initial performance: 5872 initial simulated performance: 262
episode: 0 training return: tensor(14.4597, device='cuda:0', grad_fn=<AddBackward0>)
episode: 1 training return: tensor(0.0360, device='cuda:0', grad_fn=<AddBackward0>)
episode: 2 training return: tensor(1.7326e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 3 training return: tensor(6.3769e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 1 test_true_pfm: 1043 test_simulate_pfm tensor(3.4061e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 4 training return: tensor(0.0007, device='cuda:0', grad_fn=<AddBackward0>)
episode: 5 training return: tensor(8.1426e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 6 training return: tensor(1.1282e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 7 training return: tensor(7.6878e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 2 test_true_pfm: -305 test_simulate_pfm tensor(5.9914e-06, device='cuda:0', grad_fn=<DivBackward0>)
episode: 8 training return: tensor(6.2016e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 9 training return: tensor(3.3382e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 10 training return: tensor(2.2780e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 11 training return: tensor(3.7403e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 3 test_true_pfm: -577 test_simulate_pfm tensor(1.0667e-09, device='cuda:0', grad_fn=<DivBackward0>)
episode: 12 training return: tensor(7.3945e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 13 training return: tensor(7.3300e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 14 training return: tensor(5.9970e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 15 training return: tensor(4.3886e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 4 test_true_pfm: -562 test_simulate_pfm tensor(3.5318e-11, device='cuda:0', grad_fn=<DivBackward0>)
episode: 16 training return: tensor(4.3277e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 17 training return: tensor(3.7231e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 18 training return: tensor(3.7592e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 19 training return: tensor(1.5247e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 5 test_true_pfm: -640 test_simulate_pfm tensor(1.1432e-12, device='cuda:0', grad_fn=<DivBackward0>)
episode: 20 training return: tensor(3.0008e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 21 training return: tensor(1.0306e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 22 training return: tensor(3.7439e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 23 training return: tensor(8.3208e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 6 test_true_pfm: 288 test_simulate_pfm tensor(4.9411e-06, device='cuda:0', grad_fn=<DivBackward0>)
episode: 24 training return: tensor(1.8093e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 25 training return: tensor(1.3920e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 26 training return: tensor(2.8763e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 27 training return: tensor(1.6768e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 7 test_true_pfm: -236 test_simulate_pfm tensor(1.6532e-09, device='cuda:0', grad_fn=<DivBackward0>)
episode: 28 training return: tensor(4.2987e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 29 training return: tensor(1.8374e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 30 training return: tensor(4.6109e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 31 training return: tensor(1.3647e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 8 test_true_pfm: -245 test_simulate_pfm tensor(7.7399e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 32 training return: tensor(1.0425e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 33 training return: tensor(3.1498e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 34 training return: tensor(7.8884e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 35 training return: tensor(6.3535e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 9 test_true_pfm: 224 test_simulate_pfm tensor(4.3124e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 36 training return: tensor(7.6052e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 37 training return: tensor(1.3723e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 38 training return: tensor(9.1254e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 39 training return: tensor(3.7532e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 10 test_true_pfm: -106 test_simulate_pfm tensor(1.6052e-06, device='cuda:0', grad_fn=<DivBackward0>)
episode: 40 training return: tensor(2.1002e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 41 training return: tensor(1.1264e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 42 training return: tensor(5.1000e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 43 training return: tensor(8.1969e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 11 test_true_pfm: 1163 test_simulate_pfm tensor(5.0961e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 44 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 45 training return: tensor(0.0040, device='cuda:0', grad_fn=<AddBackward0>)
episode: 46 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 47 training return: tensor(4.8473e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 12 test_true_pfm: 232 test_simulate_pfm tensor(9.5364e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 48 training return: tensor(7.2254e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 49 training return: tensor(1.3462e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 50 training return: tensor(7.4624e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 51 training return: tensor(1.4604e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 13 test_true_pfm: 262 test_simulate_pfm tensor(0.0005, device='cuda:0', grad_fn=<DivBackward0>)
episode: 52 training return: tensor(1.3542e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 53 training return: tensor(0.0059, device='cuda:0', grad_fn=<AddBackward0>)
episode: 54 training return: tensor(1.0893e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 55 training return: tensor(4.7502e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 14 test_true_pfm: -419 test_simulate_pfm tensor(0.0001, device='cuda:0', grad_fn=<DivBackward0>)
episode: 56 training return: tensor(4.1420e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 57 training return: tensor(8.9350e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 58 training return: tensor(4.3117e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 59 training return: tensor(2.5252e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 15 test_true_pfm: -264 test_simulate_pfm tensor(2.7777e-09, device='cuda:0', grad_fn=<DivBackward0>)
episode: 60 training return: tensor(6.4890e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 61 training return: tensor(5.6717e-15, device='cuda:0', grad_fn=<AddBackward0>)
episode: 62 training return: tensor(2.5295e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 63 training return: tensor(4.8321e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 16 test_true_pfm: -418 test_simulate_pfm tensor(2.5747e-06, device='cuda:0', grad_fn=<DivBackward0>)
episode: 64 training return: tensor(1.7344e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 65 training return: tensor(9.2276e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 66 training return: tensor(5.2578e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 67 training return: tensor(3.3650e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 17 test_true_pfm: 95 test_simulate_pfm tensor(0.0002, device='cuda:0', grad_fn=<DivBackward0>)
episode: 68 training return: tensor(1.7170e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 69 training return: tensor(6.4994e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 70 training return: tensor(4.2427e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 71 training return: tensor(3.7488e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 18 test_true_pfm: -107 test_simulate_pfm tensor(0.1155, device='cuda:0', grad_fn=<DivBackward0>)
episode: 72 training return: tensor(4.7541e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 73 training return: tensor(1.0888e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 74 training return: tensor(9.1004e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 75 training return: tensor(1.5509e-13, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 19 test_true_pfm: 600 test_simulate_pfm tensor(0.0002, device='cuda:0', grad_fn=<DivBackward0>)
episode: 76 training return: tensor(0.0018, device='cuda:0', grad_fn=<AddBackward0>)
episode: 77 training return: tensor(1.2060e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 78 training return: tensor(3.2701e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 79 training return: tensor(4.1431e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 20 test_true_pfm: 382 test_simulate_pfm tensor(0.0001, device='cuda:0', grad_fn=<DivBackward0>)
episode: 80 training return: tensor(3.7557e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 81 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
episode: 82 training return: tensor(0.0007, device='cuda:0', grad_fn=<AddBackward0>)
episode: 83 training return: tensor(5.2557e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 21 test_true_pfm: 835 test_simulate_pfm tensor(0.0001, device='cuda:0', grad_fn=<DivBackward0>)
episode: 84 training return: tensor(6.9967e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 85 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 86 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
episode: 87 training return: tensor(3.7482e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 22 test_true_pfm: 1029 test_simulate_pfm tensor(0.0005, device='cuda:0', grad_fn=<DivBackward0>)
episode: 88 training return: tensor(0.0081, device='cuda:0', grad_fn=<AddBackward0>)
episode: 89 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 90 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
episode: 91 training return: tensor(0.0007, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 23 test_true_pfm: 966 test_simulate_pfm tensor(0.0068, device='cuda:0', grad_fn=<DivBackward0>)
episode: 92 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 93 training return: tensor(6.6147e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 94 training return: tensor(2.8620e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 95 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 24 test_true_pfm: 1815 test_simulate_pfm tensor(0.0009, device='cuda:0', grad_fn=<DivBackward0>)
episode: 96 training return: tensor(8.9040e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 97 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 98 training return: tensor(0.0044, device='cuda:0', grad_fn=<AddBackward0>)
episode: 99 training return: tensor(5.8452e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 25 test_true_pfm: 1151 test_simulate_pfm tensor(0.0002, device='cuda:0', grad_fn=<DivBackward0>)
episode: 100 training return: tensor(3.4937e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 101 training return: tensor(1.6260e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 102 training return: tensor(1.1114e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 103 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 26 test_true_pfm: 1164 test_simulate_pfm tensor(0.0006, device='cuda:0', grad_fn=<DivBackward0>)
episode: 104 training return: tensor(0.0037, device='cuda:0', grad_fn=<AddBackward0>)
episode: 105 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 106 training return: tensor(0.0045, device='cuda:0', grad_fn=<AddBackward0>)
episode: 107 training return: tensor(0.0124, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 27 test_true_pfm: 1966 test_simulate_pfm tensor(0.0001, device='cuda:0', grad_fn=<DivBackward0>)
episode: 108 training return: tensor(4.0954e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 109 training return: tensor(5.8445e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 110 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 111 training return: tensor(7.8520e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 28 test_true_pfm: 1094 test_simulate_pfm tensor(6.3593e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 112 training return: tensor(4.4623e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 113 training return: tensor(3.2939e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 114 training return: tensor(0.0015, device='cuda:0', grad_fn=<AddBackward0>)
episode: 115 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 29 test_true_pfm: 929 test_simulate_pfm tensor(0.0024, device='cuda:0', grad_fn=<DivBackward0>)
episode: 116 training return: tensor(6.0524e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 117 training return: tensor(6.0625e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 118 training return: tensor(1.3280e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 119 training return: tensor(7.4850e-15, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 30 test_true_pfm: 1478 test_simulate_pfm tensor(0.0002, device='cuda:0', grad_fn=<DivBackward0>)
episode: 120 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
episode: 121 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
episode: 122 training return: tensor(0.0084, device='cuda:0', grad_fn=<AddBackward0>)
episode: 123 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 31 test_true_pfm: 1707 test_simulate_pfm tensor(0.0006, device='cuda:0', grad_fn=<DivBackward0>)
episode: 124 training return: tensor(6.5936e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 125 training return: tensor(8.9624e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 126 training return: tensor(3.7180e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 127 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 32 test_true_pfm: 1494 test_simulate_pfm tensor(0.0153, device='cuda:0', grad_fn=<DivBackward0>)
episode: 128 training return: tensor(9.9190e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 129 training return: tensor(0.0124, device='cuda:0', grad_fn=<AddBackward0>)
episode: 130 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 131 training return: tensor(0.0077, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 33 test_true_pfm: 1563 test_simulate_pfm tensor(0.0093, device='cuda:0', grad_fn=<DivBackward0>)
episode: 132 training return: tensor(0.0038, device='cuda:0', grad_fn=<AddBackward0>)
episode: 133 training return: tensor(7.8916e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 134 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 135 training return: tensor(0.0007, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 34 test_true_pfm: 1740 test_simulate_pfm tensor(0.0078, device='cuda:0', grad_fn=<DivBackward0>)
episode: 136 training return: tensor(5.1997e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 137 training return: tensor(7.3030e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 138 training return: tensor(0.0026, device='cuda:0', grad_fn=<AddBackward0>)
episode: 139 training return: tensor(0.0030, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 35 test_true_pfm: 3275 test_simulate_pfm tensor(0.0009, device='cuda:0', grad_fn=<DivBackward0>)
episode: 140 training return: tensor(0.0022, device='cuda:0', grad_fn=<AddBackward0>)
episode: 141 training return: tensor(0.0034, device='cuda:0', grad_fn=<AddBackward0>)
episode: 142 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
episode: 143 training return: tensor(0.0054, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 36 test_true_pfm: 2419 test_simulate_pfm tensor(0.0002, device='cuda:0', grad_fn=<DivBackward0>)
episode: 144 training return: tensor(0.0006, device='cuda:0', grad_fn=<AddBackward0>)
episode: 145 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 146 training return: tensor(2.3881e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 147 training return: tensor(0.0015, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 37 test_true_pfm: 382 test_simulate_pfm tensor(5.5712e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 148 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 149 training return: tensor(1.9676e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 150 training return: tensor(1.7777e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 151 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 38 test_true_pfm: 1799 test_simulate_pfm tensor(0.0015, device='cuda:0', grad_fn=<DivBackward0>)
episode: 152 training return: tensor(7.3695e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 153 training return: tensor(4.8335e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 154 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
episode: 155 training return: tensor(0.0055, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 39 test_true_pfm: 2027 test_simulate_pfm tensor(0.0011, device='cuda:0', grad_fn=<DivBackward0>)
episode: 156 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 157 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 158 training return: tensor(0.0008, device='cuda:0', grad_fn=<AddBackward0>)
episode: 159 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 40 test_true_pfm: 1626 test_simulate_pfm tensor(0.0014, device='cuda:0', grad_fn=<DivBackward0>)
episode: 160 training return: tensor(3.8370e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 161 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 162 training return: tensor(1.3365e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 163 training return: tensor(2.0039e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 41 test_true_pfm: -477 test_simulate_pfm tensor(3.5790e-06, device='cuda:0', grad_fn=<DivBackward0>)
episode: 164 training return: tensor(7.4198e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 165 training return: tensor(2.1192e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 166 training return: tensor(7.9887e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 167 training return: tensor(2.2948e-13, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 42 test_true_pfm: -415 test_simulate_pfm tensor(7.3214e-09, device='cuda:0', grad_fn=<DivBackward0>)
episode: 168 training return: tensor(1.9237e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 169 training return: tensor(9.1054e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 170 training return: tensor(7.3775e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 171 training return: tensor(9.5898e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 43 test_true_pfm: -262 test_simulate_pfm tensor(5.8725e-09, device='cuda:0', grad_fn=<DivBackward0>)
episode: 172 training return: tensor(5.1808e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 173 training return: tensor(1.1476e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 174 training return: tensor(1.6663e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 175 training return: tensor(1.0209e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 44 test_true_pfm: -397 test_simulate_pfm tensor(2.2609e-11, device='cuda:0', grad_fn=<DivBackward0>)
episode: 176 training return: tensor(1.7693e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 177 training return: tensor(2.8093e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 178 training return: tensor(8.2741e-17, device='cuda:0', grad_fn=<AddBackward0>)
episode: 179 training return: tensor(1.3377e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 45 test_true_pfm: -341 test_simulate_pfm tensor(9.0558e-09, device='cuda:0', grad_fn=<DivBackward0>)
episode: 180 training return: tensor(1.8354e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 181 training return: tensor(1.3273e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 182 training return: tensor(1.9181e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 183 training return: tensor(1.1082e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 46 test_true_pfm: -587 test_simulate_pfm tensor(1.7028e-11, device='cuda:0', grad_fn=<DivBackward0>)
episode: 184 training return: tensor(1.0944e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 185 training return: tensor(7.6133e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 186 training return: tensor(5.6630e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 187 training return: tensor(3.6318e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 47 test_true_pfm: -777 test_simulate_pfm tensor(6.2613e-12, device='cuda:0', grad_fn=<DivBackward0>)
episode: 188 training return: tensor(4.4351e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 189 training return: tensor(3.2213e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 190 training return: tensor(4.8264e-15, device='cuda:0', grad_fn=<AddBackward0>)
episode: 191 training return: tensor(9.9192e-16, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 48 test_true_pfm: -487 test_simulate_pfm tensor(1.8422e-14, device='cuda:0', grad_fn=<DivBackward0>)
episode: 192 training return: tensor(8.3520e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 193 training return: tensor(1.3589e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 194 training return: tensor(1.9170e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 195 training return: tensor(1.7048e-14, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 49 test_true_pfm: -537 test_simulate_pfm tensor(5.5272e-14, device='cuda:0', grad_fn=<DivBackward0>)
episode: 196 training return: tensor(6.0560e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 197 training return: tensor(5.1813e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 198 training return: tensor(7.8587e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 199 training return: tensor(2.2629e-14, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 50 test_true_pfm: -502 test_simulate_pfm tensor(8.5531e-12, device='cuda:0', grad_fn=<DivBackward0>)
episode: 200 training return: tensor(1.6861e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 201 training return: tensor(3.7885e-15, device='cuda:0', grad_fn=<AddBackward0>)
episode: 202 training return: tensor(1.5487e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 203 training return: tensor(5.4081e-14, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 51 test_true_pfm: -511 test_simulate_pfm tensor(1.7838e-11, device='cuda:0', grad_fn=<DivBackward0>)
episode: 204 training return: tensor(3.3541e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 205 training return: tensor(3.6104e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 206 training return: tensor(4.7066e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 207 training return: tensor(3.2120e-13, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 52 test_true_pfm: -581 test_simulate_pfm tensor(2.4042e-12, device='cuda:0', grad_fn=<DivBackward0>)
episode: 208 training return: tensor(2.0874e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 209 training return: tensor(2.7685e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 210 training return: tensor(4.3371e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 211 training return: tensor(8.4101e-15, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 53 test_true_pfm: -638 test_simulate_pfm tensor(6.6544e-13, device='cuda:0', grad_fn=<DivBackward0>)
episode: 212 training return: tensor(4.7969e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 213 training return: tensor(6.6404e-18, device='cuda:0', grad_fn=<AddBackward0>)
episode: 214 training return: tensor(1.3922e-18, device='cuda:0', grad_fn=<AddBackward0>)
episode: 215 training return: tensor(1.9869e-16, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 54 test_true_pfm: -463 test_simulate_pfm tensor(3.9108e-12, device='cuda:0', grad_fn=<DivBackward0>)
episode: 216 training return: tensor(6.0297e-18, device='cuda:0', grad_fn=<AddBackward0>)
episode: 217 training return: tensor(2.6246e-17, device='cuda:0', grad_fn=<AddBackward0>)
episode: 218 training return: tensor(2.6014e-15, device='cuda:0', grad_fn=<AddBackward0>)
episode: 219 training return: tensor(6.4878e-14, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 55 test_true_pfm: -521 test_simulate_pfm tensor(5.0713e-14, device='cuda:0', grad_fn=<DivBackward0>)
episode: 220 training return: tensor(1.0132e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 221 training return: tensor(2.6446e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 222 training return: tensor(5.4358e-15, device='cuda:0', grad_fn=<AddBackward0>)
episode: 223 training return: tensor(1.2383e-14, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 56 test_true_pfm: -422 test_simulate_pfm tensor(7.0319e-13, device='cuda:0', grad_fn=<DivBackward0>)
episode: 224 training return: tensor(5.0990e-18, device='cuda:0', grad_fn=<AddBackward0>)
episode: 225 training return: tensor(4.7117e-16, device='cuda:0', grad_fn=<AddBackward0>)
episode: 226 training return: tensor(1.5726e-16, device='cuda:0', grad_fn=<AddBackward0>)
episode: 227 training return: tensor(1.9219e-15, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 57 test_true_pfm: -578 test_simulate_pfm tensor(4.4424e-15, device='cuda:0', grad_fn=<DivBackward0>)
episode: 228 training return: tensor(1.1887e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 229 training return: tensor(6.9902e-17, device='cuda:0', grad_fn=<AddBackward0>)
episode: 230 training return: tensor(4.8064e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 231 training return: tensor(4.5223e-18, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 58 test_true_pfm: -529 test_simulate_pfm tensor(2.3751e-16, device='cuda:0', grad_fn=<DivBackward0>)
episode: 232 training return: tensor(7.6126e-17, device='cuda:0', grad_fn=<AddBackward0>)
episode: 233 training return: tensor(1.8147e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 234 training return: tensor(2.0495e-15, device='cuda:0', grad_fn=<AddBackward0>)
episode: 235 training return: tensor(2.0023e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 59 test_true_pfm: -563 test_simulate_pfm tensor(1.3678e-13, device='cuda:0', grad_fn=<DivBackward0>)
episode: 236 training return: tensor(3.4533e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 237 training return: tensor(3.9314e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 238 training return: tensor(3.3512e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 239 training return: tensor(9.1133e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 60 test_true_pfm: -215 test_simulate_pfm tensor(4.0026e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 240 training return: tensor(2.6745e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 241 training return: tensor(4.5771e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 242 training return: tensor(6.1336e-18, device='cuda:0', grad_fn=<AddBackward0>)
episode: 243 training return: tensor(1.1418e-15, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 61 test_true_pfm: -409 test_simulate_pfm tensor(6.5534e-13, device='cuda:0', grad_fn=<DivBackward0>)
episode: 244 training return: tensor(3.2186e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 245 training return: tensor(9.6825e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 246 training return: tensor(4.1914e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 247 training return: tensor(2.1740e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 62 test_true_pfm: -414 test_simulate_pfm tensor(2.8726e-09, device='cuda:0', grad_fn=<DivBackward0>)
episode: 248 training return: tensor(7.6580e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 249 training return: tensor(3.9840e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 250 training return: tensor(7.1759e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 251 training return: tensor(1.6412e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 63 test_true_pfm: -29 test_simulate_pfm tensor(2.3457e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 252 training return: tensor(5.7282e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 253 training return: tensor(2.2321e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 254 training return: tensor(1.0592e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 255 training return: tensor(1.6934e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 64 test_true_pfm: -151 test_simulate_pfm tensor(1.0414e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 256 training return: tensor(3.3240e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 257 training return: tensor(1.2159e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 258 training return: tensor(3.8857e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 259 training return: tensor(3.0904e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 65 test_true_pfm: 1155 test_simulate_pfm tensor(2.0850e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 260 training return: tensor(4.0768e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 261 training return: tensor(4.7354e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 262 training return: tensor(2.6401e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 263 training return: tensor(2.2717e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 66 test_true_pfm: 177 test_simulate_pfm tensor(4.1267e-08, device='cuda:0', grad_fn=<DivBackward0>)
episode: 264 training return: tensor(1.4342e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 265 training return: tensor(5.0265e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 266 training return: tensor(2.4558e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 267 training return: tensor(1.0449e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 67 test_true_pfm: -498 test_simulate_pfm tensor(6.2180e-10, device='cuda:0', grad_fn=<DivBackward0>)
episode: 268 training return: tensor(3.5493e-16, device='cuda:0', grad_fn=<AddBackward0>)
episode: 269 training return: tensor(1.2788e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 270 training return: tensor(4.1548e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 271 training return: tensor(2.4265e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 68 test_true_pfm: 212 test_simulate_pfm tensor(3.1963e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 272 training return: tensor(2.7296e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 273 training return: tensor(9.3145e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 274 training return: tensor(1.3432e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 275 training return: tensor(5.8487e-13, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 69 test_true_pfm: -262 test_simulate_pfm tensor(3.5553e-08, device='cuda:0', grad_fn=<DivBackward0>)
episode: 276 training return: tensor(6.3531e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 277 training return: tensor(3.2168e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 278 training return: tensor(1.0960e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 279 training return: tensor(8.6202e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 70 test_true_pfm: -169 test_simulate_pfm tensor(4.2075e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 280 training return: tensor(8.5061e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 281 training return: tensor(5.6462e-15, device='cuda:0', grad_fn=<AddBackward0>)
episode: 282 training return: tensor(1.4163e-15, device='cuda:0', grad_fn=<AddBackward0>)
episode: 283 training return: tensor(2.5505e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 71 test_true_pfm: 451 test_simulate_pfm tensor(1.8526e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 284 training return: tensor(2.3632e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 285 training return: tensor(1.5527e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 286 training return: tensor(8.5109e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 287 training return: tensor(5.7571e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 72 test_true_pfm: -187 test_simulate_pfm tensor(2.2760e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 288 training return: tensor(1.7369e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 289 training return: tensor(6.1852e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 290 training return: tensor(3.7600e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 291 training return: tensor(4.9478e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 73 test_true_pfm: -305 test_simulate_pfm tensor(1.5339e-06, device='cuda:0', grad_fn=<DivBackward0>)
episode: 292 training return: tensor(2.3163e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 293 training return: tensor(3.4770e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 294 training return: tensor(9.5810e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 295 training return: tensor(2.6654e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 74 test_true_pfm: 114 test_simulate_pfm tensor(4.4134e-08, device='cuda:0', grad_fn=<DivBackward0>)
episode: 296 training return: tensor(7.6502e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 297 training return: tensor(1.3368e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 298 training return: tensor(3.9928e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 299 training return: tensor(4.6445e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 75 test_true_pfm: 243 test_simulate_pfm tensor(3.2414e-06, device='cuda:0', grad_fn=<DivBackward0>)
episode: 300 training return: tensor(4.0608e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 301 training return: tensor(2.3603e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 302 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 303 training return: tensor(5.6695e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 76 test_true_pfm: 1051 test_simulate_pfm tensor(0.0002, device='cuda:0', grad_fn=<DivBackward0>)
episode: 304 training return: tensor(4.0154e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 305 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
episode: 306 training return: tensor(4.0677e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 307 training return: tensor(1.1849e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 77 test_true_pfm: 1415 test_simulate_pfm tensor(4.1189e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 308 training return: tensor(2.9155e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 309 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 310 training return: tensor(5.1264e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 311 training return: tensor(9.0249e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 78 test_true_pfm: 2204 test_simulate_pfm tensor(3.2608e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 312 training return: tensor(3.2214e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 313 training return: tensor(6.1316e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 314 training return: tensor(3.9618e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 315 training return: tensor(2.7262e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 79 test_true_pfm: 338 test_simulate_pfm tensor(0.0001, device='cuda:0', grad_fn=<DivBackward0>)
episode: 316 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 317 training return: tensor(3.6054e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 318 training return: tensor(5.4696e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 319 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 80 test_true_pfm: 903 test_simulate_pfm tensor(0.0006, device='cuda:0', grad_fn=<DivBackward0>)
episode: 320 training return: tensor(8.3456e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 321 training return: tensor(1.3289e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 322 training return: tensor(0.0006, device='cuda:0', grad_fn=<AddBackward0>)
episode: 323 training return: tensor(0.0012, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 81 test_true_pfm: 1164 test_simulate_pfm tensor(0.0006, device='cuda:0', grad_fn=<DivBackward0>)
episode: 324 training return: tensor(6.3964e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 325 training return: tensor(1.8756e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 326 training return: tensor(1.6997e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 327 training return: tensor(0.0037, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 82 test_true_pfm: 2596 test_simulate_pfm tensor(0.0006, device='cuda:0', grad_fn=<DivBackward0>)
episode: 328 training return: tensor(0.0362, device='cuda:0', grad_fn=<AddBackward0>)
episode: 329 training return: tensor(0.0022, device='cuda:0', grad_fn=<AddBackward0>)
episode: 330 training return: tensor(4.6216e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 331 training return: tensor(3.4341e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 83 test_true_pfm: 460 test_simulate_pfm tensor(0.0040, device='cuda:0', grad_fn=<DivBackward0>)
episode: 332 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 333 training return: tensor(2.0481e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 334 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
episode: 335 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 84 test_true_pfm: 2465 test_simulate_pfm tensor(0.0557, device='cuda:0', grad_fn=<DivBackward0>)
episode: 336 training return: tensor(0.0027, device='cuda:0', grad_fn=<AddBackward0>)
episode: 337 training return: tensor(2.2035e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 338 training return: tensor(3.6276e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 339 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 85 test_true_pfm: 1833 test_simulate_pfm tensor(0.0001, device='cuda:0', grad_fn=<DivBackward0>)
episode: 340 training return: tensor(6.9526e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 341 training return: tensor(1.9496e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 342 training return: tensor(4.9013e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 343 training return: tensor(5.1928e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 86 test_true_pfm: 829 test_simulate_pfm tensor(2.1736e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 344 training return: tensor(5.8280e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 345 training return: tensor(3.1428e-15, device='cuda:0', grad_fn=<AddBackward0>)
episode: 346 training return: tensor(3.8383e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 347 training return: tensor(2.3512e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 87 test_true_pfm: 1111 test_simulate_pfm tensor(3.0040e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 348 training return: tensor(2.2536e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 349 training return: tensor(4.8130e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 350 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 351 training return: tensor(5.1607e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 88 test_true_pfm: 1673 test_simulate_pfm tensor(9.8172e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 352 training return: tensor(1.3937e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 353 training return: tensor(3.6462e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 354 training return: tensor(1.4493e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 355 training return: tensor(1.2503e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 89 test_true_pfm: 1865 test_simulate_pfm tensor(2.2846e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 356 training return: tensor(5.4101e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 357 training return: tensor(6.2107e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 358 training return: tensor(2.5596e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 359 training return: tensor(1.5663e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 90 test_true_pfm: 1223 test_simulate_pfm tensor(9.1401e-06, device='cuda:0', grad_fn=<DivBackward0>)
episode: 360 training return: tensor(3.8456e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 361 training return: tensor(9.1387e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 362 training return: tensor(4.4413e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 363 training return: tensor(7.1603e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 91 test_true_pfm: 74 test_simulate_pfm tensor(1.3186e-08, device='cuda:0', grad_fn=<DivBackward0>)
episode: 364 training return: tensor(6.3282e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 365 training return: tensor(5.2145e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 366 training return: tensor(1.1229e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 367 training return: tensor(2.3105e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 92 test_true_pfm: -277 test_simulate_pfm tensor(9.9802e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 368 training return: tensor(2.0505e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 369 training return: tensor(1.9442e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 370 training return: tensor(2.1493e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 371 training return: tensor(8.1981e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 93 test_true_pfm: 755 test_simulate_pfm tensor(1.9250e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 372 training return: tensor(5.0780e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 373 training return: tensor(1.1665e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 374 training return: tensor(1.0180e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 375 training return: tensor(2.2650e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 94 test_true_pfm: 4 test_simulate_pfm tensor(1.1724e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 376 training return: tensor(5.0319e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 377 training return: tensor(7.2672e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 378 training return: tensor(1.5042e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 379 training return: tensor(2.4436e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 95 test_true_pfm: 148 test_simulate_pfm tensor(3.8722e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 380 training return: tensor(3.7518e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 381 training return: tensor(3.6370e-16, device='cuda:0', grad_fn=<AddBackward0>)
episode: 382 training return: tensor(6.2408e-17, device='cuda:0', grad_fn=<AddBackward0>)
episode: 383 training return: tensor(4.1700e-15, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 96 test_true_pfm: -449 test_simulate_pfm tensor(2.0296e-11, device='cuda:0', grad_fn=<DivBackward0>)
episode: 384 training return: tensor(1.6683e-16, device='cuda:0', grad_fn=<AddBackward0>)
episode: 385 training return: tensor(4.3689e-20, device='cuda:0', grad_fn=<AddBackward0>)
episode: 386 training return: tensor(4.4803e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 387 training return: tensor(8.7335e-14, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 97 test_true_pfm: -393 test_simulate_pfm tensor(1.9017e-12, device='cuda:0', grad_fn=<DivBackward0>)
episode: 388 training return: tensor(2.3300e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 389 training return: tensor(7.4130e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 390 training return: tensor(7.3396e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 391 training return: tensor(5.2355e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 98 test_true_pfm: 215 test_simulate_pfm tensor(5.0084e-11, device='cuda:0', grad_fn=<DivBackward0>)
episode: 392 training return: tensor(3.6012e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 393 training return: tensor(3.6495e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 394 training return: tensor(1.5660e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 395 training return: tensor(1.8304e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 99 test_true_pfm: 1318 test_simulate_pfm tensor(1.2650e-06, device='cuda:0', grad_fn=<DivBackward0>)
episode: 396 training return: tensor(3.5583e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 397 training return: tensor(2.4703e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 398 training return: tensor(1.3369e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 399 training return: tensor(3.9602e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 100 test_true_pfm: 1611 test_simulate_pfm tensor(9.4932e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 400 training return: tensor(2.4730e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 401 training return: tensor(1.6402e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 402 training return: tensor(9.3190e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 403 training return: tensor(2.7007e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 101 test_true_pfm: 561 test_simulate_pfm tensor(6.6543e-08, device='cuda:0', grad_fn=<DivBackward0>)
episode: 404 training return: tensor(2.0548e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 405 training return: tensor(3.9738e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 406 training return: tensor(2.0456e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 407 training return: tensor(1.5756e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 102 test_true_pfm: 972 test_simulate_pfm tensor(5.2133e-06, device='cuda:0', grad_fn=<DivBackward0>)
episode: 408 training return: tensor(1.2562e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 409 training return: tensor(2.2606e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 410 training return: tensor(0.0082, device='cuda:0', grad_fn=<AddBackward0>)
episode: 411 training return: tensor(4.9397e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 103 test_true_pfm: 270 test_simulate_pfm tensor(3.8595e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 412 training return: tensor(3.8237e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 413 training return: tensor(2.4733e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 414 training return: tensor(4.1837e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 415 training return: tensor(4.7917e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 104 test_true_pfm: 1429 test_simulate_pfm tensor(0.0003, device='cuda:0', grad_fn=<DivBackward0>)
episode: 416 training return: tensor(0.0006, device='cuda:0', grad_fn=<AddBackward0>)
episode: 417 training return: tensor(4.6250e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 418 training return: tensor(0.0007, device='cuda:0', grad_fn=<AddBackward0>)
episode: 419 training return: tensor(0.0009, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 105 test_true_pfm: 190 test_simulate_pfm tensor(0.0014, device='cuda:0', grad_fn=<DivBackward0>)
episode: 420 training return: tensor(6.7169e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 421 training return: tensor(3.2689e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 422 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 423 training return: tensor(2.9856e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 106 test_true_pfm: 688 test_simulate_pfm tensor(0.0007, device='cuda:0', grad_fn=<DivBackward0>)
episode: 424 training return: tensor(4.1524e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 425 training return: tensor(0.0009, device='cuda:0', grad_fn=<AddBackward0>)
episode: 426 training return: tensor(6.9004e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 427 training return: tensor(3.3780e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 107 test_true_pfm: 2051 test_simulate_pfm tensor(1.8766e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 428 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 429 training return: tensor(5.8475e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 430 training return: tensor(1.1629e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 431 training return: tensor(2.4698e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 108 test_true_pfm: 1641 test_simulate_pfm tensor(5.9705e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 432 training return: tensor(3.9108e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 433 training return: tensor(2.6383e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 434 training return: tensor(2.5703e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 435 training return: tensor(1.2597e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 109 test_true_pfm: 425 test_simulate_pfm tensor(6.7816e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 436 training return: tensor(5.1979e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 437 training return: tensor(5.7673e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 438 training return: tensor(1.2986e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 439 training return: tensor(4.1938e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 110 test_true_pfm: 1134 test_simulate_pfm tensor(4.9273e-06, device='cuda:0', grad_fn=<DivBackward0>)
episode: 440 training return: tensor(1.2960e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 441 training return: tensor(1.5622e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 442 training return: tensor(2.3328e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 443 training return: tensor(9.9925e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 111 test_true_pfm: 516 test_simulate_pfm tensor(5.4101e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 444 training return: tensor(2.3143e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 445 training return: tensor(9.7641e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 446 training return: tensor(1.0722e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 447 training return: tensor(1.2036e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 112 test_true_pfm: 661 test_simulate_pfm tensor(1.5103e-06, device='cuda:0', grad_fn=<DivBackward0>)
episode: 448 training return: tensor(0.0011, device='cuda:0', grad_fn=<AddBackward0>)
episode: 449 training return: tensor(1.1475e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 450 training return: tensor(8.5095e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 451 training return: tensor(9.9794e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 113 test_true_pfm: 1995 test_simulate_pfm tensor(0.0003, device='cuda:0', grad_fn=<DivBackward0>)
episode: 452 training return: tensor(9.0549e-15, device='cuda:0', grad_fn=<AddBackward0>)
episode: 453 training return: tensor(2.6513e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 454 training return: tensor(0.0190, device='cuda:0', grad_fn=<AddBackward0>)
episode: 455 training return: tensor(4.8582e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 114 test_true_pfm: 1372 test_simulate_pfm tensor(0.0009, device='cuda:0', grad_fn=<DivBackward0>)
episode: 456 training return: tensor(7.4365e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 457 training return: tensor(9.2848e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 458 training return: tensor(2.0459e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 459 training return: tensor(1.9915e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 115 test_true_pfm: -462 test_simulate_pfm tensor(8.1712e-11, device='cuda:0', grad_fn=<DivBackward0>)
episode: 460 training return: tensor(1.0192e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 461 training return: tensor(2.6819e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 462 training return: tensor(1.7094e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 463 training return: tensor(3.6270e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 116 test_true_pfm: -826 test_simulate_pfm tensor(1.3898e-09, device='cuda:0', grad_fn=<DivBackward0>)
episode: 464 training return: tensor(6.1397e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 465 training return: tensor(3.8665e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 466 training return: tensor(1.6907e-17, device='cuda:0', grad_fn=<AddBackward0>)
episode: 467 training return: tensor(1.2857e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 117 test_true_pfm: -586 test_simulate_pfm tensor(6.2896e-12, device='cuda:0', grad_fn=<DivBackward0>)
episode: 468 training return: tensor(6.3716e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 469 training return: tensor(7.0400e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 470 training return: tensor(1.2838e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 471 training return: tensor(1.2255e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 118 test_true_pfm: -607 test_simulate_pfm tensor(2.4604e-08, device='cuda:0', grad_fn=<DivBackward0>)
episode: 472 training return: tensor(6.8101e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 473 training return: tensor(8.9198e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 474 training return: tensor(3.5294e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 475 training return: tensor(3.9834e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 119 test_true_pfm: -370 test_simulate_pfm tensor(7.0659e-08, device='cuda:0', grad_fn=<DivBackward0>)
episode: 476 training return: tensor(4.4153e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 477 training return: tensor(3.1729e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 478 training return: tensor(7.7537e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 479 training return: tensor(8.2278e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 120 test_true_pfm: 264 test_simulate_pfm tensor(1.0873e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 480 training return: tensor(1.5551e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 481 training return: tensor(8.8622e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 482 training return: tensor(2.6044e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 483 training return: tensor(2.4066e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 121 test_true_pfm: 1225 test_simulate_pfm tensor(9.4834e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 484 training return: tensor(2.3363e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 485 training return: tensor(4.5440e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 486 training return: tensor(1.5876e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 487 training return: tensor(1.9341e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 122 test_true_pfm: 155 test_simulate_pfm tensor(1.3147e-06, device='cuda:0', grad_fn=<DivBackward0>)
episode: 488 training return: tensor(1.3873e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 489 training return: tensor(3.4254e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 490 training return: tensor(3.3028e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 491 training return: tensor(2.0641e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 123 test_true_pfm: 294 test_simulate_pfm tensor(0.0001, device='cuda:0', grad_fn=<DivBackward0>)
episode: 492 training return: tensor(1.1428e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 493 training return: tensor(1.8201e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 494 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
episode: 495 training return: tensor(1.1340e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 124 test_true_pfm: 671 test_simulate_pfm tensor(0.0002, device='cuda:0', grad_fn=<DivBackward0>)
episode: 496 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
episode: 497 training return: tensor(2.3614e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 498 training return: tensor(1.0637e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 499 training return: tensor(4.4868e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 125 test_true_pfm: 273 test_simulate_pfm tensor(6.2662e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 500 training return: tensor(7.9693e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 501 training return: tensor(5.5018e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 502 training return: tensor(0.0036, device='cuda:0', grad_fn=<AddBackward0>)
episode: 503 training return: tensor(2.6716e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 126 test_true_pfm: 1213 test_simulate_pfm tensor(5.1365e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 504 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 505 training return: tensor(0.0013, device='cuda:0', grad_fn=<AddBackward0>)
episode: 506 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
episode: 507 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 127 test_true_pfm: 861 test_simulate_pfm tensor(0.0002, device='cuda:0', grad_fn=<DivBackward0>)
episode: 508 training return: tensor(0.0009, device='cuda:0', grad_fn=<AddBackward0>)
episode: 509 training return: tensor(2.7720e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 510 training return: tensor(8.9847e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 511 training return: tensor(0.0030, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 128 test_true_pfm: 484 test_simulate_pfm tensor(0.0099, device='cuda:0', grad_fn=<DivBackward0>)
episode: 512 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 513 training return: tensor(0.0023, device='cuda:0', grad_fn=<AddBackward0>)
episode: 514 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 515 training return: tensor(0.0106, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 129 test_true_pfm: -163 test_simulate_pfm tensor(9.0289e-05, device='cuda:0', grad_fn=<DivBackward0>)
episode: 516 training return: tensor(0.0016, device='cuda:0', grad_fn=<AddBackward0>)
episode: 517 training return: tensor(8.8113e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 518 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 519 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 130 test_true_pfm: 411 test_simulate_pfm tensor(0.0193, device='cuda:0', grad_fn=<DivBackward0>)
episode: 520 training return: tensor(1.0533e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 521 training return: tensor(0.0025, device='cuda:0', grad_fn=<AddBackward0>)
episode: 522 training return: tensor(0.0018, device='cuda:0', grad_fn=<AddBackward0>)
episode: 523 training return: tensor(0.0016, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 131 test_true_pfm: 698 test_simulate_pfm tensor(0.0007, device='cuda:0', grad_fn=<DivBackward0>)
episode: 524 training return: tensor(4.9708e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 525 training return: tensor(0.0492, device='cuda:0', grad_fn=<AddBackward0>)
episode: 526 training return: tensor(8.6834e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 527 training return: tensor(0.0007, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 132 test_true_pfm: 999 test_simulate_pfm tensor(0.0089, device='cuda:0', grad_fn=<DivBackward0>)
episode: 528 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
episode: 529 training return: tensor(0.0011, device='cuda:0', grad_fn=<AddBackward0>)
episode: 530 training return: tensor(0.0409, device='cuda:0', grad_fn=<AddBackward0>)
episode: 531 training return: tensor(5.1210e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 133 test_true_pfm: 1108 test_simulate_pfm tensor(0.0024, device='cuda:0', grad_fn=<DivBackward0>)
episode: 532 training return: tensor(0.0147, device='cuda:0', grad_fn=<AddBackward0>)
episode: 533 training return: tensor(7.0382e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 534 training return: tensor(0.1001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 535 training return: tensor(1.1089e-19, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 134 test_true_pfm: 2239 test_simulate_pfm tensor(0.0094, device='cuda:0', grad_fn=<DivBackward0>)
episode: 536 training return: tensor(0.0250, device='cuda:0', grad_fn=<AddBackward0>)
episode: 537 training return: tensor(0.0033, device='cuda:0', grad_fn=<AddBackward0>)
episode: 538 training return: tensor(0.0009, device='cuda:0', grad_fn=<AddBackward0>)
episode: 539 training return: tensor(0.0010, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 135 test_true_pfm: 1052 test_simulate_pfm tensor(0.0072, device='cuda:0', grad_fn=<DivBackward0>)
episode: 540 training return: tensor(0.0010, device='cuda:0', grad_fn=<AddBackward0>)
episode: 541 training return: tensor(0.0009, device='cuda:0', grad_fn=<AddBackward0>)
episode: 542 training return: tensor(1.5499e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 543 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 136 test_true_pfm: 1366 test_simulate_pfm tensor(0.0019, device='cuda:0', grad_fn=<DivBackward0>)
episode: 544 training return: tensor(0.0050, device='cuda:0', grad_fn=<AddBackward0>)
episode: 545 training return: tensor(0.0007, device='cuda:0', grad_fn=<AddBackward0>)
episode: 546 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 547 training return: tensor(9.3655e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 137 test_true_pfm: 1809 test_simulate_pfm tensor(0.0006, device='cuda:0', grad_fn=<DivBackward0>)
episode: 548 training return: tensor(0.0022, device='cuda:0', grad_fn=<AddBackward0>)
episode: 549 training return: tensor(1.1822e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 550 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
episode: 551 training return: tensor(1.1456e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 138 test_true_pfm: 1689 test_simulate_pfm tensor(2.2541e-06, device='cuda:0', grad_fn=<DivBackward0>)
episode: 552 training return: tensor(3.8006e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 553 training return: tensor(2.2986e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 554 training return: tensor(5.5633e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 555 training return: tensor(1.4876e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 139 test_true_pfm: 241 test_simulate_pfm tensor(9.6631e-07, device='cuda:0', grad_fn=<DivBackward0>)
episode: 556 training return: tensor(3.0162e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 557 training return: tensor(8.0372e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 558 training return: tensor(3.8594e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 559 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 140 test_true_pfm: 1085 test_simulate_pfm tensor(0.0012, device='cuda:0', grad_fn=<DivBackward0>)
episode: 560 training return: tensor(5.7204e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 561 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 562 training return: tensor(1.2395e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 563 training return: tensor(0.0028, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 141 test_true_pfm: 844 test_simulate_pfm tensor(0.0003, device='cuda:0', grad_fn=<DivBackward0>)
episode: 564 training return: tensor(4.4733e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 565 training return: tensor(0.0017, device='cuda:0', grad_fn=<AddBackward0>)
episode: 566 training return: tensor(3.4787e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 567 training return: tensor(0.0063, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 142 test_true_pfm: 399 test_simulate_pfm tensor(0.0105, device='cuda:0', grad_fn=<DivBackward0>)
episode: 568 training return: tensor(0.0011, device='cuda:0', grad_fn=<AddBackward0>)
episode: 569 training return: tensor(0.0296, device='cuda:0', grad_fn=<AddBackward0>)
episode: 570 training return: tensor(0.0161, device='cuda:0', grad_fn=<AddBackward0>)
episode: 571 training return: tensor(0.0080, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 143 test_true_pfm: 3121 test_simulate_pfm tensor(0.0223, device='cuda:0', grad_fn=<DivBackward0>)
episode: 572 training return: tensor(0.2568, device='cuda:0', grad_fn=<AddBackward0>)
episode: 573 training return: tensor(0.2419, device='cuda:0', grad_fn=<AddBackward0>)
episode: 574 training return: tensor(0.0047, device='cuda:0', grad_fn=<AddBackward0>)
episode: 575 training return: tensor(1.4555e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 144 test_true_pfm: 3950 test_simulate_pfm tensor(0.0447, device='cuda:0', grad_fn=<DivBackward0>)
episode: 576 training return: tensor(0.0828, device='cuda:0', grad_fn=<AddBackward0>)
episode: 577 training return: tensor(0.0085, device='cuda:0', grad_fn=<AddBackward0>)
episode: 578 training return: tensor(0.0385, device='cuda:0', grad_fn=<AddBackward0>)
episode: 579 training return: tensor(0.0025, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 145 test_true_pfm: 3778 test_simulate_pfm tensor(0.0173, device='cuda:0', grad_fn=<DivBackward0>)
episode: 580 training return: tensor(0.0051, device='cuda:0', grad_fn=<AddBackward0>)
episode: 581 training return: tensor(0.0589, device='cuda:0', grad_fn=<AddBackward0>)
episode: 582 training return: tensor(0.0442, device='cuda:0', grad_fn=<AddBackward0>)
episode: 583 training return: tensor(0.0306, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 146 test_true_pfm: 4037 test_simulate_pfm tensor(0.0061, device='cuda:0', grad_fn=<DivBackward0>)
episode: 584 training return: tensor(0.0033, device='cuda:0', grad_fn=<AddBackward0>)
episode: 585 training return: tensor(0.0045, device='cuda:0', grad_fn=<AddBackward0>)
episode: 586 training return: tensor(0.0156, device='cuda:0', grad_fn=<AddBackward0>)
episode: 587 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 147 test_true_pfm: 3797 test_simulate_pfm tensor(0.1655, device='cuda:0', grad_fn=<DivBackward0>)
episode: 588 training return: tensor(0.0007, device='cuda:0', grad_fn=<AddBackward0>)
episode: 589 training return: tensor(2.5148e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 590 training return: tensor(0.0014, device='cuda:0', grad_fn=<AddBackward0>)
episode: 591 training return: tensor(0.0017, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 148 test_true_pfm: 4011 test_simulate_pfm tensor(0.0044, device='cuda:0', grad_fn=<DivBackward0>)
episode: 592 training return: tensor(0.0012, device='cuda:0', grad_fn=<AddBackward0>)
episode: 593 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 594 training return: tensor(2.0514e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 595 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 149 test_true_pfm: 1082 test_simulate_pfm tensor(0.0002, device='cuda:0', grad_fn=<DivBackward0>)
episode: 596 training return: tensor(5.7991e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 597 training return: tensor(8.8910e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 598 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
episode: 599 training return: tensor(4.5946e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 150 test_true_pfm: 1998 test_simulate_pfm tensor(0.0024, device='cuda:0', grad_fn=<DivBackward0>)
