initial performance: 10330
episode: 0 training return: tensor(117.3437, device='cuda:0', grad_fn=<AddBackward0>)
episode: 1 training return: tensor(0.0033, device='cuda:0', grad_fn=<AddBackward0>)
episode: 2 training return: tensor(9.7161e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 3 training return: tensor(2.5476e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 1 test_pfm: -832
episode: 4 training return: tensor(2.6284e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 5 training return: tensor(3.4203e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 6 training return: tensor(4.8770e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 7 training return: tensor(3.7297e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 2 test_pfm: 1016
episode: 8 training return: tensor(1.2299e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 9 training return: tensor(3.5302e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 10 training return: tensor(4.5944e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 11 training return: tensor(2.0775e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 3 test_pfm: -177
episode: 12 training return: tensor(1.5068e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 13 training return: tensor(1.1663e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 14 training return: tensor(2.5048e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 15 training return: tensor(3.2521e-13, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 4 test_pfm: -434
episode: 16 training return: tensor(9.6446e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 17 training return: tensor(9.7274e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 18 training return: tensor(1.1040e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 19 training return: tensor(8.0351e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 5 test_pfm: -436
episode: 20 training return: tensor(3.8205e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 21 training return: tensor(4.8871e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 22 training return: tensor(2.0927e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 23 training return: tensor(6.5884e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 6 test_pfm: -337
episode: 24 training return: tensor(1.9006e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 25 training return: tensor(6.4407e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 26 training return: tensor(1.5012e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 27 training return: tensor(3.1591e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 7 test_pfm: -576
episode: 28 training return: tensor(8.0255e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 29 training return: tensor(1.4101e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 30 training return: tensor(3.6485e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 31 training return: tensor(9.2403e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 8 test_pfm: -526
episode: 32 training return: tensor(1.0929e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 33 training return: tensor(1.6801e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 34 training return: tensor(7.9039e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 35 training return: tensor(2.2276e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 9 test_pfm: -580
episode: 36 training return: tensor(1.8968e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 37 training return: tensor(2.8216e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 38 training return: tensor(1.0803e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 39 training return: tensor(6.6898e-13, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 10 test_pfm: -485
episode: 40 training return: tensor(4.9869e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 41 training return: tensor(6.0881e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 42 training return: tensor(5.9005e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 43 training return: tensor(5.2768e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 11 test_pfm: -404
episode: 44 training return: tensor(5.1856e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 45 training return: tensor(7.9185e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 46 training return: tensor(3.7443e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 47 training return: tensor(5.4315e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 12 test_pfm: -549
episode: 48 training return: tensor(4.9399e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 49 training return: tensor(2.0124e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 50 training return: tensor(1.0348e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 51 training return: tensor(3.0380e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 13 test_pfm: -395
episode: 52 training return: tensor(1.9205e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 53 training return: tensor(6.8820e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 54 training return: tensor(3.0207e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 55 training return: tensor(2.5733e-13, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 14 test_pfm: -254
episode: 56 training return: tensor(7.6857e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 57 training return: tensor(1.0305e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 58 training return: tensor(1.5946e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 59 training return: tensor(5.7601e-14, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 15 test_pfm: 31
episode: 60 training return: tensor(1.0625e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 61 training return: tensor(1.5433e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 62 training return: tensor(9.7744e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 63 training return: tensor(2.3260e-13, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 16 test_pfm: -339
episode: 64 training return: tensor(3.0844e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 65 training return: tensor(7.0392e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 66 training return: tensor(1.6084e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 67 training return: tensor(1.4195e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 17 test_pfm: -100
episode: 68 training return: tensor(1.2857e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 69 training return: tensor(0.0017, device='cuda:0', grad_fn=<AddBackward0>)
episode: 70 training return: tensor(2.2217e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 71 training return: tensor(2.4033e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 18 test_pfm: 77
episode: 72 training return: tensor(2.2285e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 73 training return: tensor(1.6565e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 74 training return: tensor(2.4739e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 75 training return: tensor(2.5830e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 19 test_pfm: -75
episode: 76 training return: tensor(0.0049, device='cuda:0', grad_fn=<AddBackward0>)
episode: 77 training return: tensor(1.2298e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 78 training return: tensor(3.4917e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 79 training return: tensor(1.7616e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 20 test_pfm: -290
episode: 80 training return: tensor(9.3436e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 81 training return: tensor(1.4057e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 82 training return: tensor(3.8825e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 83 training return: tensor(4.0131e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 21 test_pfm: -366
episode: 84 training return: tensor(2.2628e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 85 training return: tensor(8.6286e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 86 training return: tensor(8.0196e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 87 training return: tensor(7.8021e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 22 test_pfm: -121
episode: 88 training return: tensor(2.0624e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 89 training return: tensor(8.4250e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 90 training return: tensor(4.5329e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 91 training return: tensor(1.1426e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 23 test_pfm: -237
episode: 92 training return: tensor(9.2771e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 93 training return: tensor(1.1593e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 94 training return: tensor(7.3901e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 95 training return: tensor(1.0267e-14, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 24 test_pfm: 120
episode: 96 training return: tensor(2.4170e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 97 training return: tensor(6.3367e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 98 training return: tensor(2.5792e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 99 training return: tensor(3.5801e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 25 test_pfm: -188
episode: 100 training return: tensor(4.3613e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 101 training return: tensor(7.6969e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 102 training return: tensor(3.2885e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 103 training return: tensor(1.3242e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 26 test_pfm: 53
episode: 104 training return: tensor(1.9553e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 105 training return: tensor(4.3149e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 106 training return: tensor(2.0468e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 107 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 27 test_pfm: 165
episode: 108 training return: tensor(5.3450e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 109 training return: tensor(1.3597e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 110 training return: tensor(1.2160e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 111 training return: tensor(2.8525e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 28 test_pfm: 401
episode: 112 training return: tensor(5.5324e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 113 training return: tensor(3.8063e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 114 training return: tensor(4.0744e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 115 training return: tensor(0.0006, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 29 test_pfm: -18
episode: 116 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 117 training return: tensor(6.8760e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 118 training return: tensor(3.6609e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 119 training return: tensor(9.2612e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 30 test_pfm: 313
episode: 120 training return: tensor(0.0013, device='cuda:0', grad_fn=<AddBackward0>)
episode: 121 training return: tensor(6.8610e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 122 training return: tensor(9.0537e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 123 training return: tensor(8.5298e-13, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 31 test_pfm: 246
episode: 124 training return: tensor(1.5927e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 125 training return: tensor(2.7761e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 126 training return: tensor(1.1208e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 127 training return: tensor(4.1162e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 32 test_pfm: 94
episode: 128 training return: tensor(2.0966e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 129 training return: tensor(8.0971e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 130 training return: tensor(1.1042e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 131 training return: tensor(4.0865e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 33 test_pfm: 499
episode: 132 training return: tensor(4.7201e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 133 training return: tensor(8.5147e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 134 training return: tensor(3.2620e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 135 training return: tensor(8.8504e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 34 test_pfm: 179
episode: 136 training return: tensor(1.5636e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 137 training return: tensor(1.1186e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 138 training return: tensor(8.1026e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 139 training return: tensor(3.3371e-13, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 35 test_pfm: -259
episode: 140 training return: tensor(2.4290e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 141 training return: tensor(5.7499e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 142 training return: tensor(1.0098e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 143 training return: tensor(8.1856e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 36 test_pfm: -193
episode: 144 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 145 training return: tensor(4.8600e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 146 training return: tensor(4.0424e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 147 training return: tensor(3.8258e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 37 test_pfm: -328
episode: 148 training return: tensor(4.3282e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 149 training return: tensor(4.7250e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 150 training return: tensor(4.8398e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 151 training return: tensor(4.7549e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 38 test_pfm: -375
episode: 152 training return: tensor(1.6215e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 153 training return: tensor(6.1031e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 154 training return: tensor(1.8844e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 155 training return: tensor(2.3501e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 39 test_pfm: -90
episode: 156 training return: tensor(1.2495e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 157 training return: tensor(8.7534e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 158 training return: tensor(6.4628e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 159 training return: tensor(5.2042e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 40 test_pfm: 136
episode: 160 training return: tensor(2.6508e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 161 training return: tensor(2.0093e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 162 training return: tensor(5.8938e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 163 training return: tensor(1.0428e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 41 test_pfm: -44
episode: 164 training return: tensor(1.2255e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 165 training return: tensor(1.0143e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 166 training return: tensor(1.2040e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 167 training return: tensor(5.5460e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 42 test_pfm: -332
episode: 168 training return: tensor(2.8078e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 169 training return: tensor(2.5815e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 170 training return: tensor(7.9144e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 171 training return: tensor(1.4410e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 43 test_pfm: -148
episode: 172 training return: tensor(2.0330e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 173 training return: tensor(9.8139e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 174 training return: tensor(3.2822e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 175 training return: tensor(2.5150e-12, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 44 test_pfm: -289
episode: 176 training return: tensor(1.1362e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 177 training return: tensor(1.3376e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 178 training return: tensor(6.5640e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 179 training return: tensor(1.6799e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 45 test_pfm: -14
episode: 180 training return: tensor(4.0821e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 181 training return: tensor(1.1708e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 182 training return: tensor(4.9068e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 183 training return: tensor(1.9501e-11, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 46 test_pfm: 96
episode: 184 training return: tensor(1.9090e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 185 training return: tensor(3.5062e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 186 training return: tensor(5.1346e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 187 training return: tensor(1.4826e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 47 test_pfm: -90
episode: 188 training return: tensor(2.7470e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 189 training return: tensor(2.9640e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 190 training return: tensor(3.3531e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 191 training return: tensor(8.4592e-10, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 48 test_pfm: 110
episode: 192 training return: tensor(8.5564e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 193 training return: tensor(1.3660e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 194 training return: tensor(3.7085e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 195 training return: tensor(1.6784e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 49 test_pfm: 641
episode: 196 training return: tensor(6.8443e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 197 training return: tensor(5.3822e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 198 training return: tensor(8.2212e-08, device='cuda:0', grad_fn=<AddBackward0>)
episode: 199 training return: tensor(9.6312e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 50 test_pfm: 673
episode: 200 training return: tensor(1.3456e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 201 training return: tensor(2.9168e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 202 training return: tensor(9.3418e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 203 training return: tensor(2.7108e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 51 test_pfm: 906
episode: 204 training return: tensor(7.8980e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 205 training return: tensor(3.7850e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 206 training return: tensor(2.7659e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 207 training return: tensor(0.0009, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 52 test_pfm: 1005
episode: 208 training return: tensor(4.3411e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 209 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
episode: 210 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 211 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 53 test_pfm: 965
episode: 212 training return: tensor(0.0020, device='cuda:0', grad_fn=<AddBackward0>)
episode: 213 training return: tensor(1.6211e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 214 training return: tensor(4.1592e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 215 training return: tensor(9.4410e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 54 test_pfm: 728
episode: 216 training return: tensor(0.0073, device='cuda:0', grad_fn=<AddBackward0>)
episode: 217 training return: tensor(3.4641e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 218 training return: tensor(2.5561e-14, device='cuda:0', grad_fn=<AddBackward0>)
episode: 219 training return: tensor(4.3661e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 55 test_pfm: 248
episode: 220 training return: tensor(7.7971e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 221 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 222 training return: tensor(0.0025, device='cuda:0', grad_fn=<AddBackward0>)
episode: 223 training return: tensor(0.0023, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 56 test_pfm: 318
episode: 224 training return: tensor(7.3926e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 225 training return: tensor(8.8033e-15, device='cuda:0', grad_fn=<AddBackward0>)
episode: 226 training return: tensor(5.9834e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 227 training return: tensor(3.9113e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 57 test_pfm: 209
episode: 228 training return: tensor(1.5401e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 229 training return: tensor(4.2550e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 230 training return: tensor(4.4956e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 231 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 58 test_pfm: 480
episode: 232 training return: tensor(0.0061, device='cuda:0', grad_fn=<AddBackward0>)
episode: 233 training return: tensor(6.8607e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 234 training return: tensor(4.5608e-11, device='cuda:0', grad_fn=<AddBackward0>)
episode: 235 training return: tensor(3.9355e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 59 test_pfm: 558
episode: 236 training return: tensor(1.4192e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 237 training return: tensor(5.8402e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 238 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 239 training return: tensor(3.0376e-16, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 60 test_pfm: 594
episode: 240 training return: tensor(4.6098e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 241 training return: tensor(5.5694e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 242 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 243 training return: tensor(1.7707e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 61 test_pfm: 634
episode: 244 training return: tensor(2.9922e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 245 training return: tensor(1.0021e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 246 training return: tensor(1.3784e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 247 training return: tensor(1.3977e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 62 test_pfm: 460
episode: 248 training return: tensor(8.5532e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 249 training return: tensor(1.2531e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 250 training return: tensor(1.3572e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 251 training return: tensor(2.6105e-06, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 63 test_pfm: 695
episode: 252 training return: tensor(2.5181e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 253 training return: tensor(9.4072e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 254 training return: tensor(1.3455e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 255 training return: tensor(6.5884e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 64 test_pfm: 89
episode: 256 training return: tensor(1.2304e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 257 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 258 training return: tensor(1.3437e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 259 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 65 test_pfm: 318
episode: 260 training return: tensor(1.7367e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 261 training return: tensor(3.6130e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 262 training return: tensor(1.6511e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 263 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 66 test_pfm: 640
episode: 264 training return: tensor(2.0425e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 265 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 266 training return: tensor(8.6436e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 267 training return: tensor(0.0006, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 67 test_pfm: 912
episode: 268 training return: tensor(5.6670e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 269 training return: tensor(7.0798e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 270 training return: tensor(2.4928e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 271 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 68 test_pfm: 797
episode: 272 training return: tensor(1.3293e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 273 training return: tensor(3.5501e-12, device='cuda:0', grad_fn=<AddBackward0>)
episode: 274 training return: tensor(6.6476e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 275 training return: tensor(4.8425e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 69 test_pfm: 907
episode: 276 training return: tensor(1.3300e-06, device='cuda:0', grad_fn=<AddBackward0>)
episode: 277 training return: tensor(9.7620e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 278 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 279 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 70 test_pfm: 848
episode: 280 training return: tensor(3.0324e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 281 training return: tensor(4.2552e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 282 training return: tensor(0.0094, device='cuda:0', grad_fn=<AddBackward0>)
episode: 283 training return: tensor(0.0079, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 71 test_pfm: 1104
episode: 284 training return: tensor(0.0007, device='cuda:0', grad_fn=<AddBackward0>)
episode: 285 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 286 training return: tensor(7.7415e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 287 training return: tensor(0.0008, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 72 test_pfm: 915
episode: 288 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 289 training return: tensor(8.1271e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 290 training return: tensor(1.0446e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 291 training return: tensor(0.0007, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 73 test_pfm: 1590
episode: 292 training return: tensor(0.0013, device='cuda:0', grad_fn=<AddBackward0>)
episode: 293 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 294 training return: tensor(6.5132e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 295 training return: tensor(6.9472e-08, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 74 test_pfm: 1914
episode: 296 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 297 training return: tensor(0.0024, device='cuda:0', grad_fn=<AddBackward0>)
episode: 298 training return: tensor(0.0010, device='cuda:0', grad_fn=<AddBackward0>)
episode: 299 training return: tensor(1.7090e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 75 test_pfm: 1688
episode: 300 training return: tensor(6.2101e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 301 training return: tensor(0.0007, device='cuda:0', grad_fn=<AddBackward0>)
episode: 302 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 303 training return: tensor(5.7113e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 76 test_pfm: 1388
episode: 304 training return: tensor(1.4166e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 305 training return: tensor(0.0015, device='cuda:0', grad_fn=<AddBackward0>)
episode: 306 training return: tensor(0.0008, device='cuda:0', grad_fn=<AddBackward0>)
episode: 307 training return: tensor(0.0010, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 77 test_pfm: 1284
episode: 308 training return: tensor(9.1086e-13, device='cuda:0', grad_fn=<AddBackward0>)
episode: 309 training return: tensor(2.6644e-10, device='cuda:0', grad_fn=<AddBackward0>)
episode: 310 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 311 training return: tensor(0.0018, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 78 test_pfm: 1570
episode: 312 training return: tensor(0.0218, device='cuda:0', grad_fn=<AddBackward0>)
episode: 313 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 314 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 315 training return: tensor(0.0044, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 79 test_pfm: 1364
episode: 316 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 317 training return: tensor(0.0017, device='cuda:0', grad_fn=<AddBackward0>)
episode: 318 training return: tensor(0.0389, device='cuda:0', grad_fn=<AddBackward0>)
episode: 319 training return: tensor(4.3637e-09, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 80 test_pfm: 3032
episode: 320 training return: tensor(0.0017, device='cuda:0', grad_fn=<AddBackward0>)
episode: 321 training return: tensor(0.0021, device='cuda:0', grad_fn=<AddBackward0>)
episode: 322 training return: tensor(0.0022, device='cuda:0', grad_fn=<AddBackward0>)
episode: 323 training return: tensor(0.0022, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 81 test_pfm: 2560
episode: 324 training return: tensor(0.0013, device='cuda:0', grad_fn=<AddBackward0>)
episode: 325 training return: tensor(0.0014, device='cuda:0', grad_fn=<AddBackward0>)
episode: 326 training return: tensor(0.0034, device='cuda:0', grad_fn=<AddBackward0>)
episode: 327 training return: tensor(0.0006, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 82 test_pfm: 1764
episode: 328 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
episode: 329 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 330 training return: tensor(0.0013, device='cuda:0', grad_fn=<AddBackward0>)
episode: 331 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 83 test_pfm: 1639
episode: 332 training return: tensor(0.0006, device='cuda:0', grad_fn=<AddBackward0>)
episode: 333 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 334 training return: tensor(0.0024, device='cuda:0', grad_fn=<AddBackward0>)
episode: 335 training return: tensor(3.1367e-07, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 84 test_pfm: 1773
episode: 336 training return: tensor(0.0108, device='cuda:0', grad_fn=<AddBackward0>)
episode: 337 training return: tensor(5.8344e-09, device='cuda:0', grad_fn=<AddBackward0>)
episode: 338 training return: tensor(0.0007, device='cuda:0', grad_fn=<AddBackward0>)
episode: 339 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 85 test_pfm: 2834
episode: 340 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 341 training return: tensor(0.0037, device='cuda:0', grad_fn=<AddBackward0>)
episode: 342 training return: tensor(1.8583e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 343 training return: tensor(0.0009, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 86 test_pfm: 2661
episode: 344 training return: tensor(0.0216, device='cuda:0', grad_fn=<AddBackward0>)
episode: 345 training return: tensor(0.0044, device='cuda:0', grad_fn=<AddBackward0>)
episode: 346 training return: tensor(0.0024, device='cuda:0', grad_fn=<AddBackward0>)
episode: 347 training return: tensor(0.0078, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 87 test_pfm: 3009
episode: 348 training return: tensor(0.0007, device='cuda:0', grad_fn=<AddBackward0>)
episode: 349 training return: tensor(0.0015, device='cuda:0', grad_fn=<AddBackward0>)
episode: 350 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 351 training return: tensor(0.0033, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 88 test_pfm: 2774
episode: 352 training return: tensor(0.0036, device='cuda:0', grad_fn=<AddBackward0>)
episode: 353 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
episode: 354 training return: tensor(0.0321, device='cuda:0', grad_fn=<AddBackward0>)
episode: 355 training return: tensor(0.0085, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 89 test_pfm: 3546
episode: 356 training return: tensor(0.0027, device='cuda:0', grad_fn=<AddBackward0>)
episode: 357 training return: tensor(0.0008, device='cuda:0', grad_fn=<AddBackward0>)
episode: 358 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
episode: 359 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 90 test_pfm: 3072
episode: 360 training return: tensor(0.0012, device='cuda:0', grad_fn=<AddBackward0>)
episode: 361 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 362 training return: tensor(7.1243e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 363 training return: tensor(7.7110e-05, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 91 test_pfm: 1883
episode: 364 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
episode: 365 training return: tensor(0.0008, device='cuda:0', grad_fn=<AddBackward0>)
episode: 366 training return: tensor(7.5010e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 367 training return: tensor(0.0046, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 92 test_pfm: 3876
episode: 368 training return: tensor(0.0048, device='cuda:0', grad_fn=<AddBackward0>)
episode: 369 training return: tensor(0.0062, device='cuda:0', grad_fn=<AddBackward0>)
episode: 370 training return: tensor(0.0007, device='cuda:0', grad_fn=<AddBackward0>)
episode: 371 training return: tensor(0.0006, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 93 test_pfm: 3440
episode: 372 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
episode: 373 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 374 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
episode: 375 training return: tensor(0.0028, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 94 test_pfm: 2451
episode: 376 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 377 training return: tensor(1.0136e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 378 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
episode: 379 training return: tensor(0.0040, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 95 test_pfm: 3936
episode: 380 training return: tensor(0.0035, device='cuda:0', grad_fn=<AddBackward0>)
episode: 381 training return: tensor(0.0008, device='cuda:0', grad_fn=<AddBackward0>)
episode: 382 training return: tensor(0.0076, device='cuda:0', grad_fn=<AddBackward0>)
episode: 383 training return: tensor(0.0038, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 96 test_pfm: 3549
episode: 384 training return: tensor(4.0244e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 385 training return: tensor(2.2494e-07, device='cuda:0', grad_fn=<AddBackward0>)
episode: 386 training return: tensor(2.6861e-05, device='cuda:0', grad_fn=<AddBackward0>)
episode: 387 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 97 test_pfm: 3375
episode: 388 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
episode: 389 training return: tensor(0.0007, device='cuda:0', grad_fn=<AddBackward0>)
episode: 390 training return: tensor(0.0006, device='cuda:0', grad_fn=<AddBackward0>)
episode: 391 training return: tensor(0.0001, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 98 test_pfm: 3684
episode: 392 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
episode: 393 training return: tensor(0.0028, device='cuda:0', grad_fn=<AddBackward0>)
episode: 394 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
episode: 395 training return: tensor(0.0004, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 99 test_pfm: 3763
episode: 396 training return: tensor(0.0002, device='cuda:0', grad_fn=<AddBackward0>)
episode: 397 training return: tensor(0.0012, device='cuda:0', grad_fn=<AddBackward0>)
episode: 398 training return: tensor(0.0003, device='cuda:0', grad_fn=<AddBackward0>)
episode: 399 training return: tensor(0.0005, device='cuda:0', grad_fn=<AddBackward0>)
epoch: 100 test_pfm: 2166
