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" @@ -3,250 +3,225 @@ "optimizer_state_dict": { "state": { "0": { - "step": "tensor(8764.)", - "exp_avg": "tensor([[ 2.0182e-05, -4.0774e-05, -1.8981e-06, ..., 2.2015e-05,\n 1.9074e-05, 1.4373e-05],\n [-7.9625e-06, -3.4462e-05, 2.4047e-05, ..., 1.1974e-05,\n 5.1719e-05, 6.0935e-06],\n [-1.0442e-04, -2.1260e-04, -5.8897e-05, ..., 3.3500e-05,\n -1.6321e-05, 4.1818e-05],\n ...,\n [ 1.4544e-04, -1.4001e-04, -6.1162e-06, ..., 1.9627e-05,\n 1.2254e-05, -1.5506e-05],\n [-2.5096e-06, -7.7651e-06, -2.3486e-06, ..., -8.7364e-07,\n -2.1224e-06, 2.1338e-06],\n [-1.1176e-04, 1.1404e-04, 2.5317e-06, ..., -3.0614e-05,\n -4.9245e-05, 3.8300e-05]], device='cuda:0')", - "exp_avg_sq": "tensor([[3.3660e-08, 4.8834e-08, 6.4916e-09, ..., 1.5737e-08, 1.4136e-08,\n 3.8758e-09],\n [1.2246e-07, 7.6356e-08, 3.1421e-08, ..., 4.7450e-08, 2.6371e-08,\n 4.2753e-08],\n [5.4321e-08, 7.3815e-08, 1.8063e-08, ..., 2.0440e-08, 1.2274e-08,\n 1.6688e-08],\n ...,\n [1.4520e-07, 8.5617e-08, 1.7459e-08, ..., 2.0662e-08, 1.5681e-08,\n 1.0392e-08],\n [7.5145e-11, 6.6761e-11, 1.3205e-11, ..., 1.8387e-11, 1.7622e-11,\n 2.4248e-11],\n [9.9692e-08, 6.6532e-08, 9.6312e-09, ..., 1.4973e-08, 2.2145e-08,\n 1.0498e-08]], device='cuda: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')" }, "1": { - "step": "tensor(8764.)", - "exp_avg": "tensor([ 8.2773e-04, -6.4425e-04, 1.7857e-03, 4.9067e-04, 6.1624e-04,\n 8.7618e-05, 4.9512e-04, -1.6068e-03, 1.5713e-03, -2.4647e-05,\n 9.6222e-04, 5.7114e-04, 5.6052e-45, -2.0018e-03, -1.0832e-04,\n 4.1191e-04, -1.6788e-03, -1.2133e-03, -3.9914e-04, 5.6052e-45,\n 1.1034e-03, -1.3365e-03, -1.4586e-03, -4.6580e-04, 5.6052e-45,\n 2.1290e-03, 4.8886e-04, -4.8179e-03, -2.7341e-03, 2.3249e-03,\n 3.2989e-04, -8.8534e-04, -3.4335e-19, 5.6052e-45, 2.0127e-03,\n 9.6239e-04, -2.5377e-03, 1.0100e-03, 8.5631e-04, -1.5793e-03,\n 1.2097e-03, -2.7516e-04, 2.5896e-03, 5.9248e-04, -8.9859e-04,\n 2.9885e-04, -1.0231e-04, 2.4279e-03, -2.9709e-03, -1.1404e-03,\n -7.4071e-04, 5.6052e-45, 1.4141e-17, 5.6052e-45, -1.8489e-03,\n 1.2444e-03, 3.0435e-03, -1.6353e-03, -1.4592e-05, -1.0750e-03,\n -5.5469e-04, 8.2039e-04, 1.6815e-03, -7.3230e-05, -8.3017e-04,\n -8.9741e-05, -3.7664e-04, 5.6052e-45, -2.7964e-03, 1.6682e-03,\n 6.2612e-04, 5.6052e-45, 1.9980e-03, 1.0338e-04, 1.7844e-03,\n 3.5141e-03, 1.9415e-03, 1.9057e-03, -2.0060e-03, -2.1909e-03,\n -1.3067e-03, 9.9421e-04, 1.9729e-04, 3.6982e-04, -1.3752e-04,\n 1.2777e-03, -9.4964e-04, -8.6883e-04, 9.8645e-04, -8.9502e-04,\n 5.6052e-45, 2.0321e-03, 3.6609e-04, -1.8435e-04, -2.0463e-03,\n -5.6449e-04, -8.4193e-04, -8.1765e-06, 8.0381e-05, 3.6658e-03,\n -1.2114e-03, 2.0239e-03, -7.1849e-04, 2.1659e-04, -8.7701e-03,\n 8.4257e-04, 5.4097e-04, -1.1808e-03, 4.3671e-03, -3.8656e-04,\n -6.4111e-04, 1.7248e-04, 5.6052e-45, 1.1811e-03, -5.5449e-04,\n -4.1107e-04, 1.1029e-04, 2.6941e-04, 5.6052e-45, -8.7269e-05,\n 5.6052e-45, -4.0957e-04, 5.6052e-45, -1.5132e-03, -8.9312e-04,\n 1.8585e-03, -7.7790e-04, 5.6052e-45, 2.3466e-04, 9.1805e-04,\n 2.1684e-04, -5.6645e-04, -1.2484e-03, 5.6052e-45, 1.6146e-03,\n 5.1343e-03, -3.0490e-03, -7.0139e-04, -1.7743e-03, 2.3718e-04,\n -2.9958e-03, -4.7922e-04, -8.0391e-04, 3.3454e-04, -5.8078e-04,\n 1.1812e-03, -7.1774e-17, -8.4783e-05, 9.4378e-04, -1.2443e-03,\n 5.6052e-45, -9.2922e-04, -9.2161e-04, 5.6052e-45, -1.1006e-03,\n 1.3668e-03, 5.6052e-45, -2.5135e-03, 1.4426e-04, -1.3429e-05,\n 3.8922e-03, 1.7803e-03, -8.6296e-04, 4.4729e-04, 1.7739e-03,\n 1.1162e-03, -3.7004e-03, -1.9289e-04, 3.5782e-03, 5.6052e-45,\n 5.6052e-45, 1.9803e-03, 2.3266e-04, 4.9124e-04, -5.1448e-04,\n -3.1422e-03, 9.1015e-04, -7.3558e-04, 5.6052e-45, 3.3636e-04,\n 5.3761e-04, -4.1913e-04, 6.1199e-04, 5.6052e-45, -2.2210e-03,\n -2.7276e-03, -3.1889e-03, 2.8493e-04, -2.0855e-04, 5.6052e-45,\n 5.6052e-45, 9.5652e-04, -1.2630e-03, 1.6360e-03, 5.5614e-04,\n 3.4465e-04, -3.5330e-04, 5.2866e-04, 5.6052e-45, -1.0250e-03,\n 1.5011e-04, -6.9857e-04, -2.7772e-05, 1.0855e-03, 2.5180e-03,\n 5.6052e-45, 2.4834e-04, -7.4751e-04, 1.5066e-03, 3.8407e-04,\n 4.4421e-04, 6.1384e-04, -2.5369e-03, -2.2903e-05, -5.2461e-04,\n -7.2133e-04, 1.0224e-03, -5.3143e-04, 1.5448e-03, -3.0465e-03,\n 5.8001e-04, 1.8682e-03, -3.8420e-04, 1.6195e-03, 6.9232e-04,\n -1.9160e-03, -4.5354e-05, 1.8747e-03, -1.4597e-04, -6.8441e-04,\n -3.8738e-04, 6.0303e-03, -2.7899e-05, 5.6052e-45, 5.2764e-03,\n -7.4054e-04, -2.0248e-03, 3.5693e-03, 3.1313e-03, -7.8526e-04,\n -1.5008e-03, 3.2735e-04, -1.8656e-03, 9.5574e-04, 2.3722e-03,\n -3.9908e-04, -1.5825e-03, 4.1585e-03, -8.0897e-04, 2.0960e-03,\n -2.3298e-03, -1.4655e-03, -6.2848e-03, 6.5966e-04, -6.1024e-04,\n 8.7266e-04, 5.6052e-45, -3.4838e-04, 7.1056e-04, -2.7258e-04,\n -1.4859e-03, 2.6273e-03, 3.0573e-03, 3.0832e-04, -6.4224e-03,\n -1.5697e-03, 1.3539e-04, 5.6052e-45, 2.7669e-03, -3.4179e-03,\n 6.6493e-04, -2.7624e-03, -4.4543e-04, -5.2986e-04, -1.0633e-03,\n 1.9116e-08, 2.8892e-04, 5.6052e-45, 4.2836e-04, -1.3736e-05,\n 4.0174e-03, 2.5140e-04, 1.4610e-04, 1.1332e-03, -1.1089e-03,\n 1.8500e-03, -1.7069e-06, 1.6849e-03, 1.8568e-03, -1.4822e-03,\n -2.7685e-04, 1.0891e-04, 1.2516e-03, 1.0307e-03, -4.8170e-05,\n -1.9083e-03, -1.5284e-03, 4.2632e-04, 8.9915e-04, 2.1877e-04,\n -3.4887e-04, 1.0388e-03, -3.4039e-04, -1.1169e-04, 1.5134e-03,\n 5.9226e-04, -2.5771e-04, 5.6052e-45, 5.6052e-45, -2.9321e-03,\n 2.5862e-03, -2.0084e-03, 1.3321e-05, -2.5144e-03, 1.3686e-03,\n 4.1642e-03, -8.5512e-04, 7.2002e-04, 1.4807e-03, -2.0000e-03,\n 7.3819e-04, -3.1331e-03, 2.9554e-03, 2.8573e-04, 2.9124e-05,\n 1.1793e-03, -6.0067e-04, -4.9977e-04, -7.4153e-05, 1.0614e-03,\n 1.5445e-04, 1.2792e-03, 4.8051e-04, 1.9615e-04, 5.6052e-45,\n -4.8989e-04, 3.6085e-03, -1.4915e-04, 9.9903e-04, 1.5752e-03,\n -1.4566e-03, -1.4576e-03, -8.7608e-04, -5.3462e-04, -3.4671e-04,\n 1.1216e-05, 5.6052e-45, -1.2838e-04, -3.7601e-03, 7.8891e-04,\n -1.8381e-04, 1.7889e-05, -2.3339e-04, -1.0192e-03, -3.1364e-03,\n -4.2878e-04, 1.4757e-03, 3.7549e-05, 1.0976e-03, -1.2838e-03,\n 3.9614e-04, 9.8725e-04, 1.0959e-03, 1.6716e-03, -1.6409e-04,\n 1.9207e-05, 1.7592e-03, -2.2722e-03, -1.0977e-03, 5.2207e-05,\n -1.4760e-03, 1.5352e-04, -5.7351e-04, 1.1521e-03, -4.9552e-05,\n 1.5729e-03, -7.5888e-04, 8.1303e-04, 5.6052e-45, 9.0218e-04,\n -1.7156e-03, 5.6052e-45, 1.6691e-03, 1.3558e-03, 2.4062e-03,\n -2.5692e-04, -1.1860e-03, -2.4563e-03, -1.2424e-03, 1.2869e-03,\n 2.1723e-03, -1.6746e-03, 4.9060e-04, -5.6052e-45, 5.6052e-45,\n 2.0775e-03, -1.8036e-03, -1.3034e-03, -2.1443e-03, 4.2014e-04,\n -6.9643e-04, -1.1622e-03, -1.0378e-04, 8.2443e-04, 7.4210e-04,\n 1.3883e-03, -1.1441e-03, 5.6052e-45, 4.8680e-04, 4.2605e-04,\n 5.6052e-45, -1.4322e-03, 5.6052e-45, -2.5965e-03, -7.5789e-04,\n 1.0123e-03, -8.5900e-04, 5.6052e-45, -1.9630e-03, 1.2519e-03,\n -1.3044e-04, -1.1691e-03, -5.2099e-04, 1.0979e-03, -2.3067e-03,\n 4.6410e-04, -1.4142e-03, -4.5494e-04, 2.1748e-03, 5.6052e-45,\n -1.1997e-03, 1.5133e-03, -3.0294e-04, 5.6052e-45, -8.1899e-04,\n -1.0962e-03, -2.4025e-03, -1.5583e-03, 1.5534e-03, 1.5569e-03,\n -1.4238e-04, -1.7844e-03, 2.1826e-03, -1.2957e-03, -3.8465e-04,\n 1.1225e-03, -1.2652e-03, 5.6052e-45, -2.0127e-04, -8.5212e-04,\n 5.6052e-45, -1.9310e-03, -2.0809e-03, 7.3990e-04, -2.5793e-04,\n -1.4591e-06, 1.3225e-03, -6.5221e-04, 7.3424e-04, 5.6052e-45,\n 5.5580e-04, 1.9511e-04, 5.6052e-45, 1.0027e-04, -4.1564e-04,\n -1.6945e-03, 9.8967e-04, 5.6052e-45, -3.2336e-03, 4.6944e-04,\n 2.9962e-03, -2.5836e-04, -1.4509e-05, -1.3102e-03, 2.8218e-04,\n 6.0319e-04, 5.6052e-45, -1.2067e-03, 5.6052e-45, -6.4428e-04,\n 3.9529e-05, 1.3031e-03, -2.5292e-03, -6.4129e-06, -1.7618e-04,\n 6.7476e-04, -1.1076e-04, -5.8217e-05, 6.1865e-04, 3.1321e-04,\n 5.6052e-45, -8.4158e-04, -7.0308e-04, 7.9133e-03, 3.1855e-04,\n 1.5999e-03, 7.1572e-04, 1.5979e-03, 3.3382e-03, 4.2100e-04,\n 1.3982e-03, 5.6052e-45, -1.0131e-03, 1.4660e-03, 1.3862e-03,\n 1.9460e-03, -1.5197e-03, -1.0353e-03, -1.1254e-06, -4.0425e-04,\n 4.0114e-04, 4.9384e-04, -1.6642e-03, -2.5886e-03, 5.6052e-45,\n 5.6052e-45, 1.2701e-03, -6.7478e-04, 1.3946e-03, 2.7300e-05,\n -1.7730e-03, -1.9198e-03, 6.0817e-04, 1.3600e-03, 9.2329e-04,\n 1.2708e-04, -1.0352e-03, -2.6717e-04, -1.0693e-03, 5.0040e-04,\n 2.5503e-04, 2.3313e-04, -1.5503e-03, -1.1250e-04, 1.1253e-13,\n 3.6064e-04, -1.4253e-03, 5.6052e-45, 5.6052e-45, -3.8738e-03,\n -1.1088e-03, 2.7068e-04, -1.3287e-03, -4.9581e-03, 1.3900e-04,\n -1.9448e-03, -2.2386e-03, 1.5938e-05, -3.1080e-04, -1.0636e-03,\n -3.6972e-04, 1.5126e-03, 5.0351e-04, 2.2495e-04, 3.5768e-04,\n 1.7465e-03, -7.7100e-05, 3.7583e-03, 1.4316e-03, 6.8150e-04,\n -5.8458e-04, -9.2893e-04, -4.9790e-04, 6.9771e-04, 2.4666e-04,\n -6.7207e-04, 4.7628e-04, 4.9579e-04, -1.3267e-03, 2.5109e-04,\n 3.8340e-03, 1.4504e-04, 1.5863e-03, -2.2466e-03, 9.5885e-04,\n 5.6052e-45, 5.6052e-45, 8.5443e-04, -1.4534e-03, 1.3633e-04,\n -9.7699e-04, 5.6052e-45, -2.0840e-03, 3.8862e-04, -4.0500e-04,\n 4.1766e-03, -1.6476e-04, 7.7546e-04, -1.5928e-04, 1.0407e-03,\n -4.1593e-03, -7.5137e-05, 2.3522e-03, 1.8350e-03, -5.5561e-04,\n -6.0348e-04, -9.6258e-04, -8.1638e-04, -3.2454e-03, -1.2969e-03,\n 7.4895e-04, 1.1299e-03, 1.5143e-03, 1.0670e-03, -1.6771e-03,\n -2.6141e-03, 2.0824e-04, 1.9134e-04, 4.3226e-04, -1.6366e-03,\n -2.3573e-04, -1.0413e-03, -3.3944e-04, -1.9002e-03, 4.5692e-04,\n 1.1255e-03, 3.1364e-04, 3.4930e-03, 3.0225e-04, 1.1595e-06,\n 5.6052e-45, -2.7086e-05, 1.5793e-03, 8.4390e-04, -6.3016e-04,\n -4.9585e-03, 1.9688e-04, 5.4001e-05, 1.6109e-03, -2.3454e-03,\n -1.4736e-03, 2.4908e-09, 1.5868e-03, 1.8527e-03, -4.3517e-04,\n -7.9860e-04, -2.0151e-03, 1.8394e-03, 2.5244e-05, -1.4951e-03],\n device='cuda:0')", - "exp_avg_sq": "tensor([1.1273e-05, 5.1797e-05, 1.9440e-05, 1.7217e-05, 1.9449e-05, 4.4492e-06,\n 2.9770e-05, 2.8510e-05, 2.0807e-05, 2.5146e-05, 2.5313e-05, 5.3386e-05,\n 3.5182e-09, 2.4866e-05, 3.0669e-05, 3.3956e-05, 4.3389e-05, 2.8850e-05,\n 2.2814e-05, 9.1601e-10, 2.4757e-05, 1.8483e-05, 1.9234e-05, 3.0473e-05,\n 3.3436e-09, 3.8172e-05, 2.5830e-05, 4.0012e-05, 3.2905e-05, 3.2854e-05,\n 3.7264e-05, 2.9129e-05, 1.1742e-08, 5.9053e-10, 3.0114e-05, 2.2832e-05,\n 3.3176e-05, 1.8133e-05, 2.1886e-05, 2.8002e-05, 3.0998e-05, 1.7940e-05,\n 2.7404e-05, 2.4656e-05, 2.6293e-05, 3.0807e-05, 2.7662e-05, 4.4733e-05,\n 5.2698e-05, 2.4782e-05, 3.0028e-05, 3.5078e-10, 5.6919e-10, 5.3269e-10,\n 4.2056e-05, 5.6650e-05, 2.2027e-05, 2.2191e-05, 7.7930e-06, 3.4976e-05,\n 3.0242e-05, 2.6698e-05, 5.3060e-05, 1.5441e-05, 3.1232e-05, 3.2492e-05,\n 2.7819e-05, 6.5197e-10, 4.7832e-05, 2.9770e-05, 1.1872e-05, 2.6632e-09,\n 3.5276e-05, 2.5343e-05, 1.2339e-05, 3.7042e-05, 3.8841e-05, 3.8538e-05,\n 2.1571e-05, 3.7756e-05, 4.9254e-05, 2.4929e-05, 3.5599e-05, 2.4263e-05,\n 2.8473e-05, 3.0795e-05, 2.6701e-05, 1.9337e-05, 2.7760e-05, 2.7242e-05,\n 4.1576e-09, 3.8606e-05, 2.6639e-05, 2.1874e-05, 4.3723e-05, 1.2501e-05,\n 3.8363e-05, 3.1196e-05, 2.6740e-05, 2.3173e-05, 2.2998e-05, 1.2007e-05,\n 3.9679e-05, 4.0403e-05, 7.9596e-05, 2.3888e-05, 2.6269e-05, 4.7644e-05,\n 1.6745e-05, 1.7231e-05, 4.1824e-05, 4.0237e-05, 2.8110e-10, 1.9598e-05,\n 1.0910e-05, 2.3334e-05, 3.1267e-05, 4.7966e-05, 2.2597e-09, 1.0256e-07,\n 4.5591e-10, 2.6255e-05, 3.6304e-10, 5.2553e-05, 2.7509e-05, 3.5435e-05,\n 3.8970e-05, 6.8291e-11, 1.2914e-05, 3.0851e-05, 2.6186e-05, 3.1513e-05,\n 3.2568e-05, 1.2183e-10, 3.2759e-05, 3.1417e-05, 2.0791e-05, 2.2869e-05,\n 2.7484e-05, 2.2866e-05, 5.5160e-05, 2.8264e-05, 4.1226e-05, 2.9687e-05,\n 3.2551e-05, 3.6347e-05, 5.3766e-09, 3.0543e-05, 2.9276e-05, 2.4711e-05,\n 4.6160e-09, 1.7543e-05, 2.5490e-05, 2.4740e-09, 2.4890e-05, 3.8092e-05,\n 4.9008e-09, 4.4546e-05, 2.4380e-05, 3.1733e-05, 2.6021e-05, 2.8000e-05,\n 7.2468e-05, 3.1023e-05, 3.0183e-05, 2.3696e-05, 4.6504e-05, 1.8179e-05,\n 3.3737e-05, 5.8220e-10, 8.9719e-10, 2.9084e-05, 4.6298e-05, 2.1423e-05,\n 2.6156e-05, 4.2278e-05, 1.3106e-05, 3.5007e-05, 1.6283e-10, 1.8783e-05,\n 3.1891e-05, 3.1791e-05, 9.3305e-06, 1.1981e-09, 3.0791e-05, 3.2247e-05,\n 4.4904e-05, 2.0346e-05, 2.7734e-05, 3.4790e-10, 3.9861e-09, 2.8048e-05,\n 1.3671e-05, 3.4568e-05, 2.4713e-05, 3.2567e-05, 1.8238e-05, 3.3351e-05,\n 6.2828e-10, 2.4609e-05, 2.3380e-05, 1.3767e-05, 4.1968e-05, 1.4906e-05,\n 6.0364e-05, 3.0989e-11, 2.3023e-05, 2.8761e-05, 2.1595e-05, 2.4372e-05,\n 1.3465e-05, 3.0027e-05, 3.2928e-05, 2.9476e-05, 2.0268e-05, 2.2353e-05,\n 2.1550e-05, 3.8122e-05, 3.0614e-05, 2.0302e-05, 1.6543e-05, 2.9863e-05,\n 2.3059e-05, 2.4036e-05, 2.8494e-05, 2.8400e-05, 3.9922e-05, 4.1975e-05,\n 1.8396e-05, 3.9321e-05, 3.3105e-05, 6.0085e-05, 3.0780e-05, 7.7029e-10,\n 3.4863e-05, 1.4814e-05, 2.4222e-05, 1.7474e-05, 7.0323e-05, 3.0378e-05,\n 2.5957e-05, 3.4401e-05, 2.5851e-05, 3.1982e-05, 4.1810e-05, 1.7741e-05,\n 1.7698e-05, 4.2603e-05, 2.0581e-05, 1.8894e-05, 2.8759e-05, 2.4901e-05,\n 2.6695e-05, 2.5855e-05, 2.9800e-05, 2.2438e-05, 3.3190e-09, 2.3369e-05,\n 1.8397e-05, 1.1830e-05, 2.9601e-05, 4.0504e-05, 3.6252e-05, 2.7336e-05,\n 3.1382e-05, 2.4348e-05, 2.4407e-05, 4.9198e-10, 3.7481e-05, 3.0397e-05,\n 2.2471e-05, 4.0952e-05, 2.2840e-05, 1.8426e-05, 3.4652e-05, 8.6132e-10,\n 2.8678e-05, 2.4261e-09, 1.9666e-05, 1.7287e-05, 2.5201e-05, 2.5379e-05,\n 4.0488e-05, 2.8273e-05, 2.8031e-05, 3.3987e-05, 7.6536e-06, 4.5791e-05,\n 3.9669e-05, 1.8340e-05, 9.2352e-06, 2.9238e-05, 3.8767e-05, 3.1807e-05,\n 2.1820e-05, 5.4615e-05, 5.3549e-05, 2.9483e-05, 2.7077e-05, 6.9100e-05,\n 2.7941e-05, 9.1014e-06, 9.2246e-06, 2.4754e-05, 2.7202e-05, 1.3513e-05,\n 3.2050e-05, 1.0212e-10, 2.0119e-09, 3.1486e-05, 4.2039e-05, 1.7291e-05,\n 2.5225e-05, 3.8511e-05, 4.2790e-05, 2.5054e-05, 3.6087e-05, 1.9927e-05,\n 2.1730e-05, 3.4800e-05, 4.0779e-05, 4.6161e-05, 3.6755e-05, 1.7430e-05,\n 1.2628e-05, 2.6797e-05, 2.5732e-05, 1.0281e-05, 2.2095e-05, 2.9692e-05,\n 9.0620e-06, 1.9599e-05, 3.2504e-05, 4.8332e-05, 2.4002e-13, 2.5450e-05,\n 2.0987e-05, 4.0258e-05, 1.6686e-05, 2.8730e-05, 6.0505e-06, 3.0033e-05,\n 2.0023e-05, 2.4150e-05, 2.5331e-05, 3.6898e-05, 3.5125e-09, 1.0199e-05,\n 2.9095e-05, 2.9642e-05, 2.3918e-05, 3.0196e-05, 2.3201e-05, 2.3486e-05,\n 4.5781e-05, 2.7998e-05, 2.7640e-05, 2.6392e-05, 3.6736e-05, 1.0141e-05,\n 1.5378e-05, 1.7884e-05, 3.7931e-05, 4.9622e-05, 3.3265e-05, 7.5043e-05,\n 1.3823e-05, 3.6306e-05, 3.7831e-05, 3.0736e-05, 3.6577e-05, 2.0074e-05,\n 2.5152e-05, 9.4860e-06, 2.3829e-05, 4.6180e-05, 5.7710e-05, 9.7927e-06,\n 1.5291e-09, 3.0029e-05, 2.3761e-05, 8.3429e-09, 2.4965e-05, 5.2599e-05,\n 3.2481e-05, 2.4719e-05, 2.7854e-05, 3.0429e-05, 1.8981e-05, 1.6101e-05,\n 4.4055e-05, 3.8381e-05, 2.6587e-05, 8.3153e-10, 2.7515e-09, 2.6601e-05,\n 3.1561e-05, 2.7640e-05, 4.8663e-05, 2.8106e-05, 3.0985e-05, 1.7375e-05,\n 1.2416e-05, 2.2789e-05, 2.6778e-05, 4.1648e-05, 2.1579e-05, 1.0133e-09,\n 1.3722e-05, 9.3331e-06, 2.9190e-10, 1.1182e-05, 5.0925e-10, 3.5357e-05,\n 3.7015e-05, 1.4003e-05, 2.6715e-05, 1.3749e-09, 1.6317e-05, 3.5460e-05,\n 2.2291e-05, 3.1739e-05, 3.5743e-05, 3.6435e-05, 2.6887e-05, 4.8309e-06,\n 2.6141e-05, 9.2170e-06, 1.7095e-05, 6.9109e-10, 2.0324e-05, 3.4510e-05,\n 2.9027e-05, 8.8316e-10, 1.6884e-05, 2.4937e-05, 4.1931e-05, 3.0977e-05,\n 2.8933e-05, 2.1816e-05, 4.0242e-05, 1.4872e-05, 2.7699e-05, 4.1151e-05,\n 2.5864e-05, 1.2116e-05, 2.8399e-05, 2.9808e-10, 2.4562e-05, 2.0554e-05,\n 9.8897e-09, 2.3354e-05, 2.7550e-05, 5.8115e-06, 2.3847e-05, 7.3918e-09,\n 1.7782e-05, 3.7284e-06, 2.3836e-05, 1.3351e-08, 2.6618e-05, 4.7488e-05,\n 8.6804e-11, 2.3244e-05, 9.6228e-06, 2.8771e-05, 2.5232e-05, 7.0894e-10,\n 4.1529e-05, 1.8182e-05, 4.6358e-05, 3.9575e-05, 2.1524e-05, 1.5674e-05,\n 1.9491e-05, 2.5433e-05, 4.9277e-09, 3.4802e-05, 1.5646e-09, 9.3115e-06,\n 8.3634e-06, 8.9900e-06, 3.1058e-05, 3.0719e-05, 5.5941e-05, 1.8189e-05,\n 3.8369e-05, 2.0126e-05, 3.4253e-05, 2.4113e-05, 6.5646e-10, 2.6814e-05,\n 3.1083e-05, 6.0704e-05, 7.6522e-06, 1.9478e-05, 2.1187e-05, 1.9135e-05,\n 3.2790e-05, 1.6771e-05, 2.5007e-05, 4.8710e-09, 1.9982e-05, 3.6210e-05,\n 2.5072e-05, 5.1209e-05, 3.4966e-05, 2.4837e-05, 3.1782e-05, 3.2060e-05,\n 2.9836e-05, 6.3070e-05, 2.0268e-05, 3.3328e-05, 3.6561e-09, 4.0351e-09,\n 3.5810e-05, 6.6472e-05, 3.6711e-05, 2.4908e-05, 3.2167e-05, 2.9984e-05,\n 2.7308e-05, 3.8995e-05, 2.4970e-05, 6.5235e-06, 2.2411e-05, 3.0601e-05,\n 2.6079e-05, 2.5492e-05, 1.6292e-06, 2.4090e-05, 3.3386e-05, 2.0111e-05,\n 3.6733e-09, 2.4086e-05, 2.9803e-05, 2.9756e-10, 8.9527e-10, 3.0056e-05,\n 3.2283e-05, 3.5436e-05, 1.8326e-05, 3.2708e-05, 1.6139e-05, 2.7822e-05,\n 2.6405e-05, 1.6421e-05, 1.7274e-05, 3.2764e-05, 2.8293e-05, 2.3556e-05,\n 3.7383e-05, 4.1208e-05, 3.0816e-05, 2.1770e-05, 2.5777e-05, 8.4081e-05,\n 2.3228e-05, 2.6591e-05, 1.6371e-05, 2.6842e-05, 3.0430e-05, 5.2466e-05,\n 8.8297e-06, 2.0897e-05, 4.0834e-05, 2.0501e-05, 3.3538e-05, 2.2415e-05,\n 3.4438e-05, 2.1073e-05, 2.8814e-05, 3.8930e-05, 6.7094e-06, 1.0098e-08,\n 5.3417e-09, 2.3250e-05, 2.3633e-05, 3.0488e-05, 2.7050e-05, 1.0167e-08,\n 2.5310e-05, 3.1970e-05, 1.8291e-05, 3.1846e-05, 2.1763e-05, 2.0995e-05,\n 4.0601e-05, 3.3635e-05, 3.4720e-05, 5.6454e-05, 3.3595e-05, 4.1093e-05,\n 2.0942e-05, 3.3692e-05, 3.1295e-05, 2.8817e-05, 4.1319e-05, 4.8132e-05,\n 3.8749e-05, 1.9050e-05, 2.4253e-05, 2.6643e-05, 3.2381e-05, 2.0979e-05,\n 2.5738e-05, 2.9189e-05, 5.9937e-05, 4.3228e-05, 3.2698e-05, 3.2954e-05,\n 3.1785e-05, 1.3805e-05, 1.7517e-05, 2.8479e-05, 2.8112e-05, 2.8400e-05,\n 1.7905e-05, 7.2059e-09, 9.2455e-09, 7.3346e-06, 3.1698e-05, 2.8865e-05,\n 2.8884e-05, 2.9498e-05, 2.7321e-05, 2.0765e-05, 2.3601e-05, 4.0243e-05,\n 2.2241e-05, 1.7422e-09, 8.7847e-06, 3.5144e-05, 2.7513e-05, 4.7851e-05,\n 3.0824e-05, 2.6605e-05, 2.5510e-08, 2.0918e-05], device='cuda:0')" + "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')" }, "2": { - "step": "tensor(8764.)", - "exp_avg": "tensor([[-1.7785e-06, 4.6818e-06, -2.3424e-05, ..., -5.2214e-06,\n 4.3520e-06, 1.1826e-05],\n [-1.9751e-06, 1.8068e-05, -3.8495e-05, ..., -4.5789e-07,\n -5.3884e-07, 1.3391e-05],\n [ 4.2994e-06, -1.6573e-05, 8.2906e-06, ..., 1.1568e-06,\n 1.5130e-07, -8.5059e-06],\n ...,\n [-1.4670e-06, 3.3114e-05, 6.5894e-06, ..., -9.3644e-06,\n 1.5096e-06, 6.1332e-05],\n [-7.2405e-06, -1.2774e-05, 8.3824e-06, ..., 5.0912e-06,\n 3.4127e-06, -1.7737e-05],\n [ 1.0344e-05, -1.0473e-05, 4.4951e-05, ..., 2.1670e-05,\n -2.0671e-06, -1.4626e-05]], device='cuda:0')", - "exp_avg_sq": "tensor([[8.0217e-10, 3.0466e-09, 4.6082e-09, ..., 4.0297e-09, 1.3367e-11,\n 3.6041e-09],\n [1.9454e-09, 1.0702e-08, 7.7039e-09, ..., 8.7261e-09, 1.8002e-11,\n 4.8208e-09],\n [9.8640e-10, 6.4236e-09, 8.3304e-09, ..., 6.3375e-09, 2.5834e-11,\n 9.2396e-09],\n ...,\n [1.8448e-09, 6.6157e-09, 9.9721e-09, ..., 7.5061e-09, 1.8015e-11,\n 1.1495e-08],\n [1.6794e-09, 6.7727e-09, 8.4228e-09, ..., 7.6173e-09, 1.8305e-11,\n 2.5474e-08],\n [1.7717e-09, 7.4916e-09, 1.0791e-08, ..., 1.1585e-08, 1.4210e-11,\n 5.2197e-09]], device='cuda:0')" + "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')" }, "3": { - "step": "tensor(8764.)", - "exp_avg": "tensor([[ 5.6052e-45, -5.6052e-45, 5.6052e-45, ..., -5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 8.9260e-06, 1.3182e-05, 3.3777e-06, ..., -1.4793e-06,\n 4.1738e-06, -4.4678e-06],\n [ 5.6052e-45, -5.6052e-45, -5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n ...,\n [ 7.7012e-06, 3.3585e-07, 3.7704e-06, ..., -7.8435e-08,\n 3.3536e-06, 4.9736e-06],\n [-8.6478e-06, 1.3429e-05, 1.4103e-05, ..., -3.7324e-06,\n -2.7325e-06, -3.9780e-06],\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([[3.1836e-13, 4.0667e-13, 3.2931e-15, ..., 4.2779e-19, 1.5370e-14,\n 1.3125e-14],\n [5.7954e-10, 2.3036e-09, 1.5975e-10, ..., 1.7105e-10, 2.1756e-10,\n 1.6122e-10],\n [2.6404e-14, 3.9817e-14, 1.4698e-14, ..., 2.5463e-15, 3.6122e-15,\n 1.0907e-14],\n ...,\n [3.0208e-09, 4.2271e-09, 6.1356e-10, ..., 8.0961e-10, 1.4419e-09,\n 1.3131e-09],\n [3.8999e-09, 2.7353e-09, 7.2976e-10, ..., 1.1388e-09, 6.9701e-10,\n 7.0186e-10],\n [2.0522e-12, 1.7285e-12, 1.5607e-13, ..., 1.9302e-13, 3.2515e-13,\n 1.0847e-13]], device='cuda:0')" + "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')" }, "4": { - "step": "tensor(8764.)", - "exp_avg": "tensor([ 5.6052e-45, 1.4233e-05, 5.6052e-45, ..., -4.4756e-05,\n -5.2153e-05, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([7.0937e-11, 3.3981e-07, 1.3583e-11, ..., 1.1560e-06, 1.0502e-06,\n 3.8164e-10], device='cuda:0')" + "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')" }, "5": { - "step": "tensor(8764.)", - "exp_avg": "tensor([[ 5.6052e-45, -8.7324e-08, -5.6052e-45, ..., 2.6997e-06,\n -2.6697e-06, 5.6052e-45],\n [ 5.6052e-45, -1.0024e-06, -5.6052e-45, ..., -7.3641e-07,\n -1.2129e-06, 5.6052e-45],\n [ 5.6052e-45, -3.0734e-06, 5.6052e-45, ..., -2.6699e-05,\n 1.1811e-06, -5.6052e-45],\n ...,\n [-5.6052e-45, 1.0647e-06, 5.6052e-45, ..., 1.5419e-05,\n -1.0154e-06, 5.6052e-45],\n [ 5.6052e-45, -6.7561e-07, 5.6052e-45, ..., 6.2919e-07,\n 1.7473e-06, 5.6052e-45],\n [-5.6052e-45, -5.8238e-07, 5.6052e-45, ..., -1.2896e-05,\n 3.5084e-07, -5.6052e-45]], device='cuda:0')", - "exp_avg_sq": "tensor([[1.3831e-15, 1.2386e-11, 2.3150e-15, ..., 9.5324e-11, 6.7848e-11,\n 3.1917e-13],\n [6.8554e-15, 3.7551e-11, 6.1589e-15, ..., 6.5938e-11, 5.9125e-11,\n 4.7147e-13],\n [4.5785e-15, 2.5382e-11, 3.7970e-14, ..., 2.4886e-10, 7.5237e-11,\n 8.5626e-13],\n ...,\n [1.1117e-15, 1.6705e-11, 9.1478e-14, ..., 2.0298e-10, 7.8572e-11,\n 1.4390e-12],\n [2.2958e-14, 1.2923e-10, 8.6749e-15, ..., 2.2712e-10, 1.3933e-10,\n 5.5331e-13],\n [7.1020e-16, 2.4000e-11, 8.7909e-15, ..., 2.3453e-10, 9.5708e-11,\n 7.2644e-13]], device='cuda:0')" + "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')" }, - "15": { - "step": "tensor(21284.)", + "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')" + }, + "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')" + }, + "14": { + "step": "tensor(23788.)", "exp_avg": "tensor([5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([9.0768e-14], device='cuda:0')" + "exp_avg_sq": "tensor([1.2166e-15], device='cuda:0')" }, - "16": { - "step": "tensor(21284.)", + "15": { + "step": "tensor(23788.)", "exp_avg": "tensor([ 5.6052e-45, -5.6052e-45, 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([1.0103e-16, 5.7945e-16, 2.1280e-16], device='cuda:0')" + "exp_avg_sq": "tensor([1.5695e-18, 8.8848e-17, 6.6800e-17], device='cuda:0')" }, - "17": { - "step": "tensor(21284.)", - "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45],\n device='cuda:0')", - "exp_avg_sq": "tensor([7.2265e-13, 5.1028e-14, 2.9948e-14, 5.7300e-14, 4.5082e-14],\n 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')" + }, + "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')" }, "19": { - "step": "tensor(21284.)", - "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([[5.6387e-22, 2.7126e-21, 4.8342e-22, ..., 1.0935e-22, 1.2948e-21,\n 6.9808e-22],\n [3.0414e-19, 3.0555e-19, 2.3384e-22, ..., 5.6246e-20, 1.1248e-20,\n 2.3967e-20],\n [3.1717e-18, 3.6314e-18, 3.6093e-22, ..., 2.9281e-19, 3.9628e-19,\n 1.2028e-19],\n ...,\n [9.5287e-20, 4.2956e-20, 8.2281e-21, ..., 2.7393e-21, 2.5433e-20,\n 3.1289e-21],\n [1.2054e-20, 9.8385e-21, 2.5097e-22, ..., 2.2470e-22, 1.2479e-21,\n 1.4451e-22],\n [1.0893e-17, 1.2757e-17, 1.9448e-21, ..., 9.8372e-19, 1.3981e-18,\n 4.7748e-19]], device='cuda:0')" + "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')" }, "20": { - "step": "tensor(21284.)", - "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([4.3544e-19, 2.8334e-16, 1.9711e-15, 4.9700e-17, 3.3878e-16, 1.3163e-15,\n 1.4892e-17, 2.0235e-15, 1.5809e-16, 3.6737e-18, 9.1274e-15, 2.9085e-17,\n 9.2101e-18, 2.0906e-16, 2.1609e-16, 1.7134e-15, 5.7824e-16, 5.7561e-17,\n 2.8658e-15, 8.6073e-16, 8.6190e-16, 8.2873e-17, 2.9527e-16, 6.0145e-16,\n 3.5870e-16, 5.8288e-17, 2.6620e-17, 3.9776e-15, 2.6729e-18, 8.4097e-17,\n 1.6205e-16, 7.9522e-16, 1.3175e-15, 4.9840e-17, 6.9895e-17, 1.8882e-16,\n 3.7035e-18, 2.3901e-16, 1.4395e-15, 7.5845e-18, 3.0601e-17, 9.7335e-17,\n 2.6104e-15, 4.4645e-16, 3.0861e-17, 2.4176e-18, 1.0714e-17, 8.5658e-17,\n 2.1035e-17, 4.6719e-19, 1.6298e-18, 1.8917e-17, 1.0113e-16, 5.4554e-16,\n 7.1462e-18, 8.0963e-16, 1.4252e-15, 6.4063e-15, 7.2931e-16, 1.3696e-18,\n 1.0451e-14, 3.5389e-19, 3.5033e-17, 5.4118e-16, 1.5864e-15, 2.3285e-15,\n 3.7915e-17, 4.0037e-17, 8.4123e-15, 7.8646e-17, 1.2150e-15, 3.3390e-16,\n 6.2401e-16, 1.0864e-16, 1.6417e-18, 1.5641e-16, 1.1320e-15, 1.8426e-17,\n 1.5853e-20, 6.3710e-19, 1.6702e-18, 2.4876e-17, 2.6254e-15, 1.1057e-15,\n 1.7108e-16, 4.1268e-17, 8.9465e-16, 1.5336e-16, 2.3702e-16, 5.9206e-15,\n 3.4007e-17, 2.2161e-16, 4.0846e-17, 1.6944e-15, 2.6600e-16, 4.8108e-17,\n 2.2370e-19, 6.6491e-17, 2.1480e-15, 4.2667e-16, 1.1624e-18, 1.5774e-15,\n 1.2712e-17, 7.9218e-19, 6.7141e-16, 1.1860e-16, 2.1788e-18, 2.8652e-17,\n 8.9103e-17, 1.1674e-16, 1.1167e-18, 2.3967e-17, 9.4524e-18, 2.6946e-18,\n 1.3637e-16, 5.7686e-17, 1.6869e-15, 4.6674e-17, 1.0202e-15, 4.4671e-16,\n 2.1495e-17, 3.8345e-16, 5.7973e-16, 4.4960e-17, 7.6153e-16, 8.6785e-17,\n 3.3746e-15, 5.7521e-17, 1.3318e-16, 9.1965e-18, 3.9304e-18, 6.1626e-17,\n 2.1671e-17, 2.8619e-16, 5.6662e-17, 9.4181e-16, 3.4077e-18, 2.1848e-17,\n 5.9510e-16, 8.4180e-17, 1.7866e-18, 4.8850e-18, 8.7935e-15, 2.2408e-14,\n 5.4527e-16, 2.3876e-18, 7.1699e-16, 6.0982e-16, 4.7324e-16, 1.2596e-17,\n 2.3711e-18, 4.2571e-16, 1.5536e-15, 4.5304e-16, 7.6598e-16, 2.2972e-16,\n 4.4094e-18, 2.2066e-17, 3.4216e-17, 2.3467e-15, 1.1998e-17, 3.5165e-16,\n 4.4100e-16, 1.1199e-17, 1.1361e-17, 1.6022e-14, 1.4732e-15, 3.6125e-18,\n 4.5542e-15, 3.3109e-19, 4.2687e-16, 2.3006e-17, 5.6595e-16, 1.2216e-16,\n 2.2701e-16, 3.1426e-15, 5.3041e-16, 2.4888e-16, 5.4172e-17, 1.4164e-15,\n 2.1802e-16, 5.2326e-16, 1.3233e-14, 1.6898e-16, 8.5138e-18, 1.3286e-19,\n 2.2020e-17, 2.0341e-15, 1.0518e-16, 4.4458e-16, 2.2308e-15, 8.5026e-16,\n 1.7818e-17, 1.0126e-17, 1.6365e-15, 2.8465e-16, 2.9404e-18, 6.3723e-16,\n 4.6990e-18, 5.4493e-16, 1.2043e-15, 6.7564e-17, 2.9759e-16, 8.4095e-18,\n 1.3989e-16, 4.3111e-15, 6.6857e-17, 4.8898e-16, 1.5107e-18, 3.7022e-15,\n 7.0173e-16, 4.3157e-18, 3.3783e-17, 2.7786e-18, 2.6025e-15, 2.6156e-17,\n 1.8546e-16, 1.1010e-15, 5.7142e-17, 4.8736e-16, 3.9202e-16, 3.0335e-16,\n 1.7582e-17, 9.9209e-16, 7.5023e-17, 3.4757e-16, 4.0458e-15, 3.6335e-17,\n 1.5659e-18, 1.2732e-17, 1.5362e-16, 2.2371e-16, 8.7588e-18, 2.9093e-16,\n 1.2151e-16, 1.3960e-15, 1.6850e-16, 9.7226e-16, 1.0296e-16, 1.7134e-17,\n 6.1770e-17, 1.0381e-17, 2.8109e-15, 1.8611e-16, 2.5974e-17, 5.5342e-16,\n 2.0146e-17, 5.8134e-17, 1.8285e-15, 4.9938e-17, 6.3179e-16, 3.6058e-18,\n 4.0491e-16, 2.7556e-17, 3.4862e-18, 6.6983e-15], device='cuda:0')" + "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')" }, "21": { - "step": "tensor(21284.)", - "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([3.2741e-20, 3.2447e-20, 4.5126e-18, 2.8166e-20, 1.3859e-19, 1.5858e-19,\n 7.5171e-20, 5.2597e-19, 3.3831e-21, 1.8147e-20, 5.6347e-18, 1.1151e-20,\n 1.6431e-21, 7.7311e-20, 4.2337e-19, 1.4515e-18, 6.2440e-19, 4.4760e-21,\n 1.0641e-17, 5.3381e-18, 4.8102e-18, 6.3241e-21, 1.0186e-20, 3.6855e-19,\n 3.0430e-18, 4.3454e-20, 6.9355e-20, 3.8368e-18, 1.2599e-21, 9.6066e-20,\n 3.2634e-20, 6.2534e-19, 4.5803e-18, 1.2276e-20, 5.7762e-20, 1.4191e-18,\n 2.0339e-20, 1.1479e-19, 2.8169e-18, 1.7258e-22, 1.3748e-20, 5.9635e-21,\n 2.9006e-18, 1.6021e-19, 2.4720e-20, 2.0211e-21, 1.4478e-20, 1.1865e-20,\n 6.9213e-21, 2.1212e-20, 2.1453e-21, 2.3539e-20, 1.6031e-19, 1.2428e-19,\n 1.0665e-22, 1.4491e-19, 8.0756e-20, 4.8833e-18, 3.5847e-19, 3.9474e-21,\n 2.2794e-17, 1.4543e-22, 1.0178e-21, 6.6870e-20, 3.0998e-18, 1.8910e-18,\n 2.5365e-20, 4.0916e-20, 2.9689e-17, 1.1109e-20, 2.2127e-18, 6.9975e-21,\n 8.6332e-20, 6.1432e-19, 6.7485e-21, 5.1560e-20, 1.1530e-18, 9.0991e-20,\n 1.9462e-20, 1.7282e-22, 8.9958e-21, 5.8452e-20, 6.5186e-19, 3.2433e-18,\n 1.6071e-19, 2.2413e-19, 1.2086e-19, 9.3381e-21, 5.8380e-20, 1.6143e-17,\n 7.8747e-20, 6.9615e-20, 2.0042e-20, 2.0486e-18, 3.7126e-20, 1.3441e-19,\n 4.0833e-22, 1.8521e-19, 3.7351e-18, 7.5525e-19, 2.6875e-21, 4.1949e-18,\n 1.9737e-21, 5.7887e-21, 1.8943e-19, 2.0729e-19, 8.1767e-21, 2.8608e-20,\n 1.9880e-20, 1.6896e-20, 2.7746e-22, 3.0876e-20, 5.1574e-20, 7.3191e-21,\n 3.3369e-20, 3.7819e-21, 4.4597e-18, 1.0973e-19, 1.4041e-18, 6.3228e-20,\n 5.6426e-20, 3.7796e-19, 4.7715e-19, 3.1658e-20, 1.5548e-18, 6.9548e-22,\n 3.5001e-18, 8.5198e-22, 4.0166e-20, 3.2222e-21, 3.0166e-21, 3.3737e-21,\n 4.6107e-21, 1.1374e-19, 1.1842e-20, 2.4301e-18, 5.5518e-21, 2.8802e-21,\n 3.6132e-19, 1.2685e-20, 5.3678e-21, 2.2257e-20, 3.8863e-17, 4.4112e-17,\n 8.4549e-20, 1.0067e-22, 7.7443e-19, 6.6447e-19, 1.1169e-18, 5.7239e-22,\n 4.4071e-21, 8.5572e-19, 1.6398e-18, 4.6582e-20, 1.4399e-19, 4.2366e-19,\n 3.3307e-23, 7.1848e-20, 1.5964e-20, 1.6168e-19, 2.1603e-21, 5.7635e-19,\n 2.1922e-20, 6.6320e-20, 2.1039e-20, 3.7479e-17, 8.6378e-19, 2.2545e-21,\n 5.9693e-18, 4.2831e-23, 4.6958e-20, 3.5738e-20, 3.8162e-18, 1.6283e-19,\n 1.0015e-19, 1.8423e-18, 1.7045e-19, 1.2170e-20, 7.8309e-20, 1.6047e-18,\n 4.1828e-21, 1.5186e-19, 3.4648e-17, 5.7423e-20, 3.1134e-21, 2.6248e-21,\n 9.9821e-22, 1.6316e-18, 2.3880e-19, 4.4327e-19, 1.3917e-18, 1.1250e-19,\n 6.9335e-21, 8.5708e-22, 9.5975e-19, 5.6873e-21, 4.6643e-22, 3.7512e-20,\n 6.8186e-21, 1.7652e-19, 3.2044e-19, 5.8641e-21, 3.1126e-20, 7.3070e-21,\n 9.4022e-21, 1.0210e-17, 7.0557e-20, 4.2763e-20, 2.3763e-20, 5.9764e-18,\n 2.2177e-19, 3.0386e-21, 2.5450e-21, 3.9475e-20, 2.5449e-19, 7.5243e-22,\n 6.5449e-21, 1.8395e-18, 2.6743e-19, 7.7865e-20, 7.2000e-19, 1.4296e-18,\n 2.2688e-21, 8.0127e-19, 4.0424e-20, 1.1619e-19, 8.1571e-19, 1.4426e-20,\n 4.6342e-22, 1.2460e-21, 1.8373e-20, 3.9634e-19, 1.3001e-21, 5.7342e-19,\n 1.3949e-20, 1.6322e-18, 1.7026e-18, 6.4399e-19, 1.0307e-19, 3.6722e-22,\n 9.6402e-21, 1.9063e-21, 1.0396e-18, 4.1979e-20, 9.8069e-21, 2.6825e-19,\n 2.3418e-21, 1.9070e-21, 9.6470e-18, 4.5387e-20, 6.6989e-20, 5.8994e-21,\n 2.9137e-19, 4.7326e-20, 6.2679e-21, 1.8736e-17], device='cuda:0')" + "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')" }, "22": { - "step": "tensor(21284.)", - "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([1.4811e-21, 3.9738e-19, 3.1492e-18, 4.4202e-20, 4.8508e-19, 1.8215e-18,\n 2.1586e-20, 2.7031e-18, 2.3359e-19, 3.1303e-20, 1.1038e-17, 1.5448e-19,\n 3.9698e-20, 3.6341e-19, 4.4664e-19, 1.7651e-18, 9.0956e-19, 1.0314e-19,\n 3.0398e-18, 1.6266e-18, 1.4318e-18, 1.3908e-19, 3.0602e-19, 1.1857e-18,\n 1.1410e-18, 8.2567e-20, 2.3630e-20, 4.5091e-18, 5.0336e-22, 1.6132e-19,\n 6.9927e-20, 1.0971e-18, 2.1080e-18, 1.5244e-19, 1.2284e-19, 8.2168e-19,\n 1.0677e-20, 1.5767e-19, 2.2498e-18, 1.4416e-21, 1.2078e-19, 1.5227e-19,\n 3.9626e-18, 6.6966e-19, 1.7098e-19, 1.5919e-20, 4.1381e-21, 1.7297e-20,\n 4.0763e-20, 1.7065e-21, 9.8388e-22, 3.2811e-20, 2.0475e-19, 4.9563e-19,\n 3.0729e-22, 8.7888e-19, 1.9327e-18, 8.6544e-18, 7.4983e-19, 1.8080e-20,\n 1.4012e-17, 4.9724e-21, 1.9290e-20, 6.9140e-19, 2.3189e-18, 3.2717e-18,\n 7.6970e-20, 1.1724e-19, 1.1459e-17, 1.3863e-19, 1.7407e-18, 3.4567e-19,\n 7.9422e-19, 5.6317e-19, 3.9282e-21, 1.0905e-19, 1.6584e-18, 6.5394e-20,\n 4.2444e-22, 4.1799e-21, 6.0578e-20, 2.0662e-20, 3.5711e-18, 2.0100e-18,\n 3.0420e-19, 2.9356e-19, 9.7270e-19, 2.3737e-19, 3.8562e-19, 7.0191e-18,\n 2.2787e-19, 3.5542e-19, 4.5898e-21, 2.3939e-18, 3.8977e-19, 5.0612e-20,\n 1.8363e-20, 6.6643e-20, 3.1102e-18, 8.2966e-19, 1.5383e-20, 2.3243e-18,\n 2.7086e-20, 9.8527e-21, 7.1226e-19, 4.8643e-19, 5.0437e-21, 1.8759e-19,\n 1.4017e-19, 1.9055e-19, 6.4290e-21, 8.4895e-21, 9.3045e-20, 2.6239e-21,\n 1.8860e-19, 1.0007e-19, 2.5678e-18, 3.2841e-20, 1.4778e-18, 6.4816e-19,\n 1.7965e-20, 4.9549e-19, 8.9549e-19, 1.0267e-19, 1.4066e-18, 6.0546e-21,\n 3.8103e-18, 1.0184e-20, 2.2013e-19, 1.4751e-21, 8.7827e-21, 8.3180e-21,\n 7.1471e-21, 3.6177e-19, 8.5624e-20, 1.5804e-18, 2.7396e-20, 1.2814e-20,\n 8.8523e-19, 1.2492e-19, 4.7318e-21, 5.6240e-22, 1.0946e-17, 2.8233e-17,\n 5.2078e-19, 3.7646e-21, 1.1781e-18, 9.4239e-19, 7.7896e-19, 2.5407e-20,\n 2.3476e-20, 1.0002e-18, 1.6249e-18, 6.6089e-19, 1.0879e-18, 3.5703e-19,\n 1.4719e-23, 1.6939e-20, 6.0217e-20, 2.8399e-18, 8.1902e-22, 5.6925e-19,\n 6.3095e-19, 1.8934e-20, 9.6258e-21, 2.1204e-17, 2.1081e-18, 8.9694e-21,\n 5.4332e-18, 1.2855e-21, 4.3563e-19, 1.2669e-20, 1.2713e-18, 2.4515e-19,\n 3.8476e-19, 3.7081e-18, 4.4099e-19, 2.5324e-19, 2.0457e-20, 2.0405e-18,\n 2.4628e-19, 5.4998e-19, 1.7673e-17, 1.3316e-19, 2.3352e-20, 7.8661e-22,\n 4.2993e-20, 2.1935e-18, 4.5232e-19, 7.1016e-19, 3.0823e-18, 9.6331e-19,\n 3.1074e-20, 2.0538e-20, 1.7248e-18, 4.0015e-19, 4.0112e-22, 8.9908e-19,\n 3.0068e-20, 4.7276e-19, 1.2564e-18, 1.1833e-19, 2.8888e-19, 7.7706e-20,\n 1.3700e-19, 6.0061e-18, 5.8120e-20, 7.0558e-19, 2.7482e-21, 5.0570e-18,\n 7.3239e-19, 1.4661e-23, 2.7666e-21, 7.2459e-21, 3.5084e-18, 1.1374e-20,\n 2.7900e-19, 1.6213e-18, 8.4803e-20, 7.2064e-19, 7.9114e-19, 6.5390e-19,\n 3.7963e-20, 1.1464e-18, 2.2254e-19, 4.0106e-19, 5.4050e-18, 1.2107e-19,\n 4.6512e-21, 5.4989e-21, 1.9998e-19, 3.9571e-19, 1.4637e-21, 6.9035e-19,\n 4.6637e-20, 2.1506e-18, 7.8051e-19, 1.4089e-18, 4.0320e-19, 1.1890e-22,\n 9.3613e-20, 2.3829e-20, 3.2683e-18, 3.5204e-19, 1.0498e-19, 8.3277e-19,\n 3.4597e-20, 9.8123e-20, 3.0812e-18, 1.0304e-19, 9.0145e-19, 2.0791e-21,\n 3.6121e-19, 3.7494e-20, 2.0994e-21, 9.4672e-18], device='cuda:0')" + "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')" }, "23": { - "step": "tensor(21284.)", - "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([[6.9732e-18, 7.3947e-18, 5.3214e-21, ..., 5.2132e-19, 9.3647e-19,\n 1.7348e-19],\n [5.4768e-18, 6.1990e-18, 3.8805e-23, ..., 3.4548e-19, 7.2323e-19,\n 1.5048e-19],\n [1.6351e-18, 1.9222e-18, 4.0434e-21, ..., 1.4721e-19, 1.8121e-19,\n 5.7660e-20],\n ...,\n [1.6112e-19, 1.0114e-19, 7.8634e-22, ..., 9.9859e-21, 2.7828e-20,\n 4.2132e-21],\n [5.5175e-19, 5.8692e-19, 7.5906e-22, ..., 5.6449e-20, 6.8712e-20,\n 2.7273e-20],\n [4.9631e-19, 7.2169e-19, 6.3544e-22, ..., 6.8182e-20, 6.4534e-20,\n 3.8013e-20]], device='cuda:0')" + "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')" }, "24": { - "step": "tensor(21284.)", - "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([3.7140e-15, 2.9102e-15, 9.5445e-16, 6.1029e-16, 5.8037e-17, 3.8480e-15,\n 3.2501e-17, 3.4466e-15, 3.3770e-15, 3.0791e-18, 4.1963e-15, 2.8637e-16,\n 3.8195e-16, 2.6844e-17, 4.1677e-18, 2.4430e-17, 5.9172e-20, 4.2960e-17,\n 8.4249e-16, 9.9926e-16, 2.2613e-16, 4.1630e-16, 5.0013e-16, 5.1993e-18,\n 2.9961e-16, 3.1716e-16, 1.2143e-16, 7.8636e-16, 1.0198e-16, 4.2978e-18,\n 5.1298e-17, 2.8983e-16, 4.6513e-16, 5.9297e-16, 3.5883e-19, 2.3250e-19,\n 3.1193e-16, 4.3564e-18, 3.2493e-16, 2.5192e-16, 6.7441e-19, 1.5838e-15,\n 4.0322e-15, 1.0530e-18, 1.0782e-16, 3.6457e-18, 5.1789e-16, 3.0593e-18,\n 1.1341e-16, 1.0480e-15, 5.6460e-18, 8.3376e-17, 1.7097e-18, 4.9410e-16,\n 2.8041e-17, 1.1606e-17, 7.6014e-15, 6.6333e-15, 2.3008e-17, 8.3920e-19,\n 9.9158e-15, 6.4325e-17, 7.6767e-16, 2.6814e-15, 2.1331e-16, 4.6135e-15,\n 1.0800e-17, 4.1473e-17, 4.8827e-16, 8.0435e-17, 9.7119e-16, 1.8175e-16,\n 1.6300e-16, 6.2261e-17, 3.6742e-17, 1.0380e-16, 5.0634e-16, 2.3399e-19,\n 3.2058e-16, 9.5712e-17, 4.7634e-16, 1.0971e-15, 4.1126e-17, 1.8668e-18,\n 5.5418e-18, 2.8617e-16, 1.1522e-15, 6.5771e-18, 1.7585e-15, 1.0772e-16,\n 3.9336e-17, 2.8926e-18, 5.2282e-17, 9.8665e-16, 2.1761e-15, 1.1548e-15,\n 7.4999e-18, 5.6351e-18, 9.4904e-16, 7.4075e-17, 4.8677e-16, 5.5569e-18,\n 1.2351e-16, 1.1981e-18, 6.6973e-17, 2.2983e-18, 4.4947e-16, 1.4451e-19,\n 5.2855e-16, 1.3629e-16, 2.5272e-17, 2.4519e-17, 2.3487e-16, 2.5340e-18,\n 5.8398e-16, 5.4751e-16, 4.1315e-17, 5.4562e-15, 1.8705e-19, 2.0470e-15,\n 1.2485e-16, 1.8594e-17, 3.3125e-16, 2.9443e-18, 6.0292e-18, 5.6974e-18,\n 3.5305e-15, 2.1248e-16, 1.0269e-18, 1.6119e-17, 7.1064e-17, 1.1316e-16,\n 2.6990e-16, 4.1553e-16, 1.3435e-17, 3.3465e-16, 7.4703e-16, 6.7393e-17,\n 1.6205e-18, 3.5501e-17, 3.1085e-16, 7.4299e-16, 9.4571e-16, 1.8671e-14,\n 1.3111e-17, 3.4425e-19, 5.4280e-16, 7.1368e-17, 3.7810e-18, 4.9507e-17,\n 1.2140e-16, 3.2167e-17, 5.5225e-17, 3.5054e-16, 2.8325e-17, 2.7763e-17,\n 1.3289e-16, 3.0262e-15, 2.5987e-17, 9.9450e-15, 1.5392e-16, 1.5857e-17,\n 6.8880e-19, 2.2738e-15, 5.7009e-16, 1.1352e-15, 1.1425e-17, 2.1618e-16,\n 3.9675e-15, 9.0207e-18, 2.0801e-15, 5.2361e-16, 3.0799e-16, 5.4619e-17,\n 4.4784e-16, 3.1473e-15, 1.1292e-15, 5.1662e-16, 9.2950e-17, 1.3998e-15,\n 3.4885e-15, 3.3879e-16, 3.9390e-15, 3.8845e-19, 4.4494e-16, 2.3993e-19,\n 5.9872e-16, 3.0072e-17, 9.5162e-17, 1.0513e-17, 2.3766e-15, 1.5407e-18,\n 3.7953e-16, 8.6036e-18, 1.4243e-16, 1.7780e-16, 1.1510e-16, 3.5199e-15,\n 1.9229e-16, 8.7055e-18, 3.3281e-16, 1.5095e-17, 1.9182e-17, 7.2440e-17,\n 1.3597e-16, 6.5357e-16, 6.3205e-17, 5.9887e-17, 4.6926e-17, 8.7731e-18,\n 2.1959e-17, 6.8085e-16, 1.2156e-18, 2.4573e-15, 2.1180e-16, 1.1400e-16,\n 1.0714e-15, 2.3628e-16, 2.3032e-15, 6.5573e-16, 2.3471e-18, 9.1752e-16,\n 2.4017e-17, 1.1602e-18, 1.0285e-16, 2.9710e-17, 4.8922e-16, 1.9110e-15,\n 7.7117e-19, 1.6843e-17, 2.7162e-16, 1.7680e-16, 2.9555e-18, 2.2882e-17,\n 5.5293e-17, 1.1066e-15, 1.2716e-16, 8.2681e-17, 2.9849e-16, 7.7047e-17,\n 1.6324e-15, 1.4894e-16, 1.7750e-15, 6.2024e-16, 6.3018e-17, 2.8599e-17,\n 6.2218e-17, 6.2468e-16, 1.0193e-15, 1.3208e-17, 1.0749e-15, 1.0875e-17,\n 8.4775e-16, 6.5040e-17, 3.6352e-16, 4.5596e-16], device='cuda:0')" + "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')" }, "25": { - "step": "tensor(21284.)", - "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([3.4891e-18, 2.9829e-18, 5.9382e-18, 1.0202e-20, 8.3261e-20, 9.5702e-18,\n 9.6129e-20, 3.6531e-18, 1.5787e-18, 5.3270e-21, 1.2488e-18, 4.1064e-18,\n 4.6870e-19, 1.7912e-21, 4.1891e-21, 3.8833e-20, 1.1888e-20, 3.0288e-21,\n 2.8956e-19, 3.1185e-18, 1.8490e-18, 4.8956e-20, 4.4145e-20, 6.0527e-20,\n 2.9410e-18, 1.0631e-19, 5.7048e-20, 1.0441e-19, 8.2384e-20, 6.6811e-21,\n 1.6461e-21, 3.1685e-19, 6.0019e-20, 8.6457e-19, 1.8873e-21, 2.8987e-21,\n 1.2234e-19, 2.7075e-23, 1.8667e-19, 1.5735e-19, 2.2199e-21, 7.4965e-18,\n 2.1285e-17, 8.4019e-21, 8.7660e-19, 5.8689e-22, 2.2283e-19, 1.1366e-22,\n 2.9872e-20, 2.5175e-19, 9.8579e-22, 7.8533e-21, 2.1642e-22, 5.1939e-19,\n 2.0656e-20, 7.0094e-21, 3.6614e-18, 8.2850e-18, 1.0113e-20, 2.0475e-21,\n 1.3698e-17, 1.4565e-20, 2.3185e-18, 1.3042e-18, 7.6691e-20, 6.8487e-18,\n 8.9842e-21, 5.4944e-20, 7.1798e-21, 2.3223e-20, 2.8439e-19, 1.9544e-20,\n 2.5726e-20, 5.9497e-20, 3.9195e-20, 1.2938e-20, 1.8282e-19, 5.7721e-21,\n 8.8586e-20, 1.6598e-20, 1.3016e-18, 1.7545e-19, 1.1502e-19, 3.0260e-20,\n 5.4810e-21, 3.6992e-18, 5.5151e-19, 3.4513e-21, 2.4057e-18, 1.3304e-20,\n 7.6171e-21, 3.1091e-21, 1.4353e-21, 4.9221e-19, 6.2915e-18, 2.9161e-19,\n 6.4970e-21, 4.6736e-21, 5.3711e-19, 4.7639e-20, 7.8722e-19, 6.9479e-21,\n 5.4509e-20, 5.8373e-21, 2.6792e-21, 5.4745e-22, 4.0957e-19, 1.5463e-21,\n 1.4365e-19, 1.5325e-20, 1.5257e-21, 3.1584e-19, 6.3336e-19, 2.3421e-21,\n 4.1584e-19, 1.6590e-18, 4.0946e-21, 1.2622e-17, 2.0382e-22, 3.2066e-18,\n 4.7989e-21, 1.9880e-21, 1.4642e-19, 4.8348e-23, 5.5947e-21, 2.6303e-22,\n 4.4478e-18, 6.1662e-20, 3.7893e-24, 1.9786e-21, 3.3356e-21, 9.3439e-21,\n 1.2407e-19, 6.2760e-19, 6.3149e-21, 7.7225e-20, 2.7547e-18, 1.2565e-20,\n 1.7501e-21, 1.0874e-19, 2.9221e-20, 1.5702e-19, 1.5287e-19, 4.5498e-17,\n 2.5951e-20, 4.6850e-23, 3.5545e-19, 9.3916e-21, 1.3064e-21, 3.5152e-20,\n 2.7018e-19, 4.0625e-21, 4.9371e-22, 1.3557e-19, 6.3543e-21, 8.3455e-20,\n 2.6745e-20, 2.7672e-18, 3.7667e-21, 9.4504e-18, 3.6701e-20, 5.8345e-22,\n 1.3869e-20, 6.7835e-18, 1.1714e-18, 1.6376e-20, 2.6178e-20, 3.0609e-20,\n 1.1316e-17, 2.0301e-22, 3.2441e-18, 5.3349e-20, 2.8668e-19, 1.6907e-20,\n 2.0847e-19, 3.1516e-18, 4.8597e-19, 1.2264e-19, 2.0944e-20, 1.0224e-18,\n 2.4451e-18, 1.3759e-20, 1.3536e-18, 7.4094e-22, 2.6945e-19, 4.4500e-21,\n 4.4515e-19, 3.9215e-20, 1.2570e-19, 1.8609e-20, 6.8320e-19, 1.2679e-20,\n 3.5816e-20, 3.6735e-21, 7.7926e-21, 1.8412e-20, 6.5573e-20, 1.0111e-17,\n 3.1395e-20, 8.4207e-21, 5.4597e-20, 8.5987e-22, 3.7132e-21, 1.8093e-19,\n 3.3352e-19, 1.3990e-19, 5.3332e-20, 2.6730e-21, 9.3954e-21, 2.0145e-20,\n 1.7733e-20, 8.2011e-19, 1.4941e-21, 8.2952e-18, 9.6022e-21, 1.0839e-20,\n 3.5002e-19, 3.1195e-20, 4.5454e-19, 5.8497e-20, 1.6966e-20, 1.9498e-18,\n 1.6530e-20, 3.0183e-21, 3.4006e-19, 1.2944e-21, 9.0936e-21, 3.4718e-18,\n 4.3406e-22, 1.3218e-21, 3.2845e-20, 1.4210e-19, 3.0447e-22, 6.4057e-21,\n 1.6721e-21, 3.6936e-18, 7.1268e-20, 2.5517e-21, 3.7203e-18, 6.1988e-21,\n 2.3826e-18, 4.3606e-20, 8.7992e-19, 8.7527e-19, 2.6619e-20, 1.1322e-21,\n 6.1915e-22, 9.6149e-19, 1.1356e-18, 1.0933e-19, 2.8824e-19, 5.3478e-21,\n 1.4513e-18, 1.8165e-19, 4.1990e-20, 1.8936e-19], device='cuda:0')" + "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')" }, "26": { - "step": "tensor(21284.)", - "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([4.7632e-18, 3.6315e-18, 2.0320e-18, 6.7250e-19, 1.6347e-19, 4.9632e-18,\n 3.1594e-20, 4.7039e-18, 4.2337e-18, 1.5534e-21, 5.2526e-18, 1.2216e-18,\n 7.0648e-19, 3.7798e-20, 1.6840e-21, 1.4490e-20, 5.3629e-22, 5.8899e-20,\n 1.0155e-18, 2.0622e-18, 8.0851e-19, 5.3030e-19, 5.6384e-19, 2.0933e-20,\n 1.1710e-18, 3.8659e-19, 3.0565e-19, 8.6810e-19, 1.2800e-19, 2.4015e-21,\n 1.5150e-20, 6.2003e-19, 5.9303e-19, 1.3571e-18, 2.0407e-22, 1.9682e-21,\n 4.1292e-19, 1.0381e-21, 4.2944e-19, 2.9987e-19, 1.1709e-20, 2.2572e-18,\n 6.1168e-18, 3.0113e-21, 6.0713e-19, 5.8605e-21, 4.4785e-19, 4.5617e-21,\n 1.4058e-19, 1.3189e-18, 9.4441e-21, 1.1239e-19, 2.5814e-21, 3.8771e-19,\n 3.8204e-20, 2.7081e-21, 9.5579e-18, 8.3838e-18, 3.3682e-21, 1.4848e-20,\n 1.2343e-17, 9.5408e-20, 9.5264e-19, 3.4641e-18, 2.6607e-19, 5.9127e-18,\n 3.5093e-21, 7.0327e-20, 6.1995e-19, 1.0872e-19, 1.2168e-18, 1.8210e-19,\n 2.6310e-19, 2.3593e-19, 4.8537e-20, 3.2312e-20, 6.4144e-19, 2.7770e-20,\n 5.0828e-19, 1.2614e-19, 6.9152e-19, 1.3743e-18, 3.9515e-20, 1.0694e-20,\n 3.8489e-20, 8.4509e-19, 1.4480e-18, 8.0260e-21, 2.3755e-18, 1.5908e-19,\n 5.7199e-20, 2.5377e-20, 2.5400e-20, 1.2173e-18, 2.7003e-18, 1.5703e-18,\n 7.0000e-20, 6.0257e-22, 1.4527e-18, 2.5275e-19, 7.6052e-19, 1.2628e-21,\n 2.0898e-19, 7.3652e-21, 3.6589e-20, 3.3311e-20, 5.8656e-19, 7.0673e-23,\n 7.8526e-19, 1.7615e-19, 3.5000e-20, 2.8156e-19, 4.4230e-19, 1.9723e-21,\n 8.0586e-19, 7.9358e-19, 5.9017e-20, 7.3001e-18, 9.8561e-22, 2.4756e-18,\n 9.5463e-20, 2.1468e-20, 4.1127e-19, 2.0910e-21, 5.4222e-20, 4.5867e-23,\n 4.2641e-18, 1.7742e-19, 2.6906e-23, 7.3583e-21, 9.6721e-20, 5.8142e-20,\n 2.1361e-19, 7.2973e-19, 9.3063e-21, 4.3374e-19, 1.3904e-18, 7.8858e-20,\n 1.7092e-21, 4.2432e-20, 4.0214e-19, 7.6884e-19, 1.4478e-18, 2.4245e-17,\n 4.9387e-21, 4.5059e-22, 7.6381e-19, 9.5284e-20, 1.5512e-22, 4.1899e-20,\n 4.9132e-19, 4.9909e-20, 3.7724e-20, 4.4278e-19, 3.7566e-20, 1.6274e-19,\n 3.3407e-20, 3.7543e-18, 3.0094e-20, 1.2514e-17, 6.5126e-20, 2.3495e-20,\n 2.2283e-21, 3.2663e-18, 6.6707e-19, 1.4448e-18, 1.6931e-20, 2.7619e-19,\n 5.0073e-18, 1.3078e-20, 2.5157e-18, 5.9089e-19, 7.3171e-19, 6.9822e-20,\n 6.2308e-19, 4.0301e-18, 1.2426e-18, 6.9503e-19, 7.5083e-20, 1.7392e-18,\n 4.3359e-18, 3.7169e-19, 4.9606e-18, 2.2090e-20, 5.5552e-19, 1.7032e-21,\n 7.8967e-19, 1.3654e-20, 3.0645e-19, 7.5090e-21, 2.9844e-18, 1.3406e-21,\n 4.8216e-19, 1.1199e-20, 1.0692e-19, 2.3221e-19, 9.3424e-20, 4.4148e-18,\n 2.5119e-19, 2.2990e-21, 2.8013e-19, 2.2228e-20, 1.1027e-20, 3.0731e-19,\n 1.1216e-19, 8.2475e-19, 9.4162e-20, 7.9379e-20, 1.0959e-19, 4.2645e-20,\n 6.5428e-21, 7.8291e-19, 1.7919e-21, 3.5810e-18, 2.7362e-19, 1.0420e-19,\n 1.3653e-18, 2.9795e-19, 2.8923e-18, 8.4252e-19, 6.3465e-21, 1.5863e-18,\n 1.4160e-20, 1.1549e-21, 4.4507e-19, 2.8187e-20, 6.2942e-19, 2.7597e-18,\n 2.3235e-22, 3.6624e-21, 4.0839e-19, 2.1869e-19, 5.1266e-22, 1.1033e-19,\n 1.2647e-20, 2.0661e-18, 2.4146e-19, 1.0882e-19, 1.1658e-18, 2.0735e-20,\n 2.0448e-18, 2.0482e-19, 2.0738e-18, 1.1379e-18, 1.8568e-19, 3.9360e-20,\n 8.4374e-20, 7.9189e-19, 1.3793e-18, 1.7836e-19, 1.3779e-18, 1.7287e-21,\n 9.9139e-19, 5.9042e-20, 4.5915e-19, 1.0168e-18], device='cuda:0')" + "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')" }, "27": { - "step": "tensor(21284.)", - "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([[3.2939e-20, 2.0719e-22, 4.1859e-21, ..., 3.2328e-22, 1.1658e-21,\n 5.7196e-23],\n [7.1671e-18, 7.0922e-18, 4.9330e-21, ..., 4.9991e-19, 9.6064e-19,\n 2.0103e-19],\n [5.7435e-19, 6.7349e-19, 2.7388e-22, ..., 5.3126e-20, 7.8801e-20,\n 3.3085e-20],\n ...,\n [5.9405e-18, 6.3290e-18, 4.3392e-21, ..., 3.4559e-19, 8.1912e-19,\n 1.4184e-19],\n [8.2572e-20, 1.1379e-19, 1.6296e-21, ..., 9.2632e-21, 7.2502e-21,\n 2.3152e-21],\n [1.0454e-17, 1.0831e-17, 2.9731e-21, ..., 8.0526e-19, 1.3395e-18,\n 3.2421e-19]], device='cuda:0')" + "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')" }, "28": { - "step": "tensor(21284.)", - "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([1.3896e-18, 3.8880e-15, 4.0122e-16, 8.2742e-16, 1.5825e-16, 1.3126e-17,\n 4.0481e-16, 2.0052e-15, 1.1719e-15, 1.4582e-17, 4.6304e-17, 2.7826e-17,\n 6.2708e-17, 6.0273e-17, 1.1933e-17, 5.8709e-17, 2.0823e-18, 2.2375e-16,\n 1.7715e-16, 1.4304e-15, 2.4708e-17, 6.3077e-16, 3.9103e-15, 7.2820e-16,\n 1.0594e-15, 7.5294e-18, 2.2520e-17, 4.2077e-15, 9.5162e-20, 6.8809e-18,\n 2.2104e-17, 8.1326e-16, 4.6610e-16, 3.3405e-16, 7.2375e-17, 1.9293e-16,\n 1.3186e-16, 4.7604e-17, 3.1990e-16, 6.5319e-17, 5.9965e-17, 3.0411e-16,\n 2.7925e-15, 2.6591e-16, 4.7451e-17, 2.4845e-16, 9.4597e-17, 2.0776e-16,\n 4.3419e-18, 4.7546e-16, 2.8738e-16, 6.1382e-15, 4.8054e-17, 1.1249e-16,\n 1.8409e-17, 1.6222e-15, 1.4348e-14, 3.1160e-15, 2.3565e-16, 3.3397e-16,\n 2.7122e-15, 4.7365e-17, 5.3711e-16, 5.7418e-15, 1.3071e-16, 2.0987e-15,\n 1.4225e-16, 1.3530e-16, 5.1182e-15, 1.0458e-15, 7.3605e-19, 6.5735e-17,\n 4.4741e-17, 6.3534e-17, 4.9648e-19, 2.0632e-17, 1.9976e-18, 7.8672e-18,\n 4.1473e-16, 3.5233e-17, 3.7506e-18, 1.0730e-15, 1.6170e-16, 7.9065e-16,\n 1.6501e-18, 1.2742e-16, 8.8761e-16, 1.4641e-17, 3.1763e-15, 1.9010e-17,\n 6.3641e-16, 2.6213e-16, 8.9665e-16, 1.3173e-16, 5.9817e-16, 6.6325e-15,\n 1.0118e-17, 4.5677e-18, 3.9615e-16, 1.6197e-16, 4.7216e-18, 6.8480e-17,\n 3.0767e-17, 3.6286e-16, 8.0128e-18, 1.1260e-16, 5.0415e-16, 1.9751e-19,\n 7.6735e-17, 3.8249e-18, 3.0240e-19, 5.4093e-17, 4.3227e-19, 3.0958e-17,\n 5.2209e-16, 8.9984e-16, 5.5647e-16, 4.2498e-15, 4.2199e-18, 2.6480e-17,\n 1.7567e-15, 3.4994e-18, 6.7786e-18, 5.5260e-19, 7.4697e-18, 6.3689e-18,\n 4.5314e-16, 7.0456e-17, 8.2802e-17, 3.4254e-16, 1.6555e-16, 2.6928e-17,\n 6.7405e-17, 2.0928e-17, 4.5042e-17, 1.2921e-19, 1.1491e-16, 9.5990e-17,\n 3.6577e-16, 6.7200e-15, 4.6668e-16, 2.6898e-18, 2.3312e-16, 1.1438e-14,\n 2.9501e-16, 5.0604e-17, 3.1252e-17, 4.0620e-17, 3.2384e-17, 8.1561e-17,\n 1.5889e-17, 1.1078e-17, 8.7061e-18, 8.4443e-16, 7.1453e-18, 3.5013e-18,\n 3.5896e-18, 1.6244e-15, 4.2812e-17, 2.7533e-16, 2.4993e-16, 2.1892e-19,\n 6.8864e-18, 2.6020e-15, 1.6112e-17, 1.6850e-15, 1.5490e-18, 3.2731e-16,\n 4.8576e-15, 7.7700e-17, 2.4948e-15, 5.4826e-17, 1.1726e-16, 4.3348e-19,\n 5.9646e-18, 6.3872e-17, 2.5311e-16, 3.3266e-19, 2.0860e-15, 4.4479e-17,\n 1.0045e-15, 7.1386e-16, 3.5104e-17, 2.4289e-17, 3.1456e-18, 1.2938e-16,\n 3.6062e-17, 1.4820e-17, 5.2806e-17, 2.3347e-16, 1.2461e-15, 7.4080e-20,\n 4.1225e-16, 6.8458e-16, 5.3951e-16, 1.6895e-14, 5.4099e-18, 2.9417e-16,\n 5.2532e-17, 4.3463e-16, 1.4939e-16, 4.8714e-16, 1.5353e-16, 8.1844e-17,\n 1.6704e-16, 1.5044e-16, 8.7217e-19, 3.2060e-16, 7.6265e-19, 3.6681e-15,\n 2.7903e-17, 6.8621e-18, 5.6492e-18, 3.6971e-15, 1.1497e-15, 2.7487e-16,\n 2.0335e-16, 3.1767e-15, 1.0451e-14, 2.8555e-18, 8.7769e-18, 9.4170e-16,\n 4.1989e-18, 6.1880e-16, 4.7957e-17, 1.3043e-17, 2.9413e-15, 2.3729e-15,\n 2.6277e-18, 9.9566e-18, 2.1987e-15, 1.2199e-16, 1.7247e-16, 9.2177e-17,\n 1.1428e-16, 2.0457e-18, 6.9661e-17, 2.2944e-18, 3.4986e-16, 5.7748e-17,\n 1.6943e-18, 2.9118e-16, 2.4277e-15, 1.1999e-16, 3.7385e-16, 9.6307e-17,\n 2.8012e-15, 2.6877e-17, 1.1264e-15, 1.9668e-16, 3.3994e-17, 1.4971e-19,\n 8.9866e-17, 2.6752e-15, 5.0567e-17, 5.5479e-15], device='cuda:0')" + "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')" }, "29": { - "step": "tensor(21284.)", - "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([2.0553e-20, 6.5309e-18, 1.5795e-19, 1.7594e-20, 2.6501e-20, 6.8182e-20,\n 9.2264e-21, 7.4587e-19, 2.0527e-19, 4.2330e-20, 1.6109e-19, 8.0644e-21,\n 3.9721e-20, 3.3968e-19, 4.3170e-20, 1.0185e-19, 9.5845e-22, 3.8167e-19,\n 7.5789e-21, 2.2991e-18, 4.5648e-20, 1.6190e-19, 1.7955e-17, 2.9914e-19,\n 3.0567e-18, 1.2545e-21, 5.5117e-21, 2.7588e-18, 3.3697e-21, 1.8285e-22,\n 5.9770e-21, 6.4070e-19, 1.0339e-19, 5.7032e-19, 8.1506e-21, 1.7687e-19,\n 1.1937e-20, 8.2712e-21, 3.2524e-19, 5.6991e-21, 1.2586e-19, 6.8121e-20,\n 3.6980e-18, 3.1274e-19, 9.8688e-20, 7.5678e-19, 4.9434e-21, 8.5045e-20,\n 7.0440e-21, 5.7193e-19, 1.5147e-19, 1.8201e-17, 2.8025e-20, 8.2026e-21,\n 4.2468e-20, 1.2962e-18, 3.1620e-17, 1.4035e-18, 8.8271e-20, 1.9261e-19,\n 3.7425e-19, 8.9817e-20, 1.4508e-19, 2.8597e-18, 6.5933e-21, 1.0087e-18,\n 1.2903e-19, 7.9786e-19, 1.9165e-17, 1.6665e-18, 8.5937e-21, 1.3405e-19,\n 1.7243e-20, 1.6477e-19, 6.3917e-20, 2.8257e-21, 2.6464e-20, 9.3213e-22,\n 1.1001e-19, 2.9722e-20, 9.6950e-23, 6.7455e-19, 7.9903e-21, 6.9794e-18,\n 2.2587e-21, 1.5355e-18, 1.2451e-19, 1.6178e-20, 6.5214e-18, 5.6411e-20,\n 1.6584e-18, 2.0021e-20, 2.5041e-19, 8.5864e-21, 1.2357e-19, 7.0359e-18,\n 3.1978e-19, 1.2865e-21, 2.7448e-19, 1.0714e-19, 3.6035e-20, 4.7863e-21,\n 1.3748e-20, 5.2322e-20, 1.0729e-20, 3.8872e-19, 1.0716e-18, 2.7055e-22,\n 1.6895e-20, 1.6113e-20, 8.3499e-21, 1.4122e-19, 8.4785e-20, 1.8737e-21,\n 2.4162e-18, 1.0789e-18, 3.8141e-19, 9.4962e-18, 2.1019e-22, 7.6883e-21,\n 1.0015e-18, 8.8584e-21, 1.0895e-21, 2.2358e-21, 3.6074e-21, 8.0249e-22,\n 8.4797e-21, 3.7764e-21, 4.7807e-21, 2.7619e-20, 2.9912e-20, 5.5996e-22,\n 2.1779e-21, 1.1692e-20, 1.6408e-20, 3.8418e-21, 6.1268e-20, 9.8060e-20,\n 7.5233e-19, 2.1295e-17, 6.9570e-20, 1.6848e-20, 4.4492e-20, 5.0290e-18,\n 5.4452e-20, 6.1033e-20, 3.2156e-21, 9.3013e-22, 3.8202e-21, 2.7762e-20,\n 5.2264e-21, 1.6359e-21, 1.2241e-20, 2.8061e-18, 1.6816e-21, 7.9929e-21,\n 3.2812e-22, 1.0703e-18, 1.5963e-22, 1.9187e-20, 1.0298e-19, 2.2592e-21,\n 6.2740e-20, 9.2777e-18, 9.3143e-22, 1.5525e-19, 1.5319e-20, 1.7489e-19,\n 6.5094e-18, 1.2398e-19, 5.2826e-18, 1.1035e-21, 1.4256e-19, 2.5343e-21,\n 4.4203e-20, 1.0675e-19, 3.1915e-20, 9.5208e-21, 4.6008e-19, 1.4291e-21,\n 1.0546e-19, 7.4686e-19, 1.9440e-19, 1.4980e-19, 1.7793e-20, 7.9211e-20,\n 9.1247e-20, 7.2940e-21, 5.4785e-20, 1.3373e-19, 6.0024e-19, 1.9114e-20,\n 7.7363e-20, 5.4010e-19, 5.0905e-20, 4.8989e-17, 1.0110e-22, 2.1055e-20,\n 2.1716e-21, 6.4103e-20, 9.0188e-21, 6.4907e-19, 1.2042e-20, 2.0466e-19,\n 4.4539e-20, 2.6089e-21, 5.8627e-21, 5.8131e-20, 6.0585e-21, 7.5135e-18,\n 1.8544e-20, 2.0626e-21, 4.2538e-21, 1.2456e-17, 4.6539e-20, 4.4024e-20,\n 2.2194e-20, 7.3134e-18, 1.9973e-17, 2.8855e-20, 4.0643e-21, 2.7340e-18,\n 5.4486e-22, 2.4220e-19, 2.4224e-20, 8.8715e-22, 5.5557e-19, 1.7457e-18,\n 9.1002e-23, 8.4513e-21, 3.9278e-19, 9.3707e-20, 8.5365e-19, 2.4114e-18,\n 8.8422e-21, 1.2035e-20, 9.5225e-20, 7.7485e-21, 6.3233e-19, 5.1422e-22,\n 2.5762e-20, 3.6920e-19, 2.8151e-18, 8.6214e-20, 3.2951e-19, 1.5521e-20,\n 2.4637e-18, 8.9158e-22, 8.3658e-19, 4.6966e-19, 4.7436e-21, 9.3977e-23,\n 1.0092e-20, 1.7772e-19, 3.8856e-21, 1.0844e-17], device='cuda:0')" + "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')" }, "30": { - "step": "tensor(21284.)", - "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([4.5436e-21, 5.3532e-18, 6.2571e-19, 9.5018e-19, 2.4809e-19, 3.1768e-20,\n 5.5398e-19, 2.7558e-18, 1.6210e-18, 6.7031e-20, 4.5718e-20, 7.4016e-20,\n 2.1715e-19, 4.0575e-19, 9.4851e-20, 3.5130e-20, 3.5785e-21, 4.0992e-19,\n 2.0813e-19, 2.1143e-18, 1.6678e-20, 9.1669e-19, 3.8931e-18, 1.2693e-18,\n 1.7595e-18, 1.4626e-20, 5.3453e-20, 4.7560e-18, 9.8616e-23, 1.1555e-20,\n 2.0916e-20, 1.2268e-18, 6.8186e-19, 1.0667e-18, 1.6998e-20, 4.8028e-19,\n 2.0662e-19, 4.5658e-20, 5.6266e-19, 3.7202e-20, 2.1848e-19, 4.6632e-19,\n 3.9967e-18, 4.9231e-19, 2.5587e-19, 4.6812e-19, 7.5079e-20, 5.1682e-20,\n 2.5094e-21, 5.7766e-19, 4.3142e-19, 8.8239e-18, 9.9406e-20, 7.8708e-20,\n 5.3734e-20, 1.7432e-18, 1.9232e-17, 4.2441e-18, 2.3783e-19, 4.9586e-19,\n 3.5805e-18, 1.4623e-19, 5.3066e-19, 6.9687e-18, 1.9429e-19, 2.9120e-18,\n 2.2376e-19, 4.2863e-19, 7.2703e-18, 1.5763e-18, 7.9378e-22, 4.9293e-20,\n 9.7216e-20, 3.5747e-19, 8.1239e-20, 7.4660e-21, 8.7020e-21, 3.5966e-22,\n 6.2559e-19, 8.6980e-20, 2.4578e-21, 1.3504e-18, 2.1638e-19, 1.8178e-18,\n 1.3399e-20, 4.9709e-19, 1.0117e-18, 8.2102e-21, 4.7072e-18, 1.4860e-20,\n 1.1578e-18, 3.7586e-19, 9.3613e-19, 1.8108e-19, 8.4780e-19, 8.0637e-18,\n 2.1360e-19, 3.7583e-22, 6.8910e-19, 3.4963e-19, 2.0783e-20, 1.0299e-19,\n 9.0889e-20, 4.5508e-19, 3.9594e-21, 5.5207e-19, 8.4860e-19, 4.8553e-21,\n 2.1797e-19, 3.4462e-21, 3.7617e-21, 1.5835e-19, 8.2214e-20, 4.4171e-20,\n 5.7536e-19, 1.3706e-18, 8.7085e-19, 5.1673e-18, 8.1815e-21, 3.3576e-20,\n 1.8770e-18, 1.4550e-21, 1.1078e-20, 1.1908e-21, 4.6472e-20, 1.3201e-20,\n 4.8354e-19, 3.8210e-20, 4.1883e-20, 2.9989e-19, 2.5008e-19, 6.3951e-21,\n 5.6205e-20, 3.1988e-21, 4.4850e-20, 2.2240e-21, 2.4844e-19, 1.5217e-19,\n 6.1865e-19, 9.3904e-18, 6.5299e-19, 1.6253e-21, 4.5823e-19, 1.4370e-17,\n 2.5899e-19, 2.5494e-20, 6.3109e-20, 6.2654e-20, 5.8121e-20, 1.3742e-19,\n 1.2084e-21, 2.7415e-20, 4.0129e-21, 1.3587e-18, 8.5321e-21, 1.2131e-21,\n 3.9648e-23, 2.2674e-18, 3.0330e-20, 3.2327e-19, 1.1646e-19, 1.8506e-22,\n 2.5951e-20, 3.1556e-18, 2.9268e-20, 2.2369e-18, 1.1027e-20, 5.2875e-19,\n 5.5869e-18, 8.8250e-20, 2.5688e-18, 3.4528e-20, 3.7697e-19, 7.6681e-22,\n 2.2361e-20, 3.1245e-20, 2.0693e-19, 2.7055e-21, 2.3230e-18, 6.4989e-20,\n 1.1867e-18, 6.3779e-19, 4.9160e-20, 7.1735e-20, 5.9041e-21, 2.3826e-19,\n 1.9226e-19, 4.4351e-21, 2.5089e-19, 3.8592e-19, 1.7357e-18, 2.8292e-23,\n 6.0376e-19, 9.9026e-19, 5.2382e-19, 2.2556e-17, 4.4540e-22, 4.2294e-19,\n 7.9738e-20, 3.7955e-19, 1.1903e-19, 8.0223e-19, 1.2509e-19, 2.4989e-19,\n 1.7141e-19, 2.1395e-19, 4.2514e-21, 4.6570e-19, 7.4671e-21, 4.7405e-18,\n 6.7788e-21, 1.1838e-21, 1.2395e-21, 4.7466e-18, 1.5334e-18, 2.5592e-19,\n 2.9301e-19, 4.5133e-18, 1.4041e-17, 1.9921e-20, 1.6632e-20, 1.3467e-18,\n 7.4356e-21, 7.2531e-19, 1.8221e-19, 5.8433e-21, 3.8961e-18, 3.0870e-18,\n 2.8317e-21, 2.3996e-20, 2.7547e-18, 2.0721e-19, 2.7433e-19, 5.5938e-19,\n 4.9793e-20, 1.6634e-20, 3.1384e-19, 8.5419e-22, 9.1757e-19, 2.6055e-20,\n 2.2804e-21, 5.6191e-19, 2.7119e-18, 2.6024e-19, 5.4530e-19, 1.5893e-19,\n 3.8818e-18, 4.6038e-20, 1.6675e-18, 3.0760e-19, 5.0455e-20, 1.3609e-22,\n 1.0662e-19, 3.5277e-18, 7.3447e-20, 7.6946e-18], device='cuda:0')" + "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')" }, "31": { - "step": "tensor(21284.)", - "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([[2.3430e-18, 2.4404e-18, 1.2337e-21, ..., 1.7104e-19, 2.7977e-19,\n 8.1303e-20],\n [3.1591e-20, 4.0813e-20, 7.3362e-23, ..., 2.9673e-21, 3.6988e-21,\n 7.9926e-22],\n [2.0711e-18, 2.2796e-18, 9.0095e-22, ..., 1.8224e-19, 2.1973e-19,\n 7.9283e-20],\n ...,\n [1.4879e-17, 1.6706e-17, 3.3315e-22, ..., 1.3508e-18, 1.6711e-18,\n 7.5887e-19],\n [1.0996e-18, 1.2591e-18, 2.6895e-21, ..., 8.8261e-20, 1.5163e-19,\n 4.4655e-20],\n [1.2672e-20, 1.3357e-20, 2.3111e-24, ..., 8.8003e-22, 1.0664e-21,\n 2.4282e-22]], device='cuda:0')" + "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')" }, "32": { - "step": "tensor(21284.)", - "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([1.2237e-15, 1.4736e-17, 1.1764e-15, 4.7395e-17, 1.0081e-18, 1.6119e-17,\n 2.3683e-16, 2.2955e-15, 8.5515e-16, 4.8423e-17, 8.4917e-15, 6.6384e-17,\n 4.3360e-19, 4.1201e-19, 6.5689e-18, 1.5843e-16, 9.8743e-18, 7.6854e-16,\n 1.2751e-15, 4.8034e-16, 8.6416e-16, 2.3355e-17, 7.7324e-17, 3.3849e-15,\n 2.2416e-17, 2.9895e-16, 1.0631e-15, 1.9175e-17, 1.4471e-18, 4.0841e-16,\n 1.6425e-17, 1.4330e-15, 3.7120e-16, 3.0783e-16, 9.1021e-17, 9.6237e-16,\n 2.2823e-17, 1.2195e-16, 8.7507e-16, 6.5174e-17, 3.0353e-17, 2.5013e-16,\n 8.4141e-16, 2.1186e-17, 1.5315e-18, 7.8902e-18, 3.9686e-16, 6.4033e-18,\n 8.1546e-16, 7.4214e-16, 1.2561e-16, 7.5126e-16, 9.9288e-18, 3.7246e-16,\n 1.3979e-16, 1.0751e-15, 8.4812e-16, 3.1707e-15, 1.6906e-16, 5.9591e-16,\n 1.1345e-14, 8.1417e-18, 1.1304e-17, 1.0754e-15, 1.3521e-17, 2.2457e-15,\n 2.7008e-17, 2.2359e-17, 1.8442e-18, 4.3091e-16, 6.2437e-16, 9.0649e-16,\n 7.4856e-16, 2.0326e-16, 5.1000e-18, 5.2650e-18, 9.7549e-17, 7.0329e-17,\n 8.6177e-16, 2.8706e-17, 4.8441e-19, 3.7093e-15, 3.1346e-15, 9.2207e-19,\n 2.7704e-16, 8.0990e-17, 1.2893e-15, 3.9138e-16, 3.7476e-15, 1.5826e-17,\n 9.8295e-16, 4.5219e-16, 5.6696e-17, 1.8429e-17, 1.6095e-16, 2.7731e-15,\n 4.5110e-16, 5.8008e-18, 2.0840e-15, 1.5708e-16, 1.1036e-15, 2.9444e-18,\n 5.2161e-17, 2.5657e-16, 4.4620e-17, 2.2507e-16, 1.1339e-16, 1.4058e-18,\n 1.5222e-16, 9.3398e-18, 6.0329e-16, 5.8212e-21, 5.8473e-18, 2.7812e-18,\n 1.8413e-15, 2.9370e-16, 6.9376e-16, 1.5608e-15, 2.9358e-17, 3.6383e-15,\n 1.6241e-16, 1.4760e-16, 4.0158e-16, 1.8545e-18, 3.3995e-17, 1.3683e-17,\n 1.3722e-17, 2.1107e-16, 4.1517e-17, 1.6402e-17, 1.0083e-17, 2.8645e-16,\n 1.1477e-17, 1.9208e-19, 5.0525e-18, 2.9892e-17, 9.9497e-18, 4.0985e-17,\n 2.1250e-16, 3.5471e-15, 6.8619e-16, 4.1141e-17, 4.4869e-15, 2.4511e-14,\n 1.8301e-15, 6.2738e-17, 2.2414e-17, 9.0104e-16, 5.1109e-19, 8.5438e-17,\n 9.8891e-17, 4.0399e-18, 3.0975e-17, 8.4370e-16, 7.7626e-16, 3.0054e-16,\n 9.9777e-17, 2.0109e-15, 2.4451e-17, 1.3886e-16, 1.2388e-17, 2.0083e-18,\n 3.8076e-16, 8.5118e-16, 9.1470e-18, 5.6599e-16, 6.2204e-16, 2.0353e-17,\n 2.3073e-15, 7.0808e-18, 3.2851e-15, 8.6725e-16, 7.5944e-16, 1.6864e-18,\n 3.1245e-17, 8.6962e-17, 3.4528e-16, 1.3742e-15, 2.0739e-15, 4.2482e-18,\n 5.5478e-15, 4.4374e-18, 4.1725e-15, 4.4254e-17, 2.6778e-17, 2.7063e-16,\n 1.3191e-17, 2.3884e-17, 3.2530e-16, 7.1763e-17, 1.2527e-16, 4.1228e-17,\n 1.8300e-15, 8.6821e-16, 1.1330e-16, 1.6424e-14, 5.5799e-17, 4.7972e-18,\n 3.1572e-17, 4.2049e-17, 2.6090e-15, 1.3312e-17, 5.2191e-16, 6.6197e-18,\n 2.5644e-17, 4.9989e-16, 1.5371e-17, 1.7716e-15, 1.5581e-17, 6.3814e-15,\n 4.3745e-18, 6.0883e-16, 1.3346e-17, 1.6347e-15, 9.4158e-17, 2.9034e-17,\n 1.2647e-18, 6.7806e-17, 1.0911e-14, 9.0110e-16, 1.4484e-16, 9.6482e-18,\n 3.3786e-18, 1.6638e-16, 8.0454e-17, 2.2068e-16, 2.8411e-15, 1.2044e-16,\n 3.7671e-17, 3.0448e-17, 2.1184e-17, 9.8285e-17, 8.3229e-18, 4.8880e-16,\n 1.3150e-16, 6.7239e-16, 6.5588e-17, 3.7303e-18, 3.6995e-16, 1.0188e-15,\n 1.5208e-15, 1.0830e-17, 9.0499e-16, 6.3441e-16, 5.9309e-18, 9.0502e-18,\n 1.0131e-15, 2.8979e-16, 1.4213e-16, 1.5497e-17, 7.8345e-16, 2.8482e-18,\n 6.7261e-16, 9.5974e-15, 7.0833e-16, 6.3694e-18], device='cuda:0')" + "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')" }, "33": { - "step": "tensor(21284.)", - "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([6.3408e-20, 5.2277e-20, 4.2004e-19, 3.0906e-20, 1.1040e-19, 7.0705e-20,\n 1.1233e-20, 2.1161e-18, 9.8692e-20, 6.6533e-20, 1.2783e-17, 2.8880e-20,\n 2.2114e-20, 1.2632e-21, 1.3582e-19, 4.4951e-21, 8.5193e-21, 2.1777e-18,\n 7.0123e-19, 2.9891e-19, 5.4880e-19, 8.4358e-20, 3.9389e-21, 6.5333e-18,\n 6.1026e-21, 8.4574e-20, 6.1506e-19, 1.1335e-20, 1.2779e-20, 5.6810e-19,\n 2.8399e-21, 2.7559e-18, 4.7833e-20, 3.5506e-19, 1.3176e-20, 5.4844e-18,\n 1.1338e-20, 2.1495e-20, 4.9634e-19, 1.9386e-21, 1.3305e-20, 4.5402e-20,\n 3.1059e-19, 4.1350e-19, 1.7801e-21, 6.2153e-22, 1.3453e-19, 1.0697e-22,\n 2.7198e-18, 2.3785e-19, 7.7094e-20, 3.4282e-20, 4.2862e-20, 9.5524e-20,\n 1.4298e-19, 2.2375e-19, 1.0504e-19, 1.3728e-18, 3.8914e-20, 2.9802e-19,\n 1.4252e-17, 2.4642e-21, 3.8858e-20, 8.1992e-20, 1.9659e-21, 1.3001e-18,\n 1.4231e-21, 2.3440e-21, 5.8575e-20, 1.0556e-19, 3.1577e-19, 4.1246e-20,\n 6.7857e-20, 5.4050e-19, 3.3942e-20, 1.3533e-21, 1.3173e-21, 1.0778e-20,\n 2.3997e-19, 5.3565e-19, 8.2371e-23, 1.0879e-17, 2.4710e-18, 2.0347e-20,\n 1.2904e-19, 8.5367e-20, 5.9768e-19, 1.5238e-19, 1.3064e-17, 1.1031e-20,\n 4.6222e-18, 3.0647e-18, 1.2523e-21, 1.0061e-19, 1.3854e-20, 6.4341e-19,\n 1.5950e-18, 1.1782e-20, 8.8317e-19, 7.2473e-20, 1.2473e-18, 8.8867e-21,\n 3.3041e-20, 3.2525e-20, 1.0649e-20, 1.9849e-18, 5.7866e-21, 1.3013e-20,\n 2.5854e-20, 3.9263e-21, 2.5672e-19, 1.1774e-21, 2.6388e-20, 1.5461e-20,\n 2.9447e-18, 1.4711e-18, 6.0854e-19, 5.4208e-19, 5.8365e-21, 4.9666e-18,\n 6.1484e-21, 4.9986e-21, 7.4251e-19, 4.8706e-22, 2.2587e-20, 2.8429e-21,\n 1.2279e-20, 4.2040e-20, 4.4892e-20, 2.5805e-20, 5.4799e-20, 1.5246e-19,\n 6.7294e-21, 1.6765e-22, 3.1338e-21, 5.8002e-21, 2.5309e-20, 1.3535e-20,\n 1.0769e-19, 2.9669e-18, 1.5626e-18, 1.4016e-20, 5.0290e-18, 9.7884e-17,\n 1.8725e-18, 1.9469e-20, 2.7794e-21, 5.8697e-19, 1.1566e-21, 1.4310e-19,\n 4.0219e-19, 5.5198e-21, 4.6692e-20, 6.5890e-19, 1.2647e-19, 1.9616e-18,\n 7.4107e-21, 8.3832e-19, 7.2845e-21, 3.9396e-19, 3.1575e-21, 4.0423e-21,\n 1.0432e-20, 8.6270e-19, 5.1871e-21, 5.2082e-21, 1.5277e-18, 1.6929e-21,\n 3.1399e-18, 5.5041e-21, 7.2461e-18, 2.6815e-19, 1.0201e-18, 9.1990e-22,\n 7.3933e-21, 3.5795e-20, 8.9911e-20, 4.9360e-19, 3.0416e-19, 1.5923e-20,\n 3.8759e-18, 9.2096e-21, 1.1883e-18, 1.1161e-19, 2.4811e-21, 8.6632e-20,\n 5.8390e-21, 3.6411e-20, 7.5060e-19, 3.1754e-21, 6.7253e-21, 1.0104e-20,\n 7.8918e-18, 1.2175e-18, 4.7634e-21, 4.0535e-17, 1.4402e-21, 1.6527e-20,\n 3.4277e-21, 5.1951e-21, 2.0554e-18, 6.3124e-20, 1.0489e-19, 2.5296e-20,\n 2.3463e-20, 1.0534e-19, 1.4895e-19, 9.1590e-19, 8.2367e-21, 2.0037e-17,\n 1.0969e-20, 1.7008e-19, 1.2263e-21, 2.3713e-18, 4.9424e-20, 5.1969e-20,\n 6.5547e-21, 4.3043e-21, 1.6859e-17, 2.4808e-18, 2.7008e-19, 2.4041e-21,\n 2.2281e-22, 2.9145e-19, 3.6178e-20, 4.3620e-20, 5.7020e-19, 1.4132e-20,\n 1.9105e-21, 7.2293e-21, 9.3889e-20, 2.9785e-21, 1.6136e-21, 7.9709e-19,\n 2.2932e-20, 2.0283e-19, 1.9817e-20, 7.2624e-22, 2.4255e-18, 1.0392e-18,\n 2.9461e-18, 1.9376e-21, 8.2948e-20, 3.0658e-19, 2.0837e-21, 3.6021e-22,\n 4.0496e-19, 1.3825e-19, 1.0670e-20, 2.0425e-20, 1.9965e-19, 9.1829e-22,\n 8.4637e-19, 7.3335e-18, 3.3877e-19, 6.6968e-20], device='cuda:0')" + "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')" }, "34": { - "step": "tensor(21284.)", - "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([1.4919e-18, 1.9517e-20, 1.6806e-18, 4.0994e-20, 1.1176e-19, 2.5702e-20,\n 3.4327e-19, 3.2789e-18, 1.1861e-18, 1.0924e-19, 1.0148e-17, 2.7138e-19,\n 7.6775e-21, 1.4331e-21, 1.5649e-19, 1.5438e-19, 2.6004e-20, 1.2552e-18,\n 1.4197e-18, 9.6013e-19, 1.3557e-18, 3.4630e-20, 6.3976e-20, 4.7146e-18,\n 1.1014e-19, 4.4327e-19, 1.4162e-18, 5.8178e-21, 4.3058e-21, 6.2860e-19,\n 1.3401e-20, 1.9342e-18, 5.4760e-19, 8.8731e-19, 7.9395e-21, 1.8393e-18,\n 7.9928e-20, 5.3875e-20, 1.2667e-18, 8.3237e-21, 8.8188e-20, 3.8662e-19,\n 1.4038e-18, 1.9821e-19, 1.7183e-20, 1.9225e-20, 3.5472e-19, 1.1555e-21,\n 1.2131e-18, 9.6410e-19, 1.9277e-19, 1.0351e-18, 3.6282e-20, 2.8406e-19,\n 2.3629e-19, 1.2148e-18, 1.1392e-18, 4.2901e-18, 1.9029e-19, 7.7217e-19,\n 1.4948e-17, 1.3973e-20, 9.9647e-21, 1.3133e-18, 2.2740e-20, 3.1022e-18,\n 3.1999e-22, 4.0879e-20, 6.1964e-21, 6.3142e-19, 8.7924e-19, 9.6491e-19,\n 9.8986e-19, 6.0380e-19, 3.9483e-20, 7.5638e-22, 1.4569e-19, 4.9126e-20,\n 1.1692e-18, 2.7433e-19, 8.8243e-22, 4.2894e-18, 4.2395e-18, 7.9988e-21,\n 2.7624e-19, 1.9209e-19, 1.5035e-18, 5.7696e-19, 5.5345e-18, 1.1299e-20,\n 1.7898e-18, 1.0778e-18, 2.7957e-20, 2.9126e-20, 2.4521e-19, 3.5250e-18,\n 8.4942e-19, 2.9768e-20, 2.8897e-18, 3.9722e-19, 1.6574e-18, 3.3658e-21,\n 2.4412e-19, 3.1167e-19, 2.4218e-20, 9.9680e-19, 1.6873e-19, 3.9373e-21,\n 2.9804e-19, 2.0448e-21, 8.8035e-19, 2.0015e-22, 5.1885e-20, 7.0105e-21,\n 2.2301e-18, 6.4087e-19, 1.0628e-18, 1.9908e-18, 5.5052e-20, 4.9382e-18,\n 1.2447e-19, 1.5723e-19, 6.7689e-19, 9.1774e-21, 1.5399e-19, 1.0093e-21,\n 4.8974e-21, 1.0993e-19, 1.8123e-20, 8.6777e-21, 1.9492e-20, 1.3534e-19,\n 1.1392e-20, 9.8566e-21, 1.4715e-20, 5.1498e-20, 9.6137e-21, 7.4301e-20,\n 3.4123e-19, 4.8244e-18, 1.1019e-18, 2.9458e-20, 5.7685e-18, 3.0443e-17,\n 1.8193e-18, 7.9743e-20, 9.5299e-20, 1.3113e-18, 1.4440e-21, 1.2378e-19,\n 4.9872e-19, 2.3332e-20, 1.5536e-20, 1.2372e-18, 1.0949e-18, 4.7980e-19,\n 4.2455e-20, 2.7451e-18, 4.5549e-20, 1.5558e-19, 9.8941e-22, 2.3808e-21,\n 5.2497e-19, 1.1801e-18, 1.3406e-21, 7.7222e-19, 1.0505e-18, 4.1820e-20,\n 2.5098e-18, 1.7504e-21, 3.5692e-18, 9.6759e-19, 1.2995e-18, 2.1966e-21,\n 6.5633e-20, 1.2405e-19, 3.0104e-19, 1.6606e-18, 2.3298e-18, 6.0128e-21,\n 6.5848e-18, 7.2301e-22, 5.5811e-18, 1.2455e-19, 5.1046e-20, 4.2928e-19,\n 7.1803e-20, 1.5497e-20, 8.3080e-19, 1.1570e-19, 1.8702e-19, 5.9386e-20,\n 3.0449e-18, 1.2327e-18, 8.6059e-20, 2.1709e-17, 2.0038e-21, 9.0466e-21,\n 5.4083e-20, 2.6259e-20, 2.8092e-18, 9.9695e-20, 4.8514e-19, 8.7271e-20,\n 2.1194e-20, 7.2695e-19, 6.7909e-20, 2.4268e-18, 4.6913e-20, 8.0346e-18,\n 4.4089e-22, 6.4857e-19, 5.8582e-21, 2.2572e-18, 1.2705e-19, 1.7336e-20,\n 2.1914e-21, 1.0515e-19, 1.4493e-17, 1.5157e-18, 3.9668e-19, 5.8150e-20,\n 1.0102e-20, 2.5351e-19, 2.6457e-19, 1.8388e-19, 3.8178e-18, 2.4041e-19,\n 6.6175e-20, 6.6893e-21, 2.7066e-20, 1.5185e-19, 8.0249e-22, 9.2363e-19,\n 6.0953e-20, 9.9674e-19, 2.5199e-19, 8.9579e-21, 1.2366e-18, 9.5119e-19,\n 2.2252e-18, 2.2628e-20, 1.0151e-18, 9.8140e-19, 4.7118e-20, 1.8735e-20,\n 1.4263e-18, 4.4874e-19, 2.0837e-19, 9.5035e-20, 1.1203e-18, 2.6667e-22,\n 7.5710e-19, 1.2654e-17, 1.0012e-18, 1.9765e-20], device='cuda:0')" + "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')" }, "35": { - "step": "tensor(21284.)", - "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([[1.2324e-19, 1.4548e-19, 7.5888e-23, ..., 1.1017e-20, 1.2693e-20,\n 5.4309e-21],\n [6.3151e-21, 3.6000e-21, 3.6848e-22, ..., 1.5549e-22, 1.8144e-21,\n 8.3753e-22],\n [4.3723e-18, 5.0218e-18, 6.1933e-21, ..., 3.5477e-19, 6.0500e-19,\n 1.5729e-19],\n ...,\n [1.4634e-17, 1.5062e-17, 1.6484e-20, ..., 9.9154e-19, 1.7737e-18,\n 4.8775e-19],\n [4.0543e-22, 1.3990e-23, 1.5329e-22, ..., 7.6923e-23, 8.9332e-23,\n 2.6076e-23],\n [8.2004e-18, 9.3390e-18, 5.9626e-21, ..., 6.9884e-19, 1.1212e-18,\n 2.9181e-19]], device='cuda:0')" + "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')" }, "36": { - "step": "tensor(21284.)", - "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([7.1271e-17, 2.3348e-19, 2.5660e-15, 1.1579e-14, 2.8285e-16, 6.0360e-18,\n 1.6927e-15, 9.9558e-16, 7.0855e-16, 1.6348e-18, 9.6103e-17, 5.2456e-19,\n 5.4493e-16, 3.2516e-17, 1.6889e-17, 2.1714e-16, 2.9157e-17, 2.7403e-16,\n 1.0417e-17, 2.3433e-16, 5.9592e-17, 1.3386e-18, 1.7741e-15, 3.5927e-15,\n 2.8479e-16, 4.2960e-17, 4.3399e-16, 3.8405e-16, 1.6063e-17, 9.1912e-17,\n 1.3915e-16, 2.1464e-16, 4.3086e-16, 3.4618e-16, 3.8432e-18, 2.9372e-16,\n 1.0192e-15, 1.3956e-16, 7.6708e-16, 5.8481e-18, 1.5630e-15, 9.0978e-18,\n 9.8953e-16, 1.1926e-15, 2.4374e-17, 3.0722e-18, 9.8187e-17, 3.9350e-17,\n 1.2169e-17, 1.5171e-15, 3.1158e-16, 1.4816e-15, 4.6968e-19, 1.8843e-16,\n 5.7950e-17, 1.7300e-17, 4.1188e-15, 6.1532e-17, 3.7166e-16, 6.3506e-19,\n 1.9435e-18, 4.4379e-18, 5.6494e-16, 2.8362e-16, 2.2050e-16, 1.3360e-15,\n 2.7041e-16, 9.7833e-18, 3.1920e-16, 1.2787e-15, 2.5993e-16, 1.5669e-16,\n 1.0476e-16, 1.5698e-18, 2.8156e-19, 1.3496e-16, 2.3930e-16, 2.7838e-17,\n 2.3518e-15, 2.3817e-19, 3.2826e-18, 2.5886e-16, 3.5340e-15, 7.4942e-19,\n 2.8634e-16, 8.6084e-18, 3.0485e-16, 3.6285e-17, 4.5647e-16, 3.1835e-15,\n 4.6456e-16, 1.1314e-16, 1.9104e-17, 3.2142e-16, 2.1253e-15, 1.1214e-14,\n 4.3472e-18, 5.8974e-18, 7.0333e-17, 2.0385e-16, 3.9618e-16, 7.6284e-17,\n 1.0256e-16, 3.5562e-16, 2.1765e-15, 5.2045e-18, 4.5244e-19, 1.2896e-16,\n 1.5794e-16, 1.2216e-15, 1.2192e-15, 1.6160e-18, 9.9470e-19, 1.2790e-17,\n 1.7384e-17, 3.9989e-16, 3.8575e-16, 1.0400e-16, 4.5469e-16, 2.0637e-17,\n 3.5414e-15, 1.2572e-15, 1.5368e-18, 1.1503e-16, 3.4664e-16, 1.0061e-17,\n 3.5139e-18, 8.6543e-17, 2.9683e-18, 2.0416e-16, 2.8656e-17, 1.2054e-16,\n 1.0396e-15, 1.5862e-16, 1.0961e-16, 9.8462e-17, 4.3234e-16, 4.5659e-17,\n 8.6320e-17, 1.0461e-18, 1.6051e-17, 2.3167e-15, 6.6777e-16, 7.2476e-17,\n 2.3486e-15, 1.2704e-18, 4.2574e-19, 5.2840e-16, 7.3187e-17, 2.3609e-17,\n 1.4624e-17, 1.6158e-16, 8.6968e-17, 3.6059e-18, 1.5628e-17, 2.3399e-17,\n 1.4695e-17, 3.9623e-15, 1.1768e-15, 4.0275e-15, 8.9912e-17, 1.2379e-16,\n 4.5264e-15, 1.2985e-19, 8.5978e-17, 1.2655e-14, 1.4530e-15, 2.7041e-16,\n 3.4115e-15, 4.7078e-18, 8.3694e-17, 7.4230e-16, 1.5065e-16, 3.2304e-20,\n 9.9945e-17, 3.8813e-15, 2.2368e-17, 4.8696e-18, 5.7755e-15, 1.7097e-15,\n 2.0386e-18, 1.7817e-17, 2.1938e-18, 1.0507e-16, 4.1852e-16, 2.3677e-19,\n 1.4803e-17, 3.3670e-15, 1.7794e-17, 4.8721e-19, 3.7486e-18, 6.7467e-16,\n 7.9058e-16, 6.5898e-16, 8.6325e-16, 9.6872e-17, 1.0832e-18, 2.7358e-15,\n 3.6040e-16, 1.4823e-16, 2.0456e-16, 9.4346e-17, 8.0919e-17, 1.6477e-17,\n 1.0270e-17, 1.4744e-17, 9.2395e-18, 5.5581e-17, 3.5187e-18, 8.3390e-16,\n 2.9543e-16, 2.3405e-17, 2.6829e-17, 1.2083e-16, 7.4155e-15, 9.8195e-18,\n 4.4550e-17, 4.5548e-18, 2.1369e-15, 3.1745e-15, 3.5823e-19, 1.6564e-16,\n 1.2275e-16, 4.3047e-16, 7.3850e-19, 1.0321e-17, 1.4218e-15, 1.6164e-15,\n 4.5159e-19, 9.9306e-19, 2.3347e-19, 2.4842e-17, 2.3388e-16, 6.9782e-18,\n 5.3847e-17, 6.2310e-16, 1.0587e-16, 1.4971e-19, 8.6277e-16, 1.0779e-17,\n 1.3720e-15, 1.0602e-15, 2.9109e-15, 2.1662e-15, 5.1656e-17, 4.5021e-18,\n 6.9885e-18, 7.5932e-16, 1.7296e-15, 2.2150e-17, 2.1980e-15, 5.5945e-18,\n 3.6526e-16, 7.4997e-15, 8.4096e-20, 4.8030e-15], device='cuda:0')" + "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')" }, "37": { - "step": "tensor(21284.)", - "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([2.0135e-19, 3.2265e-20, 6.0430e-18, 2.1501e-17, 5.0994e-20, 6.3429e-20,\n 7.8458e-19, 3.3034e-19, 1.0146e-19, 3.4314e-22, 2.6111e-19, 1.1907e-20,\n 9.9527e-19, 5.7702e-20, 1.3442e-19, 4.7665e-21, 1.9770e-19, 1.1199e-19,\n 7.9011e-21, 1.8047e-19, 2.4568e-20, 9.6635e-21, 6.0652e-19, 1.7770e-17,\n 6.9912e-19, 1.2286e-20, 3.5838e-19, 1.4397e-20, 4.4057e-20, 1.0691e-20,\n 1.2706e-19, 4.3420e-19, 4.7993e-20, 4.6042e-19, 9.3914e-22, 2.0492e-18,\n 2.5975e-18, 6.0616e-20, 1.2971e-18, 2.2509e-20, 4.5585e-18, 3.4597e-21,\n 8.1280e-19, 4.1467e-18, 1.3081e-20, 5.3774e-20, 7.4157e-21, 1.2189e-21,\n 3.6253e-21, 1.1864e-18, 3.0365e-19, 4.1317e-19, 1.7848e-21, 1.4083e-20,\n 7.2976e-21, 2.6855e-21, 1.0722e-18, 2.5955e-19, 3.1415e-20, 9.3927e-21,\n 5.8243e-20, 5.1626e-21, 3.3353e-19, 8.3847e-21, 3.7820e-20, 3.5894e-19,\n 3.2015e-19, 1.7913e-21, 1.8931e-20, 2.8302e-18, 1.1430e-19, 7.4781e-21,\n 2.8627e-20, 8.3114e-21, 1.3083e-21, 8.6402e-21, 6.3157e-20, 7.7205e-21,\n 2.8746e-18, 2.3572e-22, 3.1520e-21, 2.2352e-20, 3.7075e-18, 2.1171e-20,\n 4.8214e-19, 1.0268e-21, 1.0010e-19, 6.8419e-21, 1.7801e-19, 1.8827e-18,\n 4.2277e-19, 1.9920e-20, 1.6201e-20, 1.4232e-20, 3.3777e-18, 4.7390e-17,\n 2.4404e-20, 1.3684e-20, 2.5282e-20, 5.3455e-19, 5.7457e-20, 5.0986e-21,\n 2.8010e-19, 1.1478e-19, 1.6707e-18, 5.8821e-22, 1.3055e-21, 4.5161e-19,\n 3.2235e-20, 1.5781e-18, 4.7823e-18, 1.5538e-21, 3.9087e-22, 1.2824e-21,\n 2.3457e-20, 1.5004e-19, 2.9789e-19, 1.3513e-20, 1.6184e-19, 9.7529e-20,\n 2.6287e-18, 2.7256e-18, 3.7164e-22, 1.3141e-19, 2.5613e-18, 1.0087e-21,\n 9.0053e-21, 1.1123e-20, 1.6718e-21, 4.3731e-20, 2.3828e-21, 5.5853e-20,\n 5.7871e-19, 3.1420e-20, 1.2861e-20, 4.9095e-21, 4.6160e-19, 2.4670e-21,\n 1.6384e-19, 3.5606e-20, 1.9030e-21, 1.4951e-18, 1.6027e-19, 4.3157e-19,\n 6.6014e-18, 1.7084e-21, 1.8088e-21, 4.9956e-19, 1.4917e-19, 4.2844e-21,\n 5.8265e-21, 1.6090e-19, 3.6745e-21, 1.7227e-21, 8.4902e-22, 7.9963e-19,\n 2.4395e-21, 1.4849e-17, 5.3018e-19, 1.2026e-18, 1.1547e-20, 1.1661e-19,\n 1.9088e-17, 7.6666e-21, 5.9127e-21, 2.3435e-17, 3.5768e-18, 6.1964e-19,\n 3.7772e-18, 2.7116e-20, 1.0159e-22, 3.4355e-19, 1.1336e-19, 2.5545e-20,\n 3.0217e-21, 4.1301e-18, 1.8545e-20, 1.1635e-20, 6.5094e-18, 3.4206e-18,\n 3.1059e-20, 3.8775e-20, 1.0032e-19, 8.2285e-20, 8.2456e-20, 2.1565e-21,\n 3.5447e-21, 5.0954e-18, 5.4735e-21, 2.7653e-21, 2.6562e-20, 1.0163e-19,\n 2.6325e-18, 1.1185e-18, 3.2079e-19, 4.2051e-19, 7.6428e-23, 1.4759e-17,\n 4.1499e-19, 4.9485e-21, 2.1070e-20, 3.0587e-21, 9.3075e-21, 1.9193e-19,\n 1.3971e-21, 2.6873e-21, 1.1891e-19, 3.4139e-22, 3.8642e-20, 3.1083e-19,\n 2.0396e-19, 1.9765e-21, 1.0363e-21, 3.4876e-20, 9.9514e-18, 3.1136e-22,\n 7.9260e-22, 3.3242e-21, 2.2551e-19, 7.2428e-18, 4.4550e-21, 4.9156e-20,\n 9.3232e-20, 8.1245e-19, 1.6092e-20, 1.6817e-20, 9.0808e-20, 6.1686e-19,\n 9.7649e-21, 5.6347e-21, 3.7516e-20, 5.7562e-21, 5.0139e-19, 4.6120e-21,\n 1.7495e-21, 1.0054e-18, 1.1263e-19, 6.7916e-22, 3.0313e-18, 2.5967e-21,\n 2.4113e-19, 1.4347e-18, 3.4918e-18, 6.6389e-18, 1.9247e-20, 2.2263e-20,\n 4.1766e-20, 5.1461e-19, 5.9194e-18, 1.3551e-20, 5.2433e-18, 4.0376e-22,\n 1.9832e-19, 5.4553e-18, 2.6471e-21, 3.1809e-17], device='cuda:0')" + "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')" }, "38": { - "step": "tensor(21284.)", - "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([7.3785e-20, 9.9308e-23, 4.0610e-18, 1.4486e-17, 4.4620e-19, 1.4563e-20,\n 2.4386e-18, 1.5670e-18, 1.0605e-18, 1.1844e-21, 7.2635e-20, 4.4264e-21,\n 1.1194e-18, 2.1597e-19, 1.4956e-19, 2.0910e-19, 1.7177e-19, 4.5329e-19,\n 9.1838e-21, 5.7950e-19, 1.9673e-19, 5.5218e-21, 2.2107e-18, 5.4339e-18,\n 8.7105e-19, 5.2392e-20, 7.6653e-19, 3.9359e-19, 1.6723e-20, 1.5680e-19,\n 2.1536e-19, 5.5856e-19, 6.5050e-19, 1.0384e-18, 8.7390e-21, 9.7941e-19,\n 1.7939e-18, 5.6552e-20, 1.2808e-18, 3.3447e-20, 2.1945e-18, 1.8817e-20,\n 1.8407e-18, 1.6651e-18, 1.2513e-19, 8.9791e-20, 6.9225e-20, 4.2425e-21,\n 1.2117e-21, 1.9747e-18, 4.8573e-19, 2.1503e-18, 1.6274e-21, 1.4052e-19,\n 3.0171e-20, 1.2205e-20, 5.7405e-18, 1.0072e-19, 4.5565e-19, 8.3087e-21,\n 1.8789e-23, 4.1467e-20, 5.5434e-19, 3.7362e-19, 3.3961e-19, 1.9701e-18,\n 3.6603e-19, 1.9862e-20, 4.7385e-19, 2.0220e-18, 4.0779e-19, 1.3408e-19,\n 2.6356e-19, 2.7206e-21, 8.3302e-22, 1.0684e-19, 3.9008e-19, 6.8488e-21,\n 3.2343e-18, 9.8698e-22, 7.1637e-21, 3.2376e-19, 5.0670e-18, 5.1124e-21,\n 3.5025e-19, 2.3556e-20, 2.7586e-19, 6.6489e-20, 7.6057e-19, 4.1052e-18,\n 7.8734e-19, 2.0179e-19, 5.8748e-21, 4.8133e-19, 3.1397e-18, 1.4575e-17,\n 1.1705e-20, 3.6903e-20, 2.0604e-19, 5.8139e-19, 6.0404e-19, 1.2482e-19,\n 4.8129e-19, 3.9620e-19, 2.6130e-18, 4.2239e-20, 1.5821e-21, 4.9865e-19,\n 2.7577e-19, 1.8303e-18, 2.0365e-18, 3.0179e-20, 9.5537e-21, 2.0294e-20,\n 7.7088e-21, 6.4202e-19, 6.3695e-19, 1.7940e-19, 6.9230e-19, 3.4425e-20,\n 4.1680e-18, 1.5068e-18, 3.5491e-21, 2.4249e-19, 1.1292e-18, 2.2362e-20,\n 8.6085e-23, 7.0657e-21, 5.1161e-24, 1.3403e-19, 5.1442e-20, 1.9602e-19,\n 1.1418e-18, 2.6820e-19, 1.7705e-19, 1.5375e-19, 8.4008e-19, 1.9280e-20,\n 1.6968e-19, 1.9736e-21, 2.6969e-20, 2.6815e-18, 1.0963e-18, 1.0197e-19,\n 2.3873e-18, 3.0376e-20, 2.7618e-22, 8.4248e-19, 1.8924e-19, 1.4074e-20,\n 5.4590e-20, 4.9861e-19, 7.5659e-20, 5.9552e-21, 2.6482e-20, 2.4721e-19,\n 7.2103e-22, 5.7834e-18, 1.7291e-18, 5.0962e-18, 2.9755e-20, 2.2249e-19,\n 6.7902e-18, 8.7479e-23, 1.4290e-19, 1.7659e-17, 2.3146e-18, 5.5660e-19,\n 4.3176e-18, 3.9730e-20, 6.6576e-20, 7.5643e-19, 4.8883e-19, 5.6824e-20,\n 1.6233e-19, 4.9522e-18, 6.8902e-21, 2.4433e-21, 6.9844e-18, 2.5803e-18,\n 2.6681e-21, 1.3224e-20, 3.2026e-21, 1.5149e-19, 6.3632e-19, 1.0618e-21,\n 2.4393e-20, 3.9293e-18, 8.9076e-20, 1.0785e-21, 1.3170e-21, 7.6234e-19,\n 1.4980e-18, 1.0034e-18, 9.1623e-19, 1.5163e-19, 3.9720e-22, 4.2560e-18,\n 6.7189e-19, 1.2607e-19, 1.5284e-19, 1.5293e-19, 6.3542e-20, 2.1351e-19,\n 4.0466e-22, 2.5693e-20, 6.1304e-20, 9.3752e-20, 6.9778e-21, 1.3769e-18,\n 2.8731e-19, 3.1450e-20, 4.9498e-20, 2.8377e-19, 1.0392e-17, 2.5413e-21,\n 7.2410e-20, 8.1278e-21, 3.0042e-18, 4.7919e-18, 2.9071e-22, 3.0007e-19,\n 7.6846e-20, 6.7608e-19, 4.5099e-21, 6.0969e-21, 1.9818e-18, 2.2244e-18,\n 1.1300e-21, 2.3788e-21, 3.3115e-22, 4.2698e-20, 3.7551e-19, 7.3537e-20,\n 2.0725e-20, 1.2237e-18, 3.3660e-19, 2.6266e-23, 1.6640e-18, 1.0016e-21,\n 1.9676e-18, 1.7009e-18, 3.5172e-18, 3.0344e-18, 1.0633e-19, 1.2790e-20,\n 1.6520e-20, 1.1643e-18, 2.9148e-18, 5.7968e-20, 3.5769e-18, 5.0946e-21,\n 4.1490e-19, 1.0424e-17, 5.3204e-22, 7.7211e-18], device='cuda:0')" + "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')" }, "39": { - "step": "tensor(21284.)", - "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([[4.8239e-21, 1.0315e-19, 2.0907e-20, ..., 8.1621e-23, 9.9647e-21,\n 3.2148e-19],\n [4.0766e-20, 5.8082e-21, 1.6573e-21, ..., 4.0396e-20, 5.4935e-23,\n 4.6073e-19],\n [6.5989e-22, 1.6652e-19, 4.2024e-21, ..., 2.7940e-21, 1.0398e-20,\n 2.5881e-19],\n ...,\n [7.9495e-19, 8.5711e-20, 1.3853e-18, ..., 5.5852e-20, 1.7579e-19,\n 1.4122e-17],\n [2.4329e-19, 2.3389e-19, 2.6607e-18, ..., 3.8566e-20, 5.1791e-20,\n 4.3212e-18],\n [7.6080e-19, 1.8240e-17, 2.3021e-16, ..., 5.2884e-18, 1.6416e-19,\n 1.2789e-16]], device='cuda:0')" + "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')" }, "40": { - "step": "tensor(21284.)", - "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([1.2760e-19, 3.1608e-19, 8.5117e-20, 3.1920e-20, 2.6258e-19, 7.3217e-20,\n 2.9308e-20, 9.5638e-20, 2.5318e-20, 6.8606e-20, 2.3218e-19, 5.7597e-20,\n 2.1326e-21, 2.6998e-20, 8.6018e-21, 4.3724e-21, 3.5254e-19, 1.1915e-19,\n 5.9571e-20, 3.1990e-19, 2.7906e-20, 5.3054e-19, 7.3066e-20, 4.0047e-20,\n 1.2705e-19, 2.2551e-19, 1.5784e-21, 6.0992e-21, 1.6149e-19, 2.7518e-19,\n 1.9877e-19, 2.1836e-19, 1.2002e-19, 2.7184e-19, 4.2937e-19, 1.2989e-18,\n 3.6020e-19, 5.4936e-20, 4.1206e-19, 2.7141e-19, 1.1017e-19, 1.6991e-19,\n 2.6035e-19, 1.3829e-19, 2.6900e-19, 1.1211e-19, 2.0515e-19, 8.7436e-20,\n 5.7877e-21, 7.5377e-19, 9.8712e-20, 1.5557e-19, 1.1869e-20, 7.2008e-19,\n 2.0201e-20, 1.2519e-19, 3.0636e-19, 6.4295e-21, 3.7689e-19, 7.7510e-20,\n 1.2585e-18, 7.5896e-19, 2.6254e-20, 8.7769e-20, 5.6996e-21, 2.0725e-20,\n 7.8819e-20, 4.0894e-20, 1.1031e-19, 4.7680e-19, 1.9400e-19, 9.9393e-20,\n 5.2072e-20, 1.4494e-19, 5.1961e-19, 7.6962e-19, 1.7668e-19, 4.3357e-20,\n 2.8868e-19, 2.9315e-20, 5.7501e-19, 6.2035e-20, 1.5317e-20, 1.0541e-20,\n 1.2585e-20, 4.6696e-20, 1.1120e-18, 1.3595e-19, 4.8546e-21, 2.7777e-19,\n 3.5333e-21, 1.6597e-19, 2.2173e-20, 1.4752e-18, 2.5439e-21, 1.0936e-19,\n 2.3814e-19, 1.7557e-21, 5.2207e-21, 1.4163e-19, 4.1404e-19, 1.1716e-21,\n 6.1904e-19, 3.7770e-20, 7.0222e-20, 6.6476e-20, 9.3451e-20, 5.7146e-19,\n 3.8197e-19, 2.1168e-19, 5.2412e-20, 1.4724e-19, 9.2718e-20, 1.1202e-19,\n 6.2338e-20, 1.4264e-19, 1.1943e-19, 2.1406e-19, 3.4675e-19, 1.4328e-18,\n 1.0094e-20, 1.3404e-19, 2.4435e-19, 7.0978e-21, 4.2609e-21, 5.3194e-21,\n 1.9100e-19, 8.2748e-19, 8.0748e-21, 1.4035e-20, 7.4383e-20, 6.5544e-19,\n 7.9388e-22, 4.2255e-19, 1.0019e-19, 1.7083e-19, 3.6569e-19, 1.1894e-20,\n 1.0842e-19, 2.1778e-21, 1.2164e-19, 1.9181e-20, 2.0730e-20, 3.0739e-20,\n 4.9883e-21, 2.4012e-22, 6.9183e-20, 4.3517e-20, 1.5656e-20, 2.2179e-20,\n 4.1162e-21, 1.8324e-20, 1.3285e-19, 2.4817e-19, 1.7965e-20, 1.7576e-20,\n 5.3398e-20, 1.9066e-20, 9.0255e-20, 1.4066e-20, 3.8269e-19, 2.7304e-21,\n 3.6196e-21, 5.3177e-19, 1.8811e-19, 1.5284e-18, 1.5835e-18, 2.3064e-19,\n 3.1545e-19, 1.7780e-18, 5.0970e-20, 6.2732e-20, 1.0053e-19, 6.1696e-19,\n 4.9934e-20, 4.8173e-19, 1.1308e-20, 2.3228e-19, 2.1171e-19, 2.6379e-18,\n 6.1243e-19, 1.0069e-18, 7.0491e-20, 1.5365e-19, 1.9081e-18, 3.8205e-20,\n 4.9382e-22, 1.7879e-21, 6.4309e-19, 1.6017e-19, 6.1425e-19, 2.8680e-19,\n 1.5405e-19, 2.5282e-20, 9.7066e-19, 9.5307e-19, 4.9758e-19, 1.4956e-19,\n 5.3711e-19, 3.5583e-21, 1.4896e-19, 3.5311e-19, 2.1750e-21, 4.7469e-21,\n 8.5623e-20, 7.1551e-19, 4.9952e-19, 7.7644e-20, 1.4544e-20, 6.8555e-22,\n 8.7840e-19, 4.3733e-19, 1.1820e-20, 7.6707e-21, 2.4504e-19, 7.8809e-20,\n 1.0654e-19, 4.4887e-19, 4.0133e-19, 9.7667e-19, 1.4627e-20, 1.6633e-19,\n 3.8169e-21, 1.1318e-18, 6.8230e-20, 3.8855e-20, 1.9787e-20, 1.3824e-20,\n 1.0083e-19, 1.6146e-20, 2.1055e-19, 2.3407e-20, 7.4879e-21, 1.2237e-19,\n 6.4663e-20, 1.0401e-20, 1.0695e-20, 5.9154e-20, 4.9135e-20, 2.1136e-20,\n 1.7530e-20, 2.3558e-20, 2.7809e-20, 3.8119e-20, 1.2496e-19, 2.8944e-20,\n 3.0570e-20, 1.8001e-20, 1.0776e-20, 5.0321e-20, 1.3645e-19, 1.8035e-19,\n 1.5080e-20, 1.1838e-19, 2.0485e-20, 1.5568e-19, 3.0645e-34, 2.5511e-35,\n 7.1459e-35, 1.5796e-35, 1.5428e-36, 4.2915e-35, 2.1601e-35, 6.5708e-35,\n 4.9692e-37, 4.7626e-35, 9.9727e-36, 2.0482e-35, 9.1800e-35, 3.5339e-36,\n 2.0818e-36, 5.4542e-36, 1.3343e-35, 4.4864e-36, 1.6631e-34, 3.2647e-35,\n 3.1323e-36, 2.8768e-37, 7.4713e-36, 5.0512e-35, 4.8375e-36, 1.1257e-36,\n 1.7527e-36, 1.2389e-35, 2.6516e-37, 8.9582e-36, 3.9997e-35, 3.7772e-35,\n 2.3242e-35, 1.2970e-36, 7.6062e-36, 8.8306e-36, 2.6965e-35, 5.9626e-36,\n 1.8975e-35, 4.2054e-35, 2.6685e-35, 1.6429e-35, 5.6935e-36, 4.7495e-35,\n 4.6957e-35, 7.8680e-36, 3.6590e-36, 7.3682e-36, 4.8029e-35, 1.4474e-35,\n 5.4638e-36, 1.2779e-36, 1.7802e-35, 1.9846e-35, 4.6794e-35, 4.9351e-35,\n 3.5195e-35, 2.0477e-34, 2.6712e-35, 2.0591e-35, 7.3069e-35, 3.7796e-35,\n 4.2989e-35, 4.1732e-35, 1.0366e-35, 1.3278e-35, 2.1477e-36, 7.1449e-36,\n 2.6118e-35, 7.0615e-36, 3.5764e-35, 2.3422e-36, 6.0074e-37, 2.2337e-35,\n 5.8290e-36, 1.9332e-35, 1.2353e-35, 1.1230e-35, 3.4369e-35, 1.4448e-35,\n 1.3784e-35, 1.0880e-34, 6.2787e-36, 8.9332e-36, 1.8780e-35, 1.1818e-36,\n 1.1033e-35, 3.1374e-37, 1.0629e-35, 8.9168e-37, 4.3559e-36, 2.6717e-35,\n 1.4344e-36, 1.1720e-35, 3.7205e-36, 9.9590e-36, 1.5733e-35, 7.6283e-36,\n 6.0963e-36, 1.3086e-35, 1.2578e-35, 2.7630e-36, 4.4725e-35, 4.9455e-35,\n 7.0765e-36, 1.7968e-35, 1.2296e-35, 5.3248e-35, 2.1217e-35, 7.1266e-36,\n 2.0833e-36, 1.1394e-35, 2.7758e-35, 2.7110e-36, 1.2078e-35, 8.3394e-37,\n 4.1981e-36, 1.1774e-35, 1.9391e-36, 6.1586e-36, 7.6848e-36, 9.5662e-36,\n 3.4760e-35, 1.3913e-35, 5.5659e-35, 7.8649e-35, 4.0377e-35, 1.1181e-34,\n 2.1752e-36, 2.1641e-36, 2.1716e-35, 7.9934e-36, 8.3970e-37, 2.1057e-37,\n 7.8910e-36, 2.0153e-36, 9.3102e-36, 5.8758e-36, 3.0637e-35, 3.9924e-36,\n 7.4375e-36, 8.1385e-35, 7.9911e-36, 1.3960e-36, 3.5545e-36, 1.0958e-36,\n 1.0794e-36, 1.3273e-35, 2.2746e-35, 2.5208e-35, 1.4147e-36, 3.6018e-35,\n 3.7267e-36, 3.4747e-36, 8.6933e-35, 1.7317e-35, 2.8047e-36, 1.1737e-36,\n 7.6673e-36, 3.3306e-36, 1.7199e-35, 5.4388e-37, 1.1923e-35, 2.7350e-36,\n 4.4931e-35, 5.5840e-36, 1.0433e-35, 3.9324e-35, 1.5799e-35, 7.7734e-36,\n 3.2714e-35, 1.2506e-35, 2.1131e-35, 2.0392e-36, 1.9695e-34, 2.7228e-36,\n 4.1693e-35, 6.5395e-36, 8.7152e-36, 2.2536e-35, 8.0744e-37, 5.1446e-36,\n 2.1800e-35, 1.4831e-35, 1.3289e-35, 1.1923e-35, 2.2950e-35, 4.3396e-36,\n 1.6636e-35, 1.7946e-35, 1.8227e-35, 1.7046e-35, 3.7406e-35, 6.3552e-35,\n 1.0175e-35, 2.8044e-36, 3.1866e-35, 8.8485e-36, 7.6432e-37, 2.0672e-36,\n 2.4184e-35, 5.5090e-35, 2.2218e-35, 9.7207e-36, 2.6797e-36, 4.6866e-37,\n 1.8677e-35, 3.1652e-35, 4.7373e-36, 1.8425e-36, 5.2329e-36, 7.1708e-36,\n 9.0157e-36, 5.1982e-36, 1.8707e-35, 2.6255e-35, 5.3787e-35, 5.3582e-36,\n 5.6281e-36, 1.5110e-35, 1.9110e-35, 1.2074e-35, 1.1718e-36, 3.6038e-35,\n 1.1537e-36, 5.2699e-36, 1.9196e-35, 1.8394e-36, 2.3196e-35, 3.5497e-35,\n 5.5326e-36, 3.5826e-35, 3.5267e-35, 2.0705e-35, 1.7007e-35, 2.2427e-35,\n 5.2856e-37, 1.5503e-35, 4.7953e-36, 1.3104e-35, 1.9605e-36, 6.5460e-36,\n 1.2550e-36, 8.0670e-35, 1.2953e-35, 2.0799e-35, 2.4974e-36, 7.4675e-36,\n 2.1795e-35, 8.3723e-35, 1.9970e-35, 3.5695e-35, 4.3028e-35, 3.3407e-35,\n 5.3616e-36, 7.1206e-36, 1.2630e-15, 4.7014e-17, 2.7686e-16, 1.8194e-17,\n 7.1594e-17, 2.3837e-17, 4.8783e-18, 1.0009e-16, 5.5263e-17, 2.2751e-17,\n 7.8999e-16, 1.5407e-17, 4.6019e-17, 6.1459e-16, 9.0411e-17, 1.2268e-16,\n 2.0401e-17, 7.5671e-19, 5.1808e-17, 7.2803e-17, 1.5325e-16, 1.1093e-16,\n 6.6841e-17, 1.6519e-17, 1.6243e-18, 1.3308e-17, 3.8109e-18, 9.5675e-17,\n 8.6590e-16, 2.4771e-17, 2.7679e-16, 6.5360e-17, 6.2056e-17, 1.5693e-16,\n 4.6021e-19, 3.7879e-16, 1.9394e-16, 6.1053e-17, 3.9656e-16, 9.5265e-18,\n 1.5560e-17, 3.0049e-17, 2.8222e-17, 7.6631e-17, 4.3617e-17, 1.3105e-18,\n 9.3218e-18, 9.9599e-17, 1.6460e-17, 1.9070e-16, 6.5383e-16, 1.3771e-16,\n 1.4474e-17, 3.1564e-16, 4.0398e-17, 2.1991e-16, 2.0742e-16, 3.3642e-17,\n 2.9831e-16, 2.5046e-17, 1.0782e-17, 1.9223e-19, 3.1188e-18, 5.1517e-18,\n 1.2820e-16, 7.7118e-17, 5.8860e-17, 2.4539e-17, 7.8241e-17, 1.1614e-17,\n 1.6057e-16, 3.4677e-16, 2.1153e-16, 5.3238e-18, 9.1767e-18, 2.7772e-16,\n 1.9991e-16, 5.3611e-16, 2.4098e-16, 1.7798e-17, 9.9948e-18, 2.0629e-17,\n 9.8505e-17, 3.7593e-17, 9.3213e-17, 1.7177e-17, 1.1321e-16, 1.3112e-17,\n 3.3397e-17, 3.6311e-16, 5.0487e-16, 4.1118e-17, 1.7852e-16, 1.2235e-16,\n 7.2286e-17, 1.8027e-16, 1.7728e-16, 1.1911e-17, 7.4599e-17, 6.0456e-17,\n 5.0276e-17, 4.8237e-16, 6.5906e-17, 4.0087e-16, 2.0744e-16, 3.8439e-17,\n 7.9983e-17, 9.3810e-18, 1.9821e-17, 1.9237e-18, 2.1490e-16, 2.2356e-17,\n 2.0986e-16, 2.9318e-17, 5.0881e-16, 1.2901e-16, 8.3145e-17, 4.4193e-16,\n 7.1067e-17, 4.3658e-17, 5.1688e-18, 6.9253e-17, 1.0391e-16, 2.5264e-17,\n 2.5539e-17, 1.0652e-16, 2.1176e-17, 1.1080e-17, 2.5667e-16, 1.0417e-17,\n 1.4527e-17, 1.6572e-17, 1.8031e-17, 7.1966e-17, 1.0014e-16, 4.5077e-17,\n 4.4535e-17, 2.9324e-18, 9.4031e-17, 6.5624e-17, 1.1372e-17, 5.7735e-17,\n 2.0237e-17, 1.2658e-17, 1.3086e-16, 3.0261e-17, 3.2262e-19, 8.7175e-17,\n 6.5499e-17, 2.0235e-17, 1.8076e-16, 1.0251e-16, 3.3931e-17, 1.3944e-17,\n 3.4209e-16, 1.2179e-17, 1.2404e-17, 1.3920e-16, 5.1855e-17, 1.2575e-16,\n 1.2064e-16, 4.3106e-17, 1.8561e-16, 1.1145e-16, 9.2672e-18, 1.7818e-17,\n 2.3712e-19, 2.0671e-17, 1.3103e-16, 1.8981e-19, 6.7557e-16, 3.4577e-19,\n 4.0814e-17, 1.6268e-17, 3.6986e-16, 1.4279e-16, 6.3507e-17, 1.3382e-16,\n 3.4339e-16, 5.2290e-17, 5.9108e-18, 2.4484e-17, 4.0695e-17, 2.1865e-17,\n 7.1767e-18, 1.5696e-17, 2.8092e-17, 6.3922e-18, 2.9879e-16, 1.7965e-18,\n 6.5053e-17, 1.3462e-17, 4.7634e-16, 2.3320e-16, 9.7489e-18, 3.2948e-17,\n 1.8293e-16, 9.1225e-18, 9.1534e-18, 2.1609e-16, 5.1243e-17, 1.0573e-16,\n 3.1219e-17, 1.4381e-16, 1.6052e-17, 3.6459e-17, 1.7158e-16, 8.0536e-18,\n 1.4026e-16, 6.6022e-17, 2.3277e-17, 1.0521e-16, 5.0698e-16, 1.0313e-17,\n 4.7129e-17, 8.7693e-17, 4.0136e-19, 7.3025e-17, 6.6878e-17, 1.5823e-16,\n 6.3625e-18, 1.0899e-16, 2.3036e-17, 3.6291e-17, 1.0268e-16, 4.0809e-19,\n 1.6378e-16, 2.7501e-17, 2.2793e-16, 7.7160e-17, 1.0629e-16, 1.9222e-17,\n 5.0097e-17, 1.8096e-17, 3.1377e-16, 6.2683e-17, 4.2680e-17, 2.9859e-16,\n 2.6935e-16, 2.5757e-16, 5.4552e-17, 2.6109e-16, 9.4829e-18, 1.1140e-18,\n 2.4951e-17, 1.1771e-17, 7.2454e-18, 1.6345e-17, 2.0531e-16, 6.7744e-18,\n 1.7782e-17, 9.3072e-17, 3.4366e-17, 1.2737e-17, 5.3273e-18, 2.3973e-16],\n device='cuda:0')" + "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')" }, "41": { - "step": "tensor(21284.)", - "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([[2.5461e-17, 1.8633e-17, 6.0063e-18, ..., 9.6079e-17, 3.3172e-18,\n 3.7436e-17],\n [1.2010e-17, 4.9155e-18, 2.8307e-18, ..., 2.2587e-17, 1.0730e-18,\n 1.9614e-17],\n [2.7883e-18, 1.6481e-19, 6.9055e-19, ..., 8.8607e-19, 2.2142e-19,\n 9.2772e-19],\n ...,\n [1.0190e-17, 7.3594e-19, 2.7462e-18, ..., 4.5303e-18, 6.3329e-19,\n 2.3098e-18],\n [5.2677e-18, 8.7536e-19, 1.3082e-18, ..., 4.4444e-18, 3.3122e-19,\n 1.0727e-18],\n [1.2785e-17, 3.8317e-18, 2.9610e-18, ..., 1.6572e-17, 1.1207e-18,\n 1.1924e-17]], device='cuda:0')" + "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')" }, "42": { - "step": "tensor(21284.)", - "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([1.1315e-15, 2.8879e-16, 2.5530e-18, 1.1127e-15, 3.4365e-16, 2.9604e-17,\n 4.0294e-17, 6.2043e-17, 7.0252e-18, 1.5338e-16, 1.9719e-16, 2.2541e-16,\n 3.3576e-16, 6.7131e-17, 3.7066e-16, 1.6679e-16, 7.0844e-16, 1.0718e-16,\n 1.0903e-17, 8.6462e-16, 1.0984e-16, 7.5448e-17, 2.2653e-16, 9.0078e-16,\n 4.9228e-17, 1.2286e-15, 2.1992e-16, 2.2877e-16, 1.1964e-17, 7.1174e-17,\n 1.0243e-15, 2.5908e-17, 6.0027e-16, 1.9590e-17, 3.9670e-16, 8.1443e-18,\n 1.5290e-16, 3.2425e-16, 1.9930e-16, 1.8827e-16, 1.2555e-17, 2.6378e-16,\n 9.0766e-16, 2.4408e-17, 4.2669e-16, 3.1709e-17, 2.4344e-17, 9.1071e-18,\n 2.7631e-16, 8.6291e-17, 1.0523e-16, 5.8068e-17, 4.4569e-17, 1.1991e-16,\n 1.8366e-15, 1.9279e-17, 5.3436e-17, 2.6965e-17, 5.8036e-18, 4.6148e-16,\n 1.4358e-16, 5.2129e-18, 3.0133e-16, 1.2405e-15, 5.0816e-17, 7.7133e-17,\n 5.5547e-16, 8.6912e-18, 1.7355e-15, 1.7763e-17, 2.4916e-16, 1.7511e-16,\n 1.1853e-16, 6.1943e-18, 4.1769e-17, 5.4053e-17, 2.9298e-16, 9.9987e-17,\n 1.5983e-16, 3.0847e-16, 3.3453e-17, 3.1971e-17, 2.8719e-15, 9.6315e-17,\n 6.4019e-18, 1.1983e-17, 2.4516e-17, 1.2874e-17, 2.5116e-16, 1.9782e-17,\n 6.6809e-18, 5.0111e-16, 3.3903e-17, 9.5738e-16, 2.2721e-16, 3.0266e-17,\n 7.0200e-16, 2.5369e-16, 1.1598e-15, 8.6733e-18, 1.3190e-16, 1.1755e-15,\n 1.5437e-15, 2.5810e-15, 3.2054e-17, 1.1241e-16, 6.7279e-18, 4.1074e-16,\n 2.0089e-16, 1.0161e-17, 1.8019e-17, 1.9532e-16, 2.2036e-17, 3.3882e-17,\n 2.9778e-17, 1.6029e-16, 1.6698e-17, 1.2997e-16, 1.4637e-17, 1.1697e-16,\n 9.4565e-16, 5.4999e-18, 2.3289e-15, 8.7617e-17, 1.6929e-15, 1.5549e-17,\n 5.5442e-16, 6.8073e-16, 7.8736e-16, 6.5367e-17, 8.3576e-16, 6.3571e-16,\n 1.0183e-16, 7.6630e-18, 1.0184e-17, 9.0021e-16, 7.8651e-16, 5.4753e-16,\n 9.7009e-17, 1.5182e-17, 1.6493e-16, 1.0864e-15, 2.8200e-16, 4.2175e-16,\n 2.2054e-16, 4.9188e-18, 2.7691e-16, 1.9903e-16, 1.8769e-17, 8.0435e-18,\n 4.3667e-18, 7.2830e-16, 2.8069e-17, 8.8001e-17, 5.7279e-16, 1.2038e-15,\n 3.8534e-16, 8.7843e-16, 1.0083e-16, 7.4381e-16, 7.9227e-16, 6.7547e-16,\n 3.0913e-16, 1.0674e-15, 5.4199e-18, 1.3449e-15, 1.4766e-16, 1.7024e-16,\n 9.7243e-17, 1.3549e-17, 2.7661e-17, 6.3666e-18, 1.4196e-16, 4.1318e-17,\n 2.0030e-16, 6.3086e-16, 3.3827e-16, 1.1112e-17, 3.1996e-16, 7.1307e-18,\n 1.4774e-17, 4.2609e-16, 7.5795e-18, 1.7610e-16, 1.0938e-17, 2.7412e-16,\n 4.5454e-16, 1.2587e-16, 2.0407e-16, 4.3904e-16, 7.9514e-16, 4.8959e-16,\n 4.8051e-16, 8.7743e-18, 6.6875e-17, 1.0054e-17, 1.3873e-17, 1.1347e-16,\n 5.1337e-18, 1.5019e-16, 4.5874e-16, 5.1927e-16, 6.7246e-17, 1.9894e-16,\n 6.3372e-17, 1.0447e-15, 1.0974e-17, 7.5862e-16, 3.0818e-17, 7.0531e-17,\n 2.1648e-17, 1.8293e-16, 1.9868e-17, 1.0628e-16, 7.9518e-16, 1.2667e-16,\n 6.4963e-17, 9.0443e-18, 2.1217e-17, 6.1432e-17, 3.6039e-16, 1.5992e-17,\n 7.3415e-18, 1.6557e-16, 4.1972e-16, 2.2762e-15, 2.5131e-16, 5.3052e-17,\n 4.4109e-17, 1.2704e-15, 4.4409e-16, 8.4686e-17, 1.4107e-15, 4.1522e-17,\n 6.3050e-17, 1.5319e-16, 6.5208e-17, 3.8002e-16, 6.6417e-17, 1.1535e-18,\n 6.0656e-16, 3.8962e-17, 5.5234e-16, 7.0316e-17, 1.0854e-17, 2.0298e-17,\n 6.7897e-16, 2.5908e-17, 9.8259e-16, 1.1449e-15, 6.4033e-16, 1.1637e-16,\n 2.4214e-16, 2.0040e-17, 2.9572e-17, 2.0632e-16], device='cuda:0')" + "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')" }, "43": { - "step": "tensor(21284.)", - "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],\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([[2.2463e-16, 1.4118e-15, 1.7642e-16, ..., 3.6542e-15, 3.4432e-15,\n 1.5922e-15],\n [1.4987e-17, 9.6426e-17, 1.1402e-17, ..., 2.5405e-16, 2.4760e-16,\n 1.0666e-16],\n [1.0459e-17, 6.0679e-17, 8.1939e-18, ..., 1.5804e-16, 1.4054e-16,\n 7.3609e-17],\n [1.6403e-17, 1.0989e-16, 1.3049e-17, ..., 2.8141e-16, 2.7285e-16,\n 1.1656e-16],\n [1.6180e-17, 9.2571e-17, 1.2858e-17, ..., 2.4019e-16, 2.1231e-16,\n 1.1382e-16]], device='cuda:0')" - }, - "44": { - "step": "tensor(21284.)", - "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45],\n device='cuda:0')", - "exp_avg_sq": "tensor([8.0293e-14, 5.6698e-15, 3.3274e-15, 6.3666e-15, 5.0087e-15],\n device='cuda:0')" - }, - "45": { - "step": "tensor(21284.)", - "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],\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([[2.2500e-16, 1.4125e-15, 1.7669e-16, ..., 3.6567e-15, 3.4432e-15,\n 1.5946e-15],\n [1.4994e-17, 9.6439e-17, 1.1407e-17, ..., 2.5409e-16, 2.4760e-16,\n 1.0671e-16],\n [1.0495e-17, 6.0740e-17, 8.2199e-18, ..., 1.5827e-16, 1.4054e-16,\n 7.3838e-17],\n [1.6411e-17, 1.0990e-16, 1.3055e-17, ..., 2.8147e-16, 2.7285e-16,\n 1.1661e-16],\n [1.6241e-17, 9.2675e-17, 1.2903e-17, ..., 2.4059e-16, 2.1231e-16,\n 1.1422e-16]], device='cuda:0')" - }, - "46": { - "step": "tensor(21284.)", - "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45],\n device='cuda:0')", - "exp_avg_sq": "tensor([8.0299e-14, 5.6699e-15, 3.3280e-15, 6.3668e-15, 5.0098e-15],\n device='cuda:0')" - }, - "47": { - "step": "tensor(21284.)", - "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],\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([[2.2463e-16, 1.4118e-15, 1.7642e-16, ..., 3.6542e-15, 3.4432e-15,\n 1.5922e-15],\n [1.4987e-17, 9.6426e-17, 1.1402e-17, ..., 2.5405e-16, 2.4760e-16,\n 1.0666e-16],\n [1.0459e-17, 6.0679e-17, 8.1939e-18, ..., 1.5804e-16, 1.4054e-16,\n 7.3609e-17],\n [1.6403e-17, 1.0989e-16, 1.3049e-17, ..., 2.8141e-16, 2.7285e-16,\n 1.1656e-16],\n [1.6180e-17, 9.2571e-17, 1.2858e-17, ..., 2.4019e-16, 2.1231e-16,\n 1.1382e-16]], device='cuda:0')" - }, - "48": { - "step": "tensor(21284.)", - "exp_avg": "tensor([-5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45, 5.6052e-45],\n device='cuda:0')", - "exp_avg_sq": "tensor([8.0293e-14, 5.6698e-15, 3.3274e-15, 6.3666e-15, 5.0087e-15],\n device='cuda:0')" - }, - "6": { - "step": "tensor(7512.)", - "exp_avg": "tensor([[ 4.9854e-07, 2.1596e-06, -9.6562e-06, ..., 5.9631e-09,\n 6.4181e-06, 2.4993e-06],\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 [-6.8664e-44, -1.5554e-43, 2.3262e-43, ..., 2.2701e-43,\n 3.3351e-43, 8.5479e-44],\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([[1.4335e-08, 8.3232e-09, 1.6180e-09, ..., 9.1739e-10, 1.0913e-09,\n 1.2542e-09],\n [1.1822e-13, 4.4540e-14, 4.6942e-15, ..., 9.7048e-15, 3.2589e-14,\n 4.8714e-15],\n [8.9143e-13, 1.0339e-13, 5.6054e-14, ..., 1.7149e-14, 1.2477e-13,\n 9.5696e-15],\n ...,\n [8.1242e-12, 3.4987e-11, 6.1224e-12, ..., 4.3114e-12, 4.5763e-12,\n 3.5401e-12],\n [2.0966e-12, 6.7308e-13, 1.1468e-13, ..., 8.5360e-14, 6.7355e-13,\n 1.8351e-14],\n [6.8850e-13, 8.2457e-13, 3.0293e-14, ..., 3.5408e-14, 1.1645e-13,\n 8.7478e-15]], device='cuda:0')" - }, - "7": { - "step": "tensor(7512.)", - "exp_avg": "tensor([1.9908e-05, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45, 8.9949e-42,\n 5.6052e-45], device='cuda:0')", - "exp_avg_sq": "tensor([1.7078e-06, 1.8833e-11, 1.0545e-10, ..., 7.7286e-09, 3.1394e-10,\n 1.9187e-10], device='cuda:0')" + "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')" }, "8": { - "step": "tensor(7512.)", - "exp_avg": "tensor([[-2.8770e-07, -5.6052e-45, 5.6052e-45, ..., -5.6052e-45,\n -5.3249e-44, -5.6052e-45],\n [ 9.8273e-07, -5.6052e-45, 5.6052e-45, ..., -5.6052e-45,\n 4.2039e-44, 5.6052e-45],\n [ 1.0263e-07, -5.6052e-45, -5.6052e-45, ..., -5.6052e-45,\n 1.8217e-43, -5.6052e-45],\n ...,\n [-1.3544e-06, 5.6052e-45, -5.6052e-45, ..., -5.6052e-45,\n -1.6816e-44, -5.6052e-45],\n [-3.0603e-08, -5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n -9.3887e-44, -5.6052e-45],\n [-2.6994e-07, 5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 1.6816e-44, -5.6052e-45]], device='cuda:0')", - "exp_avg_sq": "tensor([[6.3412e-11, 6.9678e-14, 1.3855e-13, ..., 1.8502e-12, 4.0079e-13,\n 2.0643e-13],\n [6.7597e-11, 8.2975e-14, 1.5516e-13, ..., 2.2718e-12, 4.0440e-13,\n 3.4972e-13],\n [6.4600e-11, 7.0760e-14, 1.4303e-13, ..., 4.1788e-12, 4.2178e-13,\n 4.5041e-13],\n ...,\n [1.2245e-10, 1.2052e-13, 2.8781e-13, ..., 3.0939e-12, 7.2467e-13,\n 6.5419e-13],\n [7.3906e-11, 1.7032e-13, 2.7577e-13, ..., 2.0130e-12, 9.9125e-13,\n 4.9023e-13],\n [1.1118e-10, 6.4154e-14, 1.4427e-13, ..., 2.6487e-12, 4.5594e-13,\n 7.6545e-13]], device='cuda:0')" + "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')" }, "9": { - "step": "tensor(7512.)", - "exp_avg": "tensor([[ 2.6317e-05, -1.6313e-05, 4.6890e-07, ..., 1.7440e-06,\n 9.6505e-06, 1.5082e-06],\n [ 2.6435e-08, -3.9305e-08, -8.2584e-09, ..., 3.7843e-08,\n -1.9831e-08, -1.1462e-08],\n [-6.2581e-06, 1.5904e-05, 1.7100e-05, ..., -5.3792e-06,\n 6.4331e-06, 7.0246e-06],\n ...,\n [-2.4415e-07, 6.5924e-07, -4.2370e-07, ..., -3.1356e-06,\n -1.1131e-06, -3.1441e-06],\n [ 5.6052e-45, -5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, -5.6052e-45],\n [ 3.2730e-06, 5.0244e-06, 3.6937e-06, ..., 1.3829e-06,\n 1.4456e-06, -1.2209e-06]], device='cuda:0')", - "exp_avg_sq": "tensor([[6.1181e-09, 2.5779e-09, 4.6190e-10, ..., 4.6640e-10, 7.2866e-10,\n 2.8604e-10],\n [3.6632e-10, 2.5726e-10, 3.3101e-11, ..., 4.6296e-11, 4.2138e-11,\n 3.9822e-11],\n [2.6983e-09, 2.6513e-09, 6.1308e-10, ..., 2.9139e-10, 9.0128e-10,\n 2.5469e-10],\n ...,\n [7.8327e-10, 2.7699e-09, 2.4455e-10, ..., 3.1425e-10, 3.3213e-10,\n 2.6594e-10],\n [6.2069e-10, 4.1678e-10, 4.7345e-11, ..., 5.5707e-11, 6.7778e-11,\n 3.7032e-11],\n [4.5264e-09, 2.6453e-09, 5.9442e-10, ..., 1.2929e-09, 6.6125e-10,\n 5.6445e-10]], device='cuda:0')" + "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')" }, "10": { - "step": "tensor(7512.)", - "exp_avg": "tensor([ 2.3891e-04, -3.8899e-07, 1.0684e-04, ..., -5.5825e-05,\n 5.6052e-45, 4.8544e-05], device='cuda:0')", - "exp_avg_sq": "tensor([8.3291e-07, 8.4591e-08, 5.8645e-07, ..., 4.8595e-07, 8.3363e-08,\n 1.5053e-06], device='cuda:0')" + "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')" }, "11": { - "step": "tensor(7512.)", - "exp_avg": "tensor([[-6.0484e-08, -3.5157e-09, -1.6698e-06, ..., -5.4709e-07,\n -5.6052e-45, 3.0677e-08],\n [ 2.6674e-07, 1.3797e-08, -1.5694e-06, ..., -4.6201e-07,\n -5.6052e-45, -2.5403e-08],\n [-9.0853e-07, 6.9575e-09, -2.0038e-06, ..., -3.9492e-07,\n -5.6052e-45, 4.4455e-07],\n ...,\n [-7.3567e-07, 2.1666e-08, -2.9420e-06, ..., 2.0468e-07,\n 5.6052e-45, 9.2520e-09],\n [-3.5218e-07, 4.0375e-08, 6.2799e-07, ..., 3.1335e-07,\n -5.6052e-45, -4.2172e-07],\n [ 1.9479e-06, 2.3553e-09, 3.3146e-06, ..., -5.4778e-07,\n -5.6052e-45, -8.3423e-09]], device='cuda:0')", - "exp_avg_sq": "tensor([[6.0021e-11, 4.4199e-12, 1.8480e-11, ..., 8.7259e-12, 4.3408e-12,\n 3.1252e-11],\n [5.0513e-11, 7.9713e-12, 2.6129e-11, ..., 2.4337e-11, 7.8321e-12,\n 2.7669e-11],\n [9.1091e-11, 8.9976e-12, 2.9579e-11, ..., 2.1574e-11, 8.1540e-12,\n 2.7565e-11],\n ...,\n [1.4373e-10, 8.1480e-12, 2.9111e-11, ..., 2.5018e-11, 8.6969e-12,\n 2.4706e-11],\n [7.9418e-11, 1.0341e-11, 2.7521e-11, ..., 1.6803e-11, 9.8605e-12,\n 3.5564e-11],\n [6.4596e-11, 9.6854e-12, 2.9664e-11, ..., 3.0424e-11, 8.5014e-12,\n 3.3017e-11]], device='cuda:0')" + "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')" }, "12": { - "step": "tensor(10016.)", - "exp_avg": "tensor([[ 1.0733e-06, 5.9301e-07, 1.1084e-06, ..., -1.1260e-06,\n -7.9756e-07, 4.1220e-07],\n [ 5.6052e-45, -5.6052e-45, 5.6052e-45, ..., 5.6052e-45,\n 5.6052e-45, 5.6052e-45],\n [ 3.9096e-06, 1.3317e-06, 6.0891e-06, ..., 1.4191e-06,\n 4.4139e-06, 9.3660e-07],\n ...,\n [-3.6163e-06, -5.3090e-07, 1.8161e-06, ..., -2.2841e-06,\n 2.0056e-07, 2.6083e-06],\n [-4.1390e-06, 6.1189e-06, -8.3921e-07, ..., -6.2971e-07,\n -1.9904e-06, 1.1207e-06],\n [ 1.5186e-06, 4.9709e-06, 1.9487e-06, ..., -3.3802e-06,\n -1.6063e-06, -5.4591e-07]], device='cuda:0')", - "exp_avg_sq": "tensor([[2.2621e-10, 6.0628e-10, 8.1398e-11, ..., 1.7367e-10, 9.0598e-11,\n 7.0216e-11],\n [4.9119e-14, 1.3428e-13, 1.4050e-18, ..., 7.4577e-15, 1.4780e-14,\n 8.0864e-15],\n [1.0850e-09, 3.9958e-10, 7.3343e-11, ..., 1.1709e-10, 1.5914e-10,\n 4.7252e-11],\n ...,\n [1.3515e-09, 5.0402e-10, 6.0527e-11, ..., 9.8959e-11, 2.8179e-10,\n 4.7911e-11],\n [7.5568e-10, 7.8101e-10, 1.3478e-10, ..., 9.2007e-11, 2.8693e-10,\n 5.4235e-11],\n [1.2033e-09, 1.0247e-09, 1.1591e-10, ..., 1.0771e-10, 3.0513e-10,\n 1.4318e-10]], device='cuda:0')" + "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')" }, "13": { - "step": "tensor(10016.)", - "exp_avg": "tensor([-2.7623e-05, 5.6052e-45, 5.5472e-05, ..., -3.8822e-05,\n -7.6574e-05, 5.0669e-05], device='cuda:0')", - "exp_avg_sq": "tensor([1.7265e-07, 6.6402e-11, 1.3076e-07, ..., 1.4914e-07, 1.5993e-07,\n 3.9732e-07], device='cuda:0')" - }, - "14": { - "step": "tensor(10016.)", - "exp_avg": "tensor([[ 1.4559e-07, -5.6052e-45, 5.2317e-07, ..., -1.1216e-07,\n -1.7687e-08, 6.0074e-08],\n [ 1.6496e-07, -5.6052e-45, -7.7360e-07, ..., -1.5523e-08,\n -4.3459e-08, 4.4930e-08],\n [ 1.7335e-07, -5.6052e-45, 5.2840e-08, ..., 5.0578e-08,\n 2.5160e-08, 1.0040e-07],\n ...,\n [ 3.5437e-08, -5.6052e-45, 4.6285e-08, ..., 7.7685e-09,\n 8.6288e-08, -5.1922e-08],\n [-1.2871e-07, 5.6052e-45, -1.1390e-06, ..., 1.6663e-07,\n -6.7402e-08, -2.6010e-07],\n [ 4.1622e-07, -5.6052e-45, -6.5756e-07, ..., -1.7113e-07,\n 2.5219e-07, -1.0156e-07]], device='cuda:0')", - "exp_avg_sq": "tensor([[9.5707e-13, 2.2841e-14, 8.6361e-13, ..., 4.9136e-13, 5.5343e-13,\n 6.1365e-13],\n [1.3276e-12, 2.3052e-15, 2.2673e-12, ..., 1.4636e-12, 8.4700e-13,\n 2.5842e-12],\n [2.2500e-12, 3.5452e-14, 3.3538e-12, ..., 9.8571e-13, 1.3532e-12,\n 8.2118e-12],\n ...,\n [1.9485e-12, 3.9681e-14, 3.7247e-12, ..., 2.8779e-13, 1.1903e-12,\n 8.5940e-13],\n [1.7618e-12, 3.9193e-14, 3.8227e-12, ..., 1.3400e-12, 9.9202e-13,\n 2.8894e-12],\n [2.6756e-12, 3.7781e-16, 3.3051e-12, ..., 1.3900e-12, 7.3951e-13,\n 3.1421e-12]], device='cuda:0')" + "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')" } }, "param_groups": [ { "lr": 0.005000500000000001, - "name": "scale_256", + "name": "shared", "betas": [ 0.9, 0.999 @@ -263,13 +238,12 @@ "initial_lr": 0.01, "params": [ 0, - 1, - 2 + 1 ] }, { "lr": 0.005000500000000001, - "name": "scale_512", + "name": "scale_384", "betas": [ 0.9, 0.999 @@ -285,9 +259,9 @@ "decoupled_weight_decay": true, "initial_lr": 0.01, "params": [ + 2, 3, - 4, - 5 + 4 ] }, { @@ -308,9 +282,9 @@ "decoupled_weight_decay": true, "initial_lr": 0.01, "params": [ + 5, 6, - 7, - 8 + 7 ] }, { @@ -331,9 +305,9 @@ "decoupled_weight_decay": true, "initial_lr": 0.01, "params": [ + 8, 9, - 10, - 11 + 10 ] }, { @@ -354,9 +328,9 @@ "decoupled_weight_decay": true, "initial_lr": 0.01, "params": [ + 11, 12, - 13, - 14 + 13 ] }, { @@ -377,6 +351,7 @@ "decoupled_weight_decay": true, "initial_lr": 0.005, "params": [ + 14, 15, 16, 17, @@ -405,12 +380,7 @@ 40, 41, 42, - 43, - 44, - 45, - 46, - 47, - 48 + 43 ] } ] @@ -443,20 +413,26 @@ ] }, "metrics": { - "final_val_acc": 75.094 + "final_val_acc": 81.212 }, "train_config": { "name": "david_training", - "run_id": "20251012_032356", + "run_id": "20251012_041353", "dataset_name": "AbstractPhil/imagenet-clip-features-orderly", - "model_variant": "clip_vit_b16", + "model_variant": "clip_vit_l14", "num_classes": 1000, - "preset": "high_accuracy", + "preset": "clip_vit_l14", "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 + }, "num_epochs": 20, "batch_size": 1024, "learning_rate": 0.01, @@ -473,8 +449,8 @@ "gradient_clip": 5.0, "scheduler_type": "cosine_restarts", "min_lr": 1e-06, - "freeze_strategy": "performance", - "freeze_threshold": 70.0, + "freeze_strategy": "never", + "freeze_threshold": 90.0, "unfreeze_on_plateau": true, "patience": 10, "track_gradients": true,