episode: 0 training return: tensor(1.4687, device='cuda:0', grad_fn=<AddBackward0>)
episode: 1 training return: tensor(0.5228, device='cuda:0', grad_fn=<AddBackward0>)
episode: 2 training return: tensor(2.9089e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 3 training return: tensor(8.1215e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 1 test_pfm: -357
episode: 4 training return: tensor(1.4066e-16, device='cuda:0', grad_fn=<AddBackward0>)
episode: 5 training return: tensor(2.9903e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 6 training return: tensor(2.1449e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 7 training return: tensor(4.1070e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 2 test_pfm: -459
episode: 8 training return: tensor(2.6112e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 9 training return: tensor(0.0024, device='cuda:0', grad_fn=<AddBackward0>)
episode: 10 training return: tensor(8.0656e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 11 training return: tensor(1.1664e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 3 test_pfm: -267
episode: 12 training return: tensor(6.0773e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 13 training return: tensor(2.5966e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 14 training return: tensor(6.8603e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 15 training return: tensor(7.7094e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 4 test_pfm: -539
episode: 16 training return: tensor(5.7002e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 17 training return: tensor(1.3793e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 18 training return: tensor(3.7284e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 19 training return: tensor(5.5769e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 5 test_pfm: -638
episode: 20 training return: tensor(5.0153e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 21 training return: tensor(2.1308e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 22 training return: tensor(1.5691e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 23 training return: tensor(1.8221e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 6 test_pfm: -490
episode: 24 training return: tensor(5.3031e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 25 training return: tensor(5.0522e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 26 training return: tensor(2.7414e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 27 training return: tensor(1.1330e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 7 test_pfm: -538
episode: 28 training return: tensor(3.2568e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 29 training return: tensor(7.9049e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 30 training return: tensor(1.3649e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 31 training return: tensor(1.3363e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 8 test_pfm: 140
episode: 32 training return: tensor(8.9330e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 33 training return: tensor(1.8800e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 34 training return: tensor(7.0167e-15, device='cuda:0', grad_fn=<AddBackward0>)
episode: 35 training return: tensor(3.0717e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 9 test_pfm: 153
episode: 36 training return: tensor(1.8904e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 37 training return: tensor(1.7656e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 38 training return: tensor(1.3205e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 39 training return: tensor(6.0248e-13, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 10 test_pfm: -559
episode: 40 training return: tensor(1.1854e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 41 training return: tensor(5.8911e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 42 training return: tensor(1.5987e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 43 training return: tensor(1.2550e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 11 test_pfm: -704
episode: 44 training return: tensor(8.9038e-15, device='cuda:0', grad_fn=<AddBackward0>)
episode: 45 training return: tensor(4.6644e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 46 training return: tensor(3.0189e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 47 training return: tensor(1.8887e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 12 test_pfm: -311
episode: 48 training return: tensor(1.1594e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 49 training return: tensor(1.8692e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 50 training return: tensor(4.7404e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 51 training return: tensor(3.3830e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 13 test_pfm: -104
episode: 52 training return: tensor(1.1020e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 53 training return: tensor(2.9224e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 54 training return: tensor(1.0538e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 55 training return: tensor(1.3677e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 14 test_pfm: -191
episode: 56 training return: tensor(6.6034e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 57 training return: tensor(9.6243e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 58 training return: tensor(7.6074e-19, device='cuda:0', grad_fn=<AddBackward0>)
episode: 59 training return: tensor(5.7618e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 15 test_pfm: -374
episode: 60 training return: tensor(9.7546e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 61 training return: tensor(8.5458e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 62 training return: tensor(6.3125e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 63 training return: tensor(2.3401e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 16 test_pfm: -238
episode: 64 training return: tensor(7.0393e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 65 training return: tensor(6.3449e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 66 training return: tensor(6.8296e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 67 training return: tensor(1.1983e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 17 test_pfm: 419
episode: 68 training return: tensor(2.7124e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 69 training return: tensor(1.4498e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 70 training return: tensor(8.4061e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 71 training return: tensor(9.5120e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 18 test_pfm: 804
episode: 72 training return: tensor(1.2914e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 73 training return: tensor(5.2555e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 74 training return: tensor(4.3932e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 75 training return: tensor(3.7732e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 19 test_pfm: 710
episode: 76 training return: tensor(6.1565e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 77 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 78 training return: tensor(2.5562e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 79 training return: tensor(4.4839e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 20 test_pfm: 507
episode: 80 training return: tensor(1.5129e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 81 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 82 training return: tensor(7.4288e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 83 training return: tensor(1.4345e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 21 test_pfm: 808
episode: 84 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
episode: 85 training return: tensor(2.7178e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 86 training return: tensor(2.1340e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 87 training return: tensor(7.9692e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 22 test_pfm: 1056
episode: 88 training return: tensor(1.5651e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 89 training return: tensor(3.7675e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 90 training return: tensor(3.4444e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 91 training return: tensor(6.9933e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 23 test_pfm: 1214
episode: 92 training return: tensor(1.4701e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 93 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 94 training return: tensor(9.8547e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 95 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 24 test_pfm: 2427
episode: 96 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 97 training return: tensor(8.4153e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 98 training return: tensor(7.8657e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 99 training return: tensor(9.8093e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 25 test_pfm: 2337
episode: 100 training return: tensor(1.1789e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 101 training return: tensor(6.3612e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 102 training return: tensor(0.0803, device='cuda:0', grad_fn=<AddBackward0>)
episode: 103 training return: tensor(4.5306e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 26 test_pfm: 2154
episode: 104 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 105 training return: tensor(1.9257e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 106 training return: tensor(2.1981e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 107 training return: tensor(9.8818e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 27 test_pfm: 1267
episode: 108 training return: tensor(2.2587e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 109 training return: tensor(9.4084e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 110 training return: tensor(6.5454e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 111 training return: tensor(2.2829e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 28 test_pfm: 729
episode: 112 training return: tensor(4.1003e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 113 training return: tensor(2.0023e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 114 training return: tensor(6.8129e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 115 training return: tensor(6.6267e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 29 test_pfm: -52
episode: 116 training return: tensor(1.3276e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 117 training return: tensor(8.9733e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 118 training return: tensor(3.3609e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 119 training return: tensor(4.4775e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 30 test_pfm: -29
episode: 120 training return: tensor(5.0591e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 121 training return: tensor(1.4326e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 122 training return: tensor(1.4114e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 123 training return: tensor(3.5117e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 31 test_pfm: -121
episode: 124 training return: tensor(3.9489e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 125 training return: tensor(3.0782e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 126 training return: tensor(3.1866e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 127 training return: tensor(1.7651e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 32 test_pfm: 814
episode: 128 training return: tensor(3.8334e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 129 training return: tensor(3.2667e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 130 training return: tensor(8.2931e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 131 training return: tensor(9.9385e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 33 test_pfm: 518
episode: 132 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 133 training return: tensor(6.0543e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 134 training return: tensor(6.6653e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 135 training return: tensor(3.5018e-13, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 34 test_pfm: 1252
episode: 136 training return: tensor(0.0014, device='cuda:0', grad_fn=<AddBackward0>)
episode: 137 training return: tensor(4.9504e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 138 training return: tensor(1.4267e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 139 training return: tensor(2.9942e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 35 test_pfm: 1031
episode: 140 training return: tensor(1.3931e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 141 training return: tensor(8.4087e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 142 training return: tensor(3.9933e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 143 training return: tensor(4.3449e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 36 test_pfm: 1135
episode: 144 training return: tensor(1.2130e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 145 training return: tensor(9.0236e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 146 training return: tensor(1.3694e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 147 training return: tensor(5.9475e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 37 test_pfm: 5
episode: 148 training return: tensor(2.6775e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 149 training return: tensor(3.5587e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 150 training return: tensor(2.0688e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 151 training return: tensor(1.6426e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 38 test_pfm: 919
episode: 152 training return: tensor(1.6420e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 153 training return: tensor(8.5261e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 154 training return: tensor(3.1906e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 155 training return: tensor(2.0218e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 39 test_pfm: 884
episode: 156 training return: tensor(4.1007e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 157 training return: tensor(2.9022e-16, device='cuda:0', grad_fn=<AddBackward0>)
episode: 158 training return: tensor(3.1456e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 159 training return: tensor(1.2746e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 40 test_pfm: 1300
episode: 160 training return: tensor(7.2766e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 161 training return: tensor(4.8060e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 162 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 163 training return: tensor(5.9317e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 41 test_pfm: 1444
episode: 164 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
episode: 165 training return: tensor(2.5730e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 166 training return: tensor(2.3070e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 167 training return: tensor(3.9566e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 42 test_pfm: 690
episode: 168 training return: tensor(1.1614e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 169 training return: tensor(6.8676e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 170 training return: tensor(1.3262e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 171 training return: tensor(2.0737e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 43 test_pfm: 725
episode: 172 training return: tensor(1.5726e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 173 training return: tensor(6.1703e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 174 training return: tensor(1.2178e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 175 training return: tensor(7.7126e-13, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 44 test_pfm: 125
episode: 176 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 177 training return: tensor(7.1889e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 178 training return: tensor(3.6115e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 179 training return: tensor(3.4956e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 45 test_pfm: 901
episode: 180 training return: tensor(2.1531e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 181 training return: tensor(1.9299e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 182 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
episode: 183 training return: tensor(8.5945e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 46 test_pfm: 746
episode: 184 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 185 training return: tensor(8.0643e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 186 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 187 training return: tensor(5.5859e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 47 test_pfm: 1290
episode: 188 training return: tensor(5.6986e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 189 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 190 training return: tensor(6.4053e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 191 training return: tensor(2.1230e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 48 test_pfm: 1596
episode: 192 training return: tensor(4.0499e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 193 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 194 training return: tensor(1.2513e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 195 training return: tensor(5.4970e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 49 test_pfm: 1995
episode: 196 training return: tensor(0.0041, device='cuda:0', grad_fn=<AddBackward0>)
episode: 197 training return: tensor(0.0006, device='cuda:0', grad_fn=<AddBackward0>)
episode: 198 training return: tensor(2.9024e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 199 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 50 test_pfm: 1966
episode: 200 training return: tensor(0.0014, device='cuda:0', grad_fn=<AddBackward0>)
episode: 201 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
episode: 202 training return: tensor(7.4927e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 203 training return: tensor(4.6114e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 51 test_pfm: 1765
episode: 204 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 205 training return: tensor(6.5793e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 206 training return: tensor(6.2872e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 207 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 52 test_pfm: 1320
episode: 208 training return: tensor(1.8620e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 209 training return: tensor(7.5204e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 210 training return: tensor(1.2646e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 211 training return: tensor(5.5115e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 53 test_pfm: 1442
episode: 212 training return: tensor(3.2693e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 213 training return: tensor(1.3219e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 214 training return: tensor(7.8409e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 215 training return: tensor(5.3294e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 54 test_pfm: 189
episode: 216 training return: tensor(1.4426e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 217 training return: tensor(1.1582e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 218 training return: tensor(1.9731e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 219 training return: tensor(8.7424e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 55 test_pfm: 897
episode: 220 training return: tensor(1.9146e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 221 training return: tensor(1.8953e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 222 training return: tensor(4.9609e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 223 training return: tensor(3.2824e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 56 test_pfm: 1384
episode: 224 training return: tensor(1.4466e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 225 training return: tensor(4.5181e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 226 training return: tensor(3.8373e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 227 training return: tensor(5.2243e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 57 test_pfm: -91
episode: 228 training return: tensor(1.6923e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 229 training return: tensor(1.2522e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 230 training return: tensor(5.7004e-19, device='cuda:0', grad_fn=<AddBackward0>)
episode: 231 training return: tensor(8.2774e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 58 test_pfm: -512
episode: 232 training return: tensor(8.7743e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 233 training return: tensor(3.9958e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 234 training return: tensor(1.0512e-18, device='cuda:0', grad_fn=<AddBackward0>)
episode: 235 training return: tensor(5.1267e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 59 test_pfm: -548
episode: 236 training return: tensor(1.2696e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 237 training return: tensor(2.6423e-18, device='cuda:0', grad_fn=<AddBackward0>)
episode: 238 training return: tensor(7.6079e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 239 training return: tensor(2.5810e-15, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 60 test_pfm: -527
episode: 240 training return: tensor(8.8469e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 241 training return: tensor(1.7535e-17, device='cuda:0', grad_fn=<AddBackward0>)
episode: 242 training return: tensor(3.7064e-17, device='cuda:0', grad_fn=<AddBackward0>)
episode: 243 training return: tensor(3.1820e-14, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 61 test_pfm: -565
episode: 244 training return: tensor(4.9688e-15, device='cuda:0', grad_fn=<AddBackward0>)
episode: 245 training return: tensor(2.2995e-15, device='cuda:0', grad_fn=<AddBackward0>)
episode: 246 training return: tensor(1.0164e-18, device='cuda:0', grad_fn=<AddBackward0>)
episode: 247 training return: tensor(2.8435e-15, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 62 test_pfm: -573
episode: 248 training return: tensor(3.2932e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 249 training return: tensor(3.1455e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 250 training return: tensor(7.6921e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 251 training return: tensor(3.8363e-16, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 63 test_pfm: -542
episode: 252 training return: tensor(5.7417e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 253 training return: tensor(4.8775e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 254 training return: tensor(5.7274e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 255 training return: tensor(6.4371e-13, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 64 test_pfm: -460
episode: 256 training return: tensor(2.1384e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 257 training return: tensor(2.3434e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 258 training return: tensor(3.0009e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 259 training return: tensor(1.0632e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 65 test_pfm: -280
episode: 260 training return: tensor(4.5031e-08, device='cuda:0', grad_fn=<AddBackward0>)
