5920e-07, 1.2004e-07, 2.4800e-07, 4.2623e-07,\n 1.6435e-07, 2.9217e-07, 4.2009e-07, 2.9591e-07, 2.2528e-07, 3.0166e-07,\n 2.8132e-07, 2.9005e-07, 1.2789e-07, 2.5149e-07, 1.3761e-07, 8.1217e-08,\n 1.5388e-07, 2.6906e-07], device='cuda:0')" }, "4": { - "step": "tensor(25040.)", - "exp_avg": "tensor([[ 3.7872e-06, -1.1744e-06, 1.0282e-06, ..., 5.6052e-45,\n 1.1146e-06, 9.8124e-07],\n [ 1.0413e-06, -1.8975e-05, 7.9172e-07, ..., -5.6052e-45,\n -1.7463e-05, 6.1551e-07],\n [ 2.4576e-06, -8.9567e-06, 1.3082e-06, ..., -5.6052e-45,\n 1.5031e-05, -3.2102e-06],\n ...,\n [-5.8124e-06, -2.5645e-05, -4.2916e-06, ..., -5.6052e-45,\n -5.5089e-06, -3.4290e-06],\n [-4.6933e-06, 1.1324e-05, 6.9274e-07, ..., -5.6052e-45,\n 1.7607e-06, -7.3242e-06],\n [ 5.6216e-06, -8.5150e-06, -1.8661e-07, ..., -5.6052e-45,\n -2.5080e-06, -2.7810e-06]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.8074e-10, 9.9683e-11, 8.8343e-11, ..., 5.4304e-28, 2.2872e-10,\n 8.6738e-11],\n [3.0477e-10, 1.3356e-10, 1.8386e-10, ..., 2.7753e-26, 5.3439e-10,\n 3.9223e-10],\n [5.9773e-10, 1.5855e-10, 7.4319e-10, ..., 1.2104e-26, 3.4939e-10,\n 4.2039e-10],\n ...,\n [5.3243e-10, 2.0862e-10, 2.1727e-10, ..., 1.0590e-26, 2.2129e-10,\n 2.2508e-10],\n [6.1276e-10, 1.8953e-10, 3.9787e-10, ..., 1.5782e-28, 4.0911e-10,\n 4.6863e-10],\n [5.3071e-10, 1.4795e-10, 5.9625e-10, ..., 1.0114e-26, 3.1403e-10,\n 3.5189e-10]], device='cuda:0')" + "step": "tensor(12520.)", + "exp_avg": "tensor([[-6.2645e-06, 2.7620e-05, -6.3752e-06, ..., -2.2096e-07,\n 1.1970e-05, 8.1451e-06],\n [-5.5997e-06, -2.0422e-05, 5.4662e-06, ..., -2.0322e-06,\n 3.8756e-06, -6.0927e-06],\n [-1.9815e-06, -3.1985e-05, 1.1588e-05, ..., 5.0815e-07,\n -2.6333e-05, 1.2598e-06],\n ...,\n [-3.2060e-05, -3.0144e-06, 4.5311e-06, ..., -1.0798e-06,\n -5.0732e-05, -9.6518e-06],\n [ 7.6311e-06, -1.6229e-06, -7.7973e-06, ..., -1.1816e-06,\n -6.2412e-05, 9.8432e-06],\n [-1.5716e-05, -9.9179e-05, -1.7823e-05, ..., -3.2049e-06,\n 6.9133e-06, -3.0023e-06]], device='cuda:0')", + "exp_avg_sq": "tensor([[2.4652e-09, 7.9788e-10, 8.8448e-10, ..., 5.6255e-11, 3.4621e-10,\n 2.6910e-09],\n [4.6996e-09, 2.7530e-09, 3.1170e-09, ..., 6.2600e-11, 5.3921e-10,\n 4.2599e-09],\n [3.0225e-09, 2.1399e-09, 1.4565e-09, ..., 1.4477e-10, 1.0545e-09,\n 2.2051e-09],\n ...,\n [1.8733e-08, 1.7064e-09, 4.4232e-09, ..., 8.1151e-11, 2.0680e-09,\n 4.5656e-09],\n [4.0706e-09, 1.4898e-09, 3.2587e-09, ..., 7.6016e-11, 3.5762e-09,\n 5.8001e-09],\n [5.5760e-09, 3.0821e-09, 5.4045e-09, ..., 1.3457e-10, 5.7061e-10,\n 6.1333e-09]], device='cuda:0')" }, "5": { - "step": "tensor(23788.)", - "exp_avg": "tensor([[-1.2074e-08, 1.0135e-07, 5.8361e-08, ..., 9.4664e-07,\n 1.2818e-08, 5.1650e-10],\n [-1.5338e-07, 1.3816e-07, 3.9266e-15, ..., 1.9893e-07,\n -2.4520e-07, 1.5028e-09],\n [ 1.6835e-09, 5.5058e-08, 1.4529e-09, ..., 1.9924e-08,\n 4.6696e-07, -8.7692e-08],\n ...,\n [-1.3820e-06, 3.6539e-07, -1.1146e-08, ..., 1.9993e-08,\n 3.2483e-07, 3.0002e-06],\n [ 1.1244e-09, 4.3683e-08, 7.2904e-22, ..., -2.5118e-07,\n -8.8448e-07, 5.4964e-09],\n [ 2.5477e-08, 1.6106e-07, 1.7982e-08, ..., 7.5466e-07,\n 1.3451e-08, -1.9609e-07]], device='cuda:0')", - "exp_avg_sq": "tensor([[7.0025e-14, 4.8454e-12, 6.1642e-11, ..., 4.0904e-11, 1.1039e-11,\n 7.0180e-13],\n [2.5617e-11, 2.7253e-11, 1.2947e-15, ..., 3.6139e-11, 1.8147e-11,\n 2.1250e-12],\n [1.8658e-12, 4.8658e-12, 3.4748e-13, ..., 1.2107e-12, 2.6536e-11,\n 5.6692e-13],\n ...,\n [1.1914e-10, 1.2956e-11, 1.4472e-12, ..., 1.7857e-12, 6.1849e-12,\n 8.2024e-09],\n [3.1009e-13, 1.9226e-11, 1.7168e-14, ..., 5.8177e-11, 2.5147e-11,\n 6.4428e-12],\n [1.7684e-11, 5.6626e-12, 2.6619e-12, ..., 1.4303e-11, 4.1947e-12,\n 1.0183e-10]], device='cuda:0')" + "step": "tensor(11268.)", + "exp_avg": "tensor([[-1.8525e-06, 2.0083e-07, 3.1452e-07, ..., 7.3335e-07,\n 0.0000e+00, 5.2947e-07],\n [ 1.7694e-06, -9.4981e-08, 5.4940e-07, ..., -9.8179e-07,\n 0.0000e+00, 4.1946e-08],\n [ 1.5789e-07, -1.6118e-07, 1.0848e-06, ..., 3.7259e-06,\n 0.0000e+00, 2.9067e-06],\n ...,\n [-4.1457e-06, -1.7088e-07, -2.2955e-07, ..., 3.0773e-07,\n 0.0000e+00, 9.7976e-07],\n [-3.9881e-08, -1.9297e-07, -3.0809e-07, ..., -6.9249e-07,\n 0.0000e+00, 6.4987e-08],\n [ 3.2264e-07, 1.2982e-07, 1.1459e-07, ..., -1.1472e-06,\n 0.0000e+00, -1.5853e-06]], device='cuda:0')", + "exp_avg_sq": "tensor([[2.7611e-11, 3.6099e-12, 1.5492e-11, ..., 2.4540e-12, 0.0000e+00,\n 6.1661e-12],\n [1.3127e-11, 5.1008e-12, 1.4483e-11, ..., 2.0982e-11, 0.0000e+00,\n 8.0593e-12],\n [1.1453e-11, 8.4014e-12, 8.0252e-12, ..., 5.5600e-11, 0.0000e+00,\n 6.6102e-11],\n ...,\n [3.4881e-11, 7.9718e-12, 7.6355e-12, ..., 1.5760e-12, 0.0000e+00,\n 3.5873e-12],\n [9.0826e-13, 1.3714e-12, 4.4607e-11, ..., 6.1792e-11, 0.0000e+00,\n 1.2949e-11],\n [2.1942e-12, 9.8990e-12, 7.6443e-12, ..., 2.1875e-10, 0.0000e+00,\n 2.6375e-11]], device='cuda:0')" }, "6": { - "step": "tensor(23788.)", - "exp_avg": "tensor([-1.0874e-06, -2.3645e-06, -1.8390e-06, ..., 7.3688e-06,\n -6.3505e-06, -4.5159e-06], device='cuda:0')", - "exp_avg_sq": "tensor([6.5076e-10, 1.2028e-09, 7.1218e-10, ..., 5.4302e-09, 1.0333e-09,\n 9.8003e-10], device='cuda:0')" + "step": "tensor(11268.)", + "exp_avg": "tensor([ 8.4450e-06, -1.1573e-05, 2.1310e-05, ..., -5.5773e-06,\n 2.5801e-06, 1.5663e-05], device='cuda:0')", + "exp_avg_sq": "tensor([2.8729e-09, 2.0459e-09, 2.6377e-09, ..., 1.2660e-09, 1.7396e-09,\n 1.8250e-09], device='cuda:0')" }, "7": { - "step": "tensor(23788.)", - "exp_avg": "tensor([[-1.2865e-08, -1.2050e-07, 1.5276e-06, ..., 5.7586e-08,\n 8.3759e-08, -3.7446e-08],\n [ 6.4525e-08, 2.5138e-07, 3.9921e-07, ..., -2.3164e-07,\n 6.7710e-08, -2.2803e-08],\n [ 8.6614e-08, 3.6603e-07, -5.5850e-07, ..., -2.2524e-07,\n 1.6711e-07, 3.4905e-07],\n ...,\n [-3.0591e-08, -6.6691e-08, 1.7021e-06, ..., -1.2613e-07,\n 9.0767e-08, -2.7655e-07],\n [-4.9875e-08, 2.3182e-07, -2.4823e-06, ..., -3.5428e-08,\n 7.8321e-08, -2.4707e-07],\n [ 8.8252e-08, 1.6104e-07, -3.1471e-06, ..., -1.8318e-07,\n 9.5786e-08, 2.8499e-08]], device='cuda:0')", - "exp_avg_sq": "tensor([[2.7003e-13, 7.2645e-13, 4.1566e-12, ..., 4.8834e-13, 4.0241e-13,\n 5.4833e-13],\n [5.0140e-13, 2.6737e-12, 2.8453e-12, ..., 6.6368e-13, 9.8494e-13,\n 5.0204e-13],\n [5.8474e-13, 1.6803e-12, 1.3298e-12, ..., 7.1697e-13, 4.6027e-13,\n 9.0905e-13],\n ...,\n [2.5220e-13, 1.3670e-12, 2.0161e-12, ..., 1.3977e-12, 6.0946e-13,\n 1.1967e-12],\n [2.7506e-13, 2.2395e-12, 8.1588e-12, ..., 1.3201e-12, 4.2125e-13,\n 1.2410e-12],\n [4.3543e-13, 1.2576e-12, 1.7050e-11, ..., 7.4380e-13, 8.6372e-13,\n 7.1587e-13]], device='cuda:0')" + "step": "tensor(11268.)", + "exp_avg": "tensor([[ 4.5972e-07, 3.1196e-08, 5.7611e-07, ..., 8.5632e-08,\n 2.1753e-08, -4.2801e-07],\n [ 1.6404e-06, 2.7880e-07, -6.3306e-07, ..., 6.9177e-07,\n 3.6146e-07, -1.0307e-06],\n [ 4.5100e-07, 5.2791e-07, -1.6244e-07, ..., -1.9920e-07,\n 9.6318e-07, -1.0921e-06],\n ...,\n [-7.4474e-08, -2.5112e-06, 1.7036e-07, ..., -5.1389e-08,\n -3.4643e-07, 4.7856e-07],\n [ 4.3694e-07, -1.3849e-06, -3.8582e-07, ..., -1.0422e-07,\n 5.3094e-08, -2.7910e-07],\n [-5.7219e-08, 5.0992e-07, 3.4745e-08, ..., 7.7025e-07,\n -2.7766e-08, -1.0441e-06]], device='cuda:0')", + "exp_avg_sq": "tensor([[5.2503e-12, 2.0841e-12, 2.7646e-12, ..., 1.8498e-12, 2.7767e-12,\n 2.9103e-12],\n [7.1288e-12, 3.6513e-12, 3.8564e-12, ..., 4.4297e-12, 5.1838e-12,\n 6.0143e-12],\n [6.6456e-12, 4.2578e-12, 3.7655e-12, ..., 3.2873e-12, 4.3643e-12,\n 7.3331e-12],\n ...,\n [5.9260e-12, 7.8217e-12, 6.7143e-12, ..., 4.9780e-12, 7.2787e-12,\n 5.6242e-12],\n [5.6961e-12, 3.5072e-12, 4.5105e-12, ..., 2.6611e-12, 4.2406e-12,\n 6.4120e-12],\n [8.5862e-12, 3.6779e-12, 5.7739e-12, ..., 3.1581e-11, 5.1929e-12,\n 3.1326e-12]], device='cuda:0')" }, - "14": { - "step": "tensor(23788.)", + "32": { + "step": "tensor(11268.)", "exp_avg": "tensor([5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([1.2166e-15], device='cuda:0')" + "exp_avg_sq": "tensor([4.5374e-10], device='cuda:0')" }, - "15": { - "step": "tensor(23788.)", - "exp_avg": "tensor([ 5.6052e-45, -5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([1.5695e-18, 8.8848e-17, 6.6800e-17], device='cuda:0')" + "33": { + "step": "tensor(11268.)", + "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, -5.6052e-45], device='cuda:0')", + "exp_avg_sq": "tensor([4.1304e-12, 2.0078e-11, 6.1141e-12], device='cuda:0')" }, - "16": { - "step": "tensor(23788.)", - "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([9.4514e-14, 9.7385e-15, 1.0537e-14, 1.1256e-14], device='cuda:0')" + "34": { + "step": "tensor(11268.)", + "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 device='cuda:0')", + "exp_avg_sq": "tensor([2.4649e-08, 2.2347e-10, 3.0183e-10, 3.1322e-10, 3.5911e-10, 4.1252e-10,\n 3.2797e-10, 2.4288e-10, 3.0678e-10, 2.7262e-10], device='cuda:0')" }, - "18": { - "step": "tensor(23788.)", - "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([[1.0077e-19, 6.2904e-20, 0.0000e+00, ..., 4.1507e-19, 1.4928e-19,\n 4.7915e-21],\n [2.1499e-20, 5.5265e-20, 0.0000e+00, ..., 5.0809e-20, 1.7866e-19,\n 5.0351e-20],\n [6.4612e-21, 2.2802e-20, 0.0000e+00, ..., 7.6201e-21, 4.9402e-20,\n 1.6180e-20],\n ...,\n [8.1870e-22, 4.4033e-21, 0.0000e+00, ..., 6.2795e-21, 8.6960e-20,\n 4.8990e-22],\n [1.5770e-19, 1.6579e-19, 0.0000e+00, ..., 2.0066e-19, 8.4105e-19,\n 5.0474e-20],\n [4.9258e-22, 3.8282e-22, 0.0000e+00, ..., 7.5768e-22, 7.5155e-21,\n 1.7801e-21]], device='cuda:0')" + "36": { + "step": "tensor(11268.)", + "exp_avg": "tensor([[ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n [ 5.6052e-45, -5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n ...,\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, -5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n [-5.6052e-45, 5.6052e-45, -5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45]], device='cuda:0')", + "exp_avg_sq": "tensor([[3.8067e-17, 1.7153e-17, 2.5486e-17, ..., 1.1141e-16, 0.0000e+00,\n 3.7600e-17],\n [6.8525e-18, 1.6567e-17, 4.9828e-17, ..., 5.5088e-17, 0.0000e+00,\n 2.2115e-17],\n [6.1496e-15, 1.4550e-15, 8.0845e-14, ..., 2.5919e-14, 0.0000e+00,\n 7.2322e-14],\n ...,\n [1.7139e-15, 7.1278e-17, 1.3499e-14, ..., 7.5328e-15, 0.0000e+00,\n 2.0268e-14],\n [4.6196e-17, 1.6623e-17, 4.8855e-18, ..., 1.8139e-16, 0.0000e+00,\n 9.3531e-17],\n [1.5585e-18, 2.9243e-17, 3.8203e-18, ..., 9.7949e-18, 0.0000e+00,\n 3.9900e-17]], device='cuda:0')" }, - "19": { - "step": "tensor(23788.)", - "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([1.4255e-16, 6.4996e-17, 1.4851e-17, 2.6960e-17, 9.8086e-18, 9.5049e-20,\n 1.6876e-18, 2.6768e-18, 6.8408e-18, 3.8725e-19, 2.1695e-17, 1.1033e-16,\n 4.5464e-17, 5.5072e-17, 8.3145e-18, 5.1852e-17, 1.2359e-16, 1.0244e-17,\n 6.8521e-17, 1.1225e-16, 1.3398e-18, 4.7159e-17, 1.0195e-17, 1.0080e-16,\n 2.0394e-17, 4.1272e-17, 1.8394e-17, 1.1442e-17, 3.1158e-18, 1.9303e-17,\n 5.7745e-17, 2.8940e-17, 1.0515e-18, 2.6545e-16, 1.3764e-16, 6.6917e-17,\n 6.9385e-18, 7.6521e-19, 5.2593e-17, 5.1256e-17, 2.5147e-16, 6.2202e-18,\n 1.1100e-18, 3.5152e-18, 8.8742e-18, 6.4473e-17, 2.7722e-17, 1.8071e-17,\n 1.7584e-17, 2.0613e-17, 8.4325e-19, 2.9682e-17, 7.5391e-17, 4.5994e-17,\n 9.7534e-18, 5.0861e-18, 3.4661e-17, 4.4152e-19, 1.3332e-17, 4.7935e-18,\n 1.4931e-17, 3.2987e-17, 2.1274e-17, 1.2330e-17, 1.2084e-18, 7.8693e-18,\n 2.8615e-17, 1.6833e-18, 3.6068e-17, 3.3948e-17, 3.2470e-19, 1.0960e-17,\n 2.8993e-17, 1.3180e-18, 1.1632e-17, 9.9656e-20, 7.7121e-18, 1.9693e-16,\n 3.9155e-18, 3.6189e-19, 6.0370e-17, 8.0804e-17, 1.3243e-20, 4.5059e-19,\n 1.4395e-17, 4.9999e-17, 1.5044e-18, 2.5588e-17, 3.2803e-18, 5.6896e-17,\n 9.0348e-18, 8.2522e-17, 1.0519e-16, 7.4430e-18, 1.0421e-17, 4.1483e-17,\n 2.7347e-18, 1.6308e-16, 5.4398e-17, 1.9484e-16, 5.5609e-17, 3.8532e-18,\n 7.6891e-19, 2.3531e-17, 1.9556e-18, 1.9425e-16, 4.8760e-18, 1.3815e-17,\n 3.1130e-17, 2.4017e-19, 2.5061e-16, 4.6427e-18, 7.4535e-17, 3.1655e-17,\n 1.0738e-17, 1.4771e-16, 3.6802e-18, 1.7906e-17, 4.9439e-17, 2.3655e-17,\n 2.5803e-17, 4.7746e-17, 4.7413e-18, 7.9699e-19, 3.1080e-19, 1.7362e-17,\n 4.1406e-16, 3.2286e-18, 2.7178e-17, 5.3901e-18, 7.2533e-19, 1.3315e-16,\n 1.1191e-17, 1.8743e-17, 7.5856e-17, 5.4166e-17, 4.5779e-17, 1.4035e-16,\n 6.0520e-17, 1.3314e-20, 5.5205e-19, 8.9570e-19, 4.8740e-17, 2.0002e-16,\n 1.1203e-16, 7.3601e-18, 7.5279e-18, 9.7178e-18, 5.9950e-17, 1.3576e-19,\n 3.1068e-18, 6.2412e-17, 2.1546e-17, 2.3995e-17, 5.3681e-17, 1.8468e-18,\n 1.6862e-17, 4.4302e-17, 8.6068e-18, 1.4350e-16, 6.8038e-17, 1.1165e-17,\n 2.6484e-17, 7.1906e-17, 4.5456e-18, 3.2522e-17, 3.2016e-16, 4.8645e-17,\n 3.3612e-17, 1.5499e-17, 2.2417e-18, 5.3474e-17, 2.1859e-16, 5.4981e-17,\n 7.2254e-17, 5.2632e-19, 7.2852e-18, 2.5155e-17, 2.8667e-17, 3.0734e-17,\n 3.7718e-17, 1.7611e-17, 3.5362e-18, 2.1234e-17, 4.0417e-17, 1.1216e-19,\n 1.0586e-17, 4.3477e-17, 1.1028e-17, 1.7526e-16, 1.3101e-16, 1.3587e-17,\n 8.9941e-18, 3.8958e-18, 2.6000e-20, 6.2779e-16, 7.3573e-18, 1.2667e-17,\n 1.3986e-17, 3.0663e-16, 3.0549e-21, 2.8635e-16, 5.3863e-17, 8.1950e-18,\n 5.2320e-17, 6.4287e-16, 5.9107e-17, 1.8446e-17, 2.6771e-19, 8.2976e-17,\n 6.6992e-18, 1.2109e-16, 1.8854e-17, 3.8416e-17, 2.9408e-19, 4.8919e-17,\n 3.7617e-16, 2.1295e-19, 4.1462e-18, 1.2161e-17, 8.9527e-18, 2.5026e-16,\n 3.1517e-17, 1.7994e-16, 2.4946e-17, 3.0292e-20, 6.4270e-19, 3.6819e-18,\n 1.3646e-18, 5.2539e-17, 4.4919e-19, 4.8010e-19, 9.1196e-17, 3.5665e-17,\n 2.7816e-17, 6.1974e-18, 7.4184e-18, 5.6152e-17, 7.0303e-18, 6.1775e-18,\n 7.1665e-17, 2.9166e-19, 1.5778e-17, 9.6746e-17, 1.6865e-17, 4.8312e-19,\n 9.0287e-19, 4.0599e-17, 4.5838e-18, 1.4602e-16, 7.6126e-17, 7.6491e-18,\n 6.5757e-17, 7.7643e-18, 2.1961e-16, 3.8831e-18], device='cuda:0')" + "37": { + "step": "tensor(11268.)", + "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([6.7602e-14, 2.3332e-14, 8.7980e-11, 1.1674e-13, 4.2467e-11, 1.4626e-11,\n 2.7982e-12, 3.2110e-12, 8.1145e-11, 2.9692e-13, 4.1254e-11, 5.0198e-12,\n 4.4499e-14, 9.2815e-13, 6.5526e-15, 4.7067e-13, 6.7148e-12, 7.6044e-12,\n 4.9165e-13, 5.9047e-13, 7.2017e-12, 3.3226e-13, 5.0081e-12, 2.9832e-12,\n 2.5213e-12, 2.5120e-11, 4.7747e-12, 4.1576e-12, 3.1639e-12, 6.3060e-11,\n 2.3168e-11, 1.4336e-11, 1.6701e-10, 2.7160e-13, 2.2376e-12, 5.0049e-13,\n 6.9579e-12, 3.0923e-13, 2.8086e-11, 7.4181e-13, 2.7813e-11, 4.7430e-13,\n 1.4774e-12, 2.6333e-11, 4.3718e-14, 1.5957e-12, 2.4685e-12, 2.4613e-12,\n 5.0269e-13, 1.7728e-11, 6.1258e-12, 6.5688e-12, 4.4004e-11, 2.8779e-12,\n 4.8896e-13, 3.5236e-12, 7.0266e-16, 1.7161e-12, 7.5588e-12, 7.4101e-11,\n 9.1360e-12, 8.1054e-13, 2.5836e-14, 2.7593e-12, 3.1355e-12, 4.1473e-11,\n 3.7844e-13, 2.7786e-12, 9.1194e-13, 6.3740e-12, 3.9018e-12, 3.4032e-12,\n 1.0501e-12, 1.5533e-11, 2.1834e-11, 1.2329e-11, 1.4959e-11, 3.6119e-12,\n 1.0802e-12, 2.9868e-14, 3.6992e-12, 1.4372e-12, 3.6963e-13, 2.0814e-12,\n 2.9306e-12, 9.2092e-15, 3.0585e-11, 4.9078e-12, 8.9542e-13, 4.0021e-14,\n 2.8695e-13, 2.9288e-12, 1.6742e-11, 2.6210e-11, 2.4557e-11, 1.8648e-12,\n 2.1988e-11, 9.2180e-12, 6.3504e-12, 9.1528e-13, 2.6961e-12, 1.1315e-13,\n 4.8186e-13, 2.8891e-11, 1.1609e-11, 2.0279e-11, 6.6085e-12, 6.3373e-14,\n 2.7186e-11, 2.2712e-12, 3.9414e-12, 5.4501e-13, 5.7729e-12, 2.8062e-12,\n 2.5978e-12, 1.2283e-10, 8.9703e-14, 4.1454e-12, 7.9854e-13, 3.8382e-11,\n 3.9091e-11, 5.3346e-12, 1.2208e-11, 4.8093e-13, 6.0287e-13, 1.0532e-11,\n 5.6894e-12, 6.4270e-12, 8.6047e-13, 2.0158e-11, 1.5908e-14, 4.5451e-13,\n 1.0958e-10, 9.9906e-12, 1.8822e-13, 2.1680e-12, 6.7894e-12, 2.7022e-12,\n 4.4465e-11, 8.5645e-11, 2.0018e-12, 6.1932e-12, 5.2587e-13, 8.6776e-14,\n 9.5838e-13, 4.3421e-12, 4.5682e-12, 8.6140e-12, 3.2458e-13, 1.8362e-11,\n 1.6433e-13, 4.5208e-13, 1.5083e-11, 6.7263e-12, 5.6788e-12, 3.2804e-11,\n 4.7362e-13, 2.5169e-11, 2.1787e-13, 1.1762e-12, 2.8222e-14, 2.0292e-11,\n 1.8833e-14, 1.5322e-12, 2.6128e-11, 1.4830e-11, 6.9850e-13, 9.1728e-12,\n 7.8077e-14, 4.5716e-13, 1.9595e-12, 1.0872e-11, 2.5805e-11, 9.0863e-12,\n 5.4882e-12, 1.8675e-12, 2.7010e-11, 3.2724e-11, 7.8415e-14, 5.9994e-15,\n 5.9588e-14, 9.7509e-12, 7.8574e-13, 2.9725e-13, 1.0212e-12, 3.8129e-11,\n 2.3737e-12, 4.6780e-14, 8.5004e-13, 1.4081e-12, 1.6620e-11, 9.8251e-14,\n 1.5546e-11, 3.1049e-11, 9.2328e-12, 2.5145e-12, 7.5315e-13, 8.5052e-12,\n 2.3004e-12, 9.6258e-12, 2.5715e-12, 1.5466e-13, 8.0506e-11, 1.5855e-13,\n 5.5186e-12, 1.0335e-14, 6.2226e-12, 2.0699e-11, 5.4415e-14, 1.4146e-12,\n 2.3793e-11, 1.4381e-12, 1.0969e-11, 7.3829e-11, 6.8060e-11, 7.5029e-12,\n 1.1199e-13, 3.4170e-12, 1.9903e-11, 4.0915e-14, 4.6974e-12, 8.2767e-11,\n 2.5684e-12, 4.4970e-12, 1.4890e-11, 1.9154e-12, 7.3287e-13, 6.6907e-12,\n 4.0311e-12, 3.3611e-12, 1.9034e-12, 7.8647e-12, 3.5316e-12, 1.8812e-12,\n 8.1689e-12, 1.2936e-12, 2.6196e-12, 3.2810e-12, 7.9772e-12, 1.4053e-13,\n 3.9491e-11, 1.3648e-12, 3.7057e-13, 1.3470e-11, 4.6216e-12, 6.9149e-14,\n 7.2008e-12, 4.7901e-11, 3.5642e-11, 5.2234e-11, 1.7538e-12, 7.3604e-12,\n 5.5967e-12, 2.4384e-11, 1.9314e-13, 2.4293e-14], device='cuda:0')" }, - "20": { - "step": "tensor(23788.)", - "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([3.5169e-19, 3.2161e-19, 2.1628e-20, 1.4114e-19, 1.6558e-20, 1.3656e-21,\n 3.0027e-21, 1.3944e-20, 8.2799e-21, 4.8640e-22, 3.9962e-20, 2.0012e-19,\n 1.1522e-19, 2.5918e-19, 2.6680e-20, 1.2213e-19, 4.0059e-19, 1.7768e-20,\n 2.2165e-19, 3.1935e-19, 6.6181e-22, 1.9836e-19, 2.2429e-20, 4.1140e-19,\n 7.9965e-20, 8.0105e-20, 4.6519e-20, 3.8018e-20, 1.1307e-20, 3.4165e-20,\n 1.7635e-19, 8.5859e-20, 6.6970e-21, 7.1467e-19, 5.3328e-19, 1.6688e-19,\n 1.3775e-20, 1.3941e-23, 1.0929e-19, 9.2658e-20, 1.0637e-18, 1.4839e-20,\n 4.7224e-21, 1.5422e-20, 1.7729e-20, 1.4503e-19, 6.0385e-20, 3.8122e-20,\n 2.6893e-20, 3.1849e-20, 6.6070e-22, 8.1102e-20, 1.5653e-19, 1.0026e-19,\n 1.3360e-20, 2.3985e-20, 1.1214e-19, 9.2378e-22, 4.4002e-20, 1.3577e-20,\n 2.6959e-20, 2.4608e-19, 5.9671e-20, 2.9614e-20, 1.8287e-23, 1.2572e-20,\n 6.2479e-20, 9.1390e-22, 6.6597e-20, 1.0991e-19, 7.6407e-21, 2.5182e-20,\n 3.2191e-20, 4.4252e-21, 2.1274e-20, 6.4405e-21, 4.4969e-20, 9.9993e-19,\n 3.4271e-21, 1.0227e-24, 2.5286e-19, 2.0949e-19, 1.9253e-21, 2.2260e-21,\n 2.8315e-20, 1.1171e-19, 5.9950e-21, 1.1979e-19, 9.6140e-21, 1.0609e-19,\n 2.7366e-20, 1.8857e-19, 3.8237e-19, 9.4372e-21, 2.8830e-20, 1.4621e-19,\n 5.9215e-21, 3.4736e-19, 2.1170e-19, 5.4788e-19, 1.6667e-19, 1.7938e-20,\n 9.4026e-21, 4.1707e-20, 8.1576e-22, 4.5647e-19, 5.0971e-21, 3.2570e-20,\n 6.4817e-20, 3.3942e-22, 6.3966e-19, 1.4472e-20, 2.1124e-19, 6.6466e-20,\n 1.9594e-20, 5.9621e-19, 1.4500e-20, 3.8179e-20, 1.4907e-19, 5.7382e-20,\n 4.7477e-20, 9.9440e-20, 3.9847e-21, 7.6905e-23, 3.1717e-21, 2.6718e-20,\n 1.8723e-18, 1.2244e-20, 8.6804e-20, 8.2209e-21, 2.6579e-22, 3.9710e-19,\n 2.7385e-20, 5.5763e-20, 1.4561e-19, 1.2232e-19, 1.2858e-19, 3.6225e-19,\n 1.6556e-19, 2.2462e-21, 1.4157e-20, 5.1325e-21, 1.0709e-19, 5.8791e-19,\n 7.0393e-19, 2.9135e-20, 3.0240e-20, 2.2748e-20, 1.2419e-19, 3.4431e-23,\n 3.1099e-20, 1.3685e-19, 6.9563e-20, 2.7317e-20, 1.2593e-19, 1.1188e-20,\n 2.4884e-20, 7.5478e-20, 1.4158e-20, 5.6756e-19, 1.3268e-19, 1.2635e-20,\n 6.1700e-20, 2.3970e-19, 1.3209e-20, 4.2266e-20, 1.1005e-18, 9.2259e-20,\n 1.0285e-19, 3.9498e-20, 1.1488e-21, 2.1329e-19, 5.2654e-19, 2.0520e-19,\n 1.3601e-19, 2.0247e-21, 1.1385e-20, 1.1252e-19, 8.7809e-20, 8.0267e-20,\n 9.7404e-20, 3.4082e-20, 4.2942e-21, 3.1881e-20, 1.0118e-19, 6.0483e-26,\n 1.6308e-20, 9.6722e-20, 9.7122e-21, 4.3884e-19, 3.4194e-19, 1.7113e-20,\n 2.6748e-20, 2.1315e-20, 2.4508e-21, 2.4336e-18, 9.3664e-21, 2.0284e-20,\n 4.2327e-20, 8.4274e-19, 7.2601e-22, 7.6938e-19, 9.7196e-20, 1.1194e-20,\n 1.7403e-19, 1.6384e-18, 1.2803e-19, 4.6549e-20, 8.7756e-23, 1.2966e-19,\n 9.1763e-21, 3.0185e-19, 1.2802e-19, 1.1724e-19, 1.7575e-22, 8.4952e-20,\n 1.7897e-18, 7.3013e-23, 1.5566e-20, 1.4567e-20, 1.1435e-20, 1.1008e-18,\n 1.7538e-19, 4.3790e-19, 7.3751e-20, 6.4450e-22, 3.1239e-23, 1.0917e-20,\n 1.3516e-21, 1.4773e-19, 3.6307e-21, 3.3812e-22, 5.3795e-19, 1.5749e-19,\n 8.4184e-20, 1.4437e-20, 9.4885e-21, 1.6332e-19, 1.9295e-20, 6.1679e-21,\n 1.7745e-19, 6.7942e-23, 4.7340e-20, 2.4918e-19, 2.5574e-20, 8.2823e-23,\n 2.1266e-22, 7.7660e-20, 7.2296e-21, 3.9407e-19, 1.5469e-19, 1.0751e-20,\n 1.4729e-19, 1.5366e-20, 7.0330e-19, 2.0449e-20], device='cuda:0')" + "38": { + "step": "tensor(11268.)", + "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([1.5187e-16, 1.7418e-19, 1.6902e-13, 5.7739e-16, 8.9904e-14, 2.8297e-14,\n 4.9316e-15, 7.2007e-15, 1.8675e-13, 7.3868e-16, 1.4761e-13, 8.5998e-15,\n 2.6909e-16, 1.1597e-15, 2.1058e-17, 1.7048e-15, 1.4910e-14, 2.0922e-14,\n 1.7838e-15, 1.2892e-15, 1.1711e-14, 8.1664e-16, 6.6337e-15, 5.1297e-15,\n 4.7211e-15, 3.1273e-14, 7.3232e-15, 7.4417e-15, 1.8104e-14, 1.8302e-13,\n 2.8596e-14, 2.5165e-14, 3.1561e-13, 3.9863e-16, 3.8927e-15, 1.8965e-15,\n 8.8481e-15, 9.3052e-16, 7.1173e-14, 1.0419e-15, 7.8390e-14, 1.6156e-15,\n 4.3507e-15, 3.4979e-14, 4.3840e-17, 1.7567e-15, 3.8827e-15, 3.5338e-15,\n 1.3736e-15, 4.2210e-14, 8.0917e-15, 1.7268e-14, 1.1906e-13, 4.0773e-15,\n 7.9951e-16, 5.2272e-15, 4.2257e-17, 2.2060e-15, 1.0616e-14, 2.1544e-13,\n 1.5012e-14, 1.2699e-15, 2.5436e-19, 6.1730e-15, 5.6367e-15, 1.1489e-13,\n 9.5783e-16, 3.8587e-15, 1.8567e-15, 8.2341e-15, 6.5675e-15, 6.1851e-15,\n 9.1247e-16, 3.1165e-14, 3.7633e-14, 1.8773e-14, 4.1426e-14, 5.5167e-15,\n 2.8459e-15, 7.7618e-18, 6.5322e-15, 2.0749e-15, 7.1503e-16, 2.6115e-15,\n 4.2895e-15, 2.7359e-17, 4.6596e-14, 1.1949e-14, 2.0194e-15, 7.2043e-18,\n 6.9617e-16, 4.9106e-15, 3.7680e-14, 7.1836e-14, 2.9661e-14, 2.3196e-15,\n 3.0788e-14, 1.3530e-14, 9.0382e-15, 4.0221e-15, 3.7420e-15, 1.6316e-16,\n 8.0276e-16, 5.7792e-14, 3.3770e-14, 3.9143e-14, 9.5567e-15, 2.1369e-16,\n 9.7510e-14, 4.1525e-15, 8.8275e-15, 1.1524e-15, 9.2456e-15, 3.5167e-15,\n 4.8509e-15, 3.1635e-13, 2.1365e-16, 4.1961e-15, 4.1801e-15, 1.2086e-13,\n 6.7483e-14, 7.3396e-15, 1.5462e-14, 4.1012e-15, 1.7197e-15, 2.2589e-14,\n 1.1603e-14, 1.5134e-14, 1.6699e-15, 6.5712e-14, 2.6431e-16, 9.0672e-16,\n 2.9748e-13, 9.0782e-15, 2.6697e-16, 2.9641e-15, 2.3223e-14, 3.8576e-15,\n 1.2637e-13, 2.7545e-13, 3.9806e-15, 1.3201e-14, 2.0041e-15, 3.6607e-16,\n 1.6223e-15, 6.6433e-15, 9.0134e-15, 1.6151e-14, 1.2195e-15, 6.5280e-14,\n 2.5589e-16, 5.2727e-16, 3.2282e-14, 1.3350e-14, 1.2420e-14, 5.8200e-14,\n 6.8779e-16, 2.9191e-14, 3.1846e-16, 1.7646e-15, 1.0686e-16, 3.3560e-14,\n 7.2425e-17, 3.4934e-15, 4.4843e-14, 2.3288e-14, 8.6440e-16, 1.9144e-14,\n 3.0636e-16, 1.5272e-15, 1.8246e-15, 6.3593e-14, 6.3759e-14, 2.7012e-14,\n 1.5966e-14, 4.3335e-15, 5.3243e-14, 8.3904e-14, 3.3399e-16, 5.1200e-18,\n 1.2516e-16, 2.6708e-14, 2.4210e-15, 5.4076e-16, 3.5996e-15, 4.9574e-14,\n 3.9276e-15, 9.8538e-17, 1.6663e-15, 2.5180e-15, 3.7211e-14, 6.2327e-16,\n 2.9014e-14, 1.1292e-13, 1.3526e-14, 7.3219e-15, 1.3521e-15, 1.4623e-14,\n 3.2478e-15, 1.8178e-14, 4.1876e-15, 1.1207e-15, 1.3039e-13, 5.4043e-16,\n 8.4634e-15, 1.9947e-17, 1.4263e-14, 4.7389e-14, 1.8666e-16, 1.5734e-15,\n 4.1937e-14, 1.9534e-15, 1.8864e-14, 1.3659e-13, 1.0918e-13, 1.4110e-14,\n 1.8734e-16, 5.5284e-15, 3.2587e-14, 3.6043e-17, 8.3133e-15, 1.4193e-13,\n 6.1072e-15, 8.2739e-15, 2.1225e-14, 4.3891e-15, 1.7124e-15, 1.0957e-14,\n 4.3003e-15, 7.7113e-15, 3.6258e-15, 2.1638e-14, 5.2819e-15, 3.4504e-15,\n 1.2517e-14, 2.8603e-15, 3.3005e-15, 4.7209e-15, 2.7842e-14, 2.9940e-16,\n 1.0925e-13, 2.1956e-15, 3.5961e-15, 1.4457e-14, 6.4765e-15, 5.1526e-16,\n 1.5431e-14, 1.4876e-13, 5.8184e-14, 1.8268e-13, 2.8620e-15, 1.5188e-14,\n 9.0186e-15, 4.2205e-14, 3.3383e-16, 9.0653e-17], device='cuda:0')" }, - "21": { - "step": "tensor(23788.)", - "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([6.0319e-19, 3.0790e-19, 4.2391e-20, 1.3856e-19, 2.1789e-20, 9.2558e-22,\n 4.1979e-21, 2.2802e-20, 1.3918e-20, 2.7275e-22, 6.6242e-20, 3.7389e-19,\n 1.3186e-19, 2.5695e-19, 4.4903e-20, 2.3686e-19, 5.3567e-19, 2.8322e-20,\n 1.9686e-19, 4.6479e-19, 7.4822e-22, 2.2210e-19, 5.1408e-20, 4.3396e-19,\n 1.0683e-19, 1.2371e-19, 8.5664e-20, 6.0108e-20, 2.3285e-20, 4.9803e-20,\n 2.5896e-19, 1.3474e-19, 1.0472e-20, 9.1361e-19, 5.8863e-19, 2.8747e-19,\n 9.9547e-21, 2.8936e-24, 2.3212e-19, 1.7630e-19, 8.1807e-19, 3.4516e-20,\n 1.1390e-20, 2.7243e-20, 3.0396e-20, 2.0726e-19, 1.2802e-19, 8.9228e-20,\n 4.4185e-20, 6.6504e-20, 1.0798e-21, 1.3123e-19, 2.4575e-19, 1.4415e-19,\n 2.4321e-20, 3.2902e-20, 1.6020e-19, 8.3295e-22, 7.4291e-20, 2.8531e-20,\n 3.5729e-20, 1.7331e-19, 9.7324e-20, 5.6402e-20, 3.0045e-22, 1.9667e-20,\n 1.3529e-19, 2.4695e-21, 1.6036e-19, 1.6452e-19, 8.3996e-21, 5.5632e-20,\n 1.0078e-19, 9.5495e-21, 2.7659e-20, 5.9028e-21, 4.7893e-20, 8.7028e-19,\n 5.1646e-21, 8.2072e-25, 2.9162e-19, 3.6387e-19, 2.1210e-21, 4.8156e-21,\n 7.3379e-20, 2.3283e-19, 1.0587e-20, 1.3548e-19, 1.0638e-20, 2.5038e-19,\n 4.7295e-20, 3.5517e-19, 3.3874e-19, 1.9740e-20, 4.6895e-20, 1.1204e-19,\n 7.7664e-21, 5.5686e-19, 2.5507e-19, 8.1607e-19, 2.4514e-19, 2.6776e-20,\n 1.0723e-20, 6.4143e-20, 1.2050e-21, 6.4113e-19, 8.6838e-21, 7.2298e-20,\n 8.7600e-20, 2.2524e-22, 8.4160e-19, 2.8251e-20, 3.3528e-19, 1.1663e-19,\n 2.5792e-20, 4.6662e-19, 1.9933e-20, 4.8284e-20, 2.2614e-19, 1.1436e-19,\n 7.3117e-20, 1.4084e-19, 7.3308e-21, 2.1207e-22, 5.9063e-21, 4.9806e-20,\n 1.6646e-18, 1.8119e-20, 1.3722e-19, 1.5178e-20, 3.0685e-22, 5.7667e-19,\n 5.7125e-20, 8.8794e-20, 2.3911e-19, 1.7677e-19, 2.0279e-19, 4.7334e-19,\n 2.6943e-19, 2.7394e-21, 1.0705e-20, 8.3953e-21, 1.5600e-19, 6.8426e-19,\n 5.2060e-19, 4.3620e-20, 4.3179e-20, 5.0789e-20, 1.9656e-19, 7.6870e-23,\n 2.8518e-20, 1.8777e-19, 1.0922e-19, 6.8828e-20, 1.4904e-19, 1.7011e-20,\n 4.2205e-20, 1.6061e-19, 2.2826e-20, 6.0821e-19, 2.2342e-19, 3.1065e-20,\n 6.2297e-20, 3.2338e-19, 2.4426e-20, 9.2887e-20, 1.1432e-18, 1.5506e-19,\n 1.5408e-19, 7.5713e-20, 3.1575e-21, 1.3925e-19, 7.4979e-19, 2.6060e-19,\n 2.1728e-19, 4.1366e-21, 1.6936e-20, 1.3448e-19, 6.5586e-20, 1.4694e-19,\n 9.9618e-20, 4.4005e-20, 7.1611e-21, 5.8011e-20, 1.6945e-19, 1.7387e-23,\n 3.2491e-20, 1.2407e-19, 2.9644e-20, 7.1944e-19, 4.3723e-19, 3.5874e-20,\n 4.8942e-20, 2.5713e-20, 3.2191e-21, 2.5760e-18, 1.4883e-20, 2.7350e-20,\n 6.7906e-20, 1.0805e-18, 1.8158e-21, 1.1922e-18, 1.7332e-19, 1.8194e-20,\n 2.4540e-19, 2.2228e-18, 1.7714e-19, 4.3356e-20, 2.2166e-22, 2.8847e-19,\n 1.1101e-20, 5.2222e-19, 1.0747e-19, 1.8559e-19, 2.1011e-22, 1.6317e-19,\n 1.2632e-18, 4.9220e-23, 2.6200e-20, 2.7138e-20, 1.6629e-20, 1.0494e-18,\n 1.6349e-19, 7.6558e-19, 1.2819e-19, 1.1866e-21, 2.1414e-23, 2.4806e-20,\n 3.4481e-21, 2.2582e-19, 7.2633e-21, 7.3031e-22, 4.0787e-19, 1.6860e-19,\n 1.2833e-19, 3.1032e-20, 1.9210e-20, 2.5191e-19, 4.0030e-20, 1.4976e-20,\n 3.2343e-19, 6.4208e-23, 7.6795e-20, 3.0271e-19, 4.8000e-20, 1.6011e-22,\n 1.9643e-22, 1.2932e-19, 7.2197e-21, 6.2870e-19, 2.4290e-19, 1.7819e-20,\n 1.9742e-19, 2.6070e-20, 7.2824e-19, 2.8495e-20], device='cuda:0')" + "39": { + "step": "tensor(11268.)", + "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([3.0956e-16, 1.1820e-17, 2.4193e-13, 5.2110e-16, 1.1031e-13, 3.7018e-14,\n 8.7719e-15, 9.4259e-15, 2.2626e-13, 1.5741e-15, 1.1870e-13, 1.4335e-14,\n 4.2323e-16, 2.8379e-15, 8.5158e-17, 2.1552e-15, 2.0162e-14, 2.2691e-14,\n 1.5979e-15, 2.0595e-15, 2.0703e-14, 1.1366e-15, 1.4558e-14, 9.0316e-15,\n 7.5054e-15, 6.8376e-14, 1.4456e-14, 1.1152e-14, 1.2148e-14, 1.7501e-13,\n 6.0130e-14, 3.9903e-14, 4.4894e-13, 7.0159e-16, 6.7194e-15, 1.9737e-15,\n 1.7937e-14, 1.6914e-15, 7.6171e-14, 2.4868e-15, 7.1744e-14, 1.7815e-15,\n 4.9355e-15, 7.7497e-14, 3.1098e-17, 3.6067e-15, 7.2440e-15, 7.6606e-15,\n 1.5008e-15, 4.9026e-14, 1.8248e-14, 1.5199e-14, 1.1524e-13, 8.0521e-15,\n 1.6599e-15, 1.0777e-14, 1.0359e-16, 4.5878e-15, 2.0350e-14, 2.0156e-13,\n 2.4289e-14, 2.4040e-15, 2.3752e-17, 9.0057e-15, 9.7225e-15, 1.0655e-13,\n 1.2990e-15, 7.9358e-15, 2.9072e-15, 1.8333e-14, 1.1823e-14, 1.0440e-14,\n 2.7417e-15, 4.4578e-14, 6.2820e-14, 3.3854e-14, 4.2167e-14, 1.0731e-14,\n 3.6977e-15, 7.0548e-17, 1.0931e-14, 3.7794e-15, 1.4590e-15, 6.5020e-15,\n 7.4191e-15, 1.2174e-16, 8.2893e-14, 1.4741e-14, 4.0722e-15, 7.6887e-18,\n 8.9701e-16, 8.2245e-15, 4.2952e-14, 6.8075e-14, 6.1317e-14, 4.2915e-15,\n 5.8528e-14, 2.6968e-14, 1.6523e-14, 4.5581e-15, 8.0665e-15, 3.9300e-16,\n 1.6469e-15, 7.5726e-14, 3.4086e-14, 4.7739e-14, 1.9722e-14, 3.7571e-16,\n 6.7503e-14, 6.8223e-15, 1.1958e-14, 2.0580e-15, 1.4780e-14, 6.5223e-15,\n 7.0887e-15, 3.3093e-13, 4.7863e-16, 1.1197e-14, 3.7104e-15, 1.0915e-13,\n 1.0472e-13, 1.5193e-14, 3.4298e-14, 4.2535e-15, 2.0393e-15, 3.1299e-14,\n 1.4219e-14, 1.8227e-14, 2.9983e-15, 5.6726e-14, 3.5792e-16, 1.5369e-15,\n 2.9401e-13, 2.6493e-14, 5.8427e-16, 6.7892e-15, 2.0184e-14, 6.4888e-15,\n 1.1644e-13, 2.2739e-13, 6.5108e-15, 1.8804e-14, 1.5648e-15, 4.4721e-16,\n 3.0549e-15, 1.2460e-14, 1.1651e-14, 2.4019e-14, 1.5490e-15, 5.2865e-14,\n 5.5015e-16, 1.1172e-15, 3.9455e-14, 1.6704e-14, 1.6285e-14, 8.9957e-14,\n 1.4693e-15, 6.6824e-14, 6.8855e-16, 3.2606e-15, 5.4023e-16, 5.8005e-14,\n 2.6327e-16, 4.6268e-15, 7.3734e-14, 3.8475e-14, 2.1500e-15, 2.4961e-14,\n 4.1862e-16, 1.5732e-15, 4.9100e-15, 3.3354e-14, 6.6142e-14, 2.8205e-14,\n 1.4722e-14, 6.0636e-15, 7.6261e-14, 9.3259e-14, 9.3869e-16, 1.8273e-17,\n 2.0721e-16, 2.6260e-14, 2.7228e-15, 1.0536e-15, 3.1008e-15, 1.0398e-13,\n 6.5608e-15, 2.0529e-16, 3.0080e-15, 4.2414e-15, 4.5224e-14, 8.2989e-16,\n 3.9672e-14, 8.7107e-14, 2.5007e-14, 7.6775e-15, 2.3512e-15, 2.0767e-14,\n 6.2183e-15, 2.7599e-14, 7.6892e-15, 1.2547e-15, 2.1578e-13, 9.0545e-16,\n 1.4863e-14, 2.6140e-17, 1.7855e-14, 5.4181e-14, 6.4058e-16, 4.6105e-15,\n 6.7241e-14, 3.4529e-15, 3.1347e-14, 1.9336e-13, 1.8171e-13, 2.0165e-14,\n 4.6955e-16, 9.2508e-15, 5.3024e-14, 1.3607e-16, 1.1422e-14, 2.2608e-13,\n 7.4901e-15, 1.2700e-14, 4.3591e-14, 4.6260e-15, 2.5800e-15, 1.8981e-14,\n 1.1478e-14, 9.8764e-15, 6.1658e-15, 2.1306e-14, 8.2529e-15, 5.5889e-15,\n 2.3410e-14, 4.1332e-15, 6.6861e-15, 9.7498e-15, 2.0074e-14, 6.6284e-16,\n 1.0565e-13, 4.1072e-15, 2.7739e-15, 3.8933e-14, 1.2684e-14, 6.8705e-16,\n 2.1728e-14, 1.2179e-13, 1.0357e-13, 1.4818e-13, 5.4153e-15, 1.8879e-14,\n 1.3282e-14, 6.6565e-14, 8.1795e-16, 8.8488e-17], device='cuda:0')" }, - "22": { - "step": "tensor(23788.)", - "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([[4.2620e-20, 1.6010e-19, 0.0000e+00, ..., 2.3312e-19, 3.8121e-19,\n 5.8326e-20],\n [1.9549e-20, 2.1818e-20, 0.0000e+00, ..., 3.6005e-20, 1.1106e-19,\n 7.0596e-23],\n [2.6247e-20, 1.4258e-20, 0.0000e+00, ..., 2.3977e-20, 5.1228e-20,\n 4.1257e-21],\n ...,\n [4.4831e-20, 5.5274e-20, 0.0000e+00, ..., 6.8270e-20, 1.2918e-19,\n 4.1093e-20],\n [1.3650e-19, 4.3339e-20, 0.0000e+00, ..., 4.3082e-20, 2.7337e-19,\n 2.6025e-20],\n [6.6813e-22, 9.3566e-22, 0.0000e+00, ..., 1.0089e-20, 1.6080e-20,\n 3.2821e-21]], device='cuda:0')" + "40": { + "step": "tensor(11268.)", + "exp_avg": "tensor([[-5.6052e-45, 5.6052e-45, -5.6052e-45, ..., -5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n [-5.6052e-45, -5.6052e-45, -5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n ...,\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n [-5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45]], device='cuda:0')", + "exp_avg_sq": "tensor([[6.8036e-21, 2.7845e-19, 1.9085e-17, ..., 7.8653e-18, 0.0000e+00,\n 7.9333e-17],\n [1.8123e-17, 2.1025e-18, 1.8871e-16, ..., 1.1612e-16, 0.0000e+00,\n 5.0431e-17],\n [5.7999e-15, 8.9477e-17, 7.7679e-14, ..., 1.2162e-14, 0.0000e+00,\n 5.5812e-14],\n ...,\n [1.3456e-15, 3.7460e-16, 2.3735e-14, ..., 1.3488e-14, 0.0000e+00,\n 1.2141e-14],\n [7.7572e-18, 1.7120e-17, 2.8716e-16, ..., 2.4899e-17, 0.0000e+00,\n 1.3751e-16],\n [1.4298e-17, 1.8727e-16, 1.2743e-15, ..., 1.4906e-16, 0.0000e+00,\n 1.2532e-15]], device='cuda:0')" }, - "23": { - "step": "tensor(23788.)", - "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([1.6650e-16, 1.2815e-17, 2.8305e-17, 2.1302e-17, 7.2637e-18, 3.8962e-18,\n 1.6760e-17, 9.9714e-21, 6.6830e-18, 1.1335e-18, 1.2584e-16, 2.4917e-17,\n 3.3523e-17, 2.1751e-17, 6.3652e-18, 1.5709e-17, 6.5831e-17, 4.3062e-18,\n 1.2463e-17, 1.3275e-16, 5.2403e-19, 5.0401e-17, 2.1940e-17, 1.2675e-17,\n 1.6962e-18, 5.2863e-17, 1.6699e-17, 2.6550e-18, 1.8564e-18, 1.4786e-17,\n 7.3038e-17, 1.5208e-17, 1.8656e-18, 5.3438e-16, 5.2785e-17, 3.6987e-17,\n 6.3341e-18, 4.5904e-22, 4.9356e-17, 2.4892e-16, 3.1306e-17, 1.9109e-17,\n 3.3780e-18, 1.9433e-17, 1.1347e-16, 8.4196e-17, 4.7613e-17, 2.7859e-17,\n 7.0190e-18, 1.2091e-16, 1.3558e-18, 2.1854e-16, 1.5148e-16, 3.4547e-17,\n 3.5056e-18, 6.3336e-19, 8.3107e-17, 2.6423e-18, 3.7947e-17, 7.6513e-18,\n 1.7857e-17, 2.0477e-17, 1.0894e-16, 5.7825e-17, 2.6385e-18, 6.8265e-18,\n 4.0747e-17, 7.2667e-18, 5.7057e-17, 1.3718e-18, 2.0152e-21, 2.3410e-17,\n 4.3054e-17, 7.5631e-19, 1.8260e-17, 2.2315e-18, 6.3281e-18, 1.2145e-16,\n 1.7856e-18, 7.6326e-19, 9.7577e-18, 8.5574e-17, 2.4405e-21, 6.3420e-18,\n 1.7770e-18, 6.4463e-17, 2.5941e-17, 1.4536e-17, 2.5646e-17, 3.7564e-16,\n 5.0320e-18, 8.4697e-17, 1.0801e-16, 1.6336e-17, 2.0496e-16, 1.5314e-17,\n 1.8134e-17, 1.1513e-16, 3.8691e-17, 4.5117e-17, 1.0022e-16, 5.6425e-18,\n 1.0869e-18, 4.0492e-17, 1.5006e-18, 2.1613e-16, 1.2175e-17, 2.4644e-17,\n 7.5278e-18, 3.4810e-20, 1.7675e-16, 1.2647e-17, 1.7554e-16, 5.3585e-17,\n 5.8627e-18, 9.7517e-18, 1.1619e-17, 9.8735e-18, 7.4501e-18, 1.8106e-17,\n 5.4197e-17, 3.1695e-17, 5.3851e-18, 4.2956e-21, 3.1874e-20, 1.9504e-17,\n 1.0851e-16, 7.0362e-17, 5.6936e-18, 9.4537e-17, 7.4178e-20, 4.6719e-16,\n 2.1255e-17, 3.9511e-17, 6.7782e-17, 1.8199e-16, 5.7738e-17, 1.8347e-16,\n 5.0046e-17, 2.2830e-20, 6.5465e-19, 3.4182e-18, 2.1971e-17, 1.9783e-16,\n 3.4495e-17, 9.3411e-18, 8.2661e-18, 3.2038e-18, 1.9665e-16, 4.2227e-18,\n 3.4487e-18, 1.6427e-17, 1.7950e-17, 7.1444e-17, 2.4355e-17, 3.2619e-18,\n 1.7196e-17, 1.8263e-16, 5.1632e-18, 2.5144e-16, 6.3907e-17, 9.2904e-17,\n 9.6985e-18, 8.4541e-17, 6.4667e-17, 5.3282e-17, 5.1338e-16, 7.2221e-17,\n 7.9429e-17, 1.0272e-17, 3.2766e-18, 1.3816e-17, 9.5216e-17, 2.9292e-17,\n 5.4630e-17, 5.8511e-19, 4.8127e-18, 3.2102e-18, 6.5169e-18, 2.2942e-17,\n 3.3702e-18, 1.8591e-17, 1.2269e-17, 1.6444e-17, 1.2687e-16, 4.3120e-19,\n 1.2850e-17, 3.9345e-17, 7.5562e-18, 2.0396e-16, 2.1505e-17, 3.0725e-17,\n 3.3182e-17, 2.6221e-18, 1.1155e-20, 5.8352e-16, 6.9370e-18, 7.1913e-18,\n 4.2930e-17, 1.1178e-16, 4.5320e-21, 3.1732e-17, 1.2025e-16, 1.4228e-18,\n 7.2275e-17, 2.4055e-16, 3.7907e-17, 7.7863e-18, 2.9083e-20, 3.2877e-16,\n 6.7583e-18, 1.1275e-16, 3.4646e-18, 1.5491e-17, 2.6213e-18, 3.1270e-17,\n 4.1340e-17, 1.4215e-19, 3.1457e-18, 4.0719e-18, 1.3599e-17, 5.2835e-17,\n 9.7886e-18, 1.8732e-16, 3.3017e-17, 1.3488e-18, 4.1681e-19, 2.0132e-18,\n 7.4825e-18, 1.7146e-17, 5.2028e-19, 1.8271e-18, 2.9873e-17, 2.4191e-17,\n 9.7177e-17, 3.0447e-17, 7.1167e-18, 5.4270e-17, 8.1177e-18, 1.4238e-17,\n 5.2050e-17, 1.0812e-19, 6.8072e-17, 8.4460e-17, 2.9330e-17, 6.6708e-20,\n 2.3294e-21, 2.9104e-16, 1.4051e-18, 1.0695e-16, 3.6693e-17, 6.3101e-18,\n 4.8179e-17, 5.0406e-17, 1.1700e-16, 3.2994e-18], device='cuda:0')" + "41": { + "step": "tensor(11268.)", + "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([6.4470e-16, 1.3765e-14, 2.6147e-11, 1.5243e-14, 5.7434e-12, 8.3110e-12,\n 4.3151e-12, 5.5613e-13, 8.8192e-11, 2.4843e-12, 2.0417e-11, 1.8309e-11,\n 5.4829e-14, 1.7146e-12, 3.5598e-14, 7.1168e-14, 8.2841e-12, 4.6418e-12,\n 9.9192e-13, 2.5180e-13, 1.4616e-12, 1.4562e-13, 9.3085e-12, 4.4165e-12,\n 1.1776e-11, 1.5821e-11, 5.0702e-12, 7.3556e-12, 2.8794e-12, 4.1077e-11,\n 4.3386e-11, 1.4909e-12, 1.7364e-10, 1.5281e-12, 1.8384e-12, 1.6598e-13,\n 6.3454e-12, 1.3472e-13, 4.1410e-12, 1.4662e-12, 1.4158e-11, 1.3421e-13,\n 3.3858e-14, 8.6095e-11, 2.1505e-14, 3.8738e-12, 9.1394e-12, 7.3482e-12,\n 8.3147e-14, 4.8147e-12, 7.2211e-12, 9.3820e-13, 3.0588e-11, 2.0069e-12,\n 9.9352e-13, 2.0543e-11, 6.5780e-13, 1.8757e-12, 3.7982e-11, 3.7374e-11,\n 8.0010e-12, 1.0514e-11, 2.5272e-13, 5.0792e-12, 6.2380e-12, 3.4852e-11,\n 2.1087e-13, 9.1659e-12, 1.9313e-12, 4.4167e-12, 1.6110e-12, 1.7188e-12,\n 4.6064e-12, 4.9407e-11, 1.4934e-11, 4.3673e-11, 9.2591e-12, 1.3173e-12,\n 1.3889e-12, 6.3481e-14, 4.2869e-12, 6.1442e-13, 3.9543e-13, 1.1185e-11,\n 3.2481e-12, 3.2904e-13, 6.5189e-11, 6.5589e-12, 1.8829e-12, 1.5589e-14,\n 1.1811e-13, 4.1768e-12, 7.3458e-12, 2.0223e-12, 7.0228e-11, 2.6837e-13,\n 6.5557e-11, 4.5347e-11, 7.8823e-12, 5.2131e-13, 1.7714e-12, 1.0643e-12,\n 1.1771e-13, 3.3235e-11, 3.8693e-13, 5.5069e-12, 4.3653e-12, 7.3605e-13,\n 4.8007e-12, 2.1745e-11, 5.6031e-12, 3.7666e-13, 2.3362e-11, 2.3881e-12,\n 1.8336e-12, 7.0165e-12, 6.3450e-13, 5.4254e-12, 6.7410e-13, 1.8422e-11,\n 5.0649e-11, 6.4280e-12, 5.7501e-12, 7.4937e-14, 1.3050e-12, 7.3845e-11,\n 5.4200e-13, 4.0525e-12, 1.5345e-12, 4.8567e-12, 5.3847e-14, 1.5432e-12,\n 7.2395e-11, 3.2841e-12, 1.4487e-13, 6.1313e-12, 3.8962e-12, 6.0092e-13,\n 2.0061e-11, 3.5635e-11, 6.4267e-13, 1.7842e-12, 2.5973e-11, 4.3436e-14,\n 9.3879e-13, 5.6760e-12, 1.8163e-12, 2.1373e-11, 3.8714e-13, 1.2511e-11,\n 1.1616e-12, 7.0578e-14, 1.8871e-11, 5.0627e-12, 1.2997e-12, 1.8382e-11,\n 6.6442e-12, 2.5253e-11, 3.6798e-14, 3.2494e-12, 1.4301e-13, 1.7434e-11,\n 1.5953e-14, 2.7917e-11, 3.7036e-12, 1.8556e-11, 1.8980e-12, 6.5761e-12,\n 1.3436e-13, 4.7498e-15, 1.5890e-12, 6.0242e-12, 5.4561e-12, 6.7826e-12,\n 2.5232e-12, 2.1332e-12, 3.0067e-11, 6.7749e-12, 5.6151e-14, 9.9788e-15,\n 6.4753e-13, 6.8810e-12, 8.1312e-14, 7.4658e-13, 3.8487e-13, 8.4649e-11,\n 8.0310e-12, 1.1514e-13, 1.5904e-12, 1.8941e-12, 2.6567e-11, 7.6002e-14,\n 1.1273e-11, 2.8730e-11, 1.8796e-11, 2.2258e-13, 2.6514e-13, 8.2934e-13,\n 8.2395e-12, 2.3246e-11, 1.1899e-11, 1.4033e-13, 5.1835e-11, 6.4631e-15,\n 3.4428e-12, 2.3249e-14, 1.9901e-12, 1.6953e-11, 4.7523e-13, 5.3551e-13,\n 3.1906e-11, 1.0027e-12, 1.2402e-11, 6.8319e-11, 2.2984e-10, 1.7022e-12,\n 2.8141e-13, 1.0538e-11, 5.2185e-12, 6.9709e-15, 4.0920e-12, 9.4407e-11,\n 1.3244e-12, 8.6558e-12, 7.8115e-12, 9.7746e-13, 5.4680e-13, 1.2969e-11,\n 1.6014e-11, 8.0782e-13, 1.5564e-11, 2.3548e-12, 7.4927e-12, 1.0030e-11,\n 8.9276e-12, 4.5397e-12, 2.6013e-12, 5.0583e-12, 1.5273e-12, 2.9417e-13,\n 2.6316e-11, 5.5092e-12, 1.7043e-13, 2.1895e-11, 2.2478e-12, 4.5088e-14,\n 2.0562e-11, 2.5016e-11, 6.2970e-11, 1.9427e-11, 6.0388e-13, 1.4976e-11,\n 1.3641e-12, 2.0007e-11, 2.3748e-13, 1.7140e-12], device='cuda:0')" }, - "24": { - "step": "tensor(23788.)", - "exp_avg": "tensor([-5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([4.6668e-19, 3.9428e-20, 4.4712e-20, 9.6753e-20, 1.1317e-20, 1.5280e-20,\n 3.9289e-20, 1.3149e-22, 8.0396e-21, 7.7128e-22, 4.4876e-19, 6.1208e-20,\n 6.9728e-20, 5.1891e-20, 1.7830e-20, 4.2777e-20, 1.5855e-19, 8.7623e-21,\n 2.7347e-20, 2.9515e-19, 6.6651e-24, 1.7294e-19, 6.7177e-20, 3.4979e-20,\n 7.1488e-21, 9.7185e-20, 3.7270e-20, 9.2762e-21, 7.2831e-21, 3.6512e-20,\n 2.7956e-19, 2.8066e-20, 4.6134e-21, 2.1158e-18, 1.2156e-19, 7.7913e-20,\n 1.5399e-20, 1.3500e-22, 1.3590e-19, 5.5298e-19, 6.0697e-20, 6.3918e-20,\n 7.7129e-21, 1.6085e-19, 3.3871e-19, 1.6557e-19, 2.1668e-19, 7.1216e-20,\n 1.1200e-20, 4.6341e-19, 1.1918e-21, 6.5828e-19, 3.7697e-19, 5.8694e-20,\n 2.9946e-21, 2.1134e-21, 2.7279e-19, 1.7089e-21, 1.0586e-19, 1.9078e-20,\n 4.0069e-20, 5.8969e-20, 4.6959e-19, 1.5879e-19, 1.8666e-21, 9.5765e-21,\n 1.0867e-19, 8.0149e-21, 1.1754e-19, 6.8352e-21, 2.3884e-21, 4.7339e-20,\n 8.7796e-20, 4.5184e-21, 3.2740e-20, 1.5297e-20, 2.0243e-20, 3.1296e-19,\n 2.8621e-21, 8.1328e-22, 2.3199e-20, 2.2519e-19, 2.1681e-22, 5.0662e-20,\n 8.2002e-21, 1.5667e-19, 1.1658e-19, 3.3267e-20, 4.6477e-20, 1.3338e-18,\n 1.3792e-20, 2.0207e-19, 3.7796e-19, 2.4210e-20, 1.3203e-18, 2.9404e-20,\n 2.3398e-20, 2.3133e-19, 1.0064e-19, 1.1605e-19, 4.2457e-19, 4.2548e-20,\n 5.4508e-21, 1.0523e-19, 6.2782e-22, 4.0157e-19, 2.3800e-20, 6.1872e-20,\n 1.1800e-20, 1.4993e-21, 3.8464e-19, 4.8019e-20, 1.0977e-18, 1.0944e-19,\n 5.0889e-21, 1.7161e-20, 3.2212e-20, 1.4997e-20, 2.6864e-20, 6.0632e-20,\n 1.5564e-19, 8.0655e-20, 6.0063e-21, 1.9141e-23, 7.5085e-22, 3.2707e-20,\n 2.3877e-19, 4.1792e-19, 1.2706e-20, 3.1812e-19, 9.4526e-23, 1.5137e-18,\n 5.3041e-20, 9.2143e-20, 1.0055e-19, 6.6108e-19, 1.3105e-19, 5.3264e-19,\n 1.1920e-19, 1.6980e-22, 6.7208e-21, 1.2569e-20, 3.7279e-20, 6.1760e-19,\n 7.2753e-20, 3.1658e-20, 2.5679e-20, 9.1285e-21, 6.8537e-19, 6.0335e-21,\n 1.4907e-20, 2.7537e-20, 5.2316e-20, 2.4538e-19, 4.0260e-20, 1.5950e-20,\n 2.5507e-20, 5.4271e-19, 1.1586e-20, 7.7125e-19, 1.3426e-19, 3.5726e-19,\n 1.3618e-20, 2.8630e-19, 3.9324e-19, 1.1921e-19, 1.7600e-18, 1.3689e-19,\n 3.4370e-19, 3.8317e-20, 2.4307e-21, 1.9791e-20, 1.7945e-19, 8.9350e-20,\n 1.0078e-19, 1.9265e-21, 5.5343e-21, 7.6284e-21, 1.1616e-20, 3.9714e-20,\n 6.1215e-21, 4.3468e-20, 2.0773e-20, 2.8245e-20, 2.7173e-19, 2.7701e-24,\n 2.3275e-20, 9.5717e-20, 9.9710e-21, 5.5408e-19, 2.6877e-20, 7.9337e-20,\n 1.5221e-19, 9.4811e-21, 2.7573e-22, 2.3043e-18, 1.2693e-20, 1.1124e-20,\n 1.7084e-19, 1.6062e-19, 7.5732e-22, 5.3830e-20, 4.1079e-19, 1.6886e-21,\n 2.2151e-19, 4.5157e-19, 6.3166e-20, 1.2252e-20, 9.4722e-24, 1.0466e-18,\n 1.4113e-20, 2.5273e-19, 1.6443e-20, 3.3940e-20, 3.0506e-21, 6.0114e-20,\n 7.7444e-20, 5.7805e-23, 9.6651e-21, 4.9789e-21, 3.3107e-20, 1.5246e-19,\n 4.4648e-20, 6.0228e-19, 1.1466e-19, 5.2755e-21, 1.8966e-22, 2.1108e-20,\n 1.1822e-20, 4.9290e-20, 2.7026e-21, 2.9723e-21, 5.1197e-20, 9.7257e-20,\n 2.4502e-19, 9.9361e-20, 1.1152e-20, 1.1911e-19, 2.4983e-20, 2.1350e-20,\n 1.3220e-19, 4.5856e-23, 5.7552e-19, 2.0615e-19, 4.8021e-20, 4.5418e-23,\n 8.2719e-22, 1.3730e-18, 1.4842e-21, 3.3137e-19, 7.0012e-20, 9.4541e-21,\n 1.1034e-19, 1.0450e-19, 2.5513e-19, 1.2374e-20], device='cuda:0')" + "42": { + "step": "tensor(11268.)", + "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([2.1234e-17, 3.0068e-17, 4.5547e-14, 1.2846e-16, 1.0170e-14, 1.4046e-14,\n 7.2674e-15, 9.0943e-16, 1.5901e-13, 5.3391e-15, 3.2609e-14, 2.9668e-14,\n 1.6347e-16, 2.4224e-15, 1.1064e-16, 1.8819e-16, 1.5626e-14, 1.9387e-14,\n 2.5918e-15, 8.3964e-16, 3.4624e-15, 3.8523e-16, 2.4222e-14, 9.0565e-15,\n 4.2178e-14, 2.4697e-14, 1.0243e-14, 2.2203e-14, 1.1149e-14, 6.7998e-14,\n 8.0174e-14, 3.6646e-15, 3.5007e-13, 1.7042e-15, 3.0545e-15, 2.4772e-16,\n 1.0038e-14, 7.1733e-16, 5.6685e-15, 3.1607e-15, 1.8618e-14, 1.6992e-15,\n 4.4965e-17, 2.8750e-13, 8.8109e-17, 6.8234e-15, 2.5131e-14, 2.9089e-14,\n 2.5332e-16, 7.6496e-15, 1.4032e-14, 1.1750e-15, 5.3433e-14, 3.2346e-15,\n 2.1255e-15, 3.1913e-14, 2.3615e-15, 2.5333e-15, 8.9998e-14, 4.9623e-14,\n 1.1758e-14, 2.8763e-14, 6.7849e-16, 8.0920e-15, 1.5989e-14, 7.6243e-14,\n 6.2181e-16, 1.3499e-14, 3.5758e-15, 4.9398e-15, 2.6293e-15, 3.7866e-15,\n 7.5045e-15, 1.1953e-13, 2.0682e-14, 9.4545e-14, 1.8411e-14, 2.3148e-15,\n 2.7377e-15, 6.7673e-17, 5.2492e-15, 9.2049e-16, 1.2092e-15, 4.4016e-14,\n 4.2216e-15, 1.7892e-15, 1.7341e-13, 2.2926e-14, 4.7421e-15, 9.3584e-17,\n 5.7202e-16, 6.8206e-15, 9.5604e-15, 3.2373e-15, 1.8716e-13, 6.3576e-16,\n 1.6735e-13, 1.1609e-13, 1.4724e-14, 2.0721e-15, 3.4265e-15, 2.2009e-15,\n 3.6207e-16, 6.1844e-14, 1.3836e-15, 7.6293e-15, 5.1177e-15, 1.8950e-15,\n 4.8253e-15, 5.0249e-14, 1.2533e-14, 5.0184e-16, 3.5774e-14, 3.0058e-15,\n 4.6077e-15, 1.1764e-14, 2.3317e-15, 6.0778e-15, 2.5168e-15, 2.5979e-14,\n 9.6308e-14, 1.9065e-14, 6.3542e-15, 1.7208e-16, 3.9849e-15, 1.5007e-13,\n 1.0572e-15, 1.1241e-14, 3.5203e-15, 8.5509e-15, 3.0835e-16, 2.0624e-15,\n 1.3714e-13, 5.3234e-15, 3.7196e-16, 1.2431e-14, 1.0508e-14, 1.3320e-15,\n 3.2183e-14, 5.6518e-14, 1.2912e-15, 2.6943e-15, 4.9650e-14, 2.9003e-16,\n 2.4412e-15, 1.1083e-14, 2.5440e-15, 6.9374e-14, 1.6290e-15, 4.1335e-14,\n 2.2787e-15, 4.3718e-17, 3.8758e-14, 9.6424e-15, 2.4291e-15, 1.9077e-14,\n 1.0499e-14, 3.2375e-14, 5.4211e-17, 6.8756e-15, 1.6884e-15, 3.7948e-14,\n 8.9368e-17, 9.0157e-14, 6.4611e-15, 3.1781e-14, 4.5999e-15, 1.4288e-14,\n 3.0808e-16, 1.3830e-17, 2.2212e-15, 1.4977e-14, 6.8688e-15, 9.4435e-15,\n 5.6448e-15, 4.7166e-15, 6.6270e-14, 1.2626e-14, 1.9852e-16, 2.3864e-17,\n 1.0611e-15, 1.8254e-14, 7.5796e-17, 1.4464e-15, 9.6780e-16, 1.4659e-13,\n 1.0196e-14, 6.2784e-16, 7.0977e-15, 5.0403e-15, 8.1586e-14, 5.6325e-16,\n 1.6887e-14, 6.2074e-14, 3.2935e-14, 6.6749e-16, 2.5664e-16, 1.5828e-15,\n 1.2935e-14, 5.5142e-14, 2.7901e-14, 7.8857e-16, 8.8479e-14, 1.5010e-16,\n 3.7997e-15, 4.3977e-17, 2.5374e-15, 3.5998e-14, 1.9475e-15, 8.3504e-16,\n 9.6844e-14, 1.0437e-15, 1.6487e-14, 1.3950e-13, 6.6441e-13, 2.5022e-15,\n 4.9461e-16, 1.8729e-14, 7.5915e-15, 6.3076e-18, 4.0583e-15, 1.6902e-13,\n 3.4789e-15, 2.7056e-14, 1.4124e-14, 1.5937e-15, 1.4268e-15, 2.7118e-14,\n 2.7697e-14, 1.2526e-15, 1.9774e-14, 4.8577e-15, 1.5566e-14, 2.0737e-14,\n 1.4284e-14, 8.3601e-15, 4.2393e-15, 9.9480e-15, 2.3474e-15, 5.8946e-16,\n 5.4205e-14, 1.3872e-14, 1.6619e-15, 3.0417e-14, 2.3237e-15, 2.9859e-16,\n 5.5991e-14, 6.0185e-14, 1.4853e-13, 3.1719e-14, 9.7684e-16, 3.2288e-14,\n 2.0587e-15, 4.5507e-14, 6.1489e-16, 7.2878e-15], device='cuda:0')" }, - "25": { - "step": "tensor(23788.)", - "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([6.7737e-19, 5.9724e-20, 8.0585e-20, 1.0930e-19, 1.6957e-20, 2.4275e-20,\n 4.9148e-20, 1.5435e-22, 1.4542e-20, 1.3443e-21, 4.0143e-19, 9.0296e-20,\n 1.0087e-19, 1.0078e-19, 3.2260e-20, 7.8561e-20, 2.8702e-19, 1.2200e-20,\n 3.9894e-20, 5.7378e-19, 1.0466e-22, 2.2973e-19, 1.0939e-19, 5.9293e-20,\n 9.0008e-21, 1.7260e-19, 7.9203e-20, 1.6294e-20, 1.3037e-20, 4.3781e-20,\n 3.3095e-19, 6.7683e-20, 1.1553e-20, 1.8838e-18, 2.3091e-19, 1.6098e-19,\n 1.1978e-20, 1.0208e-22, 2.2155e-19, 8.7509e-19, 1.0963e-19, 9.2345e-20,\n 1.9229e-20, 1.0693e-19, 3.7107e-19, 2.9436e-19, 2.1882e-19, 1.3451e-19,\n 1.8184e-20, 3.8959e-19, 1.2066e-21, 8.9242e-19, 4.9211e-19, 1.0992e-19,\n 6.3225e-21, 5.1515e-21, 3.7064e-19, 4.7286e-21, 1.6303e-19, 4.2643e-20,\n 4.6011e-20, 9.6065e-20, 4.6812e-19, 2.5118e-19, 3.2396e-21, 1.8171e-20,\n 1.8114e-19, 1.6904e-20, 2.4250e-19, 8.4425e-21, 1.9735e-21, 1.1169e-19,\n 1.3845e-19, 4.7710e-21, 5.2246e-20, 1.7252e-20, 3.7179e-20, 5.0343e-19,\n 4.6945e-21, 7.2086e-22, 5.1666e-20, 3.8341e-19, 5.7871e-22, 4.2896e-20,\n 1.0756e-20, 2.7419e-19, 1.3027e-19, 6.5720e-20, 7.9891e-20, 1.5096e-18,\n 2.4856e-20, 3.6703e-19, 3.4398e-19, 5.0787e-20, 8.7835e-19, 4.7210e-20,\n 5.7274e-20, 4.0694e-19, 1.7245e-19, 2.1093e-19, 4.3140e-19, 3.7918e-20,\n 1.0331e-20, 1.1772e-19, 1.8191e-21, 7.3837e-19, 2.9765e-20, 1.1930e-19,\n 2.1598e-20, 2.6754e-21, 6.0685e-19, 6.4988e-20, 7.4911e-19, 1.8459e-19,\n 1.5486e-20, 3.1687e-20, 5.9872e-20, 2.9897e-20, 3.6380e-20, 9.3625e-20,\n 1.6027e-19, 9.6560e-20, 1.2088e-20, 1.2533e-23, 7.9458e-22, 5.9614e-20,\n 4.6261e-19, 3.2015e-19, 2.8792e-20, 2.8792e-19, 1.2545e-22, 1.8944e-18,\n 9.6329e-20, 1.6972e-19, 2.1612e-19, 6.0967e-19, 2.5716e-19, 6.2870e-19,\n 2.1545e-19, 3.6953e-22, 8.5243e-21, 2.2912e-20, 6.8388e-20, 6.6572e-19,\n 1.4637e-19, 5.0463e-20, 4.1579e-20, 1.6780e-20, 6.6661e-19, 6.8223e-21,\n 2.4245e-20, 5.4413e-20, 8.6913e-20, 2.1158e-19, 7.0214e-20, 2.3224e-20,\n 4.6237e-20, 6.1869e-19, 1.5373e-20, 1.0251e-18, 2.1240e-19, 2.7982e-19,\n 2.2971e-20, 3.7772e-19, 3.0016e-19, 1.6740e-19, 1.7919e-18, 2.4523e-19,\n 3.3784e-19, 5.0219e-20, 6.2694e-21, 3.9223e-20, 3.2980e-19, 1.3748e-19,\n 1.7749e-19, 4.5224e-21, 1.0694e-20, 1.5611e-20, 1.5036e-20, 1.0523e-19,\n 8.9793e-21, 4.6147e-20, 3.3617e-20, 5.0900e-20, 5.3011e-19, 4.6659e-23,\n 4.1739e-20, 1.1662e-19, 1.9843e-20, 8.4757e-19, 7.2225e-20, 8.9337e-20,\n 1.5757e-19, 1.5064e-20, 7.2630e-22, 2.3410e-18, 1.6546e-20, 1.9057e-20,\n 1.9958e-19, 3.8352e-19, 1.8323e-21, 1.4094e-19, 3.7832e-19, 3.1916e-21,\n 2.9927e-19, 8.4455e-19, 1.2470e-19, 2.0077e-20, 1.6801e-23, 1.1851e-18,\n 1.3154e-20, 4.7352e-19, 2.4476e-20, 7.5509e-20, 4.1212e-21, 1.0485e-19,\n 1.3788e-19, 5.6751e-22, 1.9155e-20, 9.2308e-21, 2.7675e-20, 2.3106e-19,\n 5.2256e-20, 7.8460e-19, 1.5618e-19, 1.0841e-20, 1.5045e-22, 1.8626e-20,\n 1.5697e-20, 8.0390e-20, 5.7766e-21, 3.2016e-21, 1.3433e-19, 1.2326e-19,\n 4.1752e-19, 1.4352e-19, 1.6534e-20, 2.2951e-19, 4.7232e-20, 4.0479e-20,\n 2.3085e-19, 1.0062e-22, 3.2042e-19, 2.8092e-19, 8.5816e-20, 3.4172e-22,\n 1.7108e-21, 9.9031e-19, 2.3858e-21, 4.3403e-19, 1.1210e-19, 1.4555e-20,\n 1.4900e-19, 1.5058e-19, 3.8048e-19, 2.2348e-20], device='cuda:0')" + "43": { + "step": "tensor(11268.)", + "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([7.3115e-17, 5.4462e-17, 7.5521e-14, 2.1279e-16, 1.5669e-14, 2.0602e-14,\n 1.2201e-14, 1.9916e-15, 2.4401e-13, 7.2189e-15, 5.7919e-14, 4.4626e-14,\n 1.6331e-16, 5.3547e-15, 1.6431e-16, 5.4033e-16, 2.4894e-14, 1.5102e-14,\n 2.3882e-15, 1.1468e-15, 4.9915e-15, 6.9343e-16, 2.4302e-14, 1.3124e-14,\n 3.4953e-14, 4.0768e-14, 1.4929e-14, 1.8953e-14, 1.0859e-14, 1.1394e-13,\n 1.0749e-13, 4.9827e-15, 4.8385e-13, 3.7130e-15, 6.2632e-15, 8.3331e-16,\n 1.6321e-14, 1.1423e-15, 1.2555e-14, 4.3924e-15, 3.4059e-14, 1.0641e-15,\n 2.0345e-16, 2.4134e-13, 9.5718e-17, 9.1456e-15, 2.3116e-14, 2.2571e-14,\n 4.5785e-16, 1.4209e-14, 2.2460e-14, 2.3802e-15, 7.9168e-14, 5.2026e-15,\n 3.3812e-15, 5.4728e-14, 2.6167e-15, 4.7797e-15, 9.7998e-14, 1.0755e-13,\n 1.9456e-14, 2.4476e-14, 1.1750e-15, 1.4712e-14, 1.8045e-14, 8.6582e-14,\n 7.7119e-16, 2.4474e-14, 5.7945e-15, 1.2996e-14, 4.8676e-15, 5.0789e-15,\n 9.6839e-15, 1.3596e-13, 4.3025e-14, 1.1113e-13, 2.6762e-14, 4.1902e-15,\n 4.7477e-15, 1.4556e-16, 1.1858e-14, 1.9019e-15, 1.6973e-15, 3.2214e-14,\n 7.9529e-15, 1.6008e-15, 1.7996e-13, 2.0660e-14, 6.6256e-15, 1.3284e-16,\n 6.0693e-16, 1.0772e-14, 1.9320e-14, 6.0839e-15, 1.7616e-13, 6.9310e-16,\n 1.6628e-13, 1.1510e-13, 1.7820e-14, 3.3223e-15, 5.3499e-15, 3.8657e-15,\n 3.4383e-16, 8.6125e-14, 1.1522e-15, 1.4429e-14, 1.2351e-14, 2.8312e-15,\n 1.2383e-14, 5.3344e-14, 1.6396e-14, 1.2877e-15, 6.2289e-14, 5.3684e-15,\n 5.9515e-15, 2.0442e-14, 2.4870e-15, 1.3704e-14, 3.3680e-15, 5.2845e-14,\n 1.3833e-13, 2.0152e-14, 1.6598e-14, 1.2038e-15, 4.7986e-15, 1.8841e-13,\n 1.4681e-15, 1.2612e-14, 4.6593e-15, 1.4016e-14, 4.7119e-16, 5.0424e-15,\n 1.8606e-13, 8.8661e-15, 6.4000e-16, 1.6090e-14, 1.2474e-14, 1.4845e-15,\n 5.0322e-14, 8.9860e-14, 2.4147e-15, 5.6877e-15, 6.4036e-14, 5.3083e-16,\n 3.2841e-15, 1.6618e-14, 4.2403e-15, 5.5542e-14, 1.4860e-15, 3.6323e-14,\n 3.8305e-15, 8.5756e-17, 4.8387e-14, 1.1995e-14, 4.4635e-15, 4.7759e-14,\n 1.6141e-14, 6.1128e-14, 2.0181e-16, 9.5804e-15, 2.3039e-15, 5.0961e-14,\n 1.7736e-16, 7.8483e-14, 1.0608e-14, 4.7326e-14, 4.7014e-15, 1.6805e-14,\n 5.0852e-16, 3.8005e-17, 4.1446e-15, 1.8409e-14, 1.5301e-14, 2.0099e-14,\n 7.1505e-15, 6.2556e-15, 8.4551e-14, 1.8859e-14, 7.8269e-16, 5.1650e-17,\n 2.4030e-15, 1.8709e-14, 3.1602e-16, 2.6198e-15, 1.5672e-15, 2.3303e-13,\n 1.9446e-14, 5.8033e-16, 5.8495e-15, 6.5993e-15, 6.5526e-14, 1.0154e-15,\n 2.8395e-14, 7.6541e-14, 4.5763e-14, 6.2262e-16, 9.8517e-16, 2.8472e-15,\n 2.1993e-14, 6.6592e-14, 3.0075e-14, 1.2854e-15, 1.3520e-13, 3.2333e-16,\n 8.9223e-15, 9.5781e-17, 6.3468e-15, 4.1802e-14, 2.4001e-15, 1.7307e-15,\n 9.0554e-14, 2.3077e-15, 3.5731e-14, 1.8333e-13, 6.2338e-13, 4.5761e-15,\n 9.3240e-16, 2.4866e-14, 1.4668e-14, 3.3469e-17, 9.4090e-15, 2.5857e-13,\n 4.4725e-15, 2.5716e-14, 2.3847e-14, 2.1817e-15, 1.9333e-15, 3.6091e-14,\n 4.4652e-14, 2.6046e-15, 4.2446e-14, 7.4593e-15, 1.7757e-14, 2.9745e-14,\n 2.5595e-14, 1.4377e-14, 6.8978e-15, 1.5954e-14, 3.9608e-15, 9.2303e-16,\n 6.8074e-14, 1.6634e-14, 2.2040e-15, 6.0434e-14, 6.0744e-15, 7.1273e-16,\n 5.6861e-14, 6.3539e-14, 1.7325e-13, 5.7116e-14, 1.8168e-15, 3.6252e-14,\n 3.2442e-15, 5.8728e-14, 1.1851e-15, 5.9057e-15], device='cuda:0')" + }, + "44": { + "step": "tensor(11268.)", + "exp_avg": "tensor([[ 5.6052e-45, -5.6052e-45, -5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, -5.6052e-45, ..., -5.6052e-45,\n 0.0000e+00, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n ...,\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n [-5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45]], device='cuda:0')", + "exp_avg_sq": "tensor([[1.8837e-18, 3.1945e-18, 1.2736e-17, ..., 2.2859e-18, 0.0000e+00,\n 5.0502e-17],\n [8.1939e-18, 1.0124e-18, 9.2588e-17, ..., 6.0307e-17, 0.0000e+00,\n 2.7694e-16],\n [1.5964e-14, 1.9042e-15, 1.2136e-13, ..., 7.5678e-14, 0.0000e+00,\n 2.0002e-13],\n ...,\n [4.2555e-15, 5.1611e-16, 2.9096e-14, ..., 1.3738e-14, 0.0000e+00,\n 1.4968e-14],\n [1.8505e-17, 1.1228e-17, 4.1032e-17, ..., 3.0705e-16, 0.0000e+00,\n 1.8519e-16],\n [4.0727e-18, 1.7285e-18, 1.5401e-16, ..., 1.1582e-16, 0.0000e+00,\n 7.8799e-18]], device='cuda:0')" + }, + "45": { + "step": "tensor(11268.)", + "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([7.9241e-15, 1.9546e-14, 9.9506e-11, 3.2360e-15, 3.3004e-11, 2.0608e-11,\n 4.9018e-12, 2.8881e-12, 1.0053e-10, 2.0831e-12, 2.6596e-12, 9.4841e-12,\n 1.1493e-13, 2.1525e-12, 4.6033e-13, 8.3460e-13, 7.5251e-12, 1.8510e-12,\n 2.5899e-13, 8.2761e-14, 7.0879e-12, 3.4270e-13, 9.5433e-12, 1.0143e-12,\n 8.6997e-12, 3.1041e-11, 3.7945e-12, 3.8995e-12, 1.3913e-12, 3.1169e-11,\n 1.0843e-12, 4.2025e-12, 1.6364e-10, 1.9875e-12, 1.0694e-12, 6.1327e-14,\n 1.0731e-11, 6.2050e-14, 6.6116e-12, 2.1062e-12, 5.0076e-12, 1.8663e-13,\n 5.5824e-14, 1.5522e-11, 5.0485e-15, 5.7787e-12, 2.5554e-12, 3.4328e-12,\n 6.5303e-13, 2.4833e-12, 2.0885e-11, 1.3242e-12, 8.1560e-12, 1.5130e-12,\n 8.4114e-13, 4.1184e-11, 2.4309e-13, 3.4977e-12, 4.3590e-12, 7.4955e-11,\n 7.7367e-12, 7.3242e-12, 3.8915e-13, 1.1700e-11, 6.8184e-12, 2.6008e-11,\n 7.5336e-14, 1.4159e-11, 1.0286e-12, 3.4379e-12, 3.5084e-12, 1.0309e-11,\n 4.2182e-13, 2.7717e-11, 2.9710e-11, 2.4216e-11, 1.6982e-12, 1.7866e-12,\n 5.1511e-13, 1.6160e-13, 7.1383e-12, 5.0577e-13, 7.8483e-13, 5.6390e-13,\n 7.1062e-12, 1.8075e-14, 2.1990e-11, 3.1039e-12, 1.6819e-12, 6.6563e-15,\n 4.5052e-14, 3.8531e-12, 1.1434e-11, 7.7605e-12, 2.2687e-11, 1.2984e-12,\n 3.6085e-11, 2.3900e-11, 1.1319e-12, 1.2441e-12, 4.0152e-12, 1.4176e-12,\n 2.5628e-12, 3.7252e-11, 5.4500e-12, 5.0948e-12, 8.0100e-12, 5.5443e-13,\n 5.5716e-12, 8.3933e-13, 2.1454e-13, 1.3549e-12, 1.3561e-12, 3.3914e-12,\n 5.5541e-13, 7.5377e-12, 8.9061e-13, 4.0688e-11, 3.8220e-13, 2.3584e-11,\n 4.3651e-11, 3.4576e-12, 3.0938e-11, 1.8135e-13, 1.5402e-12, 5.6775e-11,\n 3.0904e-12, 1.6233e-12, 1.6523e-12, 5.0673e-12, 1.4036e-13, 2.4254e-12,\n 1.2834e-10, 1.1369e-11, 1.7652e-13, 7.6975e-12, 4.5241e-12, 4.8111e-12,\n 4.1451e-11, 1.7598e-11, 3.5970e-12, 7.9793e-13, 1.1234e-11, 6.5868e-14,\n 4.5539e-14, 8.1656e-12, 5.8563e-12, 6.9893e-12, 1.1226e-13, 1.7867e-12,\n 2.3618e-12, 1.2088e-15, 9.1252e-12, 2.2106e-12, 4.2176e-13, 4.6283e-11,\n 2.7702e-12, 1.1545e-11, 3.2851e-14, 1.3724e-12, 3.1507e-13, 2.8098e-11,\n 5.2491e-14, 1.3954e-11, 4.8732e-11, 8.8889e-12, 1.4111e-12, 7.4843e-12,\n 1.1856e-13, 2.5821e-15, 6.4829e-12, 1.4943e-11, 3.6207e-11, 4.1576e-12,\n 5.0157e-12, 9.4214e-14, 1.3020e-11, 3.1314e-12, 2.6250e-13, 3.1764e-13,\n 7.7968e-13, 7.8124e-12, 3.4825e-14, 2.4092e-13, 1.0801e-12, 1.4236e-10,\n 2.2420e-12, 3.0886e-14, 3.5767e-13, 1.1357e-12, 2.2903e-11, 9.6892e-14,\n 1.6910e-11, 5.4298e-12, 2.9648e-12, 3.4732e-12, 2.9936e-13, 8.6075e-12,\n 6.0521e-12, 9.8921e-12, 8.9928e-12, 4.4497e-14, 4.5309e-11, 1.4277e-14,\n 5.8528e-12, 2.6474e-14, 3.1374e-12, 1.3721e-11, 3.4838e-13, 2.7369e-12,\n 9.4356e-12, 3.3264e-12, 3.1315e-11, 4.5349e-11, 8.1510e-11, 1.1565e-11,\n 1.4909e-12, 1.9397e-12, 5.7276e-11, 3.0898e-13, 5.5186e-12, 7.9992e-11,\n 4.4689e-13, 6.1861e-12, 7.3624e-11, 1.1372e-12, 1.6538e-12, 7.7097e-12,\n 1.2270e-11, 4.2093e-12, 4.7402e-12, 3.5530e-12, 7.1923e-12, 2.6532e-12,\n 8.0408e-12, 1.0540e-11, 1.1985e-12, 8.9123e-13, 5.4950e-12, 1.6481e-13,\n 5.0216e-11, 3.5187e-12, 2.1703e-13, 2.6493e-11, 8.1593e-12, 3.7793e-14,\n 9.8143e-12, 3.9687e-12, 8.8609e-11, 6.2742e-11, 1.3552e-12, 3.8646e-12,\n 1.5629e-12, 2.4470e-11, 4.9858e-13, 1.3616e-13], device='cuda:0')" + }, + "46": { + "step": "tensor(11268.)", + "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([3.2221e-17, 2.3780e-16, 2.2105e-13, 4.4502e-18, 5.7181e-14, 3.8518e-14,\n 8.5349e-15, 9.1733e-15, 2.1897e-13, 5.0435e-15, 5.1417e-15, 1.6252e-14,\n 2.1843e-16, 3.5596e-15, 1.0637e-15, 3.0561e-15, 1.4958e-14, 3.8395e-15,\n 6.3977e-16, 4.3558e-16, 1.0857e-14, 1.7651e-15, 2.7146e-14, 1.6526e-15,\n 2.9408e-14, 3.7856e-14, 7.7684e-15, 7.3837e-15, 7.4168e-15, 6.0152e-14,\n 3.4436e-15, 8.1554e-15, 3.3546e-13, 2.7615e-15, 1.6998e-15, 2.9954e-16,\n 1.7507e-14, 1.1380e-15, 9.9424e-15, 5.3986e-15, 7.2674e-15, 2.8193e-16,\n 1.3108e-16, 2.9988e-14, 9.7396e-18, 1.2368e-14, 3.4146e-15, 1.0136e-14,\n 2.1498e-15, 3.3743e-15, 3.7161e-14, 1.6617e-15, 9.6183e-15, 2.3121e-15,\n 4.3459e-15, 1.1476e-13, 7.9065e-16, 5.3319e-15, 7.0912e-15, 1.5180e-13,\n 1.0942e-14, 1.1974e-14, 1.2795e-15, 1.2299e-14, 1.3653e-14, 4.1547e-14,\n 1.6777e-16, 2.3135e-14, 1.5251e-15, 5.0312e-15, 6.3649e-15, 2.5431e-14,\n 4.9046e-16, 4.7807e-14, 7.5061e-14, 3.6039e-14, 2.3256e-15, 2.3971e-15,\n 6.7792e-16, 1.1010e-16, 1.0003e-14, 6.9985e-16, 1.5180e-15, 8.1027e-16,\n 1.8185e-14, 1.3120e-16, 3.0670e-14, 1.3191e-14, 3.7638e-15, 1.7384e-17,\n 6.3621e-18, 7.0876e-15, 2.0757e-14, 1.1330e-14, 2.7984e-14, 1.2147e-15,\n 6.7419e-14, 3.7497e-14, 2.2319e-15, 4.8152e-15, 8.1164e-15, 2.7867e-15,\n 5.9513e-15, 8.7768e-14, 1.0010e-14, 6.8157e-15, 1.1766e-14, 1.1253e-15,\n 5.0218e-15, 2.0101e-15, 6.8028e-16, 5.4253e-15, 4.8569e-15, 4.5639e-15,\n 6.9823e-16, 2.0864e-14, 2.5707e-15, 1.3948e-13, 1.5823e-15, 5.5167e-14,\n 7.8781e-14, 7.5435e-15, 7.3501e-14, 2.0559e-15, 4.5517e-15, 1.2382e-13,\n 2.7912e-15, 1.8520e-15, 3.6641e-15, 9.9568e-15, 4.3837e-16, 3.8521e-15,\n 4.6215e-13, 1.2873e-14, 4.6035e-16, 1.7221e-14, 1.6966e-14, 9.8406e-15,\n 1.0068e-13, 1.8308e-14, 7.3890e-15, 1.9142e-15, 1.8010e-14, 2.1470e-16,\n 1.5325e-16, 2.7532e-14, 9.3528e-15, 1.0999e-14, 3.1553e-16, 2.6652e-15,\n 7.8695e-15, 1.5046e-17, 1.1709e-14, 3.4353e-15, 6.5927e-16, 6.6015e-14,\n 3.7573e-15, 1.5267e-14, 3.2925e-17, 2.5951e-15, 6.9435e-16, 5.8913e-14,\n 5.5021e-16, 3.1213e-14, 1.1712e-13, 1.1651e-14, 3.0052e-15, 1.6543e-14,\n 4.9977e-16, 9.0270e-17, 1.2791e-14, 6.5730e-14, 8.1586e-14, 5.7642e-15,\n 1.3801e-14, 2.5489e-16, 1.7707e-14, 4.8065e-15, 1.7173e-15, 1.8022e-15,\n 1.9385e-15, 1.9680e-14, 1.1710e-16, 6.0444e-16, 4.8995e-15, 3.9662e-13,\n 2.7069e-15, 1.0599e-16, 5.5688e-16, 3.7629e-15, 6.9629e-14, 4.1602e-16,\n 2.6357e-14, 7.7124e-15, 6.5353e-15, 8.6329e-15, 4.4826e-16, 1.2106e-14,\n 8.2894e-15, 1.3775e-14, 1.7414e-14, 2.2226e-16, 5.2368e-14, 9.3278e-17,\n 9.4145e-15, 8.0152e-17, 4.6441e-15, 2.3519e-14, 2.5581e-15, 5.9930e-15,\n 1.4540e-14, 4.3575e-15, 6.9603e-14, 6.9764e-14, 1.2537e-13, 3.0654e-14,\n 3.4575e-15, 2.6722e-15, 1.1362e-13, 1.1116e-15, 8.0064e-15, 1.9330e-13,\n 1.0896e-15, 1.5530e-14, 2.0358e-13, 1.9662e-15, 7.9374e-15, 9.0728e-15,\n 2.0246e-14, 8.4919e-15, 8.6440e-15, 4.9153e-15, 1.8549e-14, 4.6651e-15,\n 1.1298e-14, 2.1960e-14, 1.8951e-15, 1.4055e-15, 1.0752e-14, 4.4425e-16,\n 1.6073e-13, 5.8103e-15, 9.0563e-16, 3.8933e-14, 8.9763e-15, 9.2670e-17,\n 1.8807e-14, 4.6141e-15, 2.2445e-13, 2.5750e-13, 2.0312e-15, 4.4503e-15,\n 2.6517e-15, 6.0416e-14, 1.2384e-15, 2.6802e-16], device='cuda:0')" + }, + "47": { + "step": "tensor(11268.)", + "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([8.1858e-17, 1.6639e-16, 2.7344e-13, 4.0303e-17, 8.0904e-14, 5.0434e-14,\n 1.4054e-14, 9.6806e-15, 2.6930e-13, 6.0823e-15, 8.4453e-15, 2.3687e-14,\n 4.9485e-16, 7.1055e-15, 2.0272e-15, 2.6819e-15, 2.1727e-14, 6.0245e-15,\n 9.6759e-16, 5.8194e-16, 2.0680e-14, 1.7442e-15, 2.4594e-14, 3.4882e-15,\n 2.7245e-14, 7.8035e-14, 1.1899e-14, 1.0575e-14, 6.1116e-15, 8.8312e-14,\n 3.1024e-15, 1.1024e-14, 4.4564e-13, 4.7860e-15, 3.3271e-15, 4.3636e-16,\n 2.4975e-14, 1.1739e-15, 2.0284e-14, 6.8328e-15, 1.2723e-14, 7.8369e-16,\n 2.3770e-16, 4.7221e-14, 2.5042e-17, 1.2960e-14, 7.3817e-15, 1.2047e-14,\n 2.4212e-15, 7.7067e-15, 6.3008e-14, 2.7911e-15, 2.0380e-14, 4.1406e-15,\n 3.5066e-15, 1.0424e-13, 1.1721e-15, 8.0569e-15, 1.1376e-14, 2.1048e-13,\n 1.7826e-14, 1.6967e-14, 1.5520e-15, 3.3078e-14, 1.9758e-14, 6.1289e-14,\n 3.7028e-16, 3.4984e-14, 3.1772e-15, 1.0662e-14, 1.1249e-14, 3.1870e-14,\n 8.9313e-16, 7.8399e-14, 8.4882e-14, 6.2899e-14, 5.4898e-15, 5.8575e-15,\n 1.7541e-15, 1.7928e-16, 2.1756e-14, 1.4455e-15, 2.4436e-15, 1.8492e-15,\n 1.6146e-14, 3.9640e-16, 6.4260e-14, 1.1268e-14, 6.0600e-15, 2.4191e-17,\n 6.2853e-17, 9.8151e-15, 2.8063e-14, 2.0515e-14, 5.8271e-14, 2.9494e-15,\n 9.3787e-14, 6.0913e-14, 3.0223e-15, 5.1438e-15, 1.2453e-14, 4.5029e-15,\n 8.5421e-15, 9.5424e-14, 1.6047e-14, 1.2753e-14, 2.3025e-14, 2.0861e-15,\n 1.2557e-14, 2.4618e-15, 6.7453e-16, 5.1971e-15, 3.7260e-15, 7.7115e-15,\n 1.9785e-15, 2.1205e-14, 3.5544e-15, 9.9211e-14, 1.9108e-15, 6.7231e-14,\n 1.2303e-13, 1.1390e-14, 7.9602e-14, 3.1206e-15, 5.6612e-15, 1.5170e-13,\n 7.2157e-15, 4.9385e-15, 5.4610e-15, 1.5062e-14, 7.3519e-16, 7.7652e-15,\n 3.2715e-13, 2.8498e-14, 7.3461e-16, 1.8604e-14, 1.3507e-14, 1.0996e-14,\n 1.0254e-13, 4.8244e-14, 1.0600e-14, 2.4725e-15, 2.7765e-14, 4.6268e-16,\n 2.1681e-16, 2.4749e-14, 1.2235e-14, 1.9781e-14, 6.9202e-16, 5.8674e-15,\n 8.1988e-15, 5.8142e-19, 2.3760e-14, 5.3514e-15, 1.3173e-15, 1.1447e-13,\n 7.3570e-15, 2.9697e-14, 4.4526e-17, 4.8173e-15, 1.3454e-15, 7.8570e-14,\n 6.9072e-16, 3.9804e-14, 1.3756e-13, 2.2732e-14, 3.3569e-15, 1.8760e-14,\n 7.5448e-16, 1.0051e-16, 1.5081e-14, 4.4393e-14, 8.9906e-14, 1.2726e-14,\n 1.2458e-14, 3.7392e-16, 3.7139e-14, 9.5609e-15, 1.7711e-15, 1.5703e-15,\n 3.1307e-15, 2.0974e-14, 1.5723e-16, 9.6011e-16, 4.3199e-15, 3.8847e-13,\n 5.5342e-15, 2.1953e-16, 1.2891e-15, 4.4739e-15, 5.7559e-14, 8.5342e-16,\n 4.2175e-14, 1.7165e-14, 7.6742e-15, 1.1292e-14, 1.0010e-15, 2.1668e-14,\n 1.5842e-14, 2.9142e-14, 2.2335e-14, 7.2067e-16, 1.2137e-13, 2.9658e-16,\n 1.3719e-14, 5.9972e-17, 9.7136e-15, 3.6628e-14, 2.1907e-15, 8.4198e-15,\n 2.6353e-14, 7.4431e-15, 8.7313e-14, 1.1890e-13, 2.2969e-13, 2.7633e-14,\n 5.4240e-15, 4.6003e-15, 1.4608e-13, 1.3887e-15, 1.2949e-14, 2.1346e-13,\n 1.6872e-15, 1.9151e-14, 2.0396e-13, 1.9906e-15, 6.0261e-15, 2.2228e-14,\n 3.5790e-14, 1.1408e-14, 1.4117e-14, 1.0162e-14, 1.6229e-14, 8.1271e-15,\n 2.2605e-14, 3.0292e-14, 3.4933e-15, 2.7743e-15, 1.2864e-14, 6.5563e-16,\n 1.2506e-13, 9.9274e-15, 1.6017e-15, 7.4896e-14, 1.9336e-14, 2.3273e-16,\n 2.6912e-14, 9.7578e-15, 2.4297e-13, 1.7713e-13, 4.4225e-15, 9.5167e-15,\n 3.5463e-15, 6.8024e-14, 1.8365e-15, 6.0661e-16], device='cuda:0')" + }, + "48": { + "step": "tensor(11268.)", + "exp_avg": "tensor([[ 5.6052e-45, -5.6052e-45, -5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n [-5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n ...,\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n [ 5.6052e-45, -5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45],\n [ 5.6052e-45, -5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 0.0000e+00, 5.6052e-45]], device='cuda:0')", + "exp_avg_sq": "tensor([[1.1352e-17, 2.0626e-18, 3.2239e-17, ..., 3.7569e-17, 0.0000e+00,\n 5.0946e-18],\n [2.2881e-18, 9.3955e-18, 1.1877e-16, ..., 1.6057e-16, 0.0000e+00,\n 2.8070e-16],\n [1.8980e-14, 3.7107e-16, 7.9353e-14, ..., 2.0159e-14, 0.0000e+00,\n 1.1383e-13],\n ...,\n [2.5054e-15, 1.0374e-16, 2.8864e-14, ..., 8.5446e-15, 0.0000e+00,\n 9.1856e-15],\n [3.1177e-16, 3.6359e-19, 9.4639e-18, ..., 3.9731e-17, 0.0000e+00,\n 4.0726e-17],\n [1.1707e-18, 6.9542e-19, 2.5897e-16, ..., 2.1996e-17, 0.0000e+00,\n 1.7356e-16]], device='cuda:0')" + }, + "49": { + "step": "tensor(11268.)", + "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([9.7074e-15, 8.4739e-14, 4.5459e-11, 4.7209e-14, 7.4147e-12, 2.3895e-11,\n 1.1208e-11, 2.5251e-12, 1.1931e-12, 2.4389e-12, 3.1478e-11, 9.2699e-12,\n 1.7307e-12, 6.1868e-12, 2.6861e-14, 1.7546e-12, 4.6655e-12, 2.2068e-12,\n 3.5258e-13, 7.9562e-13, 1.1407e-11, 2.2033e-13, 9.3112e-13, 3.5118e-12,\n 5.8817e-12, 3.0710e-12, 4.8675e-12, 9.9004e-12, 1.7665e-12, 1.7846e-11,\n 2.4188e-11, 2.8251e-11, 1.5341e-10, 4.8359e-13, 6.6895e-12, 1.3319e-13,\n 1.3697e-11, 2.2618e-13, 5.0532e-11, 1.4660e-12, 8.0680e-12, 1.8885e-13,\n 8.3608e-13, 3.2911e-11, 1.4984e-14, 1.6554e-12, 7.0559e-12, 4.8729e-12,\n 9.5765e-13, 9.2834e-12, 1.7130e-11, 5.7221e-12, 1.8691e-11, 1.5239e-12,\n 1.9374e-12, 4.2922e-11, 1.4852e-13, 6.5507e-12, 3.4238e-11, 4.2493e-11,\n 5.4331e-12, 4.8454e-12, 1.9277e-13, 3.3510e-11, 1.5998e-12, 4.4522e-11,\n 5.4091e-14, 1.1303e-11, 6.3632e-12, 6.3911e-12, 1.3135e-11, 2.1166e-11,\n 5.6491e-12, 1.5607e-11, 6.8964e-12, 2.6405e-11, 1.5133e-11, 5.9661e-12,\n 9.1863e-13, 3.5415e-14, 3.1450e-11, 2.2760e-12, 1.8144e-12, 1.2206e-11,\n 2.8732e-12, 4.2165e-13, 6.8155e-11, 2.0992e-12, 2.4031e-12, 1.0226e-13,\n 2.5338e-14, 2.5834e-12, 9.6755e-12, 2.1273e-12, 5.9385e-11, 1.8328e-12,\n 3.5148e-11, 4.8572e-11, 7.0016e-12, 4.2317e-12, 4.2796e-12, 1.8152e-12,\n 8.0959e-13, 1.4026e-11, 8.1052e-12, 5.7312e-12, 8.2805e-12, 6.3858e-13,\n 5.3826e-12, 6.4865e-12, 3.0654e-12, 7.6821e-13, 3.7586e-11, 9.4421e-13,\n 1.1766e-12, 3.7463e-11, 1.2125e-12, 7.7899e-12, 1.4163e-12, 2.3996e-12,\n 8.1095e-11, 6.7266e-12, 3.0222e-11, 5.3976e-13, 3.2111e-13, 9.4800e-11,\n 1.2372e-11, 6.4626e-13, 3.2550e-13, 1.3815e-11, 2.0089e-15, 2.8278e-12,\n 4.9681e-11, 3.1397e-11, 1.2766e-13, 3.1757e-12, 7.6783e-12, 2.5984e-12,\n 2.8098e-11, 2.0552e-11, 3.2856e-12, 2.9949e-12, 1.3961e-11, 2.1596e-14,\n 1.0695e-12, 1.6129e-12, 1.0349e-12, 1.7090e-11, 9.9508e-13, 1.6254e-12,\n 2.2522e-12, 2.9376e-12, 6.6404e-12, 5.0107e-12, 1.9254e-12, 2.2871e-11,\n 5.0864e-13, 7.4981e-11, 2.7005e-14, 1.3875e-12, 5.6270e-13, 1.9278e-12,\n 2.8955e-14, 3.5073e-12, 2.1674e-11, 1.4135e-11, 8.6892e-13, 4.0320e-12,\n 4.0693e-15, 7.0351e-13, 3.2681e-13, 6.0221e-12, 4.8332e-12, 5.9759e-12,\n 3.1472e-12, 7.8831e-13, 2.5789e-11, 1.5031e-11, 1.4833e-13, 3.7917e-14,\n 4.0953e-12, 1.1100e-11, 3.1086e-14, 2.3807e-12, 1.6592e-12, 3.2570e-11,\n 6.5686e-13, 1.6829e-13, 1.7680e-12, 7.6805e-13, 3.2768e-12, 3.8176e-14,\n 7.7533e-12, 2.5366e-12, 1.7931e-11, 2.1839e-12, 1.0009e-12, 7.2528e-12,\n 3.5241e-11, 3.4000e-12, 4.9903e-12, 1.0581e-13, 6.0200e-11, 1.3413e-14,\n 7.4542e-12, 3.0299e-14, 8.1999e-12, 8.1505e-12, 3.4362e-13, 5.0447e-12,\n 4.7551e-11, 2.8192e-12, 3.3529e-12, 4.0090e-11, 1.5893e-10, 2.0186e-12,\n 1.6566e-12, 1.6784e-12, 8.3836e-11, 9.1958e-14, 1.5563e-12, 1.5991e-10,\n 2.1606e-13, 1.2180e-11, 5.7288e-11, 9.9476e-13, 1.7140e-12, 2.5180e-11,\n 6.1308e-12, 1.8264e-12, 3.3422e-11, 7.2023e-12, 4.1228e-13, 2.6615e-12,\n 1.1448e-11, 1.1361e-11, 1.7014e-12, 2.1190e-12, 1.2374e-12, 7.7797e-13,\n 7.0346e-12, 5.2606e-12, 4.2974e-13, 7.2566e-11, 9.6976e-12, 8.9040e-14,\n 7.9561e-12, 3.2749e-11, 4.0929e-11, 3.5675e-11, 1.1271e-12, 3.8958e-12,\n 1.5570e-12, 1.4829e-11, 1.6097e-13, 4.1223e-14], device='cuda:0')" + }, + "50": { + "step": "tensor(11268.)", + "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([7.5161e-18, 3.2417e-16, 6.7381e-14, 7.1437e-16, 6.4176e-15, 5.4490e-14,\n 3.1147e-14, 5.5932e-15, 1.1664e-14, 4.2085e-15, 6.7246e-14, 1.6491e-14,\n 7.5282e-15, 1.1251e-14, 1.1000e-16, 6.2764e-15, 7.7828e-15, 4.3851e-15,\n 1.3557e-15, 1.4576e-15, 2.1276e-14, 3.5980e-16, 1.7946e-15, 4.9664e-15,\n 1.1145e-14, 6.3362e-15, 6.2196e-15, 4.3180e-14, 1.1520e-14, 2.1722e-14,\n 3.2215e-14, 6.2595e-14, 2.1653e-13, 5.7028e-16, 2.6025e-14, 1.8488e-16,\n 2.7411e-14, 1.6935e-15, 1.5950e-13, 2.0093e-15, 8.3066e-15, 5.4595e-16,\n 1.4960e-15, 4.9496e-14, 1.3414e-18, 1.9241e-15, 1.6870e-14, 1.0598e-14,\n 4.8612e-15, 1.5498e-14, 3.3978e-14, 1.0972e-14, 3.3345e-14, 1.9363e-15,\n 4.1192e-15, 9.9610e-14, 3.8548e-16, 1.8762e-14, 4.9160e-14, 5.4923e-14,\n 5.8179e-15, 7.0227e-15, 3.1932e-16, 6.2177e-14, 2.4993e-15, 1.2984e-13,\n 1.5160e-16, 1.4347e-14, 1.6139e-14, 7.5905e-15, 3.1037e-14, 4.9519e-14,\n 1.3862e-14, 2.0974e-14, 9.0383e-15, 4.0017e-14, 3.1495e-14, 8.6095e-15,\n 1.2842e-15, 1.0717e-16, 5.2886e-14, 6.5543e-15, 4.8536e-15, 4.4202e-14,\n 3.6407e-15, 1.1200e-15, 1.4523e-13, 2.3290e-15, 4.5804e-15, 1.5834e-16,\n 1.5985e-18, 4.8335e-15, 1.5532e-14, 4.0917e-15, 1.2350e-13, 2.6559e-15,\n 4.6763e-14, 1.0577e-13, 9.7787e-15, 2.4542e-14, 8.0957e-15, 3.1915e-15,\n 1.3863e-15, 2.4524e-14, 1.4708e-14, 8.4338e-15, 1.3881e-14, 1.9359e-15,\n 7.0694e-15, 7.6874e-15, 4.9027e-15, 1.0514e-15, 6.6820e-14, 8.4742e-16,\n 1.2980e-15, 5.4787e-14, 3.0449e-15, 1.3339e-14, 5.0948e-15, 3.9880e-15,\n 1.9991e-13, 9.6664e-15, 6.9496e-14, 4.5320e-15, 9.5460e-16, 2.7071e-13,\n 3.4532e-14, 1.2416e-15, 9.0878e-16, 3.5242e-14, 7.8519e-18, 6.0198e-15,\n 6.9224e-14, 5.5490e-14, 3.3889e-16, 4.8035e-15, 2.5986e-14, 3.0227e-15,\n 5.0702e-14, 2.6918e-14, 6.5423e-15, 5.1314e-15, 1.8199e-14, 1.5389e-16,\n 1.5068e-15, 2.4658e-15, 8.1754e-16, 5.4884e-14, 3.3098e-15, 2.6643e-15,\n 6.6734e-15, 6.7834e-15, 1.0744e-14, 8.8811e-15, 4.2474e-15, 4.7379e-14,\n 8.6664e-16, 1.7680e-13, 3.6993e-17, 1.9283e-15, 1.2641e-15, 4.4244e-15,\n 1.2190e-16, 5.9446e-15, 4.5334e-14, 2.5169e-14, 9.0794e-16, 6.1427e-15,\n 5.3386e-17, 1.7291e-15, 5.1797e-16, 1.3432e-14, 7.0209e-15, 9.3851e-15,\n 5.8898e-15, 1.1520e-15, 3.9713e-14, 2.4310e-14, 1.6556e-15, 1.3808e-16,\n 1.4804e-14, 3.1149e-14, 5.6152e-17, 4.5944e-15, 6.4669e-15, 3.9752e-14,\n 1.5151e-15, 3.7299e-16, 5.5258e-15, 1.3405e-15, 4.9918e-15, 1.9306e-16,\n 9.2315e-15, 3.8055e-15, 3.3350e-14, 4.0148e-15, 1.4613e-15, 1.0548e-14,\n 1.1539e-13, 5.3202e-15, 1.0442e-14, 6.4257e-16, 7.1057e-14, 1.9243e-16,\n 1.8330e-14, 2.5551e-17, 2.6726e-14, 9.3251e-15, 2.2891e-15, 1.4846e-14,\n 1.4192e-13, 4.2735e-15, 8.1576e-15, 4.1608e-14, 2.6323e-13, 3.4532e-15,\n 3.5396e-15, 1.5263e-15, 2.5176e-13, 2.8912e-16, 1.2635e-15, 3.5150e-13,\n 5.6330e-16, 4.0634e-14, 1.2275e-13, 1.6102e-15, 3.2467e-15, 6.5383e-14,\n 6.0400e-15, 2.6158e-15, 6.4990e-14, 1.4492e-14, 7.1409e-16, 3.5298e-15,\n 1.8839e-14, 2.7794e-14, 3.4056e-15, 3.5335e-15, 1.9351e-15, 1.8319e-15,\n 9.0593e-15, 1.0745e-14, 1.9642e-15, 1.6441e-13, 1.5184e-14, 8.0277e-16,\n 1.1993e-14, 7.2284e-14, 5.6527e-14, 7.3144e-14, 1.7943e-15, 6.4641e-15,\n 1.3361e-15, 2.0806e-14, 3.4331e-16, 2.2775e-16], device='cuda:0')" + }, + "51": { + "step": "tensor(11268.)", + "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([7.6379e-17, 3.7910e-16, 1.2107e-13, 4.0372e-16, 1.9721e-14, 6.0826e-14,\n 3.0604e-14, 8.0001e-15, 3.8417e-15, 7.4320e-15, 8.7357e-14, 2.2672e-14,\n 4.7440e-15, 1.7985e-14, 1.3959e-16, 5.4812e-15, 1.3528e-14, 6.3824e-15,\n 1.3710e-15, 2.4901e-15, 3.2214e-14, 6.7475e-16, 3.6235e-15, 1.0647e-14,\n 1.7002e-14, 8.7719e-15, 1.3310e-14, 2.6390e-14, 9.0026e-15, 4.6925e-14,\n 6.3866e-14, 7.3817e-14, 4.1449e-13, 1.2142e-15, 1.9512e-14, 4.2063e-16,\n 3.5068e-14, 1.9142e-15, 1.3818e-13, 4.3924e-15, 2.1632e-14, 6.6947e-16,\n 2.4540e-15, 8.9225e-14, 4.1435e-17, 4.0651e-15, 1.9237e-14, 1.4448e-14,\n 3.4017e-15, 2.6034e-14, 4.7031e-14, 1.3200e-14, 4.8777e-14, 4.1095e-15,\n 5.2877e-15, 1.0790e-13, 6.8194e-16, 1.6696e-14, 8.7767e-14, 1.1088e-13,\n 1.4360e-14, 1.1941e-14, 6.1398e-16, 8.8449e-14, 4.8028e-15, 1.1438e-13,\n 1.8384e-16, 2.9105e-14, 1.7730e-14, 1.7184e-14, 3.6616e-14, 5.7287e-14,\n 1.3561e-14, 4.2172e-14, 2.0171e-14, 6.7424e-14, 4.1020e-14, 1.6767e-14,\n 2.7007e-15, 1.0652e-16, 8.2815e-14, 6.2717e-15, 5.3213e-15, 3.3885e-14,\n 6.9593e-15, 1.4408e-15, 1.8103e-13, 5.7175e-15, 7.8887e-15, 1.2081e-16,\n 1.0556e-16, 7.0599e-15, 2.5661e-14, 6.1505e-15, 1.5272e-13, 4.4895e-15,\n 9.0875e-14, 1.2470e-13, 1.6526e-14, 1.3845e-14, 1.2400e-14, 5.6630e-15,\n 2.5456e-15, 3.8437e-14, 2.3362e-14, 1.5184e-14, 2.3161e-14, 2.4824e-15,\n 1.3503e-14, 1.7209e-14, 9.0586e-15, 2.1674e-15, 9.8422e-14, 2.2708e-15,\n 3.5653e-15, 1.0424e-13, 4.1090e-15, 2.0438e-14, 5.4788e-15, 7.0667e-15,\n 2.1977e-13, 1.9510e-14, 7.8055e-14, 5.3797e-15, 1.2535e-15, 2.4247e-13,\n 3.0464e-14, 2.0201e-15, 1.1802e-15, 3.7769e-14, 2.6220e-17, 8.6423e-15,\n 1.2439e-13, 7.8605e-14, 4.8596e-16, 8.6190e-15, 2.2642e-14, 6.2897e-15,\n 7.1699e-14, 5.3172e-14, 9.4249e-15, 8.9476e-15, 3.5783e-14, 6.4368e-16,\n 3.3266e-15, 4.5405e-15, 2.6092e-15, 4.5339e-14, 2.9314e-15, 4.6406e-15,\n 6.5842e-15, 6.7999e-15, 1.8060e-14, 1.2404e-14, 6.2464e-15, 6.3027e-14,\n 1.5201e-15, 1.9760e-13, 7.5966e-17, 4.3822e-15, 1.8717e-15, 6.0197e-15,\n 2.8033e-16, 1.0996e-14, 5.7294e-14, 3.6597e-14, 2.2991e-15, 1.2608e-14,\n 1.9122e-16, 2.3541e-15, 7.9142e-16, 1.7557e-14, 1.3386e-14, 1.7594e-14,\n 9.1923e-15, 2.4230e-15, 6.7554e-14, 4.1789e-14, 2.0684e-15, 1.8825e-16,\n 1.1562e-14, 3.2297e-14, 1.0999e-16, 6.8538e-15, 5.5994e-15, 9.0526e-14,\n 1.7395e-15, 6.9443e-16, 5.6429e-15, 2.2741e-15, 1.0691e-14, 5.9086e-16,\n 1.9020e-14, 7.4463e-15, 4.4230e-14, 6.8853e-15, 3.2918e-15, 1.8939e-14,\n 8.8149e-14, 1.0601e-14, 1.4218e-14, 1.1883e-15, 1.4972e-13, 4.6770e-16,\n 1.7809e-14, 3.5999e-17, 2.2824e-14, 2.0302e-14, 2.4885e-15, 1.5281e-14,\n 1.2895e-13, 5.9094e-15, 1.0369e-14, 1.1379e-13, 4.1966e-13, 6.3487e-15,\n 4.5560e-15, 3.8610e-15, 2.2251e-13, 4.5176e-16, 3.6409e-15, 4.2719e-13,\n 7.6927e-16, 3.2542e-14, 1.5262e-13, 2.2156e-15, 4.8735e-15, 6.8892e-14,\n 1.7903e-14, 5.3348e-15, 9.1449e-14, 1.9333e-14, 8.9982e-16, 7.4665e-15,\n 3.2703e-14, 3.0842e-14, 5.1475e-15, 6.7264e-15, 3.4670e-15, 2.1725e-15,\n 1.9687e-14, 1.5335e-14, 2.6593e-15, 1.9081e-13, 2.6676e-14, 1.1402e-15,\n 2.1925e-14, 8.8225e-14, 1.0990e-13, 9.9626e-14, 3.1576e-15, 1.0534e-14,\n 3.7910e-15, 4.0610e-14, 4.9390e-16, 1.7100e-16], device='cuda:0')" + }, + "52": { + "step": "tensor(11268.)", + "exp_avg": "tensor([[-5.6052e-45, -5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, -5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, -5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, 5.6052e-45],\n ...,\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [-5.6052e-45, -5.6052e-45, -5.6052e-45, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45]], device='cuda:0')", + "exp_avg_sq": "tensor([[1.9088e-17, 1.4973e-17, 2.7805e-17, ..., 1.8096e-18, 8.5794e-19,\n 1.3632e-16],\n [4.3612e-18, 1.2880e-17, 4.2723e-17, ..., 1.6876e-16, 1.0080e-17,\n 2.2346e-17],\n [2.6576e-17, 9.5552e-19, 1.4119e-17, ..., 8.1545e-17, 8.4221e-19,\n 1.1857e-18],\n ...,\n [3.7001e-14, 6.0222e-13, 8.1862e-13, ..., 1.4011e-12, 7.1514e-13,\n 3.5374e-13],\n [1.1024e-14, 1.8264e-13, 2.4941e-13, ..., 4.0306e-13, 2.2700e-13,\n 1.1772e-13],\n [4.5320e-15, 5.3266e-14, 7.9200e-14, ..., 1.3029e-13, 5.6265e-14,\n 2.7879e-14]], device='cuda:0')" + }, + "53": { + "step": "tensor(11268.)", + "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([1.5030e-16, 1.2033e-16, 3.6158e-16, 4.3927e-17, 1.7953e-16, 2.8655e-16,\n 8.1683e-17, 3.4874e-16, 8.5335e-16, 2.8842e-15, 1.0427e-16, 4.3297e-17,\n 1.2022e-15, 3.2846e-16, 2.2252e-15, 1.3137e-18, 2.2382e-16, 5.0257e-17,\n 5.9053e-16, 1.6287e-15, 1.3915e-16, 3.6341e-19, 6.2304e-18, 3.4628e-17,\n 2.1056e-16, 1.4621e-16, 3.1464e-16, 4.4039e-16, 4.0741e-17, 3.3485e-16,\n 9.2019e-16, 1.0668e-16, 6.3573e-16, 1.0061e-15, 2.2585e-15, 1.3883e-15,\n 3.5670e-16, 3.7955e-16, 6.9729e-16, 1.5652e-16, 2.7827e-17, 4.8357e-16,\n 5.0564e-16, 2.4417e-18, 3.0529e-15, 1.4523e-15, 1.3821e-17, 4.3511e-18,\n 1.6472e-16, 5.1954e-16, 5.6672e-16, 3.5148e-16, 4.0218e-16, 1.8270e-16,\n 1.6649e-15, 6.2191e-16, 9.2419e-17, 1.7225e-15, 1.3198e-17, 3.2820e-16,\n 4.8614e-16, 1.2252e-16, 1.0234e-17, 3.1667e-15, 7.9721e-17, 1.9019e-15,\n 1.1794e-17, 1.9913e-15, 3.6852e-16, 3.7442e-15, 4.4203e-15, 1.0713e-16,\n 1.0662e-15, 2.9442e-15, 2.0987e-15, 1.9514e-16, 5.1911e-15, 1.9877e-15,\n 7.2849e-15, 1.9621e-17, 2.0797e-16, 6.2376e-16, 2.4694e-16, 8.3332e-16,\n 2.6343e-15, 9.6957e-17, 1.9619e-16, 6.3558e-16, 7.5790e-15, 1.8197e-16,\n 1.4372e-15, 1.0229e-17, 5.9640e-15, 5.7337e-15, 9.8354e-16, 8.2919e-16,\n 6.5541e-18, 1.0609e-15, 1.0897e-15, 8.2907e-16, 2.4169e-15, 5.8989e-18,\n 3.1290e-17, 1.2344e-17, 6.3702e-18, 1.0085e-16, 2.5662e-17, 1.4676e-16,\n 5.9641e-16, 1.8918e-15, 2.9047e-16, 1.5169e-16, 8.6972e-17, 4.7852e-16,\n 1.3167e-16, 1.8321e-15, 6.8405e-16, 1.3351e-15, 2.0841e-15, 7.0428e-16,\n 2.8270e-17, 6.6832e-18, 1.3307e-15, 1.7973e-16, 1.2319e-16, 7.0940e-17,\n 2.3100e-15, 3.7267e-16, 5.7320e-17, 2.1158e-15, 9.0282e-17, 2.5541e-16,\n 3.6162e-16, 1.4009e-15, 2.6499e-15, 2.4433e-17, 1.0154e-16, 2.0166e-16,\n 1.4468e-16, 3.7966e-18, 5.9881e-16, 1.6655e-16, 4.8718e-17, 2.1894e-18,\n 4.8273e-16, 7.6731e-16, 1.1443e-15, 5.0164e-16, 8.8946e-17, 3.6526e-16,\n 8.5643e-18, 1.2143e-15, 1.2618e-17, 3.6602e-17, 1.8230e-15, 8.9932e-16,\n 2.9312e-18, 1.8087e-18, 1.2083e-16, 8.0881e-16, 3.6040e-15, 2.2355e-16,\n 2.1654e-16, 1.1455e-16, 3.1717e-16, 3.7409e-16, 3.0559e-16, 5.9733e-16,\n 6.9244e-18, 7.4796e-17, 8.5160e-16, 1.7150e-16, 2.1619e-16, 1.2649e-15,\n 7.9487e-19, 9.1414e-16, 1.1979e-15, 1.0900e-15, 9.2516e-18, 6.0697e-17,\n 6.9439e-17, 4.4130e-16, 1.1301e-18, 2.8590e-17, 2.0601e-16, 2.4062e-15,\n 2.2992e-16, 3.7554e-17, 1.0109e-15, 1.1461e-16, 9.8784e-17, 8.2893e-16,\n 5.8375e-16, 6.9525e-16, 3.7795e-16, 1.0749e-15, 1.2974e-18, 7.8222e-17,\n 5.8334e-16, 6.6346e-16, 1.1030e-17, 1.7091e-15, 8.1652e-17, 9.8791e-16,\n 5.3649e-17, 4.2069e-17, 1.6441e-17, 3.2359e-16, 2.0820e-15, 6.6790e-17,\n 3.4020e-16, 5.4958e-16, 1.3166e-15, 1.3994e-16, 7.2739e-18, 1.5676e-16,\n 9.1442e-17, 8.8781e-17, 4.4403e-17, 9.9208e-17, 2.5303e-17, 6.8817e-15,\n 1.8271e-15, 1.1357e-15, 3.4298e-18, 3.7691e-17, 3.7616e-16, 1.2152e-15,\n 3.6036e-15, 1.1500e-15, 4.0552e-17, 2.2085e-16, 1.2777e-15, 1.1181e-17,\n 7.3198e-16, 5.5802e-16, 1.5095e-15, 3.2981e-16, 2.3810e-16, 4.9127e-17,\n 9.8406e-17, 4.3789e-17, 1.1635e-15, 2.9990e-16, 2.8872e-16, 2.0398e-15,\n 3.5957e-16, 3.4429e-18, 7.8661e-16, 7.9789e-16, 3.3263e-15, 6.8238e-16,\n 9.7932e-16, 5.3557e-16, 8.3972e-19, 4.2418e-16, 2.2978e-30, 5.6272e-32,\n 3.5795e-31, 6.5112e-32, 3.6775e-31, 1.0948e-31, 5.9114e-31, 1.4252e-32,\n 3.2284e-31, 5.0209e-31, 4.6983e-31, 1.2069e-32, 3.0504e-34, 2.5644e-32,\n 1.4422e-31, 1.0778e-32, 2.5042e-31, 1.9182e-32, 2.2839e-31, 5.4957e-32,\n 3.5313e-33, 1.7712e-31, 1.0017e-32, 2.0640e-31, 2.9222e-33, 1.7662e-31,\n 3.9021e-31, 1.6875e-32, 1.0185e-30, 3.4740e-31, 5.6510e-32, 8.1403e-32,\n 1.0426e-31, 6.2391e-31, 5.8241e-32, 4.8005e-32, 1.9514e-32, 1.0576e-31,\n 2.8486e-31, 5.6322e-32, 9.9889e-33, 1.1582e-31, 2.3294e-32, 5.1514e-32,\n 4.8687e-32, 4.3910e-31, 1.7973e-31, 6.0212e-32, 6.4553e-32, 4.0085e-32,\n 8.0251e-32, 2.6591e-31, 2.6820e-31, 3.3614e-31, 1.3690e-31, 3.1483e-32,\n 2.7990e-31, 4.8179e-31, 6.6033e-31, 2.8822e-31, 2.3279e-31, 1.6134e-31,\n 1.1100e-32, 9.9905e-32, 7.5284e-34, 3.8992e-31, 1.0324e-30, 3.2184e-31,\n 8.6953e-33, 5.1672e-32, 3.1419e-31, 1.0506e-31, 1.7039e-31, 1.0248e-32,\n 8.7771e-32, 4.7469e-32, 1.6700e-31, 4.7043e-31, 1.6206e-30, 5.9368e-31,\n 6.4312e-32, 4.8721e-31, 5.9019e-32, 7.8748e-31, 9.6346e-32, 8.8921e-32,\n 7.9163e-31, 2.7524e-31, 1.4262e-31, 1.8249e-32, 3.1875e-32, 1.7352e-31,\n 1.8200e-31, 3.8298e-32, 1.4797e-31, 1.0522e-32, 1.3969e-32, 4.5730e-31,\n 2.4023e-32, 1.3766e-31, 2.0612e-31, 1.4378e-32, 1.2560e-31, 1.1868e-31,\n 4.7636e-32, 6.5361e-32, 4.0790e-32, 6.8090e-31, 2.7658e-31, 4.9247e-32,\n 3.7666e-32, 2.0032e-31, 2.1691e-31, 4.0737e-31, 1.7157e-31, 8.0285e-31,\n 2.2048e-31, 2.9869e-31, 6.9457e-31, 8.5282e-31, 5.1603e-31, 2.7181e-31,\n 3.2067e-31, 1.5703e-31, 3.5956e-32, 3.2210e-32, 1.7887e-30, 2.1014e-33,\n 3.3977e-31, 2.0228e-32, 7.9469e-32, 9.6858e-32, 2.4937e-31, 8.0578e-32,\n 2.1133e-32, 2.3733e-35, 5.4267e-33, 3.1921e-32, 7.1651e-32, 1.8620e-31,\n 1.7708e-32, 4.2068e-31, 1.6117e-30, 1.8194e-31, 5.2056e-31, 6.9364e-32,\n 2.8462e-31, 7.8503e-31, 3.6172e-31, 2.4241e-32, 2.6104e-32, 1.0572e-30,\n 1.8203e-31, 2.0464e-31, 9.4908e-33, 1.3021e-31, 4.5077e-31, 1.4760e-31,\n 7.7344e-34, 2.0358e-31, 5.9918e-32, 5.8322e-32, 3.3106e-31, 3.0603e-32,\n 2.1393e-32, 1.8568e-31, 1.2279e-31, 7.8630e-33, 7.2293e-32, 8.4171e-31,\n 4.9711e-32, 1.7274e-30, 8.7889e-31, 1.2384e-32, 2.8634e-31, 9.0804e-31,\n 8.9851e-31, 3.3965e-32, 4.2621e-32, 7.4448e-32, 5.2262e-32, 2.2482e-31,\n 4.4529e-33, 2.2671e-31, 1.1597e-31, 2.9173e-32, 1.0585e-31, 1.9580e-33,\n 3.6657e-31, 2.2449e-33, 4.5086e-31, 2.5167e-31, 2.7764e-31, 6.3070e-31,\n 5.7491e-31, 2.0839e-32, 1.1747e-30, 1.0670e-31, 2.7191e-31, 5.8718e-32,\n 1.7371e-33, 9.8444e-32, 1.3547e-31, 2.2597e-31, 2.6092e-32, 8.0985e-32,\n 5.3230e-33, 1.2284e-31, 1.8476e-31, 8.0932e-33, 3.6226e-31, 6.8792e-33,\n 2.0772e-31, 7.7433e-32, 3.9628e-31, 1.8137e-31, 1.0331e-32, 8.2072e-32,\n 1.9596e-31, 5.3454e-32, 5.9271e-32, 1.0799e-31, 2.1131e-31, 1.2406e-31,\n 1.6344e-31, 4.7275e-33, 2.1180e-31, 2.2774e-31, 6.0832e-31, 6.1756e-31,\n 1.5357e-32, 1.1305e-31, 1.6125e-32, 1.0779e-31, 2.3072e-33, 2.9735e-32,\n 3.5686e-32, 1.9041e-31, 6.5518e-33, 9.6570e-33, 4.2447e-32, 4.7439e-31,\n 3.3868e-32, 9.7281e-32, 2.5881e-32, 3.2417e-31, 6.6878e-31, 6.9631e-35,\n 6.3654e-31, 4.4801e-31, 6.1304e-31, 2.8852e-33, 1.0090e-31, 6.9146e-31,\n 3.0461e-32, 1.3644e-32, 8.7774e-13, 1.3257e-12, 1.3948e-14, 1.1140e-11,\n 3.0310e-12, 7.6036e-13, 1.9167e-12, 5.4003e-12, 1.0748e-13, 1.3425e-12,\n 8.8795e-12, 2.1071e-14, 6.2435e-13, 2.2625e-12, 1.1229e-11, 2.5434e-13,\n 3.3244e-12, 1.1169e-11, 7.6740e-13, 1.2850e-12, 3.5772e-12, 6.3275e-15,\n 4.7045e-12, 1.4372e-11, 9.0587e-13, 1.5664e-11, 3.3139e-13, 2.5054e-12,\n 2.0597e-12, 9.4364e-12, 2.0146e-12, 2.1854e-14, 3.5533e-13, 2.3128e-14,\n 4.3272e-12, 6.0001e-12, 6.0114e-12, 9.2311e-13, 1.7581e-13, 5.8776e-12,\n 1.1147e-11, 5.5990e-12, 6.6264e-12, 5.1708e-12, 2.2461e-12, 1.4722e-14,\n 4.0249e-12, 1.1462e-12, 5.4942e-12, 3.9154e-12, 2.1914e-11, 4.2620e-12,\n 7.2631e-13, 3.3762e-14, 1.8755e-12, 1.8706e-12, 1.2011e-11, 1.0250e-12,\n 2.8729e-12, 2.9027e-12, 2.9765e-12, 1.1425e-12, 1.0990e-12, 8.0734e-13,\n 3.1215e-13, 2.5147e-14, 6.5295e-12, 2.4350e-12, 3.3112e-12, 4.2946e-12,\n 3.1049e-13, 1.0072e-13, 6.1339e-14, 1.4566e-11, 4.7155e-13, 1.9892e-11,\n 3.8108e-12, 1.1045e-11, 5.0680e-12, 5.1284e-13, 3.2115e-12, 3.7494e-11,\n 1.3118e-12, 8.2886e-12, 3.0112e-13, 1.1266e-11, 4.0708e-12, 2.5800e-13,\n 7.2271e-14, 4.5091e-12, 6.5587e-12, 3.0592e-12, 2.2141e-14, 6.2799e-12,\n 1.6082e-11, 7.9256e-13, 3.3866e-13, 6.5728e-12, 6.1881e-13, 6.2210e-13,\n 8.3281e-13, 3.9140e-12, 1.0479e-13, 5.1375e-13, 1.3881e-12, 1.3860e-11,\n 8.3876e-12, 1.1941e-12, 1.2706e-13, 1.4892e-12, 6.7140e-13, 2.2397e-12,\n 4.1851e-12, 2.3187e-13, 3.2135e-12, 4.4724e-12, 1.1827e-12, 5.5869e-13,\n 7.8950e-12, 1.5476e-13, 5.5090e-13, 3.0284e-11, 5.5404e-12, 1.9130e-11,\n 6.9026e-12, 2.2893e-12, 1.7021e-13, 2.7997e-13, 3.4591e-12, 3.8778e-12,\n 2.1636e-12, 3.0590e-12, 7.6568e-13, 1.2954e-14, 3.3105e-12, 1.2383e-12,\n 1.7781e-12, 8.1996e-12, 9.2527e-13, 1.8089e-13, 1.4636e-15, 6.7807e-12,\n 2.9350e-12, 6.8403e-14, 2.0160e-12, 1.6296e-12, 4.2521e-13, 1.4501e-11,\n 6.1924e-13, 1.1190e-14, 1.9146e-12, 2.1659e-11, 1.1248e-12, 9.1855e-13,\n 8.5893e-12, 2.5763e-13, 9.7622e-14, 1.6231e-12, 1.4121e-15, 2.7524e-12,\n 6.9283e-12, 9.7786e-12, 2.8271e-13, 1.8488e-12, 2.5766e-12, 4.6350e-12,\n 2.1998e-12, 1.0678e-12, 1.7747e-13, 1.1593e-11, 7.1484e-13, 1.4957e-11,\n 9.4621e-12, 7.1391e-13, 2.4101e-12, 6.2597e-13, 3.8636e-14, 7.9300e-13,\n 6.4669e-12, 2.4944e-14, 4.4979e-12, 1.2853e-11, 3.3945e-14, 3.0572e-12,\n 8.4930e-12, 7.9902e-12, 2.1601e-12, 1.2484e-12, 9.7515e-14, 8.2193e-12,\n 1.4984e-13, 2.7828e-12, 1.2613e-11, 1.3055e-11, 2.2725e-12, 3.8277e-12,\n 6.8286e-12, 1.6506e-11, 1.5939e-13, 7.3215e-12, 9.0980e-13, 4.5850e-13,\n 1.6999e-11, 1.1662e-12, 5.3875e-13, 2.0797e-12, 6.7420e-13, 1.2232e-12,\n 1.9007e-12, 6.2310e-13, 3.5893e-12, 2.6925e-12, 8.6406e-12, 6.6996e-13,\n 9.9942e-14, 2.7192e-11, 8.5598e-13, 1.6101e-11, 3.2666e-14, 1.7603e-12,\n 1.3775e-12, 6.1187e-12, 6.9856e-12, 4.1244e-12, 8.9290e-12, 3.1637e-13,\n 5.6616e-12, 7.5024e-13, 3.4891e-11, 2.2587e-13, 8.2396e-13, 3.4299e-12,\n 3.5050e-12, 2.0482e-12, 7.1594e-12, 2.5906e-12, 7.2651e-13, 6.4472e-13,\n 4.3733e-12, 2.6201e-14, 6.3702e-12, 2.4042e-13, 5.9099e-16, 1.1925e-12,\n 6.0651e-12, 8.0081e-14, 1.6027e-12, 1.5621e-13, 2.1099e-13, 1.4163e-12,\n 5.0702e-12, 3.6620e-12, 2.7557e-12, 1.1961e-11, 3.6146e-12, 1.0546e-12],\n device='cuda:0')" + }, + "54": { + "step": "tensor(11268.)", + "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([[2.4202e-13, 7.9313e-15, 5.5237e-15, ..., 1.8325e-14, 9.7267e-14,\n 1.2686e-13],\n [1.2758e-13, 5.8437e-15, 4.1798e-15, ..., 1.5476e-14, 5.5270e-14,\n 7.2375e-14],\n [2.3338e-12, 5.2914e-14, 4.6270e-14, ..., 7.6578e-14, 9.3305e-13,\n 1.1718e-12],\n ...,\n [7.3406e-13, 1.8380e-14, 1.3340e-14, ..., 3.3773e-14, 2.9351e-13,\n 3.8971e-13],\n [1.8208e-14, 1.3790e-15, 4.2671e-16, ..., 6.9496e-15, 7.1599e-15,\n 6.6839e-15],\n [4.1155e-14, 3.3339e-15, 2.5957e-15, ..., 1.1174e-14, 2.0243e-14,\n 2.1292e-14]], device='cuda:0')" + }, + "55": { + "step": "tensor(11268.)", + "exp_avg": "tensor([-5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n -5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, -5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45,\n -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45, -5.6052e-45,\n -5.6052e-45, 5.6052e-45, -5.6052e-45, -5.6052e-45, 5.6052e-45,\n 5.6052e-45, 5.6052e-45, 5.6052e-45, -5.6052e-45, 5.6052e-45,\n -5.6052e-45], device='cuda:0')", + "exp_avg_sq": "tensor([4.3318e-12, 2.3206e-12, 4.1229e-11, 2.2206e-11, 1.6645e-11, 2.2722e-14,\n 5.5336e-12, 5.7436e-12, 9.9233e-12, 9.9039e-13, 1.8312e-11, 2.4101e-13,\n 1.4516e-11, 8.5051e-15, 1.4284e-11, 3.4878e-14, 7.6588e-13, 6.7675e-12,\n 6.3327e-12, 1.0849e-11, 1.5537e-11, 2.3745e-12, 1.8168e-11, 8.1935e-12,\n 4.5757e-11, 1.2360e-11, 9.8480e-12, 2.2967e-12, 9.1834e-12, 2.2124e-11,\n 1.1586e-11, 4.0399e-11, 4.4965e-12, 5.2049e-13, 1.8659e-12, 2.0023e-12,\n 5.1360e-12, 1.7012e-11, 1.0540e-11, 2.1692e-12, 5.8748e-12, 1.6471e-11,\n 1.7322e-13, 6.3939e-12, 1.1392e-11, 4.0098e-12, 4.2003e-12, 1.3249e-11,\n 4.9506e-11, 2.4688e-12, 1.8801e-11, 9.5968e-12, 4.6835e-12, 3.9734e-12,\n 1.0325e-13, 3.6310e-11, 6.8560e-13, 1.5249e-11, 3.7122e-12, 3.5800e-11,\n 2.0334e-11, 1.3861e-11, 4.3827e-13, 7.3843e-12, 3.1731e-12, 3.4936e-12,\n 3.8417e-12, 3.6398e-11, 3.6439e-13, 1.7270e-11, 1.3060e-13, 1.3989e-11,\n 2.3152e-13, 1.4337e-11, 8.7260e-14, 3.8302e-12, 3.1931e-11, 1.8267e-12,\n 1.3503e-11, 2.5649e-11, 1.1872e-11, 8.3936e-12, 9.6752e-12, 1.3345e-11,\n 1.2086e-11, 1.7045e-11, 5.8510e-12, 8.2138e-12, 1.0020e-11, 5.4994e-13,\n 2.8696e-12, 3.6047e-12, 6.3750e-11, 1.7385e-12, 8.5605e-12, 8.9480e-15,\n 4.5897e-13, 1.4582e-13, 3.2263e-14, 9.9206e-13, 6.3726e-11, 4.5354e-13,\n 6.0353e-12, 5.2973e-11, 4.1712e-12, 1.1056e-12, 1.7367e-12, 4.4622e-13,\n 1.2199e-11, 2.9011e-11, 1.6446e-13, 8.5281e-12, 3.3974e-12, 3.3232e-12,\n 1.0839e-11, 2.5581e-12, 6.3458e-13, 3.5879e-13, 8.2584e-12, 6.4379e-12,\n 1.2670e-13, 9.0529e-12, 6.4039e-13, 6.4557e-11, 2.6482e-11, 6.6465e-11,\n 2.4763e-11, 1.5572e-12, 1.3769e-12, 3.8532e-12, 2.9918e-11, 4.9015e-13,\n 3.0761e-12, 1.3805e-11, 4.8348e-14, 1.9384e-13, 8.5922e-12, 2.0220e-11,\n 9.8950e-12, 2.3853e-12, 4.4393e-12, 1.1657e-11, 6.0657e-12, 5.6236e-12,\n 6.1005e-12, 1.3501e-11, 5.0354e-12, 7.6882e-12, 5.4425e-12, 3.5276e-11,\n 1.9632e-14, 4.0422e-12, 4.9404e-13, 4.0842e-13, 8.5272e-11, 3.7122e-11,\n 1.8495e-12, 3.4675e-11, 7.8726e-12, 4.8381e-11, 9.7606e-13, 9.9970e-13,\n 6.4212e-12, 2.7273e-14, 4.7131e-11, 7.9059e-12, 1.0405e-11, 1.6647e-12,\n 7.4388e-12, 3.4126e-11, 4.6739e-13, 8.0808e-12, 1.0994e-11, 3.0205e-12,\n 3.3515e-11, 8.2178e-12, 6.5513e-12, 2.3063e-13, 5.8891e-12, 1.7741e-12,\n 1.5647e-11, 7.2519e-12, 1.0949e-12, 2.2346e-12, 2.1628e-13, 4.3274e-12,\n 1.8697e-12, 2.7621e-11, 6.8295e-13, 2.6183e-12, 2.2625e-12, 4.9950e-13,\n 1.6347e-12, 1.7746e-11, 6.4331e-12, 5.6453e-12, 1.7214e-11, 9.3888e-12,\n 1.0104e-11, 4.6986e-12, 5.0256e-12, 2.1142e-11, 1.2914e-11, 3.4860e-13,\n 2.2229e-11, 7.5602e-13, 1.7120e-11, 2.2648e-11, 2.3310e-11, 2.2413e-11,\n 6.4210e-12, 4.8598e-12, 2.0751e-11, 8.3251e-13, 1.8989e-11, 8.9948e-12,\n 1.2110e-14, 1.3090e-11, 3.1224e-14, 9.7709e-12, 5.6322e-12, 1.4065e-10,\n 1.6446e-11, 5.0255e-11, 4.5836e-13, 2.2725e-12, 1.0601e-11, 8.1307e-13,\n 2.9280e-12, 1.6324e-13, 5.0406e-12, 1.2879e-11, 1.3523e-11, 2.0289e-12,\n 4.8075e-11, 6.2966e-12, 2.7813e-12, 3.0357e-13, 2.7335e-12, 3.5414e-14,\n 1.5136e-12, 1.1441e-11, 8.3809e-13, 7.6591e-13, 5.5092e-12, 3.4526e-11,\n 5.7661e-12, 1.7953e-11, 8.7878e-12, 4.2853e-12, 3.6218e-11, 8.4204e-13,\n 2.8428e-11, 1.3099e-11, 2.4713e-13, 7.2117e-13], device='cuda:0')" + }, + "56": { + "step": "tensor(11268.)", + "exp_avg": "tensor([[-5.6052e-45, -5.6052e-45, -5.6052e-45, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n ...,\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45]], device='cuda:0')", + "exp_avg_sq": "tensor([[1.0625e-10, 2.3049e-11, 6.3773e-11, ..., 2.7615e-11, 2.3642e-12,\n 3.6274e-13],\n [9.5738e-13, 2.4950e-13, 5.6642e-13, ..., 2.2765e-13, 2.2551e-14,\n 4.4821e-15],\n [1.3039e-12, 2.7514e-13, 7.7805e-13, ..., 3.4471e-13, 2.9168e-14,\n 4.1816e-15],\n ...,\n [1.0660e-12, 2.2477e-13, 6.4349e-13, ..., 2.7396e-13, 2.5260e-14,\n 5.2502e-15],\n [1.3206e-12, 2.8517e-13, 7.9380e-13, ..., 3.4357e-13, 2.4486e-14,\n 2.4563e-15],\n [1.1719e-12, 2.4725e-13, 7.3292e-13, ..., 3.1371e-13, 2.8304e-14,\n 7.1725e-15]], device='cuda:0')" }, - "26": { - "step": "tensor(23788.)", - "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([[3.2663e-20, 4.5150e-20, 0.0000e+00, ..., 9.5566e-20, 1.5410e-19,\n 1.8824e-20],\n [3.0196e-21, 4.1980e-21, 0.0000e+00, ..., 1.1732e-22, 3.7821e-21,\n 1.2078e-22],\n [2.4407e-20, 1.9073e-20, 0.0000e+00, ..., 1.9632e-20, 2.1203e-20,\n 1.3791e-20],\n ...,\n [1.2154e-20, 9.9878e-21, 0.0000e+00, ..., 7.0457e-21, 6.8812e-20,\n 1.9860e-21],\n [7.1321e-20, 2.2072e-20, 0.0000e+00, ..., 1.9021e-20, 1.5589e-19,\n 3.4524e-20],\n [1.4423e-20, 2.5357e-21, 0.0000e+00, ..., 3.4985e-21, 3.0053e-20,\n 3.1596e-22]], device='cuda:0')" + "57": { + "step": "tensor(11268.)", + "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 device='cuda:0')", + "exp_avg_sq": "tensor([2.7387e-09, 2.4830e-11, 3.3537e-11, 3.4802e-11, 3.9901e-11, 4.5835e-11,\n 3.6441e-11, 2.6986e-11, 3.4086e-11, 3.0291e-11], device='cuda:0')" }, - "27": { - "step": "tensor(23788.)", - "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([5.5355e-17, 2.4826e-18, 1.9705e-17, 2.5604e-17, 7.7966e-18, 1.9653e-18,\n 2.5862e-17, 6.2940e-19, 1.1606e-17, 1.7713e-18, 5.8843e-17, 2.0366e-16,\n 3.4671e-17, 4.0773e-17, 4.2541e-17, 1.7609e-16, 1.9735e-17, 1.9565e-17,\n 6.6039e-17, 7.6779e-17, 2.4622e-19, 8.2493e-18, 1.6638e-17, 2.0629e-17,\n 1.9105e-17, 3.5418e-17, 2.8157e-17, 5.5304e-18, 8.0814e-19, 9.2050e-18,\n 4.8917e-17, 3.2318e-17, 5.0676e-18, 2.4857e-16, 9.0033e-17, 5.5182e-17,\n 1.6415e-18, 3.1112e-20, 2.3148e-17, 7.4746e-16, 1.1151e-16, 1.5676e-17,\n 4.9140e-18, 3.0756e-18, 1.2892e-16, 2.2702e-16, 3.5363e-17, 7.8341e-17,\n 9.7595e-18, 6.2715e-17, 1.2094e-18, 1.3352e-16, 3.4043e-17, 3.0669e-17,\n 4.8741e-18, 4.2516e-19, 5.8140e-17, 2.1576e-18, 3.8641e-17, 3.3816e-18,\n 1.9857e-17, 3.8015e-17, 9.5833e-17, 5.1074e-17, 1.4405e-18, 3.5642e-18,\n 3.8318e-17, 2.4946e-18, 2.4298e-17, 4.6168e-18, 8.4287e-20, 2.3548e-17,\n 9.9524e-17, 7.3096e-18, 2.0150e-17, 1.3470e-19, 1.3085e-18, 8.1290e-17,\n 5.2459e-18, 1.7322e-21, 4.8116e-17, 5.7906e-17, 8.2922e-20, 4.9153e-18,\n 3.4307e-17, 1.4654e-16, 1.1779e-17, 5.0596e-17, 1.0023e-17, 8.3578e-17,\n 5.9188e-17, 3.6405e-17, 9.2533e-17, 1.6884e-17, 2.9815e-17, 4.5698e-18,\n 1.5544e-17, 1.0642e-16, 2.5821e-17, 2.1492e-16, 1.9039e-16, 4.6583e-18,\n 8.2801e-20, 2.5570e-17, 1.4370e-18, 1.6860e-16, 1.3513e-17, 1.2341e-17,\n 3.6771e-17, 2.2931e-19, 1.8227e-17, 1.1894e-19, 1.1879e-16, 1.1636e-16,\n 9.1727e-18, 1.0172e-16, 8.6861e-18, 1.6590e-17, 2.4580e-17, 2.5246e-17,\n 8.7986e-18, 2.0019e-18, 3.6099e-18, 4.7888e-21, 4.6909e-19, 2.2594e-17,\n 1.9201e-16, 3.1714e-17, 8.0193e-17, 3.6464e-17, 2.0137e-19, 2.3139e-16,\n 3.7306e-17, 3.4483e-17, 8.4207e-17, 4.9134e-17, 5.4757e-17, 7.3476e-17,\n 4.9909e-17, 1.9229e-19, 1.4565e-18, 2.6836e-18, 1.9759e-17, 2.5144e-16,\n 4.8041e-17, 2.9348e-18, 4.5307e-17, 5.4732e-18, 4.6257e-17, 1.1443e-18,\n 2.2292e-18, 7.8386e-17, 5.7185e-17, 7.1563e-17, 1.5196e-17, 2.5619e-18,\n 4.5567e-17, 1.3490e-16, 2.3666e-17, 2.9819e-17, 3.6842e-17, 4.2298e-17,\n 2.2963e-17, 3.0662e-17, 2.8708e-17, 2.7021e-17, 1.9397e-16, 1.4940e-16,\n 2.1807e-17, 8.2280e-17, 8.1918e-18, 5.9113e-18, 1.2208e-16, 1.1615e-17,\n 8.7185e-17, 6.4848e-18, 2.4951e-18, 7.8722e-18, 8.4188e-18, 9.3612e-18,\n 1.1034e-17, 9.4975e-18, 3.4596e-18, 2.4645e-17, 3.7923e-16, 5.7928e-20,\n 1.8801e-17, 1.5765e-17, 1.1130e-17, 1.1374e-16, 3.3438e-17, 4.1006e-18,\n 1.6910e-17, 2.3315e-17, 1.7709e-19, 2.0658e-16, 1.0870e-17, 1.6554e-18,\n 5.1629e-17, 2.2909e-16, 1.9184e-22, 2.2403e-17, 1.1119e-16, 2.4747e-17,\n 2.0664e-17, 1.6133e-16, 3.9899e-17, 8.5961e-18, 8.5684e-21, 1.9348e-16,\n 3.9302e-18, 1.9387e-16, 1.1911e-19, 4.2990e-17, 6.1701e-19, 5.1654e-17,\n 7.2708e-17, 1.2360e-21, 6.1464e-18, 8.5465e-18, 7.2751e-18, 2.3660e-16,\n 1.2508e-17, 4.9629e-17, 1.1174e-17, 4.5976e-19, 2.5596e-19, 9.1184e-19,\n 8.3599e-18, 1.6657e-17, 1.3965e-18, 2.0491e-19, 8.7158e-17, 6.2209e-18,\n 4.5729e-18, 3.1890e-17, 2.7126e-17, 3.9914e-17, 1.6507e-17, 1.2808e-17,\n 4.3284e-18, 9.8052e-21, 3.2864e-17, 2.1310e-17, 9.5338e-18, 3.0454e-19,\n 1.9676e-19, 1.6280e-16, 7.5643e-18, 1.8770e-16, 5.9667e-17, 6.5132e-18,\n 8.9588e-17, 2.5321e-17, 5.2894e-17, 1.0053e-17], device='cuda:0')" + "58": { + "step": "tensor(11268.)", + "exp_avg": "tensor([[-5.6052e-45, -5.6052e-45, -5.6052e-45, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n ...,\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45]], device='cuda:0')", + "exp_avg_sq": "tensor([[1.0624e-10, 2.3045e-11, 6.3754e-11, ..., 2.7615e-11, 2.3588e-12,\n 3.5908e-13],\n [9.5722e-13, 2.4944e-13, 5.6615e-13, ..., 2.2765e-13, 2.2472e-14,\n 4.4286e-15],\n [1.3038e-12, 2.7509e-13, 7.7785e-13, ..., 3.4471e-13, 2.9108e-14,\n 4.1407e-15],\n ...,\n [1.0658e-12, 2.2470e-13, 6.4318e-13, ..., 2.7396e-13, 2.5171e-14,\n 5.1892e-15],\n [1.3206e-12, 2.8516e-13, 7.9373e-13, ..., 3.4357e-13, 2.4464e-14,\n 2.4418e-15],\n [1.1716e-12, 2.4712e-13, 7.3239e-13, ..., 3.1371e-13, 2.8146e-14,\n 7.0653e-15]], device='cuda:0')" }, - "28": { - "step": "tensor(23788.)", - "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([9.9046e-20, 1.2287e-20, 3.6076e-20, 8.3660e-20, 1.4923e-20, 9.2212e-21,\n 5.3238e-20, 3.5277e-21, 1.8047e-20, 2.1960e-21, 1.1036e-19, 5.3838e-19,\n 7.5683e-20, 1.0199e-19, 1.6359e-19, 7.3004e-19, 5.9545e-20, 4.8100e-20,\n 2.2264e-19, 1.5224e-19, 2.6926e-23, 2.1845e-20, 6.1924e-20, 4.1983e-20,\n 6.8126e-20, 8.5427e-20, 5.8613e-20, 1.7447e-20, 2.0764e-21, 1.6114e-20,\n 1.3241e-19, 7.6898e-20, 1.1927e-20, 6.2372e-19, 2.3552e-19, 1.6658e-19,\n 3.5167e-21, 1.3286e-23, 5.6159e-20, 4.5694e-18, 2.5021e-19, 5.0149e-20,\n 1.6410e-20, 9.1772e-21, 4.3641e-19, 1.0018e-18, 1.0918e-19, 3.5685e-19,\n 1.0599e-20, 1.3007e-19, 1.2051e-21, 3.3723e-19, 6.5342e-20, 6.7769e-20,\n 6.8980e-21, 1.6114e-21, 1.8632e-19, 1.8339e-21, 1.5786e-19, 1.1745e-20,\n 4.5997e-20, 2.7295e-19, 4.5693e-19, 1.7888e-19, 6.7398e-22, 4.1239e-21,\n 1.0375e-19, 2.2026e-21, 3.9782e-20, 1.4980e-20, 3.3032e-21, 7.1020e-20,\n 2.3141e-19, 1.7248e-20, 4.8305e-20, 1.4136e-21, 4.7030e-21, 2.3314e-19,\n 7.4397e-21, 1.3248e-22, 1.7880e-19, 1.2460e-19, 1.4473e-22, 2.3879e-20,\n 1.3595e-19, 4.2640e-19, 3.4108e-20, 2.1591e-19, 2.4529e-20, 1.9151e-19,\n 3.3541e-19, 9.1182e-20, 2.2862e-19, 2.4724e-20, 6.4490e-20, 4.4652e-21,\n 3.1380e-20, 1.8902e-19, 6.0771e-20, 6.0143e-19, 1.2465e-18, 2.1255e-20,\n 4.7544e-22, 5.8294e-20, 7.6166e-22, 3.7285e-19, 4.1202e-20, 3.3615e-20,\n 1.1003e-19, 9.5076e-25, 4.7911e-20, 2.4047e-21, 3.6347e-19, 2.2934e-19,\n 1.2812e-20, 2.6782e-19, 2.9060e-20, 3.1825e-20, 6.1656e-20, 6.6419e-20,\n 1.1543e-20, 8.2166e-21, 6.6827e-21, 2.8336e-23, 1.8682e-21, 4.2600e-20,\n 4.8901e-19, 9.8489e-20, 4.0810e-19, 8.2657e-20, 4.3447e-23, 6.0424e-19,\n 1.7312e-19, 6.7014e-20, 2.6251e-19, 8.9179e-20, 9.3708e-20, 1.2482e-19,\n 1.0755e-19, 1.0569e-23, 7.2611e-21, 1.6343e-20, 2.3504e-20, 7.1809e-19,\n 1.2334e-19, 8.6781e-21, 2.0550e-19, 1.4397e-20, 8.1815e-20, 2.3176e-22,\n 6.1011e-21, 2.0368e-19, 3.5262e-19, 2.8998e-19, 1.5304e-20, 2.1983e-20,\n 2.1581e-19, 3.5967e-19, 5.2028e-20, 6.4128e-20, 5.0168e-20, 7.0716e-20,\n 1.1262e-19, 5.6534e-20, 8.9307e-20, 6.3237e-20, 3.5134e-19, 4.9275e-19,\n 4.4867e-20, 1.9377e-19, 1.5101e-20, 8.6036e-21, 2.6790e-19, 2.2852e-20,\n 2.5497e-19, 3.9028e-20, 5.4094e-21, 1.8291e-20, 1.6420e-20, 3.2896e-20,\n 1.4865e-20, 2.0714e-20, 3.9002e-21, 4.5802e-20, 1.1661e-18, 3.9423e-22,\n 2.9862e-20, 2.8655e-20, 1.2396e-20, 2.4951e-19, 4.9397e-20, 6.8165e-21,\n 5.9421e-20, 1.0724e-19, 7.4112e-22, 4.5879e-19, 3.6757e-20, 2.5422e-21,\n 2.3965e-19, 5.6755e-19, 7.5761e-23, 5.1035e-20, 4.0555e-19, 8.1433e-20,\n 7.6417e-20, 3.2082e-19, 7.7660e-20, 2.2322e-20, 3.1321e-23, 4.1367e-19,\n 6.1944e-21, 5.2608e-19, 1.3724e-21, 1.1500e-19, 6.4573e-22, 1.1655e-19,\n 1.0246e-19, 5.5612e-22, 2.0604e-20, 1.5043e-20, 1.8396e-20, 1.0716e-18,\n 4.5361e-20, 1.2989e-19, 3.1037e-20, 4.9784e-21, 7.0733e-24, 5.2300e-21,\n 1.1511e-20, 3.8016e-20, 5.3778e-21, 3.9046e-24, 3.2477e-19, 2.1011e-20,\n 1.6617e-20, 1.1522e-19, 9.8581e-20, 8.2520e-20, 7.4397e-20, 2.8707e-20,\n 2.2341e-20, 2.0017e-22, 1.1031e-19, 3.5675e-20, 1.7388e-20, 3.0610e-22,\n 3.5454e-22, 5.4124e-19, 1.7321e-20, 6.1284e-19, 1.0157e-19, 1.1229e-20,\n 3.6459e-19, 4.5913e-20, 9.5416e-20, 6.4988e-20], device='cuda:0')" + "59": { + "step": "tensor(11268.)", + "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 device='cuda:0')", + "exp_avg_sq": "tensor([2.7387e-09, 2.4830e-11, 3.3537e-11, 3.4802e-11, 3.9901e-11, 4.5835e-11,\n 3.6441e-11, 2.6986e-11, 3.4086e-11, 3.0290e-11], device='cuda:0')" }, - "29": { - "step": "tensor(23788.)", - "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([2.2091e-19, 1.2139e-20, 6.5983e-20, 1.1237e-19, 1.8072e-20, 1.2553e-20,\n 7.9719e-20, 5.7111e-21, 3.2911e-20, 3.5978e-21, 1.9421e-19, 7.4323e-19,\n 1.1577e-19, 1.6594e-19, 1.9351e-19, 7.5636e-19, 1.0165e-19, 6.6022e-20,\n 2.1712e-19, 3.3129e-19, 3.7330e-24, 3.8912e-20, 8.0618e-20, 9.5326e-20,\n 9.3031e-20, 1.1758e-19, 1.2572e-19, 2.9625e-20, 5.1205e-21, 2.8712e-20,\n 2.2115e-19, 1.3901e-19, 2.6687e-20, 9.0227e-19, 3.8939e-19, 2.4973e-19,\n 2.8215e-21, 6.1046e-23, 1.0297e-19, 2.7198e-18, 3.8909e-19, 7.9052e-20,\n 2.6680e-20, 1.8115e-20, 4.4845e-19, 7.9613e-19, 1.6250e-19, 3.5415e-19,\n 2.6471e-20, 2.1885e-19, 2.1040e-21, 5.6763e-19, 1.1201e-19, 1.0523e-19,\n 1.2557e-20, 4.1450e-21, 2.5311e-19, 5.6151e-21, 1.8203e-19, 1.9066e-20,\n 5.8122e-20, 1.8662e-19, 4.1527e-19, 2.2739e-19, 2.3755e-21, 1.0239e-20,\n 1.7311e-19, 5.4043e-21, 1.0705e-19, 2.2797e-20, 2.1331e-21, 1.1281e-19,\n 3.3244e-19, 3.9611e-20, 6.1889e-20, 1.1878e-21, 7.6008e-21, 3.5478e-19,\n 1.0989e-20, 1.5414e-22, 2.1303e-19, 2.5917e-19, 5.5571e-22, 2.7815e-20,\n 1.6371e-19, 6.3415e-19, 5.3475e-20, 2.2052e-19, 3.4472e-20, 3.6854e-19,\n 2.7316e-19, 1.6229e-19, 3.2421e-19, 5.4711e-20, 1.3079e-19, 1.3366e-20,\n 5.0880e-20, 3.9223e-19, 1.2315e-19, 8.6589e-19, 8.2202e-19, 2.7964e-20,\n 8.1661e-22, 7.8756e-20, 1.1381e-21, 6.1596e-19, 3.4659e-20, 5.9427e-20,\n 1.1726e-19, 2.3036e-23, 6.4924e-20, 6.5471e-22, 4.9577e-19, 4.1502e-19,\n 2.9272e-20, 3.5490e-19, 4.2252e-20, 5.2485e-20, 1.1167e-19, 1.0660e-19,\n 2.7977e-20, 6.6464e-21, 7.6999e-21, 7.3893e-23, 4.4079e-21, 7.1431e-20,\n 8.1469e-19, 1.4652e-19, 3.5300e-19, 1.1816e-19, 2.6422e-24, 9.3155e-19,\n 1.7685e-19, 1.5519e-19, 2.8684e-19, 1.6848e-19, 2.3843e-19, 2.7058e-19,\n 2.1896e-19, 2.2739e-23, 1.0821e-20, 1.9111e-20, 6.7996e-20, 8.8058e-19,\n 2.0225e-19, 1.6412e-20, 2.0701e-19, 2.8835e-20, 1.5593e-19, 1.8902e-21,\n 1.3359e-20, 2.6485e-19, 2.5977e-19, 2.3755e-19, 4.6383e-20, 1.8623e-20,\n 1.3948e-19, 4.9266e-19, 7.3168e-20, 1.3755e-19, 1.3072e-19, 1.4350e-19,\n 6.2190e-20, 1.3116e-19, 1.2923e-19, 8.8096e-20, 7.1266e-19, 5.1721e-19,\n 1.0546e-19, 3.6741e-19, 2.0415e-20, 2.0131e-20, 4.3617e-19, 5.3731e-20,\n 2.9225e-19, 3.9914e-20, 7.1935e-21, 3.9670e-20, 2.1635e-20, 4.7246e-20,\n 3.2783e-20, 2.7657e-20, 9.2380e-21, 7.6647e-20, 1.4875e-18, 7.3903e-22,\n 6.2850e-20, 4.9457e-20, 3.2165e-20, 4.6943e-19, 1.2254e-19, 1.2707e-20,\n 7.9715e-20, 1.1718e-19, 2.7181e-21, 8.9449e-19, 2.9401e-20, 4.3572e-21,\n 2.4047e-19, 8.5628e-19, 6.9536e-22, 1.0265e-19, 3.8361e-19, 6.8065e-20,\n 1.0130e-19, 6.2904e-19, 1.3416e-19, 2.6240e-20, 2.6692e-22, 7.0484e-19,\n 8.5945e-21, 7.8053e-19, 1.2942e-21, 1.8905e-19, 5.1163e-22, 1.9137e-19,\n 2.5124e-19, 6.5362e-22, 3.4118e-20, 2.7287e-20, 1.5327e-20, 9.8176e-19,\n 6.2159e-20, 2.2968e-19, 5.8402e-20, 6.5448e-21, 1.1020e-23, 8.4959e-21,\n 2.1985e-20, 7.4153e-20, 9.6015e-21, 1.1625e-22, 3.8440e-19, 3.0992e-20,\n 2.2264e-20, 1.4764e-19, 7.8974e-20, 1.8075e-19, 8.2646e-20, 3.7704e-20,\n 1.7590e-20, 4.6939e-22, 1.5104e-19, 7.4131e-20, 3.1844e-20, 3.7238e-22,\n 1.5468e-22, 5.6402e-19, 1.7984e-20, 7.7503e-19, 2.0109e-19, 1.7678e-20,\n 2.9458e-19, 8.1491e-20, 2.0144e-19, 5.5838e-20], device='cuda:0')" + "60": { + "step": "tensor(11268.)", + "exp_avg": "tensor([[-5.6052e-45, -5.6052e-45, -5.6052e-45, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n ...,\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45]], device='cuda:0')", + "exp_avg_sq": "tensor([[1.0625e-10, 2.3049e-11, 6.3773e-11, ..., 2.7615e-11, 2.3642e-12,\n 3.6274e-13],\n [9.5738e-13, 2.4950e-13, 5.6642e-13, ..., 2.2765e-13, 2.2551e-14,\n 4.4821e-15],\n [1.3039e-12, 2.7514e-13, 7.7805e-13, ..., 3.4471e-13, 2.9168e-14,\n 4.1816e-15],\n ...,\n [1.0660e-12, 2.2477e-13, 6.4349e-13, ..., 2.7396e-13, 2.5260e-14,\n 5.2502e-15],\n [1.3206e-12, 2.8517e-13, 7.9380e-13, ..., 3.4357e-13, 2.4486e-14,\n 2.4563e-15],\n [1.1719e-12, 2.4725e-13, 7.3292e-13, ..., 3.1371e-13, 2.8304e-14,\n 7.1725e-15]], device='cuda:0')" }, - "30": { - "step": "tensor(23788.)", - "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([[2.0342e-20, 6.8698e-20, 0.0000e+00, ..., 2.3632e-19, 1.3447e-19,\n 1.4376e-19],\n [3.2496e-20, 1.2137e-20, 0.0000e+00, ..., 1.0151e-20, 1.1758e-19,\n 5.3023e-21],\n [2.6467e-20, 4.6703e-20, 0.0000e+00, ..., 3.2487e-20, 2.2039e-19,\n 1.3030e-20],\n ...,\n [6.3986e-21, 1.7420e-20, 0.0000e+00, ..., 2.2156e-20, 3.2931e-20,\n 2.0330e-22],\n [1.4610e-19, 1.2988e-19, 0.0000e+00, ..., 1.8886e-19, 8.0559e-19,\n 1.6423e-19],\n [1.1795e-21, 1.5770e-21, 0.0000e+00, ..., 6.0829e-21, 1.4226e-20,\n 1.4803e-22]], device='cuda:0')" + "61": { + "step": "tensor(11268.)", + "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 device='cuda:0')", + "exp_avg_sq": "tensor([2.7387e-09, 2.4830e-11, 3.3537e-11, 3.4802e-11, 3.9901e-11, 4.5835e-11,\n 3.6441e-11, 2.6986e-11, 3.4086e-11, 3.0291e-11], device='cuda:0')" }, - "31": { - "step": "tensor(23788.)", - "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([1.2605e-16, 3.9183e-17, 4.9822e-17, 1.5505e-18, 1.4088e-17, 1.4902e-18,\n 1.5284e-17, 1.3071e-18, 8.9311e-18, 5.7242e-18, 1.1759e-16, 1.1591e-16,\n 1.5908e-17, 1.1484e-17, 2.2471e-17, 3.0399e-18, 5.7070e-17, 1.6976e-17,\n 5.7005e-17, 1.5872e-16, 2.6107e-19, 4.5497e-17, 1.0025e-17, 9.3139e-17,\n 6.8698e-18, 1.5086e-17, 9.9751e-17, 1.6126e-18, 5.6637e-18, 1.7293e-17,\n 2.9999e-17, 9.8750e-17, 1.3050e-19, 1.7773e-16, 1.0226e-16, 1.2139e-16,\n 4.6044e-18, 1.6411e-19, 6.5965e-17, 4.3556e-16, 1.8306e-16, 5.1550e-18,\n 4.6850e-18, 4.7224e-18, 3.3288e-17, 7.5986e-17, 4.6311e-17, 2.1403e-17,\n 4.6619e-18, 6.4388e-17, 1.7258e-18, 1.6124e-16, 8.0813e-17, 5.3842e-17,\n 2.3834e-18, 4.9501e-18, 3.6238e-17, 4.2395e-18, 2.6866e-17, 8.4946e-18,\n 1.0380e-17, 2.6865e-17, 4.4221e-17, 1.0877e-16, 1.1395e-18, 2.7112e-17,\n 2.6112e-17, 6.3131e-18, 5.2017e-17, 7.5520e-18, 5.2963e-20, 2.1045e-17,\n 8.6369e-17, 8.4174e-18, 5.1543e-18, 6.6719e-20, 1.3549e-17, 1.5159e-16,\n 5.9311e-18, 4.2565e-20, 6.1765e-17, 1.9725e-16, 8.8160e-20, 1.4508e-18,\n 2.4498e-17, 1.4603e-16, 6.5931e-18, 2.9950e-17, 2.1910e-17, 5.6086e-18,\n 3.1027e-17, 1.1187e-16, 3.4079e-17, 3.9539e-17, 1.2680e-16, 3.1604e-17,\n 5.6038e-17, 1.9013e-16, 2.9636e-17, 2.2999e-16, 2.1424e-17, 4.3877e-18,\n 8.5480e-21, 2.2686e-17, 2.4903e-19, 2.7415e-16, 7.8473e-18, 3.1013e-17,\n 2.3582e-17, 1.6392e-19, 2.0787e-16, 9.1951e-18, 6.4571e-17, 9.7242e-17,\n 1.6990e-17, 1.3539e-16, 2.9183e-18, 1.0750e-17, 1.6820e-17, 9.0269e-17,\n 9.4137e-18, 7.2079e-18, 4.0130e-18, 6.7389e-20, 6.6708e-19, 4.8501e-17,\n 8.1164e-17, 1.1958e-17, 4.0221e-17, 8.2429e-17, 8.9999e-19, 1.8482e-16,\n 4.5135e-17, 6.7570e-18, 7.8236e-17, 2.9599e-17, 1.4810e-16, 2.8699e-17,\n 2.2616e-17, 1.6271e-20, 5.7674e-20, 5.5749e-19, 1.0143e-16, 3.7605e-16,\n 6.5848e-17, 1.2964e-17, 1.5908e-17, 2.5168e-17, 2.2226e-17, 1.8992e-18,\n 5.8146e-18, 5.1511e-17, 3.7355e-17, 9.1426e-18, 1.1189e-17, 6.6184e-18,\n 1.7798e-17, 2.0032e-16, 8.8640e-18, 9.6022e-18, 3.1733e-17, 9.9342e-18,\n 9.4593e-18, 2.8340e-17, 8.9945e-18, 4.2208e-17, 1.9354e-16, 1.4002e-17,\n 6.6099e-17, 1.1985e-17, 6.0750e-19, 2.1949e-17, 4.0566e-17, 6.3456e-18,\n 8.2106e-17, 1.3958e-17, 9.6920e-18, 9.2573e-18, 1.0319e-17, 2.7900e-17,\n 3.3689e-17, 1.0052e-17, 1.5039e-17, 2.4975e-17, 2.2067e-16, 1.2896e-19,\n 1.7302e-17, 6.5040e-17, 1.1550e-17, 2.1194e-16, 3.6818e-17, 2.1679e-17,\n 1.2464e-17, 4.3462e-18, 1.9068e-20, 5.8773e-16, 8.2119e-18, 1.7530e-18,\n 1.7226e-17, 8.0592e-17, 9.2498e-20, 2.3909e-16, 1.6480e-17, 2.0566e-17,\n 2.0205e-17, 5.5519e-17, 1.0079e-16, 2.0525e-18, 4.6943e-20, 3.8476e-17,\n 1.1115e-17, 2.8256e-17, 1.0530e-18, 7.0320e-17, 1.3926e-18, 1.5313e-16,\n 1.7768e-16, 6.9937e-19, 1.7498e-18, 1.7800e-17, 7.2504e-18, 9.9207e-17,\n 9.2259e-18, 1.2012e-16, 1.6201e-17, 2.2239e-21, 2.0602e-20, 1.5884e-18,\n 2.4653e-18, 2.5961e-17, 6.8332e-19, 2.3051e-18, 2.6820e-17, 1.5559e-17,\n 1.0618e-16, 8.1725e-19, 7.4622e-18, 4.9710e-17, 1.3888e-17, 1.6269e-17,\n 2.9501e-17, 1.2030e-19, 1.1681e-17, 4.1676e-17, 1.7126e-17, 4.1320e-19,\n 9.6300e-19, 7.9600e-17, 5.5170e-18, 1.7814e-17, 1.4326e-16, 2.1945e-18,\n 8.9349e-17, 7.9002e-18, 2.1272e-16, 5.0879e-18], device='cuda:0')" + "8": { + "step": "tensor(10016.)", + "exp_avg": "tensor([[ 5.2391e-07, -9.5890e-08, -1.6794e-07, ..., 1.0383e-06,\n 0.0000e+00, 1.6587e-07],\n [-2.9076e-07, -5.4219e-09, -8.0596e-06, ..., -3.0523e-07,\n 0.0000e+00, -1.1051e-07],\n [ 3.0308e-07, 1.0616e-08, 2.8124e-07, ..., -8.6895e-08,\n 0.0000e+00, -2.3080e-07],\n ...,\n [-7.1909e-08, 1.7884e-08, -1.0837e-07, ..., 4.0920e-06,\n 0.0000e+00, 2.0505e-07],\n [-7.4948e-07, 1.5425e-07, 8.0185e-07, ..., 2.5868e-06,\n 0.0000e+00, 1.0694e-06],\n [-2.2005e-07, 6.5157e-08, 1.4630e-06, ..., 1.3091e-06,\n 0.0000e+00, 4.5262e-06]], device='cuda:0')", + "exp_avg_sq": "tensor([[3.7811e-12, 3.4363e-13, 2.4449e-11, ..., 1.9803e-11, 0.0000e+00,\n 5.4632e-12],\n [1.7406e-11, 1.2091e-13, 2.4161e-11, ..., 2.5411e-12, 0.0000e+00,\n 1.0146e-11],\n [3.4680e-12, 7.8053e-13, 2.0143e-11, ..., 3.2588e-12, 0.0000e+00,\n 2.1691e-12],\n ...,\n [2.2858e-12, 7.1243e-13, 6.1150e-12, ..., 3.3772e-11, 0.0000e+00,\n 8.1645e-12],\n [9.0069e-12, 5.4013e-13, 7.2687e-11, ..., 5.8154e-11, 0.0000e+00,\n 4.7426e-11],\n [1.5617e-11, 3.0288e-12, 3.6350e-11, ..., 3.1824e-11, 0.0000e+00,\n 2.1300e-11]], device='cuda:0')" }, - "32": { - "step": "tensor(23788.)", - "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([2.7169e-19, 1.2559e-19, 1.3917e-19, 8.2366e-21, 3.1206e-20, 4.4236e-21,\n 2.2293e-20, 1.0638e-20, 1.2186e-20, 1.3740e-20, 2.1649e-19, 1.9057e-19,\n 3.2926e-20, 2.9102e-20, 7.4197e-20, 2.2074e-20, 1.3619e-19, 2.9543e-20,\n 1.3948e-19, 3.7740e-19, 1.2561e-22, 1.5972e-19, 4.7782e-20, 1.9619e-19,\n 2.3059e-20, 1.7458e-20, 4.7634e-19, 4.9650e-21, 1.9698e-20, 3.3192e-20,\n 4.5388e-20, 7.1928e-19, 1.2272e-21, 3.0388e-19, 3.0411e-19, 4.6232e-19,\n 5.7660e-21, 1.8592e-22, 1.9245e-19, 1.1025e-18, 5.4139e-19, 8.5733e-21,\n 1.2483e-20, 2.1498e-20, 4.3155e-20, 1.3629e-19, 1.9104e-19, 5.1539e-20,\n 5.9502e-21, 1.4953e-19, 1.8848e-21, 4.9397e-19, 1.6795e-19, 1.1728e-19,\n 2.7889e-21, 3.5299e-20, 7.8072e-20, 4.6915e-21, 1.3601e-19, 2.9289e-20,\n 2.3871e-20, 1.1018e-19, 1.4039e-19, 4.8845e-19, 1.3738e-21, 7.8898e-20,\n 7.1148e-20, 1.0123e-20, 1.2082e-19, 1.6373e-20, 1.5341e-23, 7.6751e-20,\n 2.2078e-19, 2.7833e-20, 7.7925e-21, 2.5230e-22, 9.6315e-20, 4.4063e-19,\n 1.2350e-20, 2.2798e-24, 3.6056e-19, 1.0856e-18, 5.3537e-22, 8.1488e-21,\n 1.1023e-19, 4.2644e-19, 2.5246e-20, 1.1687e-19, 4.6669e-20, 2.9454e-20,\n 8.4438e-20, 2.6808e-19, 4.9348e-20, 1.0442e-19, 4.3344e-19, 6.3058e-20,\n 1.4988e-19, 4.3517e-19, 7.5544e-20, 6.6616e-19, 4.7370e-20, 2.6588e-20,\n 1.2197e-21, 4.3457e-20, 1.2640e-22, 8.4406e-19, 1.5592e-20, 1.5716e-19,\n 3.7899e-20, 6.0298e-26, 6.3663e-19, 2.7074e-20, 1.6271e-19, 1.6924e-19,\n 3.0252e-20, 4.2532e-19, 9.1333e-21, 1.7590e-20, 3.3506e-20, 3.9016e-19,\n 1.1472e-20, 1.0327e-20, 4.2579e-21, 3.1528e-21, 6.1918e-21, 1.6057e-19,\n 1.4533e-19, 3.0085e-20, 1.2379e-19, 3.1077e-19, 4.6900e-23, 3.4686e-19,\n 1.7416e-19, 2.0684e-20, 1.2967e-19, 4.2595e-20, 5.0546e-19, 6.4746e-20,\n 5.7283e-20, 2.5056e-24, 2.3699e-23, 2.7438e-21, 2.4375e-19, 1.5563e-18,\n 2.2821e-19, 4.9287e-20, 3.9041e-20, 1.5029e-19, 2.6306e-20, 1.8929e-21,\n 3.8123e-20, 1.0518e-19, 1.4782e-19, 1.4914e-20, 1.4754e-20, 5.2890e-20,\n 3.1497e-20, 5.0893e-19, 1.3392e-20, 4.1801e-20, 7.0324e-20, 1.4770e-20,\n 1.1448e-20, 9.6581e-20, 2.1572e-20, 9.1644e-20, 3.6271e-19, 2.6829e-20,\n 3.1827e-19, 3.9476e-20, 1.9140e-22, 4.1195e-20, 6.5909e-20, 1.9369e-20,\n 2.3305e-19, 9.6763e-20, 1.4277e-20, 2.9879e-20, 2.1557e-20, 5.9651e-20,\n 7.6313e-20, 1.5440e-20, 3.2082e-20, 4.0925e-20, 5.1234e-19, 8.3211e-24,\n 2.9163e-20, 2.3724e-19, 1.3212e-20, 4.6239e-19, 6.1972e-20, 5.3370e-20,\n 4.8227e-20, 1.3660e-20, 4.9193e-23, 2.4078e-18, 8.8217e-21, 1.2990e-21,\n 3.5947e-20, 1.5446e-19, 3.1873e-21, 1.0499e-18, 2.8915e-20, 5.6764e-20,\n 4.0294e-20, 1.2392e-19, 2.6001e-19, 3.6069e-21, 3.3868e-24, 6.6959e-20,\n 2.2924e-20, 6.4017e-20, 3.3233e-21, 2.4627e-19, 1.0056e-21, 4.4384e-19,\n 4.1866e-19, 8.4740e-23, 4.8704e-21, 3.8155e-20, 1.4128e-20, 2.4117e-19,\n 3.1186e-20, 2.5544e-19, 4.4394e-20, 2.0168e-22, 1.4955e-22, 5.8111e-21,\n 3.5025e-21, 7.8379e-20, 2.0460e-21, 1.3246e-21, 9.2817e-20, 5.4432e-20,\n 3.3469e-19, 5.5431e-21, 1.1026e-20, 1.0617e-19, 5.9340e-20, 2.9796e-20,\n 5.7950e-20, 6.5492e-23, 3.3165e-20, 1.1881e-19, 2.8005e-20, 3.2633e-23,\n 4.1387e-22, 1.3024e-19, 8.7178e-21, 6.0864e-20, 4.2130e-19, 2.0514e-21,\n 3.2038e-19, 1.4888e-20, 5.6727e-19, 1.6037e-20], device='cuda:0')" + "9": { + "step": "tensor(10016.)", + "exp_avg": "tensor([ 3.6428e-06, -1.5594e-05, 4.7698e-06, ..., 5.7057e-06,\n 5.1658e-06, 3.8920e-05], device='cuda:0')", + "exp_avg_sq": "tensor([1.4399e-09, 1.6370e-09, 1.0969e-09, ..., 1.1355e-09, 1.5990e-09,\n 1.5532e-09], device='cuda:0')" }, - "33": { - "step": "tensor(23788.)", - "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([5.2110e-19, 1.8373e-19, 1.4600e-19, 9.4857e-21, 3.0470e-20, 1.1847e-20,\n 4.1951e-20, 1.3738e-20, 2.0549e-20, 9.5579e-21, 3.8392e-19, 4.1249e-19,\n 4.9525e-20, 5.4892e-20, 1.1672e-19, 1.2634e-20, 2.5575e-19, 4.3511e-20,\n 1.8332e-19, 6.6175e-19, 1.0354e-22, 2.1100e-19, 5.8984e-20, 3.9137e-19,\n 3.8216e-20, 4.7690e-20, 4.4772e-19, 9.4546e-21, 3.4298e-20, 4.5341e-20,\n 1.4244e-19, 4.6233e-19, 1.3802e-21, 6.0864e-19, 4.3362e-19, 5.3678e-19,\n 9.6907e-21, 3.3372e-22, 2.9399e-19, 1.5253e-18, 6.1286e-19, 3.0358e-20,\n 2.7775e-20, 2.9448e-20, 1.1226e-19, 2.5363e-19, 2.2361e-19, 1.0446e-19,\n 1.2366e-20, 2.1798e-19, 1.6229e-21, 6.8560e-19, 2.7467e-19, 1.6334e-19,\n 4.7417e-21, 3.7842e-20, 1.6091e-19, 1.0419e-20, 1.3877e-19, 4.6549e-20,\n 2.3105e-20, 1.3722e-19, 2.1077e-19, 4.8416e-19, 1.8953e-21, 7.0933e-20,\n 1.2985e-19, 1.3530e-20, 2.2495e-19, 3.8473e-20, 3.8379e-25, 1.0545e-19,\n 2.7866e-19, 4.7816e-20, 1.4305e-20, 1.1596e-21, 7.8501e-20, 6.2653e-19,\n 9.0462e-21, 6.2130e-23, 2.9253e-19, 8.5711e-19, 7.3105e-22, 1.3076e-20,\n 1.2438e-19, 6.1024e-19, 3.7705e-20, 1.4830e-19, 7.1392e-20, 2.4722e-20,\n 1.4274e-19, 4.9429e-19, 1.1093e-19, 1.2607e-19, 5.5312e-19, 9.0995e-20,\n 1.7132e-19, 6.4958e-19, 1.4804e-19, 9.6625e-19, 1.0330e-19, 2.8760e-20,\n 2.5848e-21, 6.3403e-20, 7.6599e-23, 9.2509e-19, 1.6085e-20, 1.5440e-19,\n 6.8975e-20, 7.9873e-23, 6.9976e-19, 4.9991e-20, 2.9088e-19, 3.3434e-19,\n 4.2242e-20, 4.4930e-19, 1.5600e-20, 3.0379e-20, 7.5954e-20, 4.1507e-19,\n 2.8614e-20, 2.0619e-20, 7.5596e-21, 1.6687e-21, 7.6944e-21, 1.4259e-19,\n 3.4150e-19, 6.0255e-20, 1.8624e-19, 2.4841e-19, 2.5314e-22, 7.5724e-19,\n 2.1202e-19, 3.5421e-20, 2.5683e-19, 9.5663e-20, 6.2655e-19, 9.9008e-20,\n 1.0630e-19, 1.0203e-22, 2.2093e-22, 4.8454e-21, 3.1129e-19, 1.2674e-18,\n 2.8930e-19, 6.8556e-20, 7.5822e-20, 1.3473e-19, 7.3815e-20, 2.0139e-21,\n 3.9226e-20, 1.6379e-19, 1.8172e-19, 2.7640e-20, 3.4310e-20, 4.7875e-20,\n 5.0015e-20, 6.8420e-19, 2.5287e-20, 4.3082e-20, 1.0603e-19, 2.9140e-20,\n 2.3258e-20, 1.3024e-19, 4.6243e-20, 1.3222e-19, 6.8151e-19, 4.6303e-20,\n 3.1405e-19, 5.7876e-20, 1.0564e-21, 6.2592e-20, 1.4310e-19, 3.4651e-20,\n 2.6573e-19, 8.5301e-20, 2.4391e-20, 5.2763e-20, 2.1696e-20, 1.2676e-19,\n 9.2298e-20, 2.6461e-20, 3.6043e-20, 7.1761e-20, 9.3201e-19, 6.8855e-23,\n 5.4442e-20, 1.8746e-19, 3.1762e-20, 8.5946e-19, 1.2412e-19, 6.2928e-20,\n 6.3996e-20, 2.7821e-20, 4.6681e-23, 2.3742e-18, 1.9315e-20, 3.2228e-21,\n 8.5376e-20, 2.8905e-19, 5.5977e-21, 1.0293e-18, 5.8415e-20, 5.1882e-20,\n 9.6019e-20, 2.0940e-19, 3.0982e-19, 5.8664e-21, 1.6054e-22, 1.3794e-19,\n 2.4087e-20, 1.1836e-19, 7.9813e-21, 3.0959e-19, 1.1289e-21, 5.0531e-19,\n 6.0141e-19, 1.0368e-22, 9.9281e-21, 4.4489e-20, 1.1101e-20, 4.3267e-19,\n 5.1924e-20, 5.1450e-19, 7.7661e-20, 5.3449e-22, 1.2244e-22, 1.2269e-20,\n 5.8584e-21, 1.2562e-19, 4.5439e-21, 2.6966e-21, 1.2180e-19, 8.0367e-20,\n 4.6112e-19, 4.1888e-21, 1.7956e-20, 2.2307e-19, 7.5950e-20, 4.3071e-20,\n 1.3257e-19, 6.8388e-23, 6.2382e-20, 1.4713e-19, 5.0968e-20, 1.4819e-24,\n 5.7536e-22, 2.5850e-19, 1.2303e-20, 8.1913e-20, 4.8158e-19, 4.4638e-21,\n 2.6687e-19, 2.4524e-20, 7.3949e-19, 3.1321e-20], device='cuda:0')" + "10": { + "step": "tensor(10016.)", + "exp_avg": "tensor([[ 3.9524e-08, -9.2433e-08, 1.0221e-07, ..., -4.4198e-07,\n 7.6784e-07, 2.0899e-07],\n [ 3.2126e-07, 3.1007e-07, 2.6399e-07, ..., 3.7226e-07,\n -5.8755e-07, 5.4598e-07],\n [-2.2047e-07, 3.0856e-07, 1.3871e-07, ..., -1.7447e-07,\n -9.6529e-07, -2.3211e-07],\n ...,\n [-1.7317e-07, -1.6380e-07, -9.4509e-08, ..., -3.1332e-08,\n 3.2863e-08, 1.8801e-07],\n [-2.4867e-07, -1.4015e-07, -1.4686e-07, ..., 6.6397e-07,\n -3.8798e-07, -2.3568e-07],\n [ 2.6323e-07, -3.6754e-07, -1.6236e-07, ..., 4.9175e-07,\n -6.2566e-07, -2.3958e-07]], device='cuda:0')", + "exp_avg_sq": "tensor([[6.7076e-13, 7.1408e-13, 1.1768e-12, ..., 1.8106e-12, 1.0345e-12,\n 6.2712e-13],\n [1.3839e-12, 1.6290e-12, 1.3338e-12, ..., 2.2275e-12, 1.6904e-12,\n 1.4582e-12],\n [1.5684e-12, 2.2788e-12, 1.8632e-12, ..., 1.9984e-12, 2.4658e-12,\n 1.2703e-12],\n ...,\n [1.9778e-12, 1.1253e-12, 1.7830e-12, ..., 2.8008e-12, 2.9287e-12,\n 1.3667e-12],\n [1.6394e-12, 1.3856e-12, 2.5231e-12, ..., 2.5221e-12, 3.2398e-12,\n 1.3361e-12],\n [2.2460e-12, 1.5797e-12, 1.4107e-12, ..., 2.1398e-12, 1.8176e-12,\n 2.2789e-12]], device='cuda:0')" }, - "34": { - "step": "tensor(23788.)", - "exp_avg": "tensor([[-5.6052e-45, -5.6052e-45, -5.6052e-45, ..., -5.6052e-45,\n -5.6052e-45, -5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n ...,\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.8574e-21, 6.7067e-21, 4.0915e-21, ..., 9.2602e-22, 5.0701e-21,\n 3.0467e-21],\n [1.3220e-22, 1.9777e-23, 2.1555e-22, ..., 5.7523e-22, 6.6871e-22,\n 2.3834e-23],\n [6.4438e-22, 2.8340e-21, 8.9360e-22, ..., 4.5726e-22, 2.7536e-21,\n 5.2688e-22],\n ...,\n [1.6124e-20, 3.8247e-20, 3.9637e-20, ..., 7.0432e-21, 5.9469e-20,\n 8.6067e-20],\n [1.0767e-19, 1.7030e-19, 2.6285e-19, ..., 5.8368e-20, 2.5718e-19,\n 4.7219e-19],\n [1.4773e-18, 3.2767e-18, 3.9155e-18, ..., 8.3355e-19, 4.9318e-18,\n 7.7376e-18]], device='cuda:0')" + "11": { + "step": "tensor(8764.)", + "exp_avg": "tensor([[-4.5763e-07, -5.1912e-09, -6.4911e-08, ..., -4.0983e-08,\n 0.0000e+00, -3.3231e-07],\n [-7.6240e-09, -1.2870e-07, 7.0829e-08, ..., 4.1722e-07,\n 0.0000e+00, 3.3970e-06],\n [ 1.5098e-07, 8.2674e-10, 4.2594e-07, ..., 4.2718e-07,\n 0.0000e+00, 3.4151e-07],\n ...,\n [ 3.4169e-08, 6.1919e-09, -2.5748e-07, ..., 1.3139e-07,\n 0.0000e+00, -2.2114e-07],\n [-2.9416e-07, 4.2277e-07, 9.9381e-07, ..., 1.4964e-06,\n 0.0000e+00, -3.6192e-07],\n [-2.7604e-07, 1.4446e-10, -8.1898e-08, ..., 3.2305e-08,\n 0.0000e+00, 3.6308e-08]], device='cuda:0')", + "exp_avg_sq": "tensor([[1.3432e-11, 8.3001e-12, 4.1771e-12, ..., 4.1365e-12, 0.0000e+00,\n 7.5649e-12],\n [2.2236e-12, 8.0298e-12, 7.6686e-13, ..., 2.2507e-12, 0.0000e+00,\n 5.9823e-11],\n [3.2761e-12, 6.3425e-13, 1.8211e-11, ..., 7.7775e-11, 0.0000e+00,\n 5.9067e-12],\n ...,\n [3.2106e-13, 3.9059e-13, 7.5758e-12, ..., 2.1005e-12, 0.0000e+00,\n 1.0711e-11],\n [7.7567e-12, 7.0507e-12, 2.6024e-11, ..., 7.7393e-11, 0.0000e+00,\n 5.7496e-11],\n [1.3668e-12, 5.0223e-15, 1.4851e-12, ..., 7.4668e-12, 0.0000e+00,\n 2.2968e-12]], device='cuda:0')" }, - "35": { - "step": "tensor(23788.)", - "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([3.7448e-20, 1.2035e-21, 1.0744e-20, 7.0702e-22, 4.7341e-21, 6.9988e-21,\n 1.0335e-21, 8.9402e-22, 4.3552e-23, 7.2966e-21, 5.9183e-21, 1.0236e-20,\n 2.5423e-22, 2.1522e-22, 2.6186e-21, 1.7175e-21, 5.6655e-22, 2.2424e-20,\n 7.6524e-21, 9.8012e-23, 3.8588e-22, 3.1313e-21, 7.2299e-21, 2.7889e-22,\n 1.3507e-20, 1.5841e-21, 2.2430e-21, 2.2871e-22, 5.2128e-22, 1.0663e-21,\n 1.0643e-21, 1.7701e-20, 2.5652e-21, 1.3731e-22, 5.1799e-24, 5.2922e-21,\n 1.8832e-21, 7.5870e-22, 2.2686e-23, 1.0115e-24, 3.0734e-21, 3.8461e-22,\n 1.8780e-21, 8.8674e-22, 1.5169e-22, 1.7269e-21, 2.1867e-21, 2.2186e-21,\n 1.1714e-21, 4.9777e-21, 1.9317e-21, 5.4941e-22, 7.1689e-22, 4.2217e-22,\n 1.5837e-23, 2.1532e-23, 4.9283e-22, 3.9556e-23, 1.5857e-21, 6.7437e-23,\n 5.6003e-21, 1.8677e-21, 4.4076e-24, 3.3979e-21, 6.5713e-21, 2.2382e-21,\n 3.3558e-21, 6.6446e-22, 8.5100e-23, 5.2456e-23, 3.4383e-21, 3.6320e-22,\n 2.8900e-23, 1.4199e-21, 8.2732e-22, 2.2301e-21, 1.6403e-21, 4.2478e-22,\n 7.0350e-27, 1.5404e-21, 1.1731e-21, 1.0326e-20, 1.5801e-22, 1.4598e-21,\n 1.9859e-22, 1.0020e-21, 5.8410e-22, 1.6163e-24, 9.9387e-22, 1.2899e-22,\n 4.6137e-24, 7.4857e-22, 6.6938e-22, 1.9076e-20, 6.2852e-21, 4.3708e-21,\n 4.7618e-22, 2.2470e-20, 1.5521e-21, 2.8784e-21, 2.6753e-22, 3.0775e-22,\n 1.5157e-22, 1.6462e-21, 2.0010e-23, 5.7408e-21, 9.4272e-21, 1.0479e-21,\n 1.1898e-21, 1.7888e-22, 1.8095e-21, 2.0548e-20, 1.5544e-22, 2.3540e-20,\n 3.2652e-21, 1.4546e-20, 1.4275e-27, 1.3343e-20, 9.7199e-21, 4.8105e-21,\n 1.8410e-20, 8.2361e-23, 9.6114e-21, 3.6242e-23, 5.2455e-21, 8.7370e-21,\n 7.9180e-21, 5.7351e-21, 1.0586e-20, 1.6020e-20, 1.8424e-21, 3.5933e-21,\n 1.0822e-22, 6.1749e-21, 1.7740e-21, 1.6226e-21, 3.7519e-21, 5.0801e-21,\n 4.2684e-24, 9.9112e-22, 1.2461e-20, 4.5857e-23, 2.6209e-21, 9.6700e-22,\n 6.0534e-21, 2.5181e-20, 5.0069e-21, 6.4306e-23, 1.5881e-20, 2.4083e-22,\n 5.1219e-22, 4.1609e-21, 1.7277e-20, 1.4200e-21, 6.1012e-21, 9.1533e-21,\n 6.7402e-23, 1.6998e-23, 6.2306e-22, 3.0631e-21, 2.9224e-21, 8.1048e-22,\n 1.2201e-20, 6.3108e-21, 7.0274e-22, 6.7808e-24, 1.3541e-22, 1.9184e-23,\n 3.8307e-22, 1.7250e-21, 3.4881e-22, 2.1161e-22, 1.4608e-21, 1.7697e-21,\n 1.6108e-22, 4.5194e-23, 1.6934e-21, 3.9097e-22, 2.4281e-22, 3.4606e-23,\n 2.8215e-21, 8.2070e-24, 6.1025e-22, 2.6344e-21, 7.4999e-22, 3.4971e-22,\n 2.6134e-21, 8.7735e-22, 8.8035e-21, 2.8743e-22, 2.4549e-21, 2.0235e-22,\n 2.2235e-21, 2.2080e-20, 3.1164e-21, 6.6408e-21, 3.3968e-21, 9.9290e-21,\n 1.1356e-24, 1.2058e-22, 5.7756e-22, 8.5752e-22, 1.4784e-21, 1.2179e-20,\n 3.0574e-22, 9.1543e-22, 6.8457e-23, 1.0311e-21, 1.1876e-22, 2.7579e-21,\n 4.7196e-25, 1.8546e-21, 1.2530e-22, 1.3660e-21, 1.2568e-20, 6.3713e-21,\n 9.3597e-21, 2.5313e-21, 3.2032e-21, 9.9649e-21, 2.5306e-21, 4.8323e-23,\n 1.2812e-21, 1.1397e-20, 7.2899e-23, 1.5046e-21, 9.1109e-21, 3.2392e-21,\n 2.8575e-21, 2.7646e-21, 2.3927e-21, 7.1240e-23, 1.0508e-21, 5.5510e-22,\n 2.5752e-21, 7.0677e-22, 1.7798e-23, 1.6327e-22, 6.5931e-22, 6.9434e-23,\n 6.3870e-21, 1.7616e-21, 2.6969e-22, 2.1349e-21, 6.5390e-23, 2.4416e-21,\n 6.5581e-21, 8.4148e-22, 2.7882e-21, 8.6917e-22, 6.6307e-22, 6.8190e-21,\n 2.6484e-21, 7.5456e-21, 7.0587e-21, 1.5300e-21, 2.4274e-37, 1.0731e-38,\n 4.3779e-36, 3.0781e-37, 4.9015e-36, 1.9264e-37, 1.2636e-37, 2.6768e-36,\n 1.7542e-36, 2.5533e-36, 7.7080e-38, 9.5473e-38, 2.6776e-36, 2.4064e-36,\n 1.3266e-37, 2.9232e-37, 1.0769e-36, 6.8608e-40, 6.8077e-36, 1.0775e-36,\n 3.5458e-37, 2.4220e-36, 1.0772e-36, 1.1058e-36, 8.0854e-37, 2.5825e-37,\n 6.3449e-37, 7.2615e-38, 5.4867e-36, 1.1277e-36, 9.7958e-37, 8.7247e-37,\n 4.7224e-36, 1.4305e-36, 1.5431e-37, 5.4681e-37, 8.6894e-37, 1.2399e-38,\n 1.1935e-37, 3.7439e-38, 2.5775e-39, 1.1520e-37, 5.7046e-37, 8.1893e-39,\n 9.3560e-37, 2.7506e-37, 1.4535e-36, 7.5101e-39, 1.1888e-36, 1.0241e-38,\n 3.1339e-37, 3.5793e-41, 4.4010e-37, 2.0951e-36, 6.0790e-37, 1.1667e-37,\n 4.5515e-36, 5.8554e-37, 8.4576e-36, 6.8054e-37, 1.9551e-38, 1.0875e-37,\n 9.8728e-37, 5.8535e-37, 1.5081e-37, 7.8599e-37, 2.8766e-36, 3.7694e-36,\n 2.0074e-36, 8.2489e-38, 7.0232e-36, 5.1195e-36, 5.3801e-36, 8.9797e-36,\n 1.8578e-35, 5.8416e-37, 1.1577e-35, 2.4621e-37, 8.5268e-37, 1.7004e-37,\n 2.5059e-36, 9.9260e-36, 5.9306e-36, 3.8642e-36, 1.0281e-36, 1.8595e-36,\n 9.0494e-36, 2.0352e-35, 1.1398e-36, 3.1218e-36, 1.7970e-36, 9.4099e-36,\n 9.8734e-37, 5.7694e-38, 6.5472e-37, 6.2508e-37, 5.8132e-36, 1.9763e-38,\n 1.1852e-36, 8.5707e-39, 6.7388e-37, 3.1819e-36, 4.6780e-36, 4.0091e-37,\n 1.0576e-37, 4.6199e-37, 7.9938e-37, 4.5844e-38, 2.8159e-38, 1.3163e-37,\n 1.5216e-37, 2.6019e-38, 6.0223e-37, 7.5274e-37, 1.0648e-36, 9.6570e-37,\n 6.8679e-37, 9.0692e-39, 3.0634e-40, 1.7648e-38, 1.2500e-38, 1.2985e-36,\n 1.8003e-36, 1.3177e-36, 6.2218e-37, 1.6881e-36, 8.3671e-39, 2.8289e-36,\n 3.0119e-38, 3.4594e-37, 1.4091e-37, 4.4903e-36, 1.7948e-36, 2.8063e-36,\n 2.3347e-36, 4.9450e-38, 4.3984e-38, 5.1921e-37, 1.0395e-36, 3.2173e-36,\n 3.4008e-37, 1.6134e-37, 9.2574e-38, 1.4022e-38, 5.2780e-36, 4.9643e-37,\n 1.4920e-36, 1.1242e-37, 1.7676e-37, 1.1433e-35, 1.9435e-36, 2.2098e-38,\n 1.5773e-36, 1.8724e-36, 5.4345e-36, 3.2305e-36, 4.1379e-36, 1.9795e-36,\n 2.9873e-38, 2.3295e-35, 5.2778e-36, 6.4728e-38, 6.3840e-36, 1.6218e-36,\n 1.4115e-36, 8.3120e-38, 9.5828e-37, 4.5679e-37, 4.7581e-37, 1.8701e-36,\n 1.4957e-36, 1.7798e-40, 1.3877e-36, 1.4027e-37, 2.1596e-38, 3.3005e-37,\n 3.3796e-36, 2.3916e-36, 1.2365e-38, 1.8837e-36, 5.4682e-37, 3.9274e-37,\n 6.5814e-36, 6.5919e-36, 6.9771e-37, 3.5856e-38, 6.7357e-36, 6.2409e-36,\n 2.9101e-36, 3.7013e-37, 3.7770e-37, 1.5000e-36, 1.7591e-36, 1.9700e-36,\n 9.6940e-36, 6.8235e-37, 2.1374e-37, 3.3747e-36, 6.5141e-37, 3.2557e-36,\n 5.1380e-37, 5.5694e-37, 5.2819e-37, 2.5159e-36, 4.9090e-36, 3.5400e-36,\n 2.3581e-38, 8.9415e-38, 1.1490e-36, 4.7474e-40, 3.4847e-37, 2.1864e-36,\n 2.3336e-37, 6.7714e-37, 9.9904e-37, 1.9940e-36, 6.2301e-37, 9.9126e-37,\n 2.0516e-36, 1.1228e-36, 9.6950e-38, 2.2479e-37, 7.1609e-38, 1.3014e-37,\n 8.8452e-37, 5.8657e-37, 1.6545e-36, 5.6822e-36, 3.5827e-36, 2.4586e-36,\n 1.7596e-36, 2.1015e-36, 7.5902e-38, 4.3534e-36, 1.3175e-37, 1.9854e-37,\n 9.3655e-37, 5.4307e-37, 9.8947e-37, 1.9989e-37, 1.5335e-37, 3.0782e-41,\n 9.2892e-38, 7.6868e-38, 1.4299e-36, 2.5419e-36, 1.9775e-37, 1.7085e-36,\n 7.5169e-37, 2.2466e-36, 1.8564e-38, 6.3087e-37, 1.7980e-37, 2.6616e-36,\n 4.1276e-37, 2.3801e-37, 4.9393e-18, 7.7323e-17, 1.3611e-17, 1.2980e-17,\n 1.0270e-16, 9.1250e-17, 1.2447e-18, 6.6345e-19, 1.6633e-18, 1.6647e-17,\n 2.2050e-20, 4.4381e-18, 8.0565e-17, 2.7100e-18, 4.4591e-18, 2.0417e-17,\n 2.9974e-17, 6.1571e-18, 5.4881e-17, 2.0707e-17, 2.2966e-20, 6.4497e-20,\n 1.1138e-17, 7.3379e-17, 6.1819e-17, 2.1537e-19, 2.7845e-17, 1.4976e-18,\n 7.8349e-18, 1.5895e-18, 3.4651e-17, 2.8934e-17, 1.1215e-16, 1.2038e-18,\n 2.2943e-17, 3.2935e-17, 1.9429e-17, 1.9664e-17, 1.7343e-17, 6.8546e-19,\n 1.0621e-16, 3.4781e-19, 1.7181e-18, 3.0251e-18, 5.1433e-17, 4.3983e-18,\n 9.6185e-18, 4.0137e-18, 7.7202e-19, 7.3511e-18, 1.4736e-17, 2.6321e-18,\n 4.2981e-17, 4.0565e-17, 4.3718e-18, 3.0636e-18, 4.7406e-18, 3.4359e-19,\n 2.7758e-17, 1.0530e-17, 2.2444e-17, 2.2213e-17, 2.3387e-18, 6.0562e-19,\n 1.2504e-21, 6.0958e-20, 1.8829e-19, 5.6310e-19, 8.1209e-18, 5.7874e-19,\n 7.2316e-18, 1.8859e-17, 2.7142e-17, 4.3263e-18, 9.8042e-18, 1.2379e-19,\n 5.5290e-18, 1.6372e-16, 8.5905e-19, 8.2536e-18, 1.1081e-17, 7.5783e-17,\n 2.7813e-20, 2.1494e-17, 6.1933e-18, 2.9913e-17, 1.9098e-17, 2.2176e-17,\n 1.8120e-21, 7.0735e-17, 4.7550e-17, 4.3288e-20, 2.1296e-18, 5.1225e-17,\n 2.3792e-18, 9.9372e-18, 1.4045e-19, 3.5296e-17, 6.7959e-18, 3.3031e-17,\n 1.7216e-17, 5.9700e-17, 7.5189e-19, 1.4436e-17, 2.9597e-17, 1.0785e-16,\n 4.7693e-17, 8.4176e-20, 1.8083e-18, 1.8108e-17, 1.1318e-17, 2.0902e-18,\n 3.1889e-18, 3.7698e-18, 1.1180e-17, 1.6628e-18, 1.4088e-17, 2.7958e-19,\n 4.3271e-17, 5.4689e-19, 4.2397e-17, 4.8570e-17, 1.4965e-17, 4.4102e-19,\n 5.1145e-18, 5.9874e-19, 1.4002e-17, 3.3185e-17, 3.3135e-18, 8.5959e-19,\n 3.4503e-17, 1.7120e-18, 2.0151e-17, 1.1354e-17, 5.8494e-18, 3.9839e-20,\n 8.1283e-19, 8.6092e-17, 1.1095e-18, 5.9729e-17, 1.0313e-20, 2.9564e-17,\n 1.1202e-17, 1.1190e-17, 3.3026e-18, 4.0239e-17, 1.5137e-17, 5.8788e-18,\n 2.4402e-17, 7.6160e-20, 5.5461e-17, 3.4933e-18, 4.3010e-19, 5.7670e-18,\n 1.1402e-17, 4.4721e-17, 8.6180e-18, 2.0484e-17, 3.1774e-19, 5.1523e-17,\n 1.5064e-17, 1.5395e-17, 1.2081e-17, 9.4484e-17, 8.6051e-20, 2.6732e-18,\n 8.1985e-17, 6.3989e-18, 7.5593e-18, 3.9325e-19, 1.3099e-17, 3.3890e-17,\n 2.9448e-17, 4.4785e-17, 1.0867e-18, 1.3329e-17, 7.7783e-18, 6.9605e-18,\n 7.1138e-18, 4.4812e-18, 1.1339e-17, 2.8447e-17, 2.1849e-18, 4.8510e-17,\n 5.4173e-17, 1.3163e-17, 1.2928e-17, 1.2305e-17, 8.8996e-20, 2.6331e-18,\n 2.6817e-18, 5.7834e-18, 2.6163e-17, 3.3138e-18, 1.4419e-18, 1.9945e-17,\n 5.5038e-21, 4.8487e-17, 4.1000e-19, 5.3126e-17, 2.2602e-17, 6.3533e-17,\n 1.6400e-18, 1.0867e-17, 4.2674e-17, 1.6057e-17, 1.5590e-18, 4.1188e-17,\n 3.1381e-18, 1.9965e-17, 1.0142e-19, 3.0760e-18, 8.9844e-17, 1.9600e-17,\n 9.0172e-17, 1.0146e-17, 1.4486e-17, 3.5881e-20, 5.0105e-18, 1.2635e-17,\n 1.1619e-19, 8.5054e-19, 1.0915e-17, 3.4949e-18, 5.0214e-18, 2.8602e-18,\n 3.2407e-17, 3.2461e-18, 8.1974e-19, 2.7059e-19, 8.9601e-20, 1.5605e-19,\n 1.9553e-18, 2.6800e-17, 1.8697e-18, 3.8230e-18, 4.7781e-19, 4.5627e-17,\n 2.2374e-17, 4.3385e-17, 9.8718e-17, 2.0446e-18, 5.3432e-17, 2.1806e-17,\n 4.8424e-18, 1.8137e-16, 3.9140e-18, 1.8664e-17, 6.2578e-18, 4.4905e-18,\n 2.3360e-18, 2.0066e-17, 1.0810e-17, 5.2211e-19, 2.7040e-18, 4.6465e-17],\n device='cuda:0')" + "12": { + "step": "tensor(8764.)", + "exp_avg": "tensor([-3.1734e-06, 4.8233e-06, 4.5898e-06, ..., -3.0027e-05,\n 3.4156e-07, -4.7405e-06], device='cuda:0')", + "exp_avg_sq": "tensor([1.0698e-09, 7.4755e-10, 9.0855e-10, ..., 1.2911e-09, 1.4461e-09,\n 1.0259e-09], device='cuda:0')" }, - "36": { - "step": "tensor(23788.)", - "exp_avg": "tensor([[ 5.6052e-45, -5.6052e-45, 5.6052e-45, ..., -5.6052e-45,\n 5.6052e-45, -5.6052e-45],\n [-5.6052e-45, -5.6052e-45, -5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, 5.6052e-45],\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., -5.6052e-45,\n 5.6052e-45, -5.6052e-45],\n ...,\n [ 5.6052e-45, 5.6052e-45, 5.6052e-45, ..., -5.6052e-45,\n 5.6052e-45, -5.6052e-45],\n [-5.6052e-45, 5.6052e-45, -5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, 5.6052e-45],\n [-5.6052e-45, 5.6052e-45, -5.6052e-45, ..., 5.6052e-45,\n -5.6052e-45, 5.6052e-45]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.6545e-18, 9.2388e-22, 9.2389e-18, ..., 2.3926e-20, 2.4087e-19,\n 6.6582e-19],\n [2.2426e-18, 2.3372e-22, 1.3104e-17, ..., 3.7506e-20, 3.5592e-19,\n 9.3998e-19],\n [1.2664e-18, 1.6108e-22, 7.4065e-18, ..., 1.9569e-20, 2.1079e-19,\n 4.7893e-19],\n ...,\n [9.5824e-19, 1.3677e-22, 5.8115e-18, ..., 8.9885e-21, 1.4499e-19,\n 3.8548e-19],\n [1.2589e-19, 3.3669e-22, 6.9343e-19, ..., 2.4118e-21, 1.3630e-20,\n 5.2260e-20],\n [6.6609e-19, 8.0262e-22, 3.9085e-18, ..., 1.1375e-20, 9.1026e-20,\n 2.6747e-19]], device='cuda:0')" + "13": { + "step": "tensor(8764.)", + "exp_avg": "tensor([[ 2.2182e-07, -2.0739e-09, 7.5835e-07, ..., -3.3445e-07,\n 2.1949e-08, 3.0941e-07],\n [ 1.6808e-07, 1.8245e-07, 3.3214e-07, ..., 2.7568e-07,\n 2.7083e-07, 4.0770e-07],\n [-2.0915e-07, 6.5567e-08, -1.1137e-06, ..., 6.5655e-07,\n 2.2354e-07, -9.9601e-08],\n ...,\n [-1.6749e-07, 2.1184e-07, -2.2033e-07, ..., -6.8793e-09,\n 2.4235e-07, -1.6973e-07],\n [ 1.4642e-07, 2.4461e-08, -4.6586e-07, ..., -8.2878e-10,\n -3.2903e-09, 3.3803e-09],\n [-1.5389e-07, 9.5534e-08, -1.2836e-07, ..., -8.0069e-08,\n 2.5866e-07, -2.3752e-07]], device='cuda:0')", + "exp_avg_sq": "tensor([[4.5088e-13, 3.1527e-13, 6.2491e-13, ..., 1.0108e-12, 5.9876e-13,\n 1.5150e-12],\n [8.1732e-13, 4.4879e-13, 1.9033e-12, ..., 2.6162e-12, 1.5164e-12,\n 1.9223e-12],\n [8.9293e-13, 7.2005e-13, 2.2119e-12, ..., 2.6867e-12, 1.7467e-12,\n 7.9776e-13],\n ...,\n [1.0520e-12, 1.0296e-12, 2.2339e-12, ..., 1.5686e-12, 1.9143e-12,\n 2.7351e-12],\n [7.8992e-13, 9.9624e-13, 7.6716e-13, ..., 1.3802e-12, 1.2542e-12,\n 3.2748e-12],\n [9.0855e-13, 6.6180e-13, 1.1952e-12, ..., 3.4612e-12, 1.5513e-12,\n 1.2847e-12]], device='cuda:0')" }, - "37": { - "step": "tensor(23788.)", - "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([5.9854e-17, 8.3269e-17, 4.4531e-17, 2.3643e-16, 7.1753e-19, 6.8574e-18,\n 1.2558e-16, 2.3581e-16, 1.9073e-16, 9.4977e-17, 4.1167e-18, 1.9831e-16,\n 9.7972e-18, 4.8857e-19, 2.5789e-17, 2.4640e-19, 1.0771e-16, 2.0398e-16,\n 5.7047e-17, 1.2187e-17, 2.2211e-17, 1.6348e-17, 7.7821e-17, 9.9759e-19,\n 2.5931e-18, 2.2003e-17, 2.2688e-18, 3.3171e-17, 1.5720e-17, 4.9353e-17,\n 1.0792e-19, 2.6483e-17, 2.6798e-17, 9.3332e-18, 1.2016e-18, 5.2540e-17,\n 1.7990e-18, 3.0210e-16, 5.0661e-17, 3.6055e-18, 1.1381e-16, 7.4138e-18,\n 2.1784e-17, 1.9699e-18, 3.9750e-17, 6.8174e-17, 8.5602e-19, 8.9055e-20,\n 5.5924e-17, 5.0810e-17, 7.2585e-17, 3.0251e-18, 3.1546e-17, 1.1824e-16,\n 1.5105e-18, 1.3212e-17, 1.1951e-16, 4.7416e-19, 5.4845e-18, 9.8464e-18,\n 4.1332e-17, 3.2351e-17, 6.6555e-18, 6.3237e-17, 3.9755e-19, 9.9677e-17,\n 1.3577e-17, 2.5956e-17, 2.9227e-17, 1.0702e-17, 5.0279e-18, 9.8503e-17,\n 1.2069e-16, 2.5285e-16, 1.6317e-18, 1.9742e-19, 9.9571e-18, 8.5014e-17,\n 1.2882e-18, 2.4449e-19, 1.9984e-17, 1.2672e-18, 5.0667e-17, 8.7033e-18,\n 9.5770e-18, 4.9504e-18, 9.6732e-18, 1.3907e-17, 1.4337e-17, 1.6726e-16,\n 1.4576e-19, 9.9441e-17, 4.1237e-17, 1.1080e-16, 2.7003e-16, 1.7211e-16,\n 1.7477e-17, 2.3563e-16, 8.4701e-18, 2.9781e-18, 1.0488e-16, 3.6054e-20,\n 4.5721e-17, 2.7622e-17, 2.6039e-18, 7.3078e-17, 4.3921e-20, 4.3037e-16,\n 8.9165e-17, 9.2501e-17, 3.2690e-19, 2.6757e-18, 8.2860e-18, 8.2659e-17,\n 2.6056e-17, 6.9461e-17, 1.4689e-19, 8.5854e-17, 1.0796e-16, 4.3279e-17,\n 1.4434e-16, 5.7020e-17, 4.4567e-17, 6.3870e-18, 3.0196e-16, 1.5835e-16,\n 4.0246e-17, 1.6632e-17, 3.4481e-17, 1.2273e-19, 3.5907e-16, 3.4897e-17,\n 4.5332e-19, 1.3387e-19, 2.3323e-16, 8.9966e-17, 1.3547e-17, 5.9526e-18,\n 5.5047e-19, 1.8024e-19, 1.4869e-16, 3.0927e-19, 2.2310e-16, 4.3412e-18,\n 5.0365e-18, 1.0677e-17, 8.8840e-17, 7.4517e-17, 2.1641e-17, 2.6061e-19,\n 4.2909e-18, 1.3692e-18, 1.6247e-19, 1.4690e-16, 1.2998e-17, 1.9272e-16,\n 4.6239e-18, 1.0803e-16, 1.4617e-17, 1.4640e-16, 3.8271e-17, 6.2294e-18,\n 6.5722e-17, 5.7176e-17, 7.9326e-17, 1.7483e-16, 3.6732e-16, 1.7355e-19,\n 6.8921e-19, 3.3424e-17, 1.9328e-17, 2.9426e-19, 1.1853e-16, 1.5257e-17,\n 1.7136e-16, 2.1370e-19, 8.0081e-17, 8.9361e-18, 3.8555e-17, 3.8693e-17,\n 4.1252e-16, 4.1074e-17, 2.6523e-17, 3.9431e-20, 7.7279e-18, 1.1978e-16,\n 4.8928e-19, 4.3762e-17, 1.1149e-16, 9.9337e-18, 1.1391e-16, 9.9049e-18,\n 2.8715e-17, 2.6284e-17, 1.7368e-16, 9.2940e-17, 1.0935e-16, 8.0274e-19,\n 3.4836e-17, 2.2311e-16, 2.3854e-19, 2.0694e-17, 1.0667e-17, 1.5643e-16,\n 1.5283e-16, 2.1859e-17, 7.2303e-17, 1.4915e-17, 1.0684e-18, 4.0754e-19,\n 4.8399e-19, 2.4339e-17, 1.2541e-16, 1.1162e-16, 4.0278e-17, 1.4012e-16,\n 1.7019e-17, 8.4186e-17, 6.1327e-19, 2.2090e-18, 4.2183e-20, 1.1890e-17,\n 1.3189e-16, 9.8286e-18, 3.3146e-18, 6.7881e-17, 1.9109e-17, 5.9668e-17,\n 4.2432e-18, 8.4167e-17, 1.1779e-18, 4.6057e-17, 1.4316e-16, 6.0385e-19,\n 1.1921e-16, 2.6748e-18, 2.8305e-17, 1.0625e-19, 2.2475e-16, 4.5299e-18,\n 2.0602e-17, 4.6507e-18, 5.4411e-18, 3.1870e-17, 4.6903e-17, 3.2502e-17,\n 1.1749e-16, 4.4679e-17, 1.9020e-16, 1.8276e-17, 9.8923e-17, 2.1445e-21,\n 7.7970e-18, 3.6737e-17, 4.7487e-18, 2.5277e-17], device='cuda:0')" + "14": { + "step": "tensor(7512.)", + "exp_avg": "tensor([[-9.9144e-08, -3.9194e-08, 6.6145e-06, ..., 3.6352e-05,\n 0.0000e+00, 4.3041e-08],\n [ 8.1803e-08, 4.6771e-10, 9.3593e-06, ..., -1.5705e-07,\n 0.0000e+00, 7.0697e-07],\n [ 1.4673e-07, 3.3228e-08, 3.6580e-07, ..., -1.3127e-08,\n 0.0000e+00, 4.7572e-07],\n ...,\n [ 5.0273e-07, 1.8915e-09, -3.1544e-08, ..., -7.1909e-09,\n 0.0000e+00, 1.7110e-07],\n [-5.6052e-09, -9.2777e-10, 2.9593e-07, ..., 4.5681e-06,\n 0.0000e+00, 3.3314e-06],\n [ 4.4287e-08, 5.6212e-09, 1.1217e-05, ..., 8.2747e-05,\n 0.0000e+00, 6.8293e-08]], device='cuda:0')", + "exp_avg_sq": "tensor([[5.2447e-13, 2.2746e-13, 9.2144e-11, ..., 3.8680e-10, 0.0000e+00,\n 6.9366e-12],\n [9.2418e-12, 2.7513e-13, 7.3857e-11, ..., 1.5449e-11, 0.0000e+00,\n 2.8060e-11],\n [2.1221e-11, 4.9112e-13, 3.5593e-12, ..., 7.8030e-13, 0.0000e+00,\n 1.2318e-12],\n ...,\n [2.3565e-11, 1.7063e-13, 7.0745e-12, ..., 3.5663e-12, 0.0000e+00,\n 6.4724e-12],\n [2.8506e-12, 5.2280e-13, 5.3315e-12, ..., 1.2502e-10, 0.0000e+00,\n 1.3957e-10],\n [3.5639e-12, 1.4660e-13, 1.5281e-10, ..., 1.6086e-09, 0.0000e+00,\n 1.5458e-11]], device='cuda:0')" }, - "38": { - "step": "tensor(23788.)", - "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([[4.1878e-16, 1.1085e-16, 1.5868e-16, ..., 2.0211e-17, 1.2336e-16,\n 2.2780e-16],\n [4.3241e-17, 1.1240e-17, 1.6827e-17, ..., 2.2408e-18, 1.2455e-17,\n 2.2583e-17],\n [4.7105e-17, 1.2144e-17, 1.7533e-17, ..., 2.1576e-18, 1.3982e-17,\n 2.6119e-17],\n [4.9352e-17, 1.3626e-17, 1.8554e-17, ..., 2.3403e-18, 1.4734e-17,\n 2.7355e-17]], device='cuda:0')" + "15": { + "step": "tensor(7512.)", + "exp_avg": "tensor([ 3.3027e-05, 2.2930e-05, 3.7146e-06, ..., 1.8131e-05,\n -3.2139e-06, 6.8807e-05], device='cuda:0')", + "exp_avg_sq": "tensor([2.3151e-09, 2.0855e-09, 1.9548e-09, ..., 2.8112e-09, 3.1657e-09,\n 3.5868e-09], device='cuda:0')" }, - "39": { - "step": "tensor(23788.)", - "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([1.0502e-14, 1.0821e-15, 1.1708e-15, 1.2507e-15], device='cuda:0')" + "16": { + "step": "tensor(7512.)", + "exp_avg": "tensor([[-2.1174e-09, -8.2970e-07, -1.1059e-07, ..., 5.8443e-08,\n -1.8422e-07, -4.0374e-07],\n [ 4.0283e-07, -2.6375e-07, 1.5211e-07, ..., -1.8824e-07,\n -1.4319e-06, 1.5411e-06],\n [-1.3496e-07, 5.9090e-07, 5.6911e-08, ..., 3.6908e-07,\n 3.0063e-07, -6.0522e-07],\n ...,\n [ 8.0747e-08, -2.1092e-07, -1.2430e-07, ..., -2.7408e-07,\n 2.8949e-07, 2.6415e-07],\n [ 6.3674e-08, -6.9401e-07, -2.4110e-07, ..., -3.4040e-07,\n -2.9516e-06, 6.0260e-08],\n [-3.6565e-07, -1.6630e-08, 6.8491e-08, ..., -1.8997e-07,\n -1.0694e-06, -1.1596e-06]], device='cuda:0')", + "exp_avg_sq": "tensor([[6.8353e-13, 7.9268e-13, 6.1169e-13, ..., 4.0575e-13, 9.0677e-13,\n 4.7225e-13],\n [9.2926e-13, 1.0657e-12, 1.2535e-12, ..., 2.1452e-12, 2.0690e-12,\n 1.1186e-12],\n [7.6799e-13, 9.2466e-13, 7.5990e-13, ..., 7.1637e-13, 2.1274e-12,\n 7.4540e-13],\n ...,\n [1.6328e-12, 1.1935e-12, 6.1102e-13, ..., 1.2074e-12, 7.5510e-13,\n 7.7320e-13],\n [1.0021e-12, 1.3796e-12, 1.2161e-12, ..., 1.5376e-12, 6.0830e-12,\n 6.1298e-13],\n [1.0035e-12, 2.1793e-12, 1.0239e-12, ..., 7.3054e-13, 1.6506e-12,\n 1.3283e-12]], device='cuda:0')" }, - "40": { - "step": "tensor(23788.)", - "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([[4.1878e-16, 1.1085e-16, 1.5868e-16, ..., 2.0211e-17, 1.2336e-16,\n 2.2780e-16],\n [4.3241e-17, 1.1240e-17, 1.6827e-17, ..., 2.2408e-18, 1.2455e-17,\n 2.2583e-17],\n [4.7105e-17, 1.2144e-17, 1.7533e-17, ..., 2.1576e-18, 1.3982e-17,\n 2.6119e-17],\n [4.9352e-17, 1.3626e-17, 1.8554e-17, ..., 2.3403e-18, 1.4734e-17,\n 2.7355e-17]], device='cuda:0')" + "17": { + "step": "tensor(6260.)", + "exp_avg": "tensor([[ 9.5503e-07, -1.3270e-09, -2.9824e-07, ..., 5.8676e-08,\n 0.0000e+00, -1.4706e-07],\n [-8.1059e-08, 1.4618e-07, 1.1680e-08, ..., 2.0458e-06,\n 0.0000e+00, 2.1162e-07],\n [ 1.4763e-07, 1.2673e-09, 6.5117e-08, ..., -6.0637e-08,\n 0.0000e+00, -2.7041e-07],\n ...,\n [-4.1496e-08, -2.2324e-07, 1.0280e-07, ..., 8.3597e-07,\n 0.0000e+00, -9.0796e-08],\n [-1.2783e-07, -6.0111e-08, 7.7144e-08, ..., -5.2068e-06,\n 0.0000e+00, 1.2393e-07],\n [ 1.1435e-07, -8.7357e-09, 1.9023e-06, ..., 2.4482e-07,\n 0.0000e+00, 8.5112e-08]], device='cuda:0')", + "exp_avg_sq": "tensor([[6.9995e-12, 6.4075e-14, 6.3755e-12, ..., 2.2819e-12, 0.0000e+00,\n 2.2964e-12],\n [1.0422e-12, 7.2287e-12, 1.5739e-11, ..., 4.0577e-10, 0.0000e+00,\n 3.0019e-10],\n [5.9075e-11, 5.2013e-13, 2.8558e-11, ..., 1.0987e-12, 0.0000e+00,\n 3.4598e-12],\n ...,\n [5.3187e-12, 6.8257e-12, 6.3911e-11, ..., 1.2294e-11, 0.0000e+00,\n 1.1448e-11],\n [1.1078e-11, 1.3477e-11, 2.5968e-11, ..., 6.5423e-11, 0.0000e+00,\n 1.0950e-11],\n [1.8585e-12, 1.4802e-12, 7.0148e-11, ..., 9.6244e-12, 0.0000e+00,\n 2.7618e-12]], device='cuda:0')" }, - "41": { - "step": "tensor(23788.)", - "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([1.0502e-14, 1.0821e-15, 1.1708e-15, 1.2507e-15], device='cuda:0')" + "18": { + "step": "tensor(6260.)", + "exp_avg": "tensor([-7.1890e-07, -5.6708e-05, -8.4577e-06, ..., -1.9044e-05,\n -2.6201e-05, 1.1328e-05], device='cuda:0')", + "exp_avg_sq": "tensor([2.2720e-09, 3.6701e-09, 2.6689e-09, ..., 1.4793e-09, 2.5373e-09,\n 2.0937e-09], device='cuda:0')" }, - "42": { - "step": "tensor(23788.)", - "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([[4.1878e-16, 1.1085e-16, 1.5868e-16, ..., 2.0211e-17, 1.2336e-16,\n 2.2780e-16],\n [4.3241e-17, 1.1240e-17, 1.6827e-17, ..., 2.2408e-18, 1.2455e-17,\n 2.2583e-17],\n [4.7105e-17, 1.2144e-17, 1.7533e-17, ..., 2.1576e-18, 1.3982e-17,\n 2.6119e-17],\n [4.9352e-17, 1.3626e-17, 1.8554e-17, ..., 2.3403e-18, 1.4734e-17,\n 2.7355e-17]], device='cuda:0')" + "19": { + "step": "tensor(6260.)", + "exp_avg": "tensor([[-2.6417e-07, 4.7218e-08, 1.9952e-08, ..., 1.2112e-07,\n -4.0797e-08, -1.0241e-07],\n [-1.0129e-07, 2.9652e-07, -7.9153e-09, ..., -7.3317e-08,\n 2.1145e-07, -4.3682e-08],\n [-9.1283e-07, -1.3733e-07, 3.2584e-07, ..., -3.1132e-08,\n 1.0390e-07, 4.0456e-08],\n ...,\n [-4.9108e-08, 3.5919e-08, -1.1573e-07, ..., 5.2719e-09,\n -7.5300e-08, 3.9253e-10],\n [ 1.8438e-07, 3.0469e-08, -1.7891e-07, ..., -1.2010e-08,\n -4.4716e-07, -1.3029e-07],\n [-2.0279e-08, -8.9886e-08, 4.3602e-08, ..., 1.6823e-07,\n 2.5753e-07, -6.5618e-08]], device='cuda:0')", + "exp_avg_sq": "tensor([[2.9577e-13, 1.0110e-12, 3.0728e-13, ..., 3.3254e-13, 5.0185e-13,\n 5.0305e-13],\n [4.8619e-13, 2.3776e-12, 8.1337e-13, ..., 7.3788e-13, 1.3377e-12,\n 7.5397e-13],\n [9.2196e-13, 1.8124e-12, 1.3355e-12, ..., 8.1659e-13, 1.4627e-12,\n 6.2089e-13],\n ...,\n [4.7633e-13, 8.1497e-13, 6.6300e-13, ..., 7.5343e-13, 1.2362e-12,\n 7.8757e-13],\n [2.6809e-13, 2.0698e-12, 1.0591e-12, ..., 6.5731e-13, 1.5546e-12,\n 9.0076e-13],\n [3.8023e-13, 6.2662e-13, 7.4862e-13, ..., 8.9159e-13, 1.3479e-12,\n 7.8253e-13]], device='cuda:0')" }, - "43": { - "step": "tensor(23788.)", - "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([1.0502e-14, 1.0821e-15, 1.1708e-15, 1.2507e-15], device='cuda:0')" + "20": { + "step": "tensor(5008.)", + "exp_avg": "tensor([[-6.1845e-07, 1.4413e-07, -1.9318e-08, ..., -1.1891e-05,\n 0.0000e+00, 6.4408e-09],\n [-5.6226e-08, 5.8805e-07, -2.1868e-07, ..., -2.5101e-07,\n 0.0000e+00, -5.6194e-06],\n [-3.4456e-09, -1.6632e-08, -1.7002e-06, ..., -5.1660e-07,\n 0.0000e+00, -5.5233e-08],\n ...,\n [ 9.0172e-08, 6.4309e-08, 1.4471e-07, ..., 2.6056e-07,\n 0.0000e+00, -4.0211e-06],\n [-4.8459e-07, 1.5604e-08, 1.1054e-07, ..., -1.5318e-07,\n 0.0000e+00, -1.9394e-07],\n [ 2.0508e-07, -8.9246e-09, -2.7086e-07, ..., -5.4503e-08,\n 0.0000e+00, 2.0587e-07]], device='cuda:0')", + "exp_avg_sq": "tensor([[2.9100e-11, 5.6381e-12, 7.9570e-12, ..., 1.5905e-10, 0.0000e+00,\n 3.4304e-12],\n [2.2399e-11, 5.8809e-11, 2.5600e-11, ..., 2.7774e-11, 0.0000e+00,\n 7.5856e-11],\n [1.1651e-12, 5.0924e-13, 6.7110e-11, ..., 1.0249e-11, 0.0000e+00,\n 4.4959e-11],\n ...,\n [1.4415e-12, 3.9656e-13, 9.8264e-11, ..., 1.1758e-10, 0.0000e+00,\n 1.2689e-10],\n [4.0739e-11, 2.2677e-12, 4.7581e-12, ..., 6.9080e-12, 0.0000e+00,\n 2.6612e-11],\n [6.0083e-11, 2.7452e-12, 2.6890e-11, ..., 1.6219e-11, 0.0000e+00,\n 4.1770e-11]], device='cuda:0')" }, - "8": { - "step": "tensor(22536.)", - "exp_avg": "tensor([[ 7.2263e-08, -3.6041e-08, 1.1662e-12, ..., -2.8868e-08,\n -2.3718e-07, -5.0621e-08],\n [-5.4148e-06, 9.4806e-08, 6.0259e-12, ..., -1.1911e-08,\n 2.6591e-06, 6.7227e-07],\n [ 4.2784e-08, 3.6937e-08, -3.4431e-11, ..., -1.1785e-08,\n -3.2087e-07, 9.1184e-07],\n ...,\n [ 1.0986e-07, -1.3670e-07, 9.4910e-21, ..., -1.0643e-08,\n 2.9263e-06, -1.4940e-08],\n [ 4.8732e-07, -1.4769e-06, 1.2920e-10, ..., -3.9437e-08,\n 3.2940e-07, -8.5371e-07],\n [-2.7998e-08, 6.3856e-10, -9.6040e-09, ..., -3.0761e-07,\n 8.1022e-10, 1.5774e-11]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.5062e-11, 6.2034e-12, 1.8354e-14, ..., 2.8134e-12, 1.1295e-11,\n 1.5963e-11],\n [2.2792e-11, 1.0663e-11, 9.4406e-15, ..., 5.1454e-12, 1.1282e-10,\n 1.3395e-11],\n [1.1317e-11, 3.8089e-12, 6.5031e-13, ..., 4.8379e-12, 1.3181e-11,\n 2.2365e-11],\n ...,\n [2.4781e-11, 5.6226e-12, 1.3364e-15, ..., 4.4366e-13, 1.3990e-10,\n 4.6709e-12],\n [1.9508e-11, 1.0931e-11, 1.3128e-12, ..., 9.4605e-12, 5.4853e-11,\n 6.1568e-12],\n [5.3941e-11, 1.2638e-11, 2.1197e-12, ..., 2.9432e-10, 1.4580e-12,\n 1.8870e-12]], device='cuda:0')" + "21": { + "step": "tensor(5008.)", + "exp_avg": "tensor([-3.7956e-05, -5.4031e-06, 7.0594e-06, ..., 2.7980e-05,\n -2.8727e-05, 1.1770e-05], device='cuda:0')", + "exp_avg_sq": "tensor([5.6596e-09, 4.7770e-09, 3.7082e-09, ..., 5.6955e-09, 6.1637e-09,\n 4.7482e-09], device='cuda:0')" }, - "9": { - "step": "tensor(22536.)", - "exp_avg": "tensor([-7.3518e-07, 5.3155e-06, 7.7293e-07, ..., 2.1352e-06,\n -1.0879e-06, 4.6073e-06], device='cuda:0')", - "exp_avg_sq": "tensor([7.4122e-10, 9.6338e-10, 6.7978e-10, ..., 1.0804e-09, 6.7205e-10,\n 2.0181e-09], device='cuda:0')" + "22": { + "step": "tensor(5008.)", + "exp_avg": "tensor([[-6.8812e-08, 3.2808e-07, -1.0458e-07, ..., -2.6516e-07,\n 1.2057e-07, -1.0668e-07],\n [ 3.3064e-07, 3.1144e-07, -1.5562e-07, ..., 1.8608e-07,\n -1.5108e-07, -1.7357e-08],\n [ 6.3645e-07, -5.7058e-08, -1.4178e-07, ..., -5.1062e-07,\n -1.1225e-07, -4.3968e-08],\n ...,\n [-1.6826e-07, -3.9908e-07, 6.1119e-07, ..., 2.9228e-07,\n 9.4299e-09, 7.5585e-08],\n [-5.2122e-08, -9.3836e-08, 9.5097e-08, ..., -4.7426e-07,\n -3.3345e-07, -2.4999e-08],\n [ 1.1571e-06, -4.2969e-07, -3.5850e-08, ..., 4.9074e-07,\n -1.7465e-07, -5.1722e-08]], device='cuda:0')", + "exp_avg_sq": "tensor([[9.1614e-13, 1.0121e-12, 4.9026e-13, ..., 2.4729e-12, 3.4346e-13,\n 6.9624e-13],\n [7.2384e-13, 9.3973e-13, 8.1467e-13, ..., 1.8977e-12, 7.2797e-13,\n 8.4920e-13],\n [1.3063e-12, 1.5015e-12, 1.0908e-12, ..., 3.6295e-12, 3.9936e-13,\n 1.2586e-12],\n ...,\n [1.1700e-12, 1.3447e-12, 1.3369e-12, ..., 1.5498e-12, 6.1317e-13,\n 2.0993e-12],\n [6.6169e-13, 1.5391e-12, 9.4662e-13, ..., 1.9525e-12, 1.1179e-12,\n 6.5295e-13],\n [2.1305e-12, 1.8973e-12, 1.0005e-12, ..., 2.1663e-12, 8.8708e-13,\n 2.3615e-12]], device='cuda:0')" }, - "10": { - "step": "tensor(22536.)", - "exp_avg": "tensor([[ 3.9438e-08, -7.2389e-08, -4.1392e-07, ..., -1.0747e-08,\n -1.0121e-07, 4.8199e-08],\n [ 4.3504e-09, 3.4073e-07, 3.6320e-07, ..., 5.1105e-09,\n 1.7838e-07, 1.3611e-07],\n [ 8.2347e-08, -2.2692e-07, 4.8652e-07, ..., 4.9861e-09,\n -1.7447e-07, 1.3872e-07],\n ...,\n [-3.6323e-08, -1.3611e-07, 4.6246e-07, ..., -1.2686e-07,\n -1.0579e-08, 6.8958e-08],\n [ 9.8628e-08, -5.8662e-08, -4.4099e-07, ..., 1.2570e-08,\n -9.3949e-08, -1.5358e-07],\n [ 3.7083e-08, -5.5424e-08, 4.7562e-07, ..., -6.5599e-08,\n 8.2806e-08, -2.7306e-08]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.5144e-13, 2.3601e-13, 5.9843e-13, ..., 2.5199e-13, 2.8761e-13,\n 2.6898e-13],\n [2.4644e-13, 6.9390e-13, 3.8157e-13, ..., 6.0781e-13, 5.1399e-13,\n 3.8013e-13],\n [4.2209e-13, 9.8489e-13, 4.3207e-13, ..., 6.6150e-13, 8.3530e-13,\n 7.2758e-13],\n ...,\n [3.5139e-13, 7.1662e-13, 4.2745e-13, ..., 7.1861e-13, 6.3588e-13,\n 6.5552e-13],\n [2.8537e-13, 9.0931e-13, 4.6134e-13, ..., 8.3460e-13, 4.5643e-13,\n 3.7040e-13],\n [2.8991e-13, 9.6541e-13, 6.3302e-13, ..., 1.3326e-12, 7.0475e-13,\n 2.8795e-13]], device='cuda:0')" + "23": { + "step": "tensor(3756.)", + "exp_avg": "tensor([[ 8.6625e-08, -1.8288e-07, 8.3755e-07, ..., 3.9635e-06,\n 0.0000e+00, 2.1120e-08],\n [ 4.0071e-07, 1.8793e-08, -4.0375e-07, ..., -5.6561e-07,\n 0.0000e+00, -7.3566e-08],\n [-1.4025e-06, 3.7002e-07, 9.7900e-07, ..., 1.2052e-06,\n 0.0000e+00, -4.3734e-07],\n ...,\n [-3.7614e-06, 5.9165e-08, -1.9186e-06, ..., 8.2550e-07,\n 0.0000e+00, -3.2771e-07],\n [-1.5528e-07, 1.0160e-06, -6.5283e-07, ..., -2.0301e-07,\n 0.0000e+00, -1.9045e-06],\n [ 1.6437e-07, 5.3580e-09, -2.5795e-06, ..., -3.8795e-07,\n 0.0000e+00, 3.0473e-07]], device='cuda:0')", + "exp_avg_sq": "tensor([[3.8523e-11, 2.1524e-11, 2.5717e-11, ..., 1.3093e-10, 0.0000e+00,\n 6.7806e-11],\n [1.7258e-11, 1.5388e-11, 2.2932e-11, ..., 5.8121e-11, 0.0000e+00,\n 2.5037e-11],\n [3.2654e-11, 4.8526e-11, 1.0993e-10, ..., 2.8691e-11, 0.0000e+00,\n 1.3936e-11],\n ...,\n [1.8172e-10, 1.3802e-11, 2.6416e-11, ..., 3.6502e-11, 0.0000e+00,\n 3.8433e-11],\n [2.1074e-11, 1.9554e-10, 3.4346e-11, ..., 1.7332e-11, 0.0000e+00,\n 9.0037e-10],\n [7.1665e-11, 1.8041e-12, 1.1006e-10, ..., 6.6133e-11, 0.0000e+00,\n 2.7220e-11]], device='cuda:0')" }, - "11": { - "step": "tensor(21284.)", - "exp_avg": "tensor([[-4.5010e-07, 5.7669e-08, 3.0969e-09, ..., -3.1271e-07,\n 9.6291e-08, -9.1990e-09],\n [-4.3908e-08, 1.0823e-07, 1.4397e-10, ..., -8.9437e-10,\n -5.5777e-08, 7.0789e-07],\n [ 8.7293e-08, -3.6128e-08, -9.4206e-09, ..., -3.0986e-07,\n -5.7038e-08, 2.3645e-10],\n ...,\n [ 5.4739e-07, 5.9701e-08, 2.8052e-12, ..., -1.4037e-08,\n 1.4730e-08, 3.3166e-13],\n [-4.1702e-08, -1.1623e-09, 9.2594e-12, ..., -3.6338e-07,\n 4.3222e-07, -1.5776e-13],\n [ 3.3657e-09, 2.1779e-08, 2.3710e-08, ..., -2.2895e-07,\n 1.4771e-07, 3.2875e-09]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.2381e-12, 8.1567e-12, 5.2332e-14, ..., 1.2894e-11, 2.5189e-12,\n 3.6284e-12],\n [5.5931e-12, 3.0054e-13, 1.2100e-13, ..., 5.7171e-13, 1.1091e-11,\n 5.0249e-11],\n [2.6949e-12, 5.3934e-12, 2.5327e-14, ..., 2.5192e-12, 1.4522e-10,\n 6.6708e-13],\n ...,\n [9.3287e-11, 2.8068e-13, 1.4415e-13, ..., 6.4667e-13, 3.3483e-12,\n 3.4600e-14],\n [1.6888e-12, 2.4687e-13, 1.6914e-12, ..., 3.0911e-11, 1.2164e-10,\n 6.1138e-13],\n [4.4546e-13, 1.9542e-11, 8.9744e-13, ..., 7.4024e-11, 4.6800e-12,\n 2.8392e-12]], device='cuda:0')" + "24": { + "step": "tensor(3756.)", + "exp_avg": "tensor([ 4.2499e-05, 1.5117e-06, -1.1170e-05, ..., 2.8940e-05,\n -1.3405e-05, -2.5607e-06], device='cuda:0')", + "exp_avg_sq": "tensor([1.1779e-08, 1.0526e-08, 9.5232e-09, ..., 8.4988e-09, 1.1161e-08,\n 1.1279e-08], device='cuda:0')" }, - "12": { - "step": "tensor(21284.)", - "exp_avg": "tensor([-1.0549e-05, 2.4165e-06, -2.3861e-06, ..., -7.9029e-06,\n 1.1984e-05, 1.1468e-06], device='cuda:0')", - "exp_avg_sq": "tensor([3.2005e-10, 4.9548e-10, 4.7905e-10, ..., 5.9107e-10, 9.6323e-10,\n 5.8690e-10], device='cuda:0')" + "25": { + "step": "tensor(3756.)", + "exp_avg": "tensor([[ 3.6704e-07, -6.6624e-08, 4.4290e-08, ..., -3.3636e-07,\n 1.0906e-07, -3.6141e-08],\n [ 3.8203e-07, 7.1240e-08, 4.9511e-07, ..., 4.2769e-07,\n -2.1302e-07, -1.6420e-07],\n [ 4.0396e-07, 3.3011e-07, -3.1809e-08, ..., 2.4831e-07,\n -8.7148e-08, -3.5561e-07],\n ...,\n [-4.2670e-07, -3.5578e-07, 9.9700e-07, ..., -2.2710e-07,\n -2.2648e-07, -1.0786e-07],\n [ 5.7559e-09, 3.5136e-07, -1.8028e-07, ..., -1.0323e-07,\n 3.0361e-08, 1.0445e-07],\n [ 2.6267e-07, -1.4132e-07, -2.6558e-09, ..., -1.1966e-06,\n 5.1226e-07, 1.6046e-07]], device='cuda:0')", + "exp_avg_sq": "tensor([[1.4248e-12, 7.3370e-13, 9.7379e-13, ..., 2.6822e-12, 2.3264e-12,\n 1.3054e-12],\n [2.6117e-12, 1.6776e-12, 2.0095e-12, ..., 3.0942e-12, 3.7556e-12,\n 4.9571e-12],\n [3.1881e-12, 2.2160e-12, 2.2749e-12, ..., 2.7148e-12, 4.6848e-12,\n 5.0733e-12],\n ...,\n [2.8834e-12, 2.3681e-12, 2.6422e-12, ..., 4.3040e-12, 3.3138e-12,\n 4.8472e-12],\n [3.8981e-12, 1.3963e-12, 2.5805e-12, ..., 5.9204e-12, 2.8707e-12,\n 4.9756e-12],\n [2.6075e-12, 1.4737e-12, 2.8133e-12, ..., 3.1698e-12, 3.8211e-12,\n 2.9316e-12]], device='cuda:0')" }, - "13": { - "step": "tensor(21284.)", - "exp_avg": "tensor([[-2.4656e-07, -1.4318e-07, 2.1095e-07, ..., 6.4344e-08,\n 5.7850e-08, -3.0665e-08],\n [-2.1587e-07, -1.2889e-07, -3.4456e-07, ..., 8.2379e-09,\n 7.2186e-08, 3.9188e-08],\n [-5.2824e-08, 1.2873e-07, 1.8983e-07, ..., -2.2606e-08,\n -4.0865e-08, 6.0729e-08],\n ...,\n [-3.0616e-07, -1.9940e-07, 9.6785e-08, ..., -7.2441e-09,\n -4.5459e-08, 6.8075e-08],\n [ 1.3188e-07, -7.1311e-08, 1.6023e-07, ..., -1.9035e-07,\n 1.3002e-07, 1.4693e-08],\n [ 1.2937e-08, 1.8283e-07, -5.8934e-08, ..., -2.2292e-07,\n -1.0049e-08, 2.6372e-08]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.1619e-13, 6.4513e-14, 8.0862e-13, ..., 9.3820e-14, 3.7516e-14,\n 7.8689e-14],\n [1.1912e-13, 1.2028e-13, 1.5075e-12, ..., 4.4100e-13, 7.0981e-14,\n 1.0198e-13],\n [2.6826e-13, 7.4735e-14, 1.2614e-12, ..., 9.0722e-13, 9.1381e-14,\n 1.2663e-13],\n ...,\n [2.1677e-13, 1.1629e-13, 2.6651e-13, ..., 1.9367e-13, 1.1072e-13,\n 1.0943e-13],\n [2.9368e-13, 2.4386e-13, 2.3169e-13, ..., 7.3020e-13, 1.0518e-13,\n 1.3523e-13],\n [3.2072e-13, 1.2275e-13, 2.1874e-13, ..., 3.8285e-13, 6.9521e-14,\n 1.2237e-13]], device='cuda:0')" + "26": { + "step": "tensor(2504.)", + "exp_avg": "tensor([[ 2.6818e-07, -5.6154e-07, 1.7891e-06, ..., 2.9804e-06,\n 0.0000e+00, -9.3856e-07],\n [ 1.0162e-06, 8.9720e-08, 1.3353e-06, ..., 6.9238e-07,\n 0.0000e+00, 2.1799e-07],\n [ 8.5368e-07, -2.2396e-06, -9.7303e-07, ..., -7.0409e-06,\n 0.0000e+00, -7.1518e-07],\n ...,\n [-1.0315e-07, 4.5212e-08, 1.8161e-05, ..., 1.7887e-04,\n 0.0000e+00, -1.0776e-06],\n [-4.4393e-07, 1.0571e-06, -3.6268e-07, ..., -7.2545e-06,\n 0.0000e+00, -2.9153e-06],\n [-1.3460e-07, 1.5739e-06, 1.3340e-06, ..., -4.3566e-06,\n 0.0000e+00, 3.0733e-06]], device='cuda:0')", + "exp_avg_sq": "tensor([[4.7025e-11, 6.6304e-11, 5.2935e-11, ..., 8.1728e-11, 0.0000e+00,\n 4.4678e-11],\n [6.6985e-11, 7.9299e-11, 3.3033e-10, ..., 3.8502e-11, 0.0000e+00,\n 2.0548e-10],\n [1.0850e-10, 3.2719e-10, 4.2286e-11, ..., 2.8572e-10, 0.0000e+00,\n 1.9407e-10],\n ...,\n [8.3318e-11, 7.8421e-12, 1.5768e-09, ..., 6.0843e-09, 0.0000e+00,\n 6.0611e-10],\n [1.3663e-10, 1.7230e-10, 4.3073e-11, ..., 2.6909e-10, 0.0000e+00,\n 1.2199e-10],\n [6.6905e-11, 3.0412e-11, 3.7693e-10, ..., 2.2863e-10, 0.0000e+00,\n 4.6511e-10]], device='cuda:0')" + }, + "27": { + "step": "tensor(2504.)", + "exp_avg": "tensor([ 2.4152e-05, 4.0226e-05, -2.2483e-04, ..., 1.2204e-04,\n -1.1629e-05, -3.6174e-05], device='cuda:0')", + "exp_avg_sq": "tensor([1.9226e-08, 1.8703e-08, 5.1420e-08, ..., 3.1267e-08, 1.6952e-08,\n 2.9708e-08], device='cuda:0')" + }, + "28": { + "step": "tensor(2504.)", + "exp_avg": "tensor([[ 7.0014e-08, -2.3550e-07, -1.8928e-07, ..., 9.9566e-07,\n 3.5126e-08, 8.7351e-07],\n [ 3.3054e-07, -1.3881e-07, 6.6509e-07, ..., 2.5783e-06,\n -2.5351e-07, 2.2954e-07],\n [-1.7870e-06, -8.3651e-08, -1.6035e-06, ..., 5.2402e-08,\n 4.0490e-08, 7.3153e-07],\n ...,\n [-1.4263e-06, 6.4761e-07, -3.9834e-07, ..., 4.7154e-07,\n 4.9860e-07, 4.2942e-07],\n [-1.4211e-07, -3.5462e-07, -1.1245e-06, ..., 8.8498e-07,\n -4.1586e-08, -1.3816e-06],\n [ 6.1833e-07, 2.5424e-07, -6.6536e-07, ..., -9.7480e-07,\n -6.8482e-07, 4.6943e-07]], device='cuda:0')", + "exp_avg_sq": "tensor([[3.9734e-12, 3.6468e-12, 1.1986e-11, ..., 7.9659e-12, 4.1398e-12,\n 6.7981e-12],\n [8.2350e-12, 5.4648e-12, 8.4214e-12, ..., 1.3570e-11, 9.0190e-12,\n 1.0688e-11],\n [8.6306e-12, 1.0488e-11, 1.5727e-11, ..., 1.3385e-11, 8.3640e-12,\n 1.1622e-11],\n ...,\n [1.4206e-11, 7.6523e-12, 1.6370e-11, ..., 1.7749e-11, 1.4635e-11,\n 1.3071e-11],\n [8.9737e-12, 6.8857e-12, 9.2066e-12, ..., 1.3603e-11, 9.1160e-12,\n 1.5713e-11],\n [8.0421e-12, 7.1649e-12, 9.7956e-12, ..., 1.0800e-11, 1.1423e-11,\n 1.1974e-11]], device='cuda:0')" + }, + "29": { + "step": "tensor(1252.)", + "exp_avg": "tensor([[ 4.6549e-06, 1.0463e-07, 1.1726e-05, ..., 3.6087e-06,\n 0.0000e+00, -1.1304e-05],\n [-9.7613e-06, 2.8244e-06, 1.8208e-06, ..., -1.0593e-06,\n 0.0000e+00, 2.3663e-06],\n [ 3.5186e-06, -2.2636e-06, -6.4493e-06, ..., 1.2065e-05,\n 0.0000e+00, -5.4349e-06],\n ...,\n [ 2.2149e-06, -7.3551e-07, -1.5728e-05, ..., 1.5083e-05,\n 0.0000e+00, -1.8443e-06],\n [ 4.7179e-06, -1.4825e-06, -4.6107e-07, ..., -3.2699e-06,\n 0.0000e+00, 2.1100e-06],\n [-3.4079e-06, -8.8482e-07, 7.7006e-06, ..., 4.0681e-06,\n 0.0000e+00, -1.1189e-06]], device='cuda:0')", + "exp_avg_sq": "tensor([[3.1810e-10, 1.4450e-10, 1.0865e-09, ..., 6.7395e-10, 0.0000e+00,\n 8.9202e-10],\n [5.3529e-10, 9.5948e-11, 3.6140e-10, ..., 2.2561e-10, 0.0000e+00,\n 2.8036e-10],\n [2.4466e-10, 1.3245e-10, 4.2474e-10, ..., 7.6973e-10, 0.0000e+00,\n 1.4730e-09],\n ...,\n [1.7838e-10, 1.8297e-10, 1.4252e-09, ..., 7.3969e-10, 0.0000e+00,\n 6.2450e-10],\n [3.8208e-10, 2.1327e-10, 1.6977e-10, ..., 3.3027e-10, 0.0000e+00,\n 3.0958e-10],\n [6.4949e-10, 1.5909e-10, 1.8713e-09, ..., 4.3686e-10, 0.0000e+00,\n 9.1388e-10]], device='cuda:0')" + }, + "30": { + "step": "tensor(1252.)", + "exp_avg": "tensor([ 1.0712e-04, -1.0426e-05, -2.6064e-05, ..., -6.4751e-05,\n -5.9825e-06, -4.5317e-05], device='cuda:0')", + "exp_avg_sq": "tensor([5.7841e-08, 6.2386e-08, 6.5328e-08, ..., 5.7239e-08, 5.1020e-08,\n 7.6560e-08], device='cuda:0')" + }, + "31": { + "step": "tensor(1252.)", + "exp_avg": "tensor([[-8.7127e-07, 2.3044e-06, 5.4709e-07, ..., 7.9924e-07,\n -1.5380e-06, -2.5151e-06],\n [-2.4811e-06, -9.3771e-07, 4.0047e-06, ..., -2.0239e-06,\n 5.3400e-07, -1.1437e-06],\n [-4.2748e-07, -2.8912e-06, -8.8406e-07, ..., -2.2906e-06,\n 3.9683e-06, -1.3857e-06],\n ...,\n [-5.2646e-07, -2.6523e-07, 2.5794e-07, ..., 1.8322e-06,\n 7.0432e-07, -1.9736e-06],\n [ 9.9442e-07, -2.8867e-06, -2.5019e-06, ..., -6.7045e-06,\n -3.0750e-06, -9.6310e-06],\n [-7.0356e-07, 3.3788e-06, -1.2132e-06, ..., 5.8414e-06,\n 4.4418e-06, 6.7221e-06]], device='cuda:0')", + "exp_avg_sq": "tensor([[3.7381e-11, 4.7652e-11, 6.1029e-11, ..., 3.9400e-11, 5.7003e-11,\n 4.7475e-11],\n [6.2163e-11, 7.9086e-11, 1.2563e-10, ..., 7.5610e-11, 8.1847e-11,\n 7.7077e-11],\n [5.6396e-11, 7.7227e-11, 1.3232e-10, ..., 6.4271e-11, 8.3451e-11,\n 7.6538e-11],\n ...,\n [5.1067e-11, 6.8155e-11, 1.2001e-10, ..., 6.5949e-11, 7.1325e-11,\n 8.0999e-11],\n [7.6909e-11, 8.6984e-11, 1.4603e-10, ..., 7.7356e-11, 8.8286e-11,\n 8.7655e-11],\n [6.5247e-11, 8.7666e-11, 1.1892e-10, ..., 7.5368e-11, 8.9398e-11,\n 8.1653e-11]], device='cuda:0')" } }, "param_groups": [ { - "lr": 0.005000500000000001, + "lr": 0.01, "name": "shared", "betas": [ 0.9, @@ -242,8 +332,8 @@ ] }, { - "lr": 0.005000500000000001, - "name": "scale_384", + "lr": 0.01, + "name": "scale_256", "betas": [ 0.9, 0.999 @@ -265,8 +355,8 @@ ] }, { - "lr": 0.005000500000000001, - "name": "scale_768", + "lr": 0.01, + "name": "scale_512", "betas": [ 0.9, 0.999 @@ -288,8 +378,8 @@ ] }, { - "lr": 0.005000500000000001, - "name": "scale_1024", + "lr": 0.01, + "name": "scale_768", "betas": [ 0.9, 0.999 @@ -311,8 +401,8 @@ ] }, { - "lr": 0.005000500000000001, - "name": "scale_1280", + "lr": 0.01, + "name": "scale_1024", "betas": [ 0.9, 0.999 @@ -334,8 +424,8 @@ ] }, { - "lr": 0.0025005, - "name": "fusion", + "lr": 0.01, + "name": "scale_1280", "betas": [ 0.9, 0.999 @@ -349,26 +439,146 @@ "differentiable": false, "fused": null, "decoupled_weight_decay": true, - "initial_lr": 0.005, + "initial_lr": 0.01, "params": [ 14, 15, - 16, + 16 + ] + }, + { + "lr": 0.01, + "name": "scale_1536", + "betas": [ + 0.9, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-05, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": true, + "initial_lr": 0.01, + "params": [ 17, 18, - 19, + 19 + ] + }, + { + "lr": 0.01, + "name": "scale_1792", + "betas": [ + 0.9, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-05, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": true, + "initial_lr": 0.01, + "params": [ 20, 21, - 22, + 22 + ] + }, + { + "lr": 0.01, + "name": "scale_2048", + "betas": [ + 0.9, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-05, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": true, + "initial_lr": 0.01, + "params": [ 23, 24, - 25, + 25 + ] + }, + { + "lr": 0.01, + "name": "scale_2304", + "betas": [ + 0.9, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-05, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": true, + "initial_lr": 0.01, + "params": [ 26, 27, - 28, + 28 + ] + }, + { + "lr": 0.01, + "name": "scale_2560", + "betas": [ + 0.9, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-05, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": true, + "initial_lr": 0.01, + "params": [ 29, 30, - 31, + 31 + ] + }, + { + "lr": 0.005, + "name": "fusion", + "betas": [ + 0.9, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-05, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": true, + "initial_lr": 0.005, + "params": [ 32, 33, 34, @@ -380,7 +590,25 @@ 40, 41, 42, - 43 + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61 ] } ] @@ -390,8 +618,14 @@ "T_i": 20, "T_mult": 2, "eta_min": 1e-06, - "T_cur": 10, + "T_cur": 0, "base_lrs": [ + 0.01, + 0.01, + 0.01, + 0.01, + 0.01, + 0.01, 0.01, 0.01, 0.01, @@ -399,45 +633,57 @@ 0.01, 0.005 ], - "last_epoch": 20, + "last_epoch": 10, "_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 + 0.01, + 0.01, + 0.01, + 0.01, + 0.01, + 0.01, + 0.01, + 0.01, + 0.01, + 0.01, + 0.01, + 0.005 ] }, "metrics": { - "final_val_acc": 81.212 + "final_val_acc": 82.374 }, "train_config": { "name": "david_training", - "run_id": "20251012_041353", + "run_id": "20251012_050214", "dataset_name": "AbstractPhil/imagenet-clip-features-orderly", "model_variant": "clip_vit_l14", "num_classes": 1000, - "preset": "clip_vit_l14", + "preset": "clip_vit_l14_deep", "custom_config_path": null, "num_classes_override": null, "use_belly_override": null, "belly_expand_override": null, "progressive_training_override": true, "scale_warmup_epochs_override": { - "384": 0, - "768": 1, - "1024": 2, - "1280": 3 + "256": 0, + "512": 1, + "768": 2, + "1024": 3, + "1280": 4, + "1536": 5, + "1792": 6, + "2048": 7, + "2304": 8, + "2560": 9 }, - "num_epochs": 20, + "num_epochs": 10, "batch_size": 1024, "learning_rate": 0.01, "weight_decay": 1e-05, - "warmup_epochs": 3, + "warmup_epochs": 0, "use_rose_loss": true, "rose_initial_weight": 0.1, "rose_max_weight": 0.5,