API tools faq
paste
Advertisement
lamiastella

landmark predictions

lamiastella
Oct 14th, 2020 (edited)
247
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 139.56 KB | None | 0 0
raw download clone embed print report
  1. size of train loader is: 90
  2. predictions are: tensor([[-0.0380, -0.1871, 0.0729, -0.3570, -0.2153, 0.3066, 1.1273, -0.0558],
  3. [-0.0316, -0.1876, 0.0317, -0.3613, -0.2333, 0.3023, 1.0940, -0.0665],
  4. [-0.0700, -0.1882, 0.0068, -0.3201, -0.1884, 0.2953, 1.0516, -0.0567],
  5. [-0.0844, -0.2009, 0.0573, -0.3166, -0.2597, 0.3127, 1.0343, -0.0573],
  6. [-0.0486, -0.2333, 0.0535, -0.3245, -0.2310, 0.2818, 1.0590, -0.0716],
  7. [-0.0240, -0.1989, 0.0572, -0.3135, -0.2435, 0.2912, 1.0612, -0.0560],
  8. [-0.0942, -0.2439, 0.0277, -0.3147, -0.2368, 0.2978, 1.0110, -0.0874],
  9. [-0.0356, -0.2285, 0.0064, -0.3179, -0.2432, 0.3083, 1.0300, -0.0756]],
  10. device='cuda:0', grad_fn=<AddmmBackward>)
  11. landmarks are: tensor([[501.9200, 240.1600, 691.0000, 358.0000, 295.0000, 294.0000, 488.6482,
  12. 279.6466],
  13. [495.6300, 246.0600, 692.0000, 235.0000, 286.0000, 242.0000, 464.0000,
  14. 339.0000],
  15. [488.7100, 240.8900, 613.4007, 218.3425, 281.0000, 220.0000, 415.9966,
  16. 338.4796],
  17. [502.5721, 245.4983, 640.0000, 131.0000, 360.0000, 143.0000, 542.9840,
  18. 321.8463],
  19. [505.1393, 246.4364, 700.0000, 306.0000, 303.0000, 294.0000, 569.6925,
  20. 351.8367],
  21. [501.0900, 244.0100, 724.0000, 251.0000, 302.0000, 276.0000, 504.6415,
  22. 291.7443],
  23. [495.9500, 244.2800, 608.0000, 127.0000, 323.0000, 166.0000, 491.0000,
  24. 333.0000],
  25. [490.2500, 241.3400, 699.0000, 304.0000, 398.6197, 313.8339, 429.1374,
  26. 303.8483]], device='cuda:0')
  27. loss_train_step before backward: tensor(166475.6875, device='cuda:0', grad_fn=<MseLossBackward>)
  28. loss_train_step after backward: tensor(166475.6875, device='cuda:0', grad_fn=<MseLossBackward>)
  29. loss_train: 166475.6875
  30. step: 1
  31. running loss: 166475.6875
  32. Train Steps: 1/90 Loss: 166475.6875 predictions are: tensor([[ 0.1848, -0.1262, 0.4091, -0.2304, -0.1060, 0.4372, 1.3852, 0.0739],
  33. [ 0.2051, -0.0935, 0.4141, -0.2483, -0.0607, 0.4669, 1.4023, 0.1099],
  34. [ 0.1923, -0.1570, 0.3770, -0.2265, -0.0891, 0.4168, 1.3499, 0.0978],
  35. [ 0.1799, -0.0961, 0.3741, -0.2188, -0.0891, 0.4474, 1.3451, 0.1094],
  36. [ 0.1413, -0.0899, 0.3942, -0.2284, -0.1005, 0.4032, 1.3716, 0.0928],
  37. [ 0.1852, -0.1177, 0.3888, -0.2247, -0.0988, 0.4586, 1.3553, 0.0956],
  38. [ 0.1285, -0.1051, 0.3694, -0.2425, -0.1180, 0.3835, 1.3341, 0.1009],
  39. [ 0.1393, -0.1126, 0.3555, -0.2547, -0.0628, 0.4297, 1.3280, 0.1162]],
  40. device='cuda:0', grad_fn=<AddmmBackward>)
  41. landmarks are: tensor([[506.5656, 247.0619, 739.0000, 256.0000, 321.0000, 284.0000, 601.9836,
  42. 326.1537],
  43. [507.1360, 244.8729, 674.0000, 325.0000, 308.0000, 290.0000, 586.7878,
  44. 345.5895],
  45. [491.8400, 243.1800, 700.0000, 273.0000, 387.9706, 313.0692, 469.0000,
  46. 334.0000],
  47. [486.7600, 240.2900, 672.0000, 259.0000, 301.0000, 285.0000, 438.0412,
  48. 303.4578],
  49. [503.9984, 248.0000, 683.0000, 130.0000, 447.0000, 135.0000, 591.2198,
  50. 324.7654],
  51. [490.2500, 241.3400, 699.0000, 304.0000, 398.6197, 313.8339, 429.1374,
  52. 303.8483],
  53. [ nan, nan, 682.0000, 133.0000, 433.0000, 142.0000, 589.3204,
  54. 328.9302],
  55. [489.1657, 239.8694, 565.0000, 143.0000, 323.0000, 117.0000, 425.5759,
  56. 299.5533]], device='cuda:0')
  57. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  58. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  59. loss_train: nan
  60. step: 2
  61. running loss: nan
  62. Train Steps: 2/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  63. [nan, nan, nan, nan, nan, nan, nan, nan],
  64. [nan, nan, nan, nan, nan, nan, nan, nan],
  65. [nan, nan, nan, nan, nan, nan, nan, nan],
  66. [nan, nan, nan, nan, nan, nan, nan, nan],
  67. [nan, nan, nan, nan, nan, nan, nan, nan],
  68. [nan, nan, nan, nan, nan, nan, nan, nan],
  69. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  70. grad_fn=<AddmmBackward>)
  71. landmarks are: tensor([[ nan, nan, 586.7780, 154.1319, 303.0000, 160.0000, 405.3121,
  72. 334.6727],
  73. [503.4279, 238.9313, 696.0000, 318.0000, 301.0000, 283.0000, 563.9941,
  74. 317.8241],
  75. [493.4800, 246.5000, 545.7030, 163.3048, 306.0000, 153.0000, 444.0000,
  76. 343.0000],
  77. [503.1426, 241.1203, 649.0000, 328.0000, 310.0000, 301.0000, 589.3204,
  78. 319.9065],
  79. [502.8574, 238.6186, 723.0000, 284.0000, 312.0000, 249.0000, 565.8936,
  80. 319.2123],
  81. [497.2200, 247.1100, 615.0000, 138.0000, 336.0000, 137.0000, 474.0000,
  82. 319.0000],
  83. [495.3700, 238.8200, 566.2405, 164.9726, 340.0000, 126.0000, 436.0000,
  84. 347.0000],
  85. [507.9918, 247.0619, 669.0000, 163.0000, 388.0000, 102.0000, 515.2408,
  86. 310.1886]], device='cuda:0')
  87. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  88. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  89. loss_train: nan
  90. step: 3
  91. running loss: nan
  92. Train Steps: 3/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  93. [nan, nan, nan, nan, nan, nan, nan, nan],
  94. [nan, nan, nan, nan, nan, nan, nan, nan],
  95. [nan, nan, nan, nan, nan, nan, nan, nan],
  96. [nan, nan, nan, nan, nan, nan, nan, nan],
  97. [nan, nan, nan, nan, nan, nan, nan, nan],
  98. [nan, nan, nan, nan, nan, nan, nan, nan],
  99. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  100. grad_fn=<AddmmBackward>)
  101. landmarks are: tensor([[496.2400, 244.3000, 711.3916, 211.3691, 302.6665, 181.0179, 471.8022,
  102. 328.6678],
  103. [501.1459, 243.3093, 674.0000, 166.0000, 354.0000, 166.0000, 563.9941,
  104. 335.1775],
  105. [488.6800, 242.1600, 575.0000, 105.0000, 308.0000, 153.0000, 469.0000,
  106. 334.0000],
  107. [502.8574, 243.6220, 735.0000, 260.0000, 294.0000, 250.0000, 562.7277,
  108. 331.7068],
  109. [502.2400, 255.1600, 710.0000, 301.0000, 329.0000, 165.0000, 433.0551,
  110. 371.7868],
  111. [490.0000, 239.9400, 700.0000, 293.0000, 380.0000, 282.0000, 442.6711,
  112. 337.0366],
  113. [502.2869, 242.9966, 642.0000, 132.0000, 345.0000, 164.0000, 545.6324,
  114. 319.2123],
  115. [495.9200, 243.9100, 607.8350, 143.0867, 345.3565, 118.8599, 474.6164,
  116. 313.2611]], device='cuda:0')
  117. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  118. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  119. loss_train: nan
  120. step: 4
  121. running loss: nan
  122. Train Steps: 4/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  123. [nan, nan, nan, nan, nan, nan, nan, nan],
  124. [nan, nan, nan, nan, nan, nan, nan, nan],
  125. [nan, nan, nan, nan, nan, nan, nan, nan],
  126. [nan, nan, nan, nan, nan, nan, nan, nan],
  127. [nan, nan, nan, nan, nan, nan, nan, nan],
  128. [nan, nan, nan, nan, nan, nan, nan, nan],
  129. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  130. grad_fn=<AddmmBackward>)
  131. landmarks are: tensor([[495.8200, 244.5800, 635.6002, 147.2001, 358.0301, 112.2785, 471.7639,
  132. 320.9747],
  133. [490.0000, 239.9400, 700.0000, 293.0000, 380.0000, 282.0000, 442.6711,
  134. 337.0366],
  135. [508.2770, 247.6873, 731.0000, 212.0000, 375.0000, 195.0000, 571.5920,
  136. 359.4722],
  137. [495.8100, 239.8400, 686.5782, 321.6804, 329.3477, 300.9463, 475.3770,
  138. 308.0492],
  139. [501.8300, 248.5600, 700.0000, 342.0000, 319.0000, 283.0000, 481.0000,
  140. 328.0000],
  141. [490.2000, 245.0600, 699.0000, 281.0000, 289.0000, 222.0000, 396.7645,
  142. 323.8376],
  143. [487.6300, 240.1400, 682.6197, 310.0718, 402.5010, 305.6570, 410.0000,
  144. 326.0000],
  145. [496.8000, 249.8500, 576.1289, 175.8134, 322.0000, 149.0000, 455.0000,
  146. 334.0000]], device='cuda:0')
  147. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  148. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  149. loss_train: nan
  150. step: 5
  151. running loss: nan
  152.  
  153. Train Steps: 5/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  154. [nan, nan, nan, nan, nan, nan, nan, nan],
  155. [nan, nan, nan, nan, nan, nan, nan, nan],
  156. [nan, nan, nan, nan, nan, nan, nan, nan],
  157. [nan, nan, nan, nan, nan, nan, nan, nan],
  158. [nan, nan, nan, nan, nan, nan, nan, nan],
  159. [nan, nan, nan, nan, nan, nan, nan, nan],
  160. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  161. grad_fn=<AddmmBackward>)
  162. landmarks are: tensor([[499.2800, 248.5500, 715.0000, 279.0000, 326.0000, 321.0000, 500.0000,
  163. 333.0000],
  164. [501.0600, 243.8700, 723.0000, 259.0000, 287.0000, 273.0000, 506.0000,
  165. 315.0000],
  166. [482.6800, 240.6500, 588.0000, 152.0000, 275.0000, 202.0000, 441.2465,
  167. 305.0196],
  168. [ nan, nan, 617.9645, 156.6336, 294.0000, 164.0000, 433.0000,
  169. 310.0000],
  170. [498.3700, 249.1100, 607.0000, 137.0000, 321.0000, 173.0000, 496.0000,
  171. 346.0000],
  172. [501.9300, 256.1800, 715.0000, 298.0000, 284.0000, 257.0000, 456.0000,
  173. 344.0000],
  174. [495.6600, 245.6500, 605.0000, 169.0000, 315.0000, 191.0000, 481.0000,
  175. 371.0000],
  176. [493.5900, 246.1400, 597.4271, 221.6781, 277.0000, 226.0000, 419.0314,
  177. 349.3644]], device='cuda:0')
  178. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  179. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  180. loss_train: nan
  181. step: 6
  182. running loss: nan
  183. Train Steps: 6/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  184. [nan, nan, nan, nan, nan, nan, nan, nan],
  185. [nan, nan, nan, nan, nan, nan, nan, nan],
  186. [nan, nan, nan, nan, nan, nan, nan, nan],
  187. [nan, nan, nan, nan, nan, nan, nan, nan],
  188. [nan, nan, nan, nan, nan, nan, nan, nan],
  189. [nan, nan, nan, nan, nan, nan, nan, nan],
  190. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  191. grad_fn=<AddmmBackward>)
  192. landmarks are: tensor([[500.2902, 239.5567, 719.0000, 286.0000, 319.0000, 331.0000, 556.3961,
  193. 317.1299],
  194. [502.0400, 245.7100, 659.0000, 135.0000, 373.0000, 107.0000, 493.7134,
  195. 292.1411],
  196. [498.8640, 237.9932, 694.0000, 324.0000, 309.0000, 271.0000, 466.0000,
  197. 312.0000],
  198. [495.8700, 247.7900, 701.0000, 247.0000, 292.0000, 294.0000, 456.5611,
  199. 306.1910],
  200. [502.2869, 242.9966, 642.0000, 132.0000, 345.0000, 164.0000, 545.6324,
  201. 319.2123],
  202. [489.9400, 241.8100, 692.0000, 292.0000, 399.9338, 306.3606, 410.9215,
  203. 346.3862],
  204. [ nan, nan, 567.7618, 140.7894, 340.0000, 111.0000, 414.0000,
  205. 335.0000],
  206. [491.4477, 242.0584, 704.0000, 290.0000, 361.0000, 322.0000, 423.0828,
  207. 305.8005]], device='cuda:0')
  208. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  209. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  210. loss_train: nan
  211. step: 7
  212. running loss: nan
  213. Train Steps: 7/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  214. [nan, nan, nan, nan, nan, nan, nan, nan],
  215. [nan, nan, nan, nan, nan, nan, nan, nan],
  216. [nan, nan, nan, nan, nan, nan, nan, nan],
  217. [nan, nan, nan, nan, nan, nan, nan, nan],
  218. [nan, nan, nan, nan, nan, nan, nan, nan],
  219. [nan, nan, nan, nan, nan, nan, nan, nan],
  220. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  221. grad_fn=<AddmmBackward>)
  222. landmarks are: tensor([[496.0500, 238.2000, 673.0000, 297.0000, 353.0000, 311.0000, 487.0000,
  223. 324.0000],
  224. [484.2800, 242.0800, 551.7881, 114.9384, 320.0000, 127.0000, 435.1920,
  225. 310.0955],
  226. [492.5886, 237.3677, 665.8701, 248.2718, 278.2819, 265.8849, 473.0849,
  227. 307.8573],
  228. [491.8400, 243.1800, 700.0000, 273.0000, 387.9706, 313.0692, 469.0000,
  229. 334.0000],
  230. [508.8475, 246.1237, 692.0000, 179.0000, 391.0000, 120.0000, 536.1351,
  231. 327.5420],
  232. [491.3300, 247.6000, 606.0000, 184.0000, 275.0000, 263.0000, 462.2595,
  233. 312.4382],
  234. [489.1657, 239.8694, 565.0000, 143.0000, 323.0000, 117.0000, 425.5759,
  235. 299.5533],
  236. [496.3100, 240.7500, 646.0000, 144.0000, 345.0000, 123.0000, 464.0000,
  237. 309.0000]], device='cuda:0')
  238. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  239. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  240. loss_train: nan
  241. step: 8
  242. running loss: nan
  243. Train Steps: 8/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  244. [nan, nan, nan, nan, nan, nan, nan, nan],
  245. [nan, nan, nan, nan, nan, nan, nan, nan],
  246. [nan, nan, nan, nan, nan, nan, nan, nan],
  247. [nan, nan, nan, nan, nan, nan, nan, nan],
  248. [nan, nan, nan, nan, nan, nan, nan, nan],
  249. [nan, nan, nan, nan, nan, nan, nan, nan],
  250. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  251. grad_fn=<AddmmBackward>)
  252. landmarks are: tensor([[500.0300, 247.8700, 695.0000, 179.0000, 342.0000, 142.0000, 493.0000,
  253. 322.0000],
  254. [495.8700, 247.7900, 701.0000, 247.0000, 292.0000, 294.0000, 456.5611,
  255. 306.1910],
  256. [507.1360, 246.1237, 727.0000, 286.0000, 314.0000, 317.0000, 600.7172,
  257. 323.3772],
  258. [491.6200, 238.9700, 696.0000, 301.0000, 352.0000, 288.0000, 430.0000,
  259. 345.0000],
  260. [492.9800, 236.9300, 707.0000, 271.0000, 340.0000, 311.0000, 467.0000,
  261. 330.0000],
  262. [485.8600, 235.9000, 669.0000, 349.0000, 354.0000, 307.0000, 416.3159,
  263. 289.0111],
  264. [493.1500, 247.1400, 633.0000, 159.0000, 283.0000, 210.0000, 449.0819,
  265. 302.2865],
  266. [503.1426, 241.4330, 727.0000, 274.0000, 315.0000, 338.0000, 564.6272,
  267. 336.5657]], device='cuda:0')
  268. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  269. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  270. loss_train: nan
  271. step: 9
  272. running loss: nan
  273. Train Steps: 9/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  274. [nan, nan, nan, nan, nan, nan, nan, nan],
  275. [nan, nan, nan, nan, nan, nan, nan, nan],
  276. [nan, nan, nan, nan, nan, nan, nan, nan],
  277. [nan, nan, nan, nan, nan, nan, nan, nan],
  278. [nan, nan, nan, nan, nan, nan, nan, nan],
  279. [nan, nan, nan, nan, nan, nan, nan, nan],
  280. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  281. grad_fn=<AddmmBackward>)
  282. landmarks are: tensor([[499.7197, 240.4949, 700.5271, 305.2639, 328.0690, 323.5423, 531.4897,
  283. 308.0097],
  284. [488.1700, 239.9200, 700.0000, 308.0000, 306.0000, 285.0000, 451.0000,
  285. 305.0000],
  286. [502.4100, 246.0400, 724.0000, 272.0000, 302.0000, 193.0000, 507.0098,
  287. 294.9176],
  288. [496.1400, 238.9700, 684.3768, 325.7215, 307.3357, 262.1889, 469.2918,
  289. 323.8934],
  290. [488.7100, 240.8900, 613.4007, 218.3425, 281.0000, 220.0000, 415.9966,
  291. 338.4796],
  292. [507.7065, 245.1856, 635.0000, 330.0000, 317.0000, 292.0000, 587.4209,
  293. 342.1188],
  294. [502.0016, 241.4330, 688.0000, 137.0000, 428.0000, 108.0000, 565.8936,
  295. 324.7654],
  296. [496.2100, 245.7600, 709.0000, 256.0000, 283.0000, 247.0000, 482.0000,
  297. 339.0000]], device='cuda:0')
  298. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  299. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  300. loss_train: nan
  301. step: 10
  302. running loss: nan
  303. Train Steps: 10/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  304. [nan, nan, nan, nan, nan, nan, nan, nan],
  305. [nan, nan, nan, nan, nan, nan, nan, nan],
  306. [nan, nan, nan, nan, nan, nan, nan, nan],
  307. [nan, nan, nan, nan, nan, nan, nan, nan],
  308. [nan, nan, nan, nan, nan, nan, nan, nan],
  309. [nan, nan, nan, nan, nan, nan, nan, nan],
  310. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  311. grad_fn=<AddmmBackward>)
  312. landmarks are: tensor([[ nan, nan, 700.0000, 148.0000, 411.0000, 157.0000, 590.5867,
  313. 333.0951],
  314. [507.7065, 249.8763, 731.0000, 239.0000, 310.0000, 259.0000, 597.5515,
  315. 328.2361],
  316. [ nan, nan, 601.0000, 127.0000, 343.0000, 120.0000, 448.0000,
  317. 337.0000],
  318. [498.3700, 249.1100, 607.0000, 137.0000, 321.0000, 173.0000, 496.0000,
  319. 346.0000],
  320. [490.2700, 247.0800, 691.0000, 320.0000, 370.0000, 316.0000, 415.4624,
  321. 328.5230],
  322. [497.2200, 247.1100, 615.0000, 138.0000, 336.0000, 137.0000, 474.0000,
  323. 319.0000],
  324. [506.5656, 249.8763, 728.0000, 201.0000, 335.0000, 221.0000, 595.0188,
  325. 331.7068],
  326. [497.7500, 237.4100, 707.0000, 301.0000, 315.0000, 276.0000, 472.0000,
  327. 301.0000]], device='cuda:0')
  328. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  329.  
  330. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  331. loss_train: nan
  332. step: 11
  333. running loss: nan
  334. Train Steps: 11/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  335. [nan, nan, nan, nan, nan, nan, nan, nan],
  336. [nan, nan, nan, nan, nan, nan, nan, nan],
  337. [nan, nan, nan, nan, nan, nan, nan, nan],
  338. [nan, nan, nan, nan, nan, nan, nan, nan],
  339. [nan, nan, nan, nan, nan, nan, nan, nan],
  340. [nan, nan, nan, nan, nan, nan, nan, nan],
  341. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  342. grad_fn=<AddmmBackward>)
  343. landmarks are: tensor([[487.5800, 238.2200, 695.0000, 286.0000, 388.7642, 292.7355, 415.2475,
  344. 296.4297],
  345. [492.8739, 241.4330, 693.0000, 281.0000, 394.7994, 321.8399, 485.0000,
  346. 334.0000],
  347. [489.9400, 241.8100, 692.0000, 292.0000, 399.9338, 306.3606, 410.9215,
  348. 346.3862],
  349. [494.3001, 239.5567, 714.0048, 287.3219, 310.3356, 293.9974, 483.3020,
  350. 316.7523],
  351. [497.5500, 246.8200, 654.0000, 169.0000, 314.0000, 167.0000, 472.0000,
  352. 321.0000],
  353. [496.3700, 240.9300, 668.0000, 163.0000, 319.0000, 153.0000, 463.0000,
  354. 308.0000],
  355. [489.8800, 244.1100, 665.1248, 300.0650, 299.0000, 279.0000, 413.3255,
  356. 324.1305],
  357. [501.2500, 244.2100, 697.0000, 336.0000, 297.0000, 287.0000, 462.0000,
  358. 366.0000]], device='cuda:0')
  359. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  360. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  361. loss_train: nan
  362. step: 12
  363. running loss: nan
  364. Train Steps: 12/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  365. [nan, nan, nan, nan, nan, nan, nan, nan],
  366. [nan, nan, nan, nan, nan, nan, nan, nan],
  367. [nan, nan, nan, nan, nan, nan, nan, nan],
  368. [nan, nan, nan, nan, nan, nan, nan, nan],
  369. [nan, nan, nan, nan, nan, nan, nan, nan],
  370. [nan, nan, nan, nan, nan, nan, nan, nan],
  371. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  372. grad_fn=<AddmmBackward>)
  373. landmarks are: tensor([[ nan, nan, 610.0000, 146.0000, 297.0000, 172.0000, 418.8090,
  374. 313.2191],
  375. [507.1360, 244.8729, 674.0000, 325.0000, 308.0000, 290.0000, 586.7878,
  376. 345.5895],
  377. [ nan, nan, 559.3947, 167.4743, 316.0000, 143.0000, 438.6643,
  378. 349.1448],
  379. [500.5700, 242.0600, 663.0000, 140.0000, 314.0000, 163.0000, 506.4223,
  380. 294.0870],
  381. [497.0400, 240.0600, 617.0000, 127.0000, 347.0000, 108.0000, 468.0000,
  382. 311.0000],
  383. [ nan, nan, 638.5021, 191.6575, 290.0000, 190.0000, 403.1752,
  384. 333.7941],
  385. [501.9300, 256.1800, 715.0000, 298.0000, 284.0000, 257.0000, 456.0000,
  386. 344.0000],
  387. [509.1327, 248.6254, 690.0000, 185.0000, 393.0000, 120.0000, 515.8740,
  388. 316.4358]], device='cuda:0')
  389. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  390. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  391. loss_train: nan
  392. step: 13
  393. running loss: nan
  394. Train Steps: 13/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  395. [nan, nan, nan, nan, nan, nan, nan, nan],
  396. [nan, nan, nan, nan, nan, nan, nan, nan],
  397. [nan, nan, nan, nan, nan, nan, nan, nan],
  398. [nan, nan, nan, nan, nan, nan, nan, nan],
  399. [nan, nan, nan, nan, nan, nan, nan, nan],
  400. [nan, nan, nan, nan, nan, nan, nan, nan],
  401. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  402. grad_fn=<AddmmBackward>)
  403. landmarks are: tensor([[496.0200, 244.2800, 587.0000, 115.0000, 336.0000, 147.0000, 492.0000,
  404. 331.0000],
  405. [504.5688, 242.6839, 642.0000, 350.0000, 302.0000, 292.0000, 551.3309,
  406. 327.5420],
  407. [487.4300, 239.4100, 672.0000, 260.0000, 295.0000, 278.0000, 444.8080,
  408. 339.3793],
  409. [ nan, nan, 727.0000, 227.0000, 365.0000, 157.0000, 539.3008,
  410. 334.4833],
  411. [507.1360, 249.5636, 672.0000, 337.0000, 306.0000, 249.0000, 587.4209,
  412. 344.8954],
  413. [488.5952, 242.9966, 696.0000, 291.0000, 357.5752, 290.8813, 403.4423,
  414. 325.8875],
  415. [506.6800, 242.2100, 691.0000, 344.0000, 321.0000, 283.0000, 509.5424,
  416. 296.3059],
  417. [495.9600, 245.1500, 673.1207, 178.4619, 329.3477, 136.4104, 469.4820,
  418. 323.4764]], device='cuda:0')
  419. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  420. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  421. loss_train: nan
  422. step: 14
  423. running loss: nan
  424. Train Steps: 14/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  425. [nan, nan, nan, nan, nan, nan, nan, nan],
  426. [nan, nan, nan, nan, nan, nan, nan, nan],
  427. [nan, nan, nan, nan, nan, nan, nan, nan],
  428. [nan, nan, nan, nan, nan, nan, nan, nan],
  429. [nan, nan, nan, nan, nan, nan, nan, nan],
  430. [nan, nan, nan, nan, nan, nan, nan, nan],
  431. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  432. grad_fn=<AddmmBackward>)
  433. landmarks are: tensor([[500.5754, 243.6220, 664.0000, 140.0000, 375.0000, 155.0000, 563.9941,
  434. 337.9540],
  435. [501.8700, 246.5800, 712.0000, 229.0000, 335.0000, 130.0000, 468.6673,
  436. 290.0746],
  437. [497.0800, 247.8600, 675.0000, 213.0000, 281.0000, 264.0000, 457.2733,
  438. 307.3623],
  439. [503.9984, 240.8076, 715.0000, 321.0000, 294.0000, 276.0000, 516.5071,
  440. 298.3883],
  441. [492.8739, 244.5602, 712.0000, 280.0000, 330.0000, 355.0000, 501.0000,
  442. 322.0000],
  443. [484.6700, 238.7000, 663.0000, 216.0000, 272.0000, 243.0000, 442.3150,
  444. 327.6658],
  445. [496.2700, 244.6800, 704.0000, 305.0000, 312.0000, 300.0000, 488.0000,
  446. 335.0000],
  447. [486.3200, 237.8300, 593.6238, 177.4812, 285.0000, 175.0000, 428.0689,
  448. 298.7724]], device='cuda:0')
  449. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  450. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  451. loss_train: nan
  452. step: 15
  453. running loss: nan
  454. Train Steps: 15/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  455. [nan, nan, nan, nan, nan, nan, nan, nan],
  456. [nan, nan, nan, nan, nan, nan, nan, nan],
  457. [nan, nan, nan, nan, nan, nan, nan, nan],
  458. [nan, nan, nan, nan, nan, nan, nan, nan],
  459. [nan, nan, nan, nan, nan, nan, nan, nan],
  460. [nan, nan, nan, nan, nan, nan, nan, nan],
  461. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  462. grad_fn=<AddmmBackward>)
  463. landmarks are: tensor([[498.8640, 244.8729, 686.0000, 180.0000, 297.0000, 182.0000, 444.0000,
  464. 338.0000],
  465. [ nan, nan, 578.0000, 130.0000, 319.0000, 137.0000, 434.1235,
  466. 310.8764],
  467. [487.7700, 239.9900, 586.0000, 160.0000, 276.0000, 211.0000, 422.7267,
  468. 302.6769],
  469. [489.9400, 241.8100, 692.0000, 292.0000, 399.9338, 306.3606, 410.9215,
  470. 346.3862],
  471. [498.3100, 251.9000, 613.0000, 162.0000, 376.0000, 128.0000, 454.0000,
  472. 347.0000],
  473. [497.2500, 245.9600, 578.0000, 122.0000, 335.0000, 133.0000, 478.0000,
  474. 317.0000],
  475. [499.6200, 246.9000, 696.0000, 293.0000, 370.0000, 331.0000, 488.0000,
  476. 313.0000],
  477. [489.7400, 240.3700, 708.0000, 253.0000, 327.0000, 331.0000, 485.0000,
  478. 331.0000]], device='cuda:0')
  479. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  480. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  481. loss_train: nan
  482. step: 16
  483. running loss: nan
  484. Train Steps: 16/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  485. [nan, nan, nan, nan, nan, nan, nan, nan],
  486. [nan, nan, nan, nan, nan, nan, nan, nan],
  487. [nan, nan, nan, nan, nan, nan, nan, nan],
  488. [nan, nan, nan, nan, nan, nan, nan, nan],
  489. [nan, nan, nan, nan, nan, nan, nan, nan],
  490. [nan, nan, nan, nan, nan, nan, nan, nan],
  491. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  492. grad_fn=<AddmmBackward>)
  493. landmarks are: tensor([[494.5853, 238.9313, 603.2948, 142.7309, 316.7463, 167.4912, 486.3070,
  494. 323.3411],
  495. [494.1100, 234.6700, 699.0000, 253.0000, 279.0000, 242.0000, 482.0000,
  496. 289.0000],
  497. [500.7300, 247.9500, 656.0000, 139.0000, 402.0000, 92.0000, 500.0116,
  498. 296.8201],
  499. [492.8739, 242.3712, 602.0000, 128.0000, 330.0000, 124.0000, 463.0000,
  500. 307.0000],
  501. [494.2300, 243.5200, 602.0000, 135.0000, 345.0000, 107.0000, 432.3427,
  502. 314.3904],
  503. [496.0500, 246.9500, 698.0000, 284.0000, 296.0000, 193.0000, 430.9181,
  504. 346.0170],
  505. [492.0181, 236.1169, 695.7136, 309.4855, 371.7718, 319.7672, 483.3020,
  506. 309.1750],
  507. [502.0016, 240.8076, 708.0000, 170.0000, 398.0000, 134.0000, 564.6272,
  508. 320.6006]], device='cuda:0')
  509. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  510.  
  511. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  512. loss_train: nan
  513. step: 17
  514. running loss: nan
  515. Train Steps: 17/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  516. [nan, nan, nan, nan, nan, nan, nan, nan],
  517. [nan, nan, nan, nan, nan, nan, nan, nan],
  518. [nan, nan, nan, nan, nan, nan, nan, nan],
  519. [nan, nan, nan, nan, nan, nan, nan, nan],
  520. [nan, nan, nan, nan, nan, nan, nan, nan],
  521. [nan, nan, nan, nan, nan, nan, nan, nan],
  522. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  523. grad_fn=<AddmmBackward>)
  524. landmarks are: tensor([[495.9600, 234.7900, 703.0000, 313.0000, 326.0000, 306.0000, 484.8128,
  525. 294.7599],
  526. [497.0900, 250.5100, 584.0000, 173.0000, 347.0000, 130.0000, 455.8488,
  527. 346.4075],
  528. [494.7200, 244.7300, 668.0000, 222.0000, 294.0000, 173.0000, 425.0000,
  529. 347.0000],
  530. [496.2100, 244.5800, 688.8793, 172.7032, 324.0115, 153.2296, 472.5629,
  531. 329.7797],
  532. [490.0600, 244.3900, 700.0000, 308.0000, 304.0000, 260.0000, 398.9014,
  533. 322.6663],
  534. [494.5853, 237.9932, 661.0566, 183.8918, 282.0930, 249.6721, 495.3221,
  535. 317.4111],
  536. [497.8200, 250.2600, 700.0000, 330.0000, 324.0000, 289.0000, 454.0000,
  537. 336.0000],
  538. [494.0300, 245.4600, 629.0000, 168.0000, 291.0000, 215.0000, 495.0000,
  539. 326.0000]], device='cuda:0')
  540. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  541. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  542. loss_train: nan
  543. step: 18
  544. running loss: nan
  545. Train Steps: 18/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  546. [nan, nan, nan, nan, nan, nan, nan, nan],
  547. [nan, nan, nan, nan, nan, nan, nan, nan],
  548. [nan, nan, nan, nan, nan, nan, nan, nan],
  549. [nan, nan, nan, nan, nan, nan, nan, nan],
  550. [nan, nan, nan, nan, nan, nan, nan, nan],
  551. [nan, nan, nan, nan, nan, nan, nan, nan],
  552. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  553. grad_fn=<AddmmBackward>)
  554. landmarks are: tensor([[500.1200, 249.7500, 693.0000, 268.0000, 290.0000, 214.0000, 483.0000,
  555. 332.0000],
  556. [501.7200, 242.6800, 683.0000, 354.0000, 300.0000, 265.0000, 479.1508,
  557. 278.9525],
  558. [493.5900, 246.1400, 597.4271, 221.6781, 277.0000, 226.0000, 419.0314,
  559. 349.3644],
  560. [490.3067, 235.8042, 701.4897, 306.3193, 331.7047, 338.5089, 479.9964,
  561. 304.8923],
  562. [497.5100, 245.3300, 715.0000, 288.0000, 306.0000, 267.0000, 468.0000,
  563. 312.0000],
  564. [496.2100, 244.5800, 688.8793, 172.7032, 324.0115, 153.2296, 472.5629,
  565. 329.7797],
  566. [503.4279, 238.9313, 696.0000, 318.0000, 301.0000, 283.0000, 563.9941,
  567. 317.8241],
  568. [501.9300, 247.0000, 648.0244, 348.0684, 320.0000, 275.0000, 446.5888,
  569. 367.1014]], device='cuda:0')
  570. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  571. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  572. loss_train: nan
  573. step: 19
  574. running loss: nan
  575. Train Steps: 19/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  576. [nan, nan, nan, nan, nan, nan, nan, nan],
  577. [nan, nan, nan, nan, nan, nan, nan, nan],
  578. [nan, nan, nan, nan, nan, nan, nan, nan],
  579. [nan, nan, nan, nan, nan, nan, nan, nan],
  580. [nan, nan, nan, nan, nan, nan, nan, nan],
  581. [nan, nan, nan, nan, nan, nan, nan, nan],
  582. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  583. grad_fn=<AddmmBackward>)
  584. landmarks are: tensor([[485.8600, 235.9000, 669.0000, 349.0000, 354.0000, 307.0000, 416.3159,
  585. 289.0111],
  586. [494.3001, 240.4949, 600.4067, 160.6728, 349.4514, 113.4323, 466.4738,
  587. 315.7639],
  588. [490.2800, 246.8100, 556.0000, 148.0000, 324.0000, 128.0000, 429.0000,
  589. 333.0000],
  590. [498.8700, 237.9300, 708.0000, 298.0000, 291.0000, 241.0000, 468.0000,
  591. 311.0000],
  592. [494.5853, 237.9932, 661.0566, 183.8918, 282.0930, 249.6721, 495.3221,
  593. 317.4111],
  594. [501.4500, 243.6300, 668.0000, 146.0000, 366.0000, 137.0000, 508.0000,
  595. 318.0000],
  596. [494.1500, 245.1300, 699.0000, 237.0000, 302.0000, 336.0000, 498.0000,
  597. 342.0000],
  598. [496.1000, 242.9900, 620.5920, 134.0372, 356.6960, 107.8909, 478.0393,
  599. 325.5612]], device='cuda:0')
  600. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  601. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  602. loss_train: nan
  603. step: 20
  604. running loss: nan
  605. Train Steps: 20/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  606. [nan, nan, nan, nan, nan, nan, nan, nan],
  607. [nan, nan, nan, nan, nan, nan, nan, nan],
  608. [nan, nan, nan, nan, nan, nan, nan, nan],
  609. [nan, nan, nan, nan, nan, nan, nan, nan],
  610. [nan, nan, nan, nan, nan, nan, nan, nan],
  611. [nan, nan, nan, nan, nan, nan, nan, nan],
  612. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  613. grad_fn=<AddmmBackward>)
  614. landmarks are: tensor([[495.8500, 246.0500, 707.0000, 317.0000, 301.0000, 316.0000, 473.0000,
  615. 334.0000],
  616. [494.8400, 247.7400, 712.0000, 274.0000, 315.0000, 325.0000, 458.6980,
  617. 306.5814],
  618. [ nan, nan, 578.0000, 130.0000, 319.0000, 137.0000, 434.1235,
  619. 310.8764],
  620. [497.4600, 248.2400, 581.0000, 134.0000, 326.0000, 159.0000, 497.0000,
  621. 347.0000],
  622. [502.5721, 242.0584, 626.3995, 124.7889, 362.5892, 124.7976, 512.3327,
  623. 319.3755],
  624. [488.7400, 242.4700, 558.0000, 190.0000, 281.0000, 203.0000, 412.2570,
  625. 319.1522],
  626. [ nan, nan, 715.0000, 171.0000, 373.0000, 187.0000, 592.4862,
  627. 331.7068],
  628. [501.0800, 241.7700, 720.0000, 286.0000, 304.0000, 310.0000, 513.1892,
  629. 286.2780]], device='cuda:0')
  630. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  631. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  632. loss_train: nan
  633. step: 21
  634. running loss: nan
  635. Train Steps: 21/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  636. [nan, nan, nan, nan, nan, nan, nan, nan],
  637. [nan, nan, nan, nan, nan, nan, nan, nan],
  638. [nan, nan, nan, nan, nan, nan, nan, nan],
  639. [nan, nan, nan, nan, nan, nan, nan, nan],
  640. [nan, nan, nan, nan, nan, nan, nan, nan],
  641. [nan, nan, nan, nan, nan, nan, nan, nan],
  642. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  643. grad_fn=<AddmmBackward>)
  644. landmarks are: tensor([[502.8574, 242.3712, 695.7136, 182.8364, 313.9077, 173.2258, 504.0367,
  645. 322.0233],
  646. [501.4600, 245.6700, 723.0000, 258.0000, 296.0000, 209.0000, 501.0000,
  647. 310.0000],
  648. [489.7362, 239.5567, 699.0000, 280.0000, 361.3809, 292.7355, 412.3982,
  649. 295.6488],
  650. [ nan, nan, 559.3947, 167.4743, 316.0000, 143.0000, 438.6643,
  651. 349.1448],
  652. [506.5656, 249.8763, 728.0000, 201.0000, 335.0000, 221.0000, 595.0188,
  653. 331.7068],
  654. [500.8607, 239.2440, 695.0000, 295.0000, 344.0000, 320.0000, 560.8282,
  655. 318.5182],
  656. [ nan, nan, 567.7618, 140.7894, 340.0000, 111.0000, 414.0000,
  657. 335.0000],
  658. [490.1900, 247.0600, 692.0000, 305.0000, 327.0000, 322.0000, 424.0101,
  659. 327.3517]], device='cuda:0')
  660. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  661. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  662. loss_train: nan
  663. step: 22
  664. running loss: nan
  665. Train Steps: 22/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  666. [nan, nan, nan, nan, nan, nan, nan, nan],
  667. [nan, nan, nan, nan, nan, nan, nan, nan],
  668. [nan, nan, nan, nan, nan, nan, nan, nan],
  669. [nan, nan, nan, nan, nan, nan, nan, nan],
  670. [nan, nan, nan, nan, nan, nan, nan, nan],
  671. [nan, nan, nan, nan, nan, nan, nan, nan],
  672. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  673. grad_fn=<AddmmBackward>)
  674. landmarks are: tensor([[487.6300, 240.1400, 682.6197, 310.0718, 402.5010, 305.6570, 410.0000,
  675. 326.0000],
  676. [498.4000, 246.7900, 577.0000, 119.0000, 346.0000, 142.0000, 501.0000,
  677. 324.0000],
  678. [ nan, nan, 517.5590, 116.6062, 322.0000, 120.0000, 410.0000,
  679. 332.0000],
  680. [507.9918, 248.6254, 740.0000, 246.0000, 330.0000, 225.0000, 570.3256,
  681. 356.6957],
  682. [ nan, nan, 727.0000, 227.0000, 365.0000, 157.0000, 539.3008,
  683. 334.4833],
  684. [502.2400, 255.1600, 710.0000, 301.0000, 329.0000, 165.0000, 433.0551,
  685. 371.7868],
  686. [502.8574, 238.6186, 723.0000, 284.0000, 312.0000, 249.0000, 565.8936,
  687. 319.2123],
  688. [508.8475, 249.8763, 723.0000, 301.0000, 300.0000, 227.0000, 515.8740,
  689. 318.5182]], device='cuda:0')
  690. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  691.  
  692. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  693. loss_train: nan
  694. step: 23
  695. running loss: nan
  696. Train Steps: 23/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  697. [nan, nan, nan, nan, nan, nan, nan, nan],
  698. [nan, nan, nan, nan, nan, nan, nan, nan],
  699. [nan, nan, nan, nan, nan, nan, nan, nan],
  700. [nan, nan, nan, nan, nan, nan, nan, nan],
  701. [nan, nan, nan, nan, nan, nan, nan, nan],
  702. [nan, nan, nan, nan, nan, nan, nan, nan],
  703. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  704. grad_fn=<AddmmBackward>)
  705. landmarks are: tensor([[482.0700, 238.7200, 684.0000, 254.0000, 289.0000, 314.0000, 446.5888,
  706. 297.9915],
  707. [481.7500, 239.6200, 642.0000, 201.0000, 268.0000, 264.0000, 445.1642,
  708. 301.5056],
  709. [498.3100, 251.9000, 613.0000, 162.0000, 376.0000, 128.0000, 454.0000,
  710. 347.0000],
  711. [509.7032, 245.4983, 667.0000, 351.0000, 316.0000, 307.0000, 524.7382,
  712. 315.7417],
  713. [498.1100, 240.0800, 675.0000, 344.0000, 309.0000, 255.0000, 460.0000,
  714. 317.0000],
  715. [489.8800, 245.0100, 556.3521, 184.1524, 292.0000, 165.0000, 413.0584,
  716. 329.1087],
  717. [494.5853, 239.8694, 703.4152, 251.4380, 284.1584, 257.0997, 483.0015,
  718. 318.7289],
  719. [496.0000, 245.8900, 659.0000, 175.0000, 321.0000, 178.0000, 480.0000,
  720. 341.0000]], device='cuda:0')
  721. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  722. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  723. loss_train: nan
  724. step: 24
  725. running loss: nan
  726. Train Steps: 24/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  727. [nan, nan, nan, nan, nan, nan, nan, nan],
  728. [nan, nan, nan, nan, nan, nan, nan, nan],
  729. [nan, nan, nan, nan, nan, nan, nan, nan],
  730. [nan, nan, nan, nan, nan, nan, nan, nan],
  731. [nan, nan, nan, nan, nan, nan, nan, nan],
  732. [nan, nan, nan, nan, nan, nan, nan, nan],
  733. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  734. grad_fn=<AddmmBackward>)
  735. landmarks are: tensor([[496.1000, 244.9000, 706.1387, 222.8866, 306.0016, 162.7361, 467.5804,
  736. 324.7273],
  737. [495.9800, 246.0800, 707.0000, 319.0000, 306.0000, 228.0000, 433.0000,
  738. 341.0000],
  739. [488.2000, 240.2900, 695.0000, 306.0000, 385.0000, 324.0000, 432.3427,
  740. 303.8483],
  741. [495.7000, 245.4200, 676.0000, 234.0000, 286.0000, 236.0000, 478.0000,
  742. 335.0000],
  743. [490.0600, 244.3900, 700.0000, 308.0000, 304.0000, 260.0000, 398.9014,
  744. 322.6663],
  745. [487.7395, 241.4330, 679.0000, 223.0000, 310.0000, 331.0000, 466.8895,
  746. 335.4748],
  747. [501.3200, 243.8900, 665.0000, 148.0000, 383.0000, 104.0000, 505.0000,
  748. 308.0000],
  749. [ nan, nan, 586.7780, 154.1319, 303.0000, 160.0000, 405.3121,
  750. 334.6727]], device='cuda:0')
  751. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  752. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  753. loss_train: nan
  754. step: 25
  755. running loss: nan
  756. Train Steps: 25/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  757. [nan, nan, nan, nan, nan, nan, nan, nan],
  758. [nan, nan, nan, nan, nan, nan, nan, nan],
  759. [nan, nan, nan, nan, nan, nan, nan, nan],
  760. [nan, nan, nan, nan, nan, nan, nan, nan],
  761. [nan, nan, nan, nan, nan, nan, nan, nan],
  762. [nan, nan, nan, nan, nan, nan, nan, nan],
  763. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  764. grad_fn=<AddmmBackward>)
  765. landmarks are: tensor([[503.7131, 242.6839, 731.0000, 246.0000, 338.5284, 254.5401, 593.7525,
  766. 317.8241],
  767. [494.7800, 244.0100, 707.0000, 267.0000, 323.0000, 284.0000, 417.0282,
  768. 308.5337],
  769. [493.2300, 246.3500, 606.0000, 104.0000, 307.0000, 159.0000, 454.4241,
  770. 306.9719],
  771. [490.3400, 244.1100, 700.0000, 304.0000, 310.0000, 254.0000, 418.8311,
  772. 352.8785],
  773. [504.5688, 241.7457, 719.0000, 289.0000, 315.0000, 210.0000, 584.8883,
  774. 322.6830],
  775. [486.5300, 242.5300, 558.0000, 115.0000, 328.0000, 119.0000, 440.1781,
  776. 334.6939],
  777. [ nan, nan, 548.7455, 131.6165, 332.0000, 112.0000, 412.2570,
  778. 343.7506],
  779. [498.2935, 243.9347, 681.0000, 343.0000, 360.0000, 303.0000, 482.0000,
  780. 321.0000]], device='cuda:0')
  781. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  782. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  783. loss_train: nan
  784. step: 26
  785. running loss: nan
  786. Train Steps: 26/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  787. [nan, nan, nan, nan, nan, nan, nan, nan],
  788. [nan, nan, nan, nan, nan, nan, nan, nan],
  789. [nan, nan, nan, nan, nan, nan, nan, nan],
  790. [nan, nan, nan, nan, nan, nan, nan, nan],
  791. [nan, nan, nan, nan, nan, nan, nan, nan],
  792. [nan, nan, nan, nan, nan, nan, nan, nan],
  793. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  794. grad_fn=<AddmmBackward>)
  795. landmarks are: tensor([[496.2200, 241.2300, 702.3867, 301.8638, 293.3280, 232.9381, 467.3902,
  796. 324.3103],
  797. [495.0900, 242.9100, 711.0000, 265.0000, 337.0000, 312.0000, 479.0000,
  798. 338.0000],
  799. [490.7800, 239.6200, 633.0000, 183.0000, 290.0000, 183.0000, 467.0000,
  800. 303.0000],
  801. [495.6300, 246.0600, 692.0000, 235.0000, 286.0000, 242.0000, 464.0000,
  802. 339.0000],
  803. [499.7197, 240.4949, 700.5271, 305.2639, 328.0690, 323.5423, 531.4897,
  804. 308.0097],
  805. [490.0800, 243.9900, 691.0000, 323.0000, 335.0000, 291.0000, 401.3054,
  806. 323.5448],
  807. [501.3800, 245.6600, 697.0000, 185.0000, 352.0000, 136.0000, 500.0000,
  808. 312.0000],
  809. [498.2935, 246.4364, 651.0000, 173.0000, 380.0000, 103.0000, 465.0000,
  810. 324.0000]], device='cuda:0')
  811. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  812. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  813. loss_train: nan
  814. step: 27
  815. running loss: nan
  816. Train Steps: 27/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  817. [nan, nan, nan, nan, nan, nan, nan, nan],
  818. [nan, nan, nan, nan, nan, nan, nan, nan],
  819. [nan, nan, nan, nan, nan, nan, nan, nan],
  820. [nan, nan, nan, nan, nan, nan, nan, nan],
  821. [nan, nan, nan, nan, nan, nan, nan, nan],
  822. [nan, nan, nan, nan, nan, nan, nan, nan],
  823. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  824. grad_fn=<AddmmBackward>)
  825. landmarks are: tensor([[489.3200, 241.0900, 525.0000, 118.0000, 299.0000, 153.0000, 422.3705,
  826. 306.1910],
  827. [495.7400, 245.4600, 704.0000, 287.0000, 283.0000, 286.0000, 476.0000,
  828. 333.0000],
  829. [501.0900, 244.0100, 724.0000, 251.0000, 302.0000, 276.0000, 504.6415,
  830. 291.7443],
  831. [486.8400, 238.8300, 696.0000, 285.0000, 361.0000, 317.0000, 425.9320,
  832. 302.2865],
  833. [490.9700, 242.1100, 708.0000, 265.0000, 312.0000, 257.0000, 435.9043,
  834. 337.0366],
  835. [500.7400, 249.3600, 705.0000, 191.0000, 382.0000, 112.0000, 497.5185,
  836. 297.6010],
  837. [500.1400, 249.4700, 719.0000, 245.0000, 303.0000, 287.0000, 498.0000,
  838. 338.0000],
  839. [495.9800, 246.0800, 707.0000, 319.0000, 306.0000, 228.0000, 433.0000,
  840. 341.0000]], device='cuda:0')
  841. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  842. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  843. loss_train: nan
  844. step: 28
  845. running loss: nan
  846. Train Steps: 28/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  847. [nan, nan, nan, nan, nan, nan, nan, nan],
  848. [nan, nan, nan, nan, nan, nan, nan, nan],
  849. [nan, nan, nan, nan, nan, nan, nan, nan],
  850. [nan, nan, nan, nan, nan, nan, nan, nan],
  851. [nan, nan, nan, nan, nan, nan, nan, nan],
  852. [nan, nan, nan, nan, nan, nan, nan, nan],
  853. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  854. grad_fn=<AddmmBackward>)
  855. landmarks are: tensor([[496.8673, 236.1169, 690.9001, 290.4881, 353.3458, 307.5619, 506.4407,
  856. 316.0934],
  857. [496.7200, 235.1800, 692.0000, 322.0000, 352.0000, 304.0000, 482.0000,
  858. 297.0000],
  859. [491.4477, 243.3093, 652.0000, 166.0000, 306.0000, 154.0000, 413.8228,
  860. 294.0870],
  861. [498.2935, 246.4364, 651.0000, 173.0000, 380.0000, 103.0000, 465.0000,
  862. 324.0000],
  863. [504.1500, 240.4100, 708.0000, 330.0000, 289.0000, 271.0000, 506.7784,
  864. 300.7246],
  865. [498.2935, 243.9347, 681.0000, 343.0000, 360.0000, 303.0000, 482.0000,
  866. 321.0000],
  867. [494.7600, 244.7600, 707.0000, 277.0000, 387.0000, 339.0000, 494.0000,
  868. 351.0000],
  869. [492.8600, 245.9100, 699.0000, 263.0000, 303.0000, 329.0000, 448.3696,
  870. 301.1151]], device='cuda:0')
  871. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  872.  
  873. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  874. loss_train: nan
  875. step: 29
  876. running loss: nan
  877. Train Steps: 29/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  878. [nan, nan, nan, nan, nan, nan, nan, nan],
  879. [nan, nan, nan, nan, nan, nan, nan, nan],
  880. [nan, nan, nan, nan, nan, nan, nan, nan],
  881. [nan, nan, nan, nan, nan, nan, nan, nan],
  882. [nan, nan, nan, nan, nan, nan, nan, nan],
  883. [nan, nan, nan, nan, nan, nan, nan, nan],
  884. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  885. grad_fn=<AddmmBackward>)
  886. landmarks are: tensor([[492.7800, 245.5500, 665.1248, 265.0410, 294.0000, 275.0000, 420.0331,
  887. 348.4859],
  888. [492.8739, 241.4330, 693.0000, 281.0000, 394.7994, 321.8399, 485.0000,
  889. 334.0000],
  890. [501.9800, 249.4200, 667.0000, 348.0000, 301.0000, 252.0000, 443.7396,
  891. 367.4918],
  892. [484.5200, 240.6700, 700.0000, 256.0000, 352.0000, 348.0000, 467.6018,
  893. 335.0844],
  894. [496.2700, 244.6800, 704.0000, 305.0000, 312.0000, 300.0000, 488.0000,
  895. 335.0000],
  896. [504.2836, 241.7457, 673.0000, 313.0000, 330.0000, 337.0000, 567.1599,
  897. 340.7306],
  898. [490.2000, 245.0600, 699.0000, 281.0000, 289.0000, 222.0000, 396.7645,
  899. 323.8376],
  900. [498.9000, 245.0400, 619.0000, 128.0000, 293.0000, 194.0000, 465.0000,
  901. 334.0000]], device='cuda:0')
  902. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  903. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  904. loss_train: nan
  905. step: 30
  906. running loss: nan
  907. Train Steps: 30/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  908. [nan, nan, nan, nan, nan, nan, nan, nan],
  909. [nan, nan, nan, nan, nan, nan, nan, nan],
  910. [nan, nan, nan, nan, nan, nan, nan, nan],
  911. [nan, nan, nan, nan, nan, nan, nan, nan],
  912. [nan, nan, nan, nan, nan, nan, nan, nan],
  913. [nan, nan, nan, nan, nan, nan, nan, nan],
  914. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  915. grad_fn=<AddmmBackward>)
  916. landmarks are: tensor([[494.0148, 240.8076, 618.6979, 167.0053, 357.9991, 107.1851, 463.1682,
  917. 321.0350],
  918. [495.8500, 249.8600, 696.0000, 304.0000, 370.0000, 339.0000, 437.0986,
  919. 311.8313],
  920. [502.0200, 242.8900, 679.0000, 173.0000, 357.0000, 122.0000, 505.7100,
  921. 309.3146],
  922. [497.7500, 250.2900, 708.0000, 313.0000, 299.0000, 276.0000, 456.0000,
  923. 338.0000],
  924. [508.5623, 245.8110, 723.0000, 233.0000, 337.0000, 177.0000, 534.8688,
  925. 323.3772],
  926. [488.0700, 242.5300, 622.0000, 157.0000, 297.0000, 169.0000, 435.1920,
  927. 338.9889],
  928. [487.2800, 239.8400, 665.1248, 260.0376, 303.0000, 273.0000, 417.0651,
  929. 339.3581],
  930. [ nan, nan, 643.0000, 149.0000, 318.0000, 151.0000, 446.0000,
  931. 336.0000]], device='cuda:0')
  932. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  933. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  934. loss_train: nan
  935. step: 31
  936. running loss: nan
  937. Train Steps: 31/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  938. [nan, nan, nan, nan, nan, nan, nan, nan],
  939. [nan, nan, nan, nan, nan, nan, nan, nan],
  940. [nan, nan, nan, nan, nan, nan, nan, nan],
  941. [nan, nan, nan, nan, nan, nan, nan, nan],
  942. [nan, nan, nan, nan, nan, nan, nan, nan],
  943. [nan, nan, nan, nan, nan, nan, nan, nan],
  944. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  945. grad_fn=<AddmmBackward>)
  946. landmarks are: tensor([[495.8300, 246.2900, 636.0000, 196.0000, 294.0000, 226.0000, 483.0000,
  947. 370.0000],
  948. [488.2000, 240.2900, 695.0000, 306.0000, 385.0000, 324.0000, 432.3427,
  949. 303.8483],
  950. [499.7000, 245.3900, 557.1127, 121.6097, 314.0000, 161.0000, 487.0000,
  951. 335.0000],
  952. [505.9951, 243.9347, 675.0000, 321.0000, 314.0000, 316.0000, 569.0593,
  953. 347.6719],
  954. [500.8400, 246.3800, 642.0000, 155.0000, 364.0000, 112.0000, 502.5046,
  955. 292.1347],
  956. [499.1492, 246.4364, 652.9544, 165.8065, 290.0000, 216.0000, 479.0000,
  957. 342.0000],
  958. [494.1500, 245.1300, 699.0000, 237.0000, 302.0000, 336.0000, 498.0000,
  959. 342.0000],
  960. [489.0500, 245.2700, 548.7455, 132.4504, 349.0000, 102.0000, 415.0000,
  961. 332.0000]], device='cuda:0')
  962. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  963. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  964. loss_train: nan
  965. step: 32
  966. running loss: nan
  967. Train Steps: 32/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  968. [nan, nan, nan, nan, nan, nan, nan, nan],
  969. [nan, nan, nan, nan, nan, nan, nan, nan],
  970. [nan, nan, nan, nan, nan, nan, nan, nan],
  971. [nan, nan, nan, nan, nan, nan, nan, nan],
  972. [nan, nan, nan, nan, nan, nan, nan, nan],
  973. [nan, nan, nan, nan, nan, nan, nan, nan],
  974. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  975. grad_fn=<AddmmBackward>)
  976. landmarks are: tensor([[495.4600, 246.4600, 595.0000, 162.0000, 292.0000, 221.0000, 499.0000,
  977. 343.0000],
  978. [495.2000, 248.0900, 640.0000, 293.0000, 285.2831, 218.8381, 449.0000,
  979. 354.0000],
  980. [489.9700, 238.4500, 618.0000, 151.0000, 283.0000, 199.0000, 471.0000,
  981. 330.0000],
  982. [499.5800, 246.0100, 621.0000, 155.0000, 397.0000, 91.0000, 470.0000,
  983. 325.0000],
  984. [490.2500, 241.3400, 699.0000, 304.0000, 398.6197, 313.8339, 429.1374,
  985. 303.8483],
  986. [497.7300, 244.3100, 573.0863, 129.9487, 299.0000, 190.0000, 488.0000,
  987. 332.0000],
  988. [502.0300, 244.2400, 646.0000, 126.0000, 376.0000, 92.0000, 491.8140,
  989. 290.0587],
  990. [490.8772, 241.7457, 661.0000, 201.0000, 290.0000, 184.0000, 454.0000,
  991. 310.0000]], device='cuda:0')
  992. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  993. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  994. loss_train: nan
  995. step: 33
  996. running loss: nan
  997. Train Steps: 33/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  998. [nan, nan, nan, nan, nan, nan, nan, nan],
  999. [nan, nan, nan, nan, nan, nan, nan, nan],
  1000. [nan, nan, nan, nan, nan, nan, nan, nan],
  1001. [nan, nan, nan, nan, nan, nan, nan, nan],
  1002. [nan, nan, nan, nan, nan, nan, nan, nan],
  1003. [nan, nan, nan, nan, nan, nan, nan, nan],
  1004. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1005. grad_fn=<AddmmBackward>)
  1006. landmarks are: tensor([[503.6500, 239.3200, 720.0000, 249.0000, 289.0000, 232.0000, 512.0000,
  1007. 306.0000],
  1008. [494.7600, 244.7600, 707.0000, 277.0000, 387.0000, 339.0000, 494.0000,
  1009. 351.0000],
  1010. [489.9700, 238.4500, 618.0000, 151.0000, 283.0000, 199.0000, 471.0000,
  1011. 330.0000],
  1012. [492.8739, 244.5602, 712.0000, 280.0000, 330.0000, 355.0000, 501.0000,
  1013. 322.0000],
  1014. [495.9600, 234.7900, 703.0000, 313.0000, 326.0000, 306.0000, 484.8128,
  1015. 294.7599],
  1016. [491.1624, 242.0584, 708.0000, 259.0000, 343.0000, 304.0000, 466.0000,
  1017. 332.0000],
  1018. [495.0900, 241.7500, 670.0000, 346.0000, 379.6035, 289.7198, 444.0957,
  1019. 335.8653],
  1020. [484.8000, 235.4200, 676.0000, 343.0000, 336.0000, 313.0000, 420.2336,
  1021. 285.1066]], device='cuda:0')
  1022. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1023. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1024. loss_train: nan
  1025. step: 34
  1026. running loss: nan
  1027. Train Steps: 34/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1028. [nan, nan, nan, nan, nan, nan, nan, nan],
  1029. [nan, nan, nan, nan, nan, nan, nan, nan],
  1030. [nan, nan, nan, nan, nan, nan, nan, nan],
  1031. [nan, nan, nan, nan, nan, nan, nan, nan],
  1032. [nan, nan, nan, nan, nan, nan, nan, nan],
  1033. [nan, nan, nan, nan, nan, nan, nan, nan],
  1034. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1035. grad_fn=<AddmmBackward>)
  1036. landmarks are: tensor([[484.6600, 239.1800, 665.8854, 277.5496, 307.0000, 299.0000, 411.7228,
  1037. 327.9374],
  1038. [490.5919, 243.9347, 580.6928, 144.1250, 287.0000, 198.0000, 480.0000,
  1039. 336.0000],
  1040. [495.8700, 247.7900, 701.0000, 247.0000, 292.0000, 294.0000, 456.5611,
  1041. 306.1910],
  1042. [ nan, nan, 617.9645, 156.6336, 294.0000, 164.0000, 433.0000,
  1043. 310.0000],
  1044. [503.4279, 241.4330, 700.0000, 300.0000, 321.0000, 344.0000, 569.6925,
  1045. 337.9540],
  1046. [490.5200, 243.8100, 691.0000, 312.0000, 383.0000, 287.0000, 420.6341,
  1047. 352.0000],
  1048. [496.1200, 243.3100, 617.3571, 115.1269, 340.0202, 124.7101, 469.6721,
  1049. 302.6288],
  1050. [505.1393, 246.4364, 700.0000, 306.0000, 303.0000, 294.0000, 569.6925,
  1051. 351.8367]], device='cuda:0')
  1052. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1053.  
  1054. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1055. loss_train: nan
  1056. step: 35
  1057. running loss: nan
  1058. Train Steps: 35/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1059. [nan, nan, nan, nan, nan, nan, nan, nan],
  1060. [nan, nan, nan, nan, nan, nan, nan, nan],
  1061. [nan, nan, nan, nan, nan, nan, nan, nan],
  1062. [nan, nan, nan, nan, nan, nan, nan, nan],
  1063. [nan, nan, nan, nan, nan, nan, nan, nan],
  1064. [nan, nan, nan, nan, nan, nan, nan, nan],
  1065. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1066. grad_fn=<AddmmBackward>)
  1067. landmarks are: tensor([[496.2100, 244.5800, 688.8793, 172.7032, 324.0115, 153.2296, 472.5629,
  1068. 329.7797],
  1069. [488.7600, 236.5500, 682.0000, 297.0000, 347.0000, 288.0000, 435.9043,
  1070. 322.9804],
  1071. [490.7900, 246.9100, 707.0000, 280.0000, 343.0000, 363.0000, 462.2595,
  1072. 305.8005],
  1073. [494.3001, 239.5567, 714.0048, 287.3219, 310.3356, 293.9974, 483.3020,
  1074. 316.7523],
  1075. [493.1591, 237.3677, 700.5271, 305.2639, 343.9919, 319.1815, 481.7994,
  1076. 312.1400],
  1077. [501.7164, 242.3712, 731.0000, 225.0000, 370.0000, 157.0000, 578.5567,
  1078. 324.7654],
  1079. [495.7000, 245.4200, 676.0000, 234.0000, 286.0000, 236.0000, 478.0000,
  1080. 335.0000],
  1081. [508.2770, 247.6873, 679.0000, 156.0000, 441.9766, 96.9323, 535.5019,
  1082. 332.4009]], device='cuda:0')
  1083. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1084. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1085. loss_train: nan
  1086. step: 36
  1087. running loss: nan
  1088. Train Steps: 36/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1089. [nan, nan, nan, nan, nan, nan, nan, nan],
  1090. [nan, nan, nan, nan, nan, nan, nan, nan],
  1091. [nan, nan, nan, nan, nan, nan, nan, nan],
  1092. [nan, nan, nan, nan, nan, nan, nan, nan],
  1093. [nan, nan, nan, nan, nan, nan, nan, nan],
  1094. [nan, nan, nan, nan, nan, nan, nan, nan],
  1095. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1096. grad_fn=<AddmmBackward>)
  1097. landmarks are: tensor([[488.6800, 240.5300, 698.0000, 282.0000, 284.0000, 250.0000, 452.0000,
  1098. 307.0000],
  1099. [ nan, nan, 517.5590, 116.6062, 322.0000, 120.0000, 410.0000,
  1100. 332.0000],
  1101. [487.6300, 240.1400, 682.6197, 310.0718, 402.5010, 305.6570, 410.0000,
  1102. 326.0000],
  1103. [496.1200, 249.0500, 687.0000, 328.0000, 296.0000, 237.0000, 451.0000,
  1104. 356.0000],
  1105. [502.0016, 242.9966, 723.0000, 226.0000, 307.0000, 212.0000, 565.8936,
  1106. 334.4833],
  1107. [490.3900, 244.2900, 684.0000, 274.0000, 291.0000, 220.0000, 423.2385,
  1108. 353.5374],
  1109. [504.5688, 242.6839, 642.0000, 350.0000, 302.0000, 292.0000, 551.3309,
  1110. 327.5420],
  1111. [487.2000, 240.6200, 627.0000, 209.0000, 283.0000, 227.0000, 436.9727,
  1112. 304.6292]], device='cuda:0')
  1113. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1114. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1115. loss_train: nan
  1116. step: 37
  1117. running loss: nan
  1118. Train Steps: 37/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1119. [nan, nan, nan, nan, nan, nan, nan, nan],
  1120. [nan, nan, nan, nan, nan, nan, nan, nan],
  1121. [nan, nan, nan, nan, nan, nan, nan, nan],
  1122. [nan, nan, nan, nan, nan, nan, nan, nan],
  1123. [nan, nan, nan, nan, nan, nan, nan, nan],
  1124. [nan, nan, nan, nan, nan, nan, nan, nan],
  1125. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1126. grad_fn=<AddmmBackward>)
  1127. landmarks are: tensor([[ nan, nan, 593.0000, 132.0000, 356.0000, 91.0000, 425.0000,
  1128. 299.0000],
  1129. [502.4100, 246.0400, 724.0000, 272.0000, 302.0000, 193.0000, 507.0098,
  1130. 294.9176],
  1131. [491.6300, 240.4700, 700.0000, 323.0000, 318.0000, 279.0000, 445.0000,
  1132. 332.0000],
  1133. [499.8300, 241.9100, 619.0000, 114.0000, 385.0000, 84.0000, 475.3155,
  1134. 294.2393],
  1135. [483.3700, 239.4200, 546.4636, 172.4778, 280.0000, 188.0000, 411.4557,
  1136. 330.5729],
  1137. [504.1500, 240.4100, 708.0000, 330.0000, 289.0000, 271.0000, 506.7784,
  1138. 300.7246],
  1139. [489.7700, 242.9000, 600.4697, 180.8168, 278.0000, 200.0000, 438.9684,
  1140. 344.9220],
  1141. [507.7065, 248.6254, 727.0000, 280.0000, 308.0000, 260.0000, 569.6925,
  1142. 353.9191]], device='cuda:0')
  1143. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1144. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1145. loss_train: nan
  1146. step: 38
  1147. running loss: nan
  1148. Train Steps: 38/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1149. [nan, nan, nan, nan, nan, nan, nan, nan],
  1150. [nan, nan, nan, nan, nan, nan, nan, nan],
  1151. [nan, nan, nan, nan, nan, nan, nan, nan],
  1152. [nan, nan, nan, nan, nan, nan, nan, nan],
  1153. [nan, nan, nan, nan, nan, nan, nan, nan],
  1154. [nan, nan, nan, nan, nan, nan, nan, nan],
  1155. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1156. grad_fn=<AddmmBackward>)
  1157. landmarks are: tensor([[487.6300, 240.1400, 682.6197, 310.0718, 402.5010, 305.6570, 410.0000,
  1158. 326.0000],
  1159. [501.9800, 249.4200, 667.0000, 348.0000, 301.0000, 252.0000, 443.7396,
  1160. 367.4918],
  1161. [506.2803, 251.7526, 739.0000, 275.0000, 341.0000, 176.0000, 587.4209,
  1162. 344.8954],
  1163. [502.4100, 246.0400, 724.0000, 272.0000, 302.0000, 193.0000, 507.0098,
  1164. 294.9176],
  1165. [493.1591, 244.5602, 707.0000, 247.0000, 297.0000, 333.0000, 499.0000,
  1166. 321.0000],
  1167. [495.8300, 246.9000, 621.0000, 163.0000, 297.0000, 192.0000, 467.0000,
  1168. 341.0000],
  1169. [497.0400, 240.0600, 617.0000, 127.0000, 347.0000, 108.0000, 468.0000,
  1170. 311.0000],
  1171. [501.1459, 243.3093, 674.0000, 166.0000, 354.0000, 166.0000, 563.9941,
  1172. 335.1775]], device='cuda:0')
  1173. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1174. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1175. loss_train: nan
  1176. step: 39
  1177. running loss: nan
  1178. Train Steps: 39/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1179. [nan, nan, nan, nan, nan, nan, nan, nan],
  1180. [nan, nan, nan, nan, nan, nan, nan, nan],
  1181. [nan, nan, nan, nan, nan, nan, nan, nan],
  1182. [nan, nan, nan, nan, nan, nan, nan, nan],
  1183. [nan, nan, nan, nan, nan, nan, nan, nan],
  1184. [nan, nan, nan, nan, nan, nan, nan, nan],
  1185. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1186. grad_fn=<AddmmBackward>)
  1187. landmarks are: tensor([[490.5919, 247.0619, 656.0000, 218.0000, 285.0000, 324.0000, 462.9718,
  1188. 308.1432],
  1189. [497.4400, 245.9000, 579.0000, 111.0000, 339.0000, 151.0000, 503.0000,
  1190. 321.0000],
  1191. [507.7065, 245.4983, 617.0000, 355.0000, 323.0000, 286.0000, 587.4209,
  1192. 343.5071],
  1193. [500.8000, 249.1900, 720.0000, 272.0000, 322.0000, 158.0000, 497.8747,
  1194. 297.9915],
  1195. [488.0247, 244.2474, 608.0761, 206.6678, 272.0000, 247.0000, 450.0000,
  1196. 337.0000],
  1197. [494.5853, 235.8042, 707.2659, 233.4960, 284.4970, 297.9904, 497.7261,
  1198. 316.7523],
  1199. [490.9700, 242.1100, 708.0000, 265.0000, 312.0000, 257.0000, 435.9043,
  1200. 337.0366],
  1201. [500.5754, 241.4330, 693.7882, 325.3166, 335.6755, 311.8676, 535.5466,
  1202. 312.2101]], device='cuda:0')
  1203. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1204. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1205. loss_train: nan
  1206. step: 40
  1207. running loss: nan
  1208. Train Steps: 40/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1209. [nan, nan, nan, nan, nan, nan, nan, nan],
  1210. [nan, nan, nan, nan, nan, nan, nan, nan],
  1211. [nan, nan, nan, nan, nan, nan, nan, nan],
  1212. [nan, nan, nan, nan, nan, nan, nan, nan],
  1213. [nan, nan, nan, nan, nan, nan, nan, nan],
  1214. [nan, nan, nan, nan, nan, nan, nan, nan],
  1215. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1216. grad_fn=<AddmmBackward>)
  1217. landmarks are: tensor([[494.5600, 245.9300, 625.0000, 180.0000, 315.0000, 142.0000, 426.0000,
  1218. 345.0000],
  1219. [491.2700, 242.2900, 692.0000, 290.0000, 387.0979, 310.5822, 463.0000,
  1220. 336.0000],
  1221. [498.6600, 245.4800, 648.0000, 177.0000, 285.0000, 233.0000, 481.0000,
  1222. 312.0000],
  1223. [507.1360, 247.3746, 691.0000, 322.0000, 326.0000, 328.0000, 601.3504,
  1224. 326.1537],
  1225. [500.0049, 240.4949, 716.0000, 251.0000, 284.0000, 263.0000, 508.9154,
  1226. 295.6488],
  1227. [502.8574, 256.4433, 680.0000, 270.0000, 362.0000, 155.0000, 435.1920,
  1228. 372.5677],
  1229. [500.9500, 245.1100, 675.0000, 189.0000, 322.0000, 158.0000, 507.1346,
  1230. 288.6207],
  1231. [496.1200, 249.0500, 687.0000, 328.0000, 296.0000, 237.0000, 451.0000,
  1232. 356.0000]], device='cuda:0')
  1233. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1234.  
  1235. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1236. loss_train: nan
  1237. step: 41
  1238. running loss: nan
  1239. Train Steps: 41/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1240. [nan, nan, nan, nan, nan, nan, nan, nan],
  1241. [nan, nan, nan, nan, nan, nan, nan, nan],
  1242. [nan, nan, nan, nan, nan, nan, nan, nan],
  1243. [nan, nan, nan, nan, nan, nan, nan, nan],
  1244. [nan, nan, nan, nan, nan, nan, nan, nan],
  1245. [nan, nan, nan, nan, nan, nan, nan, nan],
  1246. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1247. grad_fn=<AddmmBackward>)
  1248. landmarks are: tensor([[487.5800, 238.2200, 695.0000, 286.0000, 388.7642, 292.7355, 415.2475,
  1249. 296.4297],
  1250. [502.0016, 240.1822, 728.0000, 227.0000, 351.0000, 188.0000, 564.6272,
  1251. 320.6006],
  1252. [496.4300, 240.2100, 715.0000, 293.0000, 293.0000, 300.0000, 508.5592,
  1253. 296.8201],
  1254. [502.8574, 243.3093, 720.0000, 283.0000, 301.0000, 281.0000, 561.4614,
  1255. 329.6244],
  1256. [490.1500, 245.0200, 696.0000, 268.0000, 319.0000, 259.0000, 401.0383,
  1257. 328.2302],
  1258. [495.7500, 245.3800, 626.0000, 150.0000, 336.0000, 149.0000, 479.0000,
  1259. 340.0000],
  1260. [501.0800, 241.7700, 720.0000, 286.0000, 304.0000, 310.0000, 513.1892,
  1261. 286.2780],
  1262. [486.6400, 237.4500, 691.0000, 297.0000, 349.0000, 305.0000, 427.7128,
  1263. 298.7724]], device='cuda:0')
  1264. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1265. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1266. loss_train: nan
  1267. step: 42
  1268. running loss: nan
  1269. Train Steps: 42/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1270. [nan, nan, nan, nan, nan, nan, nan, nan],
  1271. [nan, nan, nan, nan, nan, nan, nan, nan],
  1272. [nan, nan, nan, nan, nan, nan, nan, nan],
  1273. [nan, nan, nan, nan, nan, nan, nan, nan],
  1274. [nan, nan, nan, nan, nan, nan, nan, nan],
  1275. [nan, nan, nan, nan, nan, nan, nan, nan],
  1276. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1277. grad_fn=<AddmmBackward>)
  1278. landmarks are: tensor([[500.6500, 242.2600, 700.0000, 293.0000, 312.0000, 330.0000, 510.0000,
  1279. 313.0000],
  1280. [498.7100, 250.7200, 626.0000, 207.0000, 305.0000, 172.0000, 454.0000,
  1281. 337.0000],
  1282. [509.1327, 245.4983, 682.0000, 338.0000, 310.0000, 297.0000, 527.9040,
  1283. 317.1299],
  1284. [496.8673, 236.1169, 690.9001, 290.4881, 353.3458, 307.5619, 506.4407,
  1285. 316.0934],
  1286. [503.4279, 238.9313, 696.0000, 318.0000, 301.0000, 283.0000, 563.9941,
  1287. 317.8241],
  1288. [489.4300, 243.1300, 677.0000, 235.0000, 283.0000, 207.0000, 404.2437,
  1289. 335.5512],
  1290. [496.0200, 244.2800, 587.0000, 115.0000, 336.0000, 147.0000, 492.0000,
  1291. 331.0000],
  1292. [501.1800, 245.4400, 716.0000, 212.0000, 288.0000, 238.0000, 503.5731,
  1293. 294.0870]], device='cuda:0')
  1294. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1295. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1296. loss_train: nan
  1297. step: 43
  1298. running loss: nan
  1299. Train Steps: 43/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1300. [nan, nan, nan, nan, nan, nan, nan, nan],
  1301. [nan, nan, nan, nan, nan, nan, nan, nan],
  1302. [nan, nan, nan, nan, nan, nan, nan, nan],
  1303. [nan, nan, nan, nan, nan, nan, nan, nan],
  1304. [nan, nan, nan, nan, nan, nan, nan, nan],
  1305. [nan, nan, nan, nan, nan, nan, nan, nan],
  1306. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1307. grad_fn=<AddmmBackward>)
  1308. landmarks are: tensor([[502.0500, 240.1100, 704.0000, 332.0000, 286.0000, 264.0000, 490.5476,
  1309. 280.3407],
  1310. [503.9400, 258.1900, 637.0000, 236.0000, 388.0000, 137.0000, 438.3973,
  1311. 373.3486],
  1312. [500.9100, 247.8700, 715.0000, 213.0000, 320.0000, 161.0000, 495.0000,
  1313. 317.0000],
  1314. [506.5656, 247.6873, 736.0000, 211.0000, 352.0000, 230.0000, 596.9183,
  1315. 329.6244],
  1316. [492.5886, 237.3677, 665.8701, 248.2718, 278.2819, 265.8849, 473.0849,
  1317. 307.8573],
  1318. [497.2500, 245.9600, 578.0000, 122.0000, 335.0000, 133.0000, 478.0000,
  1319. 317.0000],
  1320. [490.3600, 243.8000, 699.0000, 315.0000, 345.0000, 284.0000, 418.4304,
  1321. 352.4393],
  1322. [492.0800, 243.5100, 565.4799, 160.8031, 272.0000, 245.0000, 462.0000,
  1323. 344.0000]], device='cuda:0')
  1324. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1325. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1326. loss_train: nan
  1327. step: 44
  1328. running loss: nan
  1329. Train Steps: 44/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1330. [nan, nan, nan, nan, nan, nan, nan, nan],
  1331. [nan, nan, nan, nan, nan, nan, nan, nan],
  1332. [nan, nan, nan, nan, nan, nan, nan, nan],
  1333. [nan, nan, nan, nan, nan, nan, nan, nan],
  1334. [nan, nan, nan, nan, nan, nan, nan, nan],
  1335. [nan, nan, nan, nan, nan, nan, nan, nan],
  1336. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1337. grad_fn=<AddmmBackward>)
  1338. landmarks are: tensor([[ nan, nan, 594.0000, 127.0000, 315.0000, 146.0000, 435.0000,
  1339. 334.0000],
  1340. [507.9918, 247.0619, 669.0000, 163.0000, 388.0000, 102.0000, 515.2408,
  1341. 310.1886],
  1342. [488.6800, 240.5300, 698.0000, 282.0000, 284.0000, 250.0000, 452.0000,
  1343. 307.0000],
  1344. [496.2400, 244.3600, 655.1108, 143.9094, 352.0268, 123.2475, 474.3377,
  1345. 330.0576],
  1346. [495.9500, 244.2800, 608.0000, 127.0000, 323.0000, 166.0000, 491.0000,
  1347. 333.0000],
  1348. [486.0900, 237.1500, 650.0000, 235.0000, 282.0000, 245.0000, 427.7128,
  1349. 297.2106],
  1350. [484.4000, 240.8700, 594.0000, 122.0000, 329.0000, 113.0000, 417.3844,
  1351. 289.4016],
  1352. [483.0400, 236.7800, 673.0000, 293.0000, 285.0000, 273.0000, 421.3020,
  1353. 281.5925]], device='cuda:0')
  1354. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1355. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1356. loss_train: nan
  1357. step: 45
  1358. running loss: nan
  1359. Train Steps: 45/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1360. [nan, nan, nan, nan, nan, nan, nan, nan],
  1361. [nan, nan, nan, nan, nan, nan, nan, nan],
  1362. [nan, nan, nan, nan, nan, nan, nan, nan],
  1363. [nan, nan, nan, nan, nan, nan, nan, nan],
  1364. [nan, nan, nan, nan, nan, nan, nan, nan],
  1365. [nan, nan, nan, nan, nan, nan, nan, nan],
  1366. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1367. grad_fn=<AddmmBackward>)
  1368. landmarks are: tensor([[493.8100, 246.8200, 586.0000, 150.0000, 337.0000, 115.0000, 427.0000,
  1369. 342.0000],
  1370. [488.9500, 241.7100, 691.0000, 288.0000, 390.0000, 305.0000, 461.0000,
  1371. 334.0000],
  1372. [487.8500, 239.4600, 691.0000, 283.0000, 341.0000, 298.0000, 417.0000,
  1373. 339.0000],
  1374. [495.3700, 247.6800, 681.0000, 337.0000, 336.0000, 316.0000, 468.0000,
  1375. 338.0000],
  1376. [488.1700, 239.9200, 700.0000, 308.0000, 306.0000, 285.0000, 451.0000,
  1377. 305.0000],
  1378. [495.9100, 243.6200, 711.0000, 280.0000, 304.0000, 303.0000, 495.0000,
  1379. 326.0000],
  1380. [497.7600, 236.2000, 668.0000, 337.0000, 331.0000, 276.0000, 464.0000,
  1381. 314.0000],
  1382. [500.9500, 245.1100, 675.0000, 189.0000, 322.0000, 158.0000, 507.1346,
  1383. 288.6207]], device='cuda:0')
  1384. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1385. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1386. loss_train: nan
  1387. step: 46
  1388. running loss: nan
  1389. Train Steps: 46/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1390. [nan, nan, nan, nan, nan, nan, nan, nan],
  1391. [nan, nan, nan, nan, nan, nan, nan, nan],
  1392. [nan, nan, nan, nan, nan, nan, nan, nan],
  1393. [nan, nan, nan, nan, nan, nan, nan, nan],
  1394. [nan, nan, nan, nan, nan, nan, nan, nan],
  1395. [nan, nan, nan, nan, nan, nan, nan, nan],
  1396. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1397. grad_fn=<AddmmBackward>)
  1398. landmarks are: tensor([[508.2770, 247.6873, 679.0000, 156.0000, 441.9766, 96.9323, 535.5019,
  1399. 332.4009],
  1400. [497.5100, 245.3300, 715.0000, 288.0000, 306.0000, 267.0000, 468.0000,
  1401. 312.0000],
  1402. [497.2900, 250.0000, 687.0000, 335.0000, 318.0000, 310.0000, 462.0000,
  1403. 340.0000],
  1404. [495.9300, 246.6900, 678.0000, 223.0000, 284.0000, 261.0000, 485.0000,
  1405. 365.0000],
  1406. [492.9800, 236.9300, 707.0000, 271.0000, 340.0000, 311.0000, 467.0000,
  1407. 330.0000],
  1408. [496.3600, 243.7000, 667.0000, 161.0000, 294.0000, 257.0000, 507.0000,
  1409. 315.0000],
  1410. [491.4000, 238.9100, 692.0000, 293.0000, 313.0000, 259.0000, 425.2197,
  1411. 321.0281],
  1412. [507.1360, 248.9381, 704.0000, 300.0000, 312.0000, 317.0000, 603.2499,
  1413. 325.4596]], device='cuda:0')
  1414. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1415.  
  1416. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1417. loss_train: nan
  1418. step: 47
  1419. running loss: nan
  1420. Train Steps: 47/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1421. [nan, nan, nan, nan, nan, nan, nan, nan],
  1422. [nan, nan, nan, nan, nan, nan, nan, nan],
  1423. [nan, nan, nan, nan, nan, nan, nan, nan],
  1424. [nan, nan, nan, nan, nan, nan, nan, nan],
  1425. [nan, nan, nan, nan, nan, nan, nan, nan],
  1426. [nan, nan, nan, nan, nan, nan, nan, nan],
  1427. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1428. grad_fn=<AddmmBackward>)
  1429. landmarks are: tensor([[501.0900, 254.8600, 715.0000, 251.0000, 310.0000, 246.0000, 432.0000,
  1430. 321.0000],
  1431. [ nan, nan, 617.9645, 156.6336, 294.0000, 164.0000, 433.0000,
  1432. 310.0000],
  1433. [498.6600, 245.4800, 648.0000, 177.0000, 285.0000, 233.0000, 481.0000,
  1434. 312.0000],
  1435. [492.5886, 237.3677, 665.8701, 248.2718, 278.2819, 265.8849, 473.0849,
  1436. 307.8573],
  1437. [483.3700, 239.4200, 546.4636, 172.4778, 280.0000, 188.0000, 411.4557,
  1438. 330.5729],
  1439. [494.0100, 239.8300, 539.0000, 150.0000, 345.0000, 116.0000, 441.0000,
  1440. 345.0000],
  1441. [488.6800, 240.5300, 698.0000, 282.0000, 284.0000, 250.0000, 452.0000,
  1442. 307.0000],
  1443. [492.4400, 247.4700, 708.0000, 290.0000, 364.0000, 349.0000, 461.1910,
  1444. 305.0196]], device='cuda:0')
  1445. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1446. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1447. loss_train: nan
  1448. step: 48
  1449. running loss: nan
  1450. Train Steps: 48/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1451. [nan, nan, nan, nan, nan, nan, nan, nan],
  1452. [nan, nan, nan, nan, nan, nan, nan, nan],
  1453. [nan, nan, nan, nan, nan, nan, nan, nan],
  1454. [nan, nan, nan, nan, nan, nan, nan, nan],
  1455. [nan, nan, nan, nan, nan, nan, nan, nan],
  1456. [nan, nan, nan, nan, nan, nan, nan, nan],
  1457. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1458. grad_fn=<AddmmBackward>)
  1459. landmarks are: tensor([[496.1000, 243.8800, 584.0000, 116.0000, 326.0000, 152.0000, 493.0000,
  1460. 329.0000],
  1461. [502.5721, 245.4983, 640.0000, 131.0000, 360.0000, 143.0000, 542.9840,
  1462. 321.8463],
  1463. [ nan, nan, 643.0000, 149.0000, 318.0000, 151.0000, 446.0000,
  1464. 336.0000],
  1465. [507.4213, 245.8110, 743.0000, 262.0000, 345.0000, 216.0000, 579.8230,
  1466. 350.4485],
  1467. [501.7700, 247.2200, 723.0000, 247.0000, 298.0000, 192.0000, 494.0000,
  1468. 315.0000],
  1469. [502.5721, 242.0584, 626.3995, 124.7889, 362.5892, 124.7976, 512.3327,
  1470. 319.3755],
  1471. [482.4700, 239.1800, 597.0000, 170.0000, 291.0000, 163.0000, 420.2336,
  1472. 283.5448],
  1473. [495.9400, 237.1000, 685.8777, 322.4308, 326.0126, 281.2020, 475.3770,
  1474. 322.6425]], device='cuda:0')
  1475. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1476. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1477. loss_train: nan
  1478. step: 49
  1479. running loss: nan
  1480. Train Steps: 49/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1481. [nan, nan, nan, nan, nan, nan, nan, nan],
  1482. [nan, nan, nan, nan, nan, nan, nan, nan],
  1483. [nan, nan, nan, nan, nan, nan, nan, nan],
  1484. [nan, nan, nan, nan, nan, nan, nan, nan],
  1485. [nan, nan, nan, nan, nan, nan, nan, nan],
  1486. [nan, nan, nan, nan, nan, nan, nan, nan],
  1487. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1488. grad_fn=<AddmmBackward>)
  1489. landmarks are: tensor([[501.4300, 243.1200, 654.0000, 135.0000, 375.0000, 115.0000, 510.0000,
  1490. 316.0000],
  1491. [496.2000, 243.9500, 671.8087, 158.8574, 314.0060, 157.6172, 467.5804,
  1492. 307.4238],
  1493. [489.4300, 243.1300, 677.0000, 235.0000, 283.0000, 207.0000, 404.2437,
  1494. 335.5512],
  1495. [503.1100, 241.8900, 673.4919, 326.7499, 326.0000, 301.0000, 505.0000,
  1496. 307.0000],
  1497. [501.0300, 253.9500, 633.9382, 277.5496, 303.0000, 173.0000, 445.8765,
  1498. 362.8064],
  1499. [499.9400, 248.5100, 668.0000, 197.0000, 289.0000, 222.0000, 495.0000,
  1500. 324.0000],
  1501. [495.3700, 247.6800, 681.0000, 337.0000, 336.0000, 316.0000, 468.0000,
  1502. 338.0000],
  1503. [495.4411, 242.9966, 585.0000, 146.0000, 326.0000, 127.0000, 451.9311,
  1504. 339.3793]], device='cuda:0')
  1505. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1506. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1507. loss_train: nan
  1508. step: 50
  1509. running loss: nan
  1510. Train Steps: 50/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1511. [nan, nan, nan, nan, nan, nan, nan, nan],
  1512. [nan, nan, nan, nan, nan, nan, nan, nan],
  1513. [nan, nan, nan, nan, nan, nan, nan, nan],
  1514. [nan, nan, nan, nan, nan, nan, nan, nan],
  1515. [nan, nan, nan, nan, nan, nan, nan, nan],
  1516. [nan, nan, nan, nan, nan, nan, nan, nan],
  1517. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1518. grad_fn=<AddmmBackward>)
  1519. landmarks are: tensor([[490.5919, 243.9347, 580.6928, 144.1250, 287.0000, 198.0000, 480.0000,
  1520. 336.0000],
  1521. [495.9600, 238.6900, 712.0000, 273.0000, 302.0000, 312.0000, 492.0000,
  1522. 322.0000],
  1523. [496.0400, 243.3100, 640.8530, 179.2846, 307.3357, 167.8550, 475.9475,
  1524. 310.9679],
  1525. [493.7296, 243.3093, 654.0000, 159.0000, 284.0000, 221.0000, 463.0000,
  1526. 333.0000],
  1527. [495.9100, 243.6200, 711.0000, 280.0000, 304.0000, 303.0000, 495.0000,
  1528. 326.0000],
  1529. [497.0000, 249.5400, 694.0000, 347.0000, 327.0000, 259.0000, 449.0000,
  1530. 355.0000],
  1531. [496.1000, 243.8800, 584.0000, 116.0000, 326.0000, 152.0000, 493.0000,
  1532. 329.0000],
  1533. [508.5623, 249.5636, 703.0000, 335.0000, 291.0000, 266.0000, 519.0398,
  1534. 317.8241]], device='cuda:0')
  1535. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1536. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1537. loss_train: nan
  1538. step: 51
  1539. running loss: nan
  1540. Train Steps: 51/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1541. [nan, nan, nan, nan, nan, nan, nan, nan],
  1542. [nan, nan, nan, nan, nan, nan, nan, nan],
  1543. [nan, nan, nan, nan, nan, nan, nan, nan],
  1544. [nan, nan, nan, nan, nan, nan, nan, nan],
  1545. [nan, nan, nan, nan, nan, nan, nan, nan],
  1546. [nan, nan, nan, nan, nan, nan, nan, nan],
  1547. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1548. grad_fn=<AddmmBackward>)
  1549. landmarks are: tensor([[486.9100, 241.1400, 622.5284, 256.7020, 290.0000, 261.0000, 411.9899,
  1550. 317.1024],
  1551. [493.9800, 237.4300, 696.0000, 289.0000, 373.0000, 308.0000, 464.0000,
  1552. 331.0000],
  1553. [491.1100, 241.1500, 571.0000, 129.0000, 350.0000, 92.0000, 423.4389,
  1554. 300.3342],
  1555. [492.0181, 245.8110, 597.4271, 191.6575, 306.0000, 158.0000, 437.0000,
  1556. 348.0000],
  1557. [498.1200, 244.9800, 715.0000, 288.0000, 304.0000, 177.0000, 459.0000,
  1558. 321.0000],
  1559. [484.6600, 239.1800, 665.8854, 277.5496, 307.0000, 299.0000, 411.7228,
  1560. 327.9374],
  1561. [490.0000, 242.0700, 626.0000, 186.0000, 277.0000, 294.0000, 466.5333,
  1562. 338.2080],
  1563. [500.6700, 248.5800, 682.0000, 157.0000, 396.0000, 100.0000, 497.5185,
  1564. 297.9915]], device='cuda:0')
  1565. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1566. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1567. loss_train: nan
  1568. step: 52
  1569. running loss: nan
  1570. Train Steps: 52/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1571. [nan, nan, nan, nan, nan, nan, nan, nan],
  1572. [nan, nan, nan, nan, nan, nan, nan, nan],
  1573. [nan, nan, nan, nan, nan, nan, nan, nan],
  1574. [nan, nan, nan, nan, nan, nan, nan, nan],
  1575. [nan, nan, nan, nan, nan, nan, nan, nan],
  1576. [nan, nan, nan, nan, nan, nan, nan, nan],
  1577. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1578. grad_fn=<AddmmBackward>)
  1579. landmarks are: tensor([[502.0400, 245.7100, 659.0000, 135.0000, 373.0000, 107.0000, 493.7134,
  1580. 292.1411],
  1581. [495.7300, 239.1300, 704.0000, 277.0000, 335.0000, 287.0000, 455.0000,
  1582. 333.0000],
  1583. [504.8541, 240.4949, 634.8398, 344.7328, 312.0000, 302.0000, 556.3961,
  1584. 321.9889],
  1585. [501.7700, 247.2200, 723.0000, 247.0000, 298.0000, 192.0000, 494.0000,
  1586. 315.0000],
  1587. [486.9100, 239.8900, 703.0000, 267.0000, 322.0000, 279.0000, 424.5074,
  1588. 306.1910],
  1589. [ nan, nan, 517.5590, 116.6062, 322.0000, 120.0000, 410.0000,
  1590. 332.0000],
  1591. [495.4411, 242.9966, 585.0000, 146.0000, 326.0000, 127.0000, 451.9311,
  1592. 339.3793],
  1593. [502.8574, 245.1856, 672.6089, 168.0607, 333.0000, 168.0000, 538.0257,
  1594. 323.5759]], device='cuda:0')
  1595. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1596.  
  1597. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1598. loss_train: nan
  1599. step: 53
  1600. running loss: nan
  1601. Train Steps: 53/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1602. [nan, nan, nan, nan, nan, nan, nan, nan],
  1603. [nan, nan, nan, nan, nan, nan, nan, nan],
  1604. [nan, nan, nan, nan, nan, nan, nan, nan],
  1605. [nan, nan, nan, nan, nan, nan, nan, nan],
  1606. [nan, nan, nan, nan, nan, nan, nan, nan],
  1607. [nan, nan, nan, nan, nan, nan, nan, nan],
  1608. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1609. grad_fn=<AddmmBackward>)
  1610. landmarks are: tensor([[497.9500, 245.8000, 595.0000, 136.0000, 308.0000, 171.0000, 479.0000,
  1611. 315.0000],
  1612. [490.9100, 237.3000, 672.0000, 196.0000, 280.0000, 252.0000, 469.0000,
  1613. 328.0000],
  1614. [500.7800, 249.6400, 720.0000, 230.0000, 356.0000, 128.0000, 498.9431,
  1615. 299.1628],
  1616. [491.8400, 239.7800, 679.0000, 232.0000, 279.0000, 244.0000, 469.0000,
  1617. 300.0000],
  1618. [493.8100, 240.7600, 697.0720, 278.3835, 384.0000, 342.0000, 482.0000,
  1619. 336.0000],
  1620. [504.5688, 241.7457, 719.0000, 289.0000, 315.0000, 210.0000, 584.8883,
  1621. 322.6830],
  1622. [499.9700, 254.1700, 691.0000, 226.0000, 324.0000, 189.0000, 451.9311,
  1623. 347.9693],
  1624. [ nan, nan, 543.4210, 126.6131, 321.0000, 130.0000, 409.0000,
  1625. 335.0000]], device='cuda:0')
  1626. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1627. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1628. loss_train: nan
  1629. step: 54
  1630. running loss: nan
  1631. Train Steps: 54/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1632. [nan, nan, nan, nan, nan, nan, nan, nan],
  1633. [nan, nan, nan, nan, nan, nan, nan, nan],
  1634. [nan, nan, nan, nan, nan, nan, nan, nan],
  1635. [nan, nan, nan, nan, nan, nan, nan, nan],
  1636. [nan, nan, nan, nan, nan, nan, nan, nan],
  1637. [nan, nan, nan, nan, nan, nan, nan, nan],
  1638. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1639. grad_fn=<AddmmBackward>)
  1640. landmarks are: tensor([[491.1624, 242.0584, 708.0000, 259.0000, 343.0000, 304.0000, 466.0000,
  1641. 332.0000],
  1642. [491.1800, 244.2000, 696.0000, 300.0000, 369.0000, 294.0000, 420.8344,
  1643. 351.1215],
  1644. [490.0100, 240.6100, 692.0000, 274.0000, 293.0000, 234.0000, 448.0000,
  1645. 334.0000],
  1646. [496.2100, 244.5800, 688.8793, 172.7032, 324.0115, 153.2296, 472.5629,
  1647. 329.7797],
  1648. [493.5900, 246.1400, 597.4271, 221.6781, 277.0000, 226.0000, 419.0314,
  1649. 349.3644],
  1650. [490.8772, 243.6220, 642.0000, 193.0000, 292.0000, 180.0000, 404.7779,
  1651. 338.7724],
  1652. [501.9300, 247.0000, 648.0244, 348.0684, 320.0000, 275.0000, 446.5888,
  1653. 367.1014],
  1654. [495.8300, 246.2900, 636.0000, 196.0000, 294.0000, 226.0000, 483.0000,
  1655. 370.0000]], device='cuda:0')
  1656. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1657. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1658. loss_train: nan
  1659. step: 55
  1660. running loss: nan
  1661. Train Steps: 55/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1662. [nan, nan, nan, nan, nan, nan, nan, nan],
  1663. [nan, nan, nan, nan, nan, nan, nan, nan],
  1664. [nan, nan, nan, nan, nan, nan, nan, nan],
  1665. [nan, nan, nan, nan, nan, nan, nan, nan],
  1666. [nan, nan, nan, nan, nan, nan, nan, nan],
  1667. [nan, nan, nan, nan, nan, nan, nan, nan],
  1668. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1669. grad_fn=<AddmmBackward>)
  1670. landmarks are: tensor([[498.8640, 238.6186, 718.8183, 288.3773, 306.7708, 312.9795, 529.0106,
  1671. 309.9863],
  1672. [497.9500, 245.8000, 595.0000, 136.0000, 308.0000, 171.0000, 479.0000,
  1673. 315.0000],
  1674. [501.4300, 246.1100, 715.0000, 220.0000, 322.0000, 170.0000, 502.0000,
  1675. 311.0000],
  1676. [ nan, nan, 724.0000, 210.0000, 411.0000, 138.0000, 588.6873,
  1677. 342.1188],
  1678. [503.1426, 241.4330, 727.0000, 274.0000, 315.0000, 338.0000, 564.6272,
  1679. 336.5657],
  1680. [495.0900, 241.9900, 692.0000, 316.0000, 359.0000, 309.0000, 474.0000,
  1681. 303.0000],
  1682. [509.1327, 248.6254, 690.0000, 185.0000, 393.0000, 120.0000, 515.8740,
  1683. 316.4358],
  1684. [493.5600, 243.1200, 699.3539, 286.7225, 343.0000, 295.0000, 461.0000,
  1685. 337.0000]], device='cuda:0')
  1686. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1687. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1688. loss_train: nan
  1689. step: 56
  1690. running loss: nan
  1691. Train Steps: 56/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1692. [nan, nan, nan, nan, nan, nan, nan, nan],
  1693. [nan, nan, nan, nan, nan, nan, nan, nan],
  1694. [nan, nan, nan, nan, nan, nan, nan, nan],
  1695. [nan, nan, nan, nan, nan, nan, nan, nan],
  1696. [nan, nan, nan, nan, nan, nan, nan, nan],
  1697. [nan, nan, nan, nan, nan, nan, nan, nan],
  1698. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1699. grad_fn=<AddmmBackward>)
  1700. landmarks are: tensor([[504.5688, 243.6220, 716.0000, 311.0000, 300.0000, 231.0000, 507.0098,
  1701. 296.3059],
  1702. [499.7000, 245.3900, 557.1127, 121.6097, 314.0000, 161.0000, 487.0000,
  1703. 335.0000],
  1704. [500.8000, 249.1900, 720.0000, 272.0000, 322.0000, 158.0000, 497.8747,
  1705. 297.9915],
  1706. [499.1492, 252.3780, 700.0000, 324.0000, 295.0000, 291.0000, 459.0000,
  1707. 342.0000],
  1708. [504.3800, 238.9900, 716.0000, 290.0000, 295.0000, 281.0000, 510.0000,
  1709. 307.0000],
  1710. [485.6100, 238.7500, 686.0000, 305.0000, 348.0000, 324.0000, 414.3940,
  1711. 327.9374],
  1712. [ nan, nan, 679.0000, 138.0000, 445.0000, 126.0000, 591.2198,
  1713. 340.7306],
  1714. [501.7164, 242.3712, 731.0000, 225.0000, 370.0000, 157.0000, 578.5567,
  1715. 324.7654]], device='cuda:0')
  1716. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1717. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1718. loss_train: nan
  1719. step: 57
  1720. running loss: nan
  1721. Train Steps: 57/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1722. [nan, nan, nan, nan, nan, nan, nan, nan],
  1723. [nan, nan, nan, nan, nan, nan, nan, nan],
  1724. [nan, nan, nan, nan, nan, nan, nan, nan],
  1725. [nan, nan, nan, nan, nan, nan, nan, nan],
  1726. [nan, nan, nan, nan, nan, nan, nan, nan],
  1727. [nan, nan, nan, nan, nan, nan, nan, nan],
  1728. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1729. grad_fn=<AddmmBackward>)
  1730. landmarks are: tensor([[493.3600, 242.8800, 546.4636, 218.3425, 301.0000, 181.0000, 456.0000,
  1731. 353.0000],
  1732. [495.8900, 246.1100, 716.0000, 265.0000, 321.0000, 322.0000, 489.0000,
  1733. 358.0000],
  1734. [488.7100, 240.8900, 613.4007, 218.3425, 281.0000, 220.0000, 415.9966,
  1735. 338.4796],
  1736. [489.9400, 241.8100, 692.0000, 292.0000, 399.9338, 306.3606, 410.9215,
  1737. 346.3862],
  1738. [488.0247, 244.2474, 608.0761, 206.6678, 272.0000, 247.0000, 450.0000,
  1739. 337.0000],
  1740. [495.3700, 238.8200, 566.2405, 164.9726, 340.0000, 126.0000, 436.0000,
  1741. 347.0000],
  1742. [499.2800, 248.5500, 715.0000, 279.0000, 326.0000, 321.0000, 500.0000,
  1743. 333.0000],
  1744. [484.2800, 242.0800, 551.7881, 114.9384, 320.0000, 127.0000, 435.1920,
  1745. 310.0955]], device='cuda:0')
  1746. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1747. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1748. loss_train: nan
  1749. step: 58
  1750. running loss: nan
  1751. Train Steps: 58/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1752. [nan, nan, nan, nan, nan, nan, nan, nan],
  1753. [nan, nan, nan, nan, nan, nan, nan, nan],
  1754. [nan, nan, nan, nan, nan, nan, nan, nan],
  1755. [nan, nan, nan, nan, nan, nan, nan, nan],
  1756. [nan, nan, nan, nan, nan, nan, nan, nan],
  1757. [nan, nan, nan, nan, nan, nan, nan, nan],
  1758. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1759. grad_fn=<AddmmBackward>)
  1760. landmarks are: tensor([[495.8300, 246.9000, 621.0000, 163.0000, 297.0000, 192.0000, 467.0000,
  1761. 341.0000],
  1762. [497.2500, 245.9600, 578.0000, 122.0000, 335.0000, 133.0000, 478.0000,
  1763. 317.0000],
  1764. [485.9900, 240.3900, 662.0000, 295.0000, 324.0000, 306.0000, 413.3255,
  1765. 316.8095],
  1766. [495.7400, 243.0700, 625.0000, 159.0000, 281.0000, 243.0000, 489.0000,
  1767. 330.0000],
  1768. [502.0016, 240.1822, 728.0000, 227.0000, 351.0000, 188.0000, 564.6272,
  1769. 320.6006],
  1770. [486.9100, 239.8900, 703.0000, 267.0000, 322.0000, 279.0000, 424.5074,
  1771. 306.1910],
  1772. [498.2500, 240.2200, 700.0000, 315.0000, 306.0000, 314.0000, 509.0000,
  1773. 300.0000],
  1774. [508.8475, 244.2474, 728.0000, 287.0000, 299.0000, 238.0000, 533.6024,
  1775. 319.9065]], device='cuda:0')
  1776. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1777.  
  1778. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1779. loss_train: nan
  1780. step: 59
  1781. running loss: nan
  1782. Train Steps: 59/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1783. [nan, nan, nan, nan, nan, nan, nan, nan],
  1784. [nan, nan, nan, nan, nan, nan, nan, nan],
  1785. [nan, nan, nan, nan, nan, nan, nan, nan],
  1786. [nan, nan, nan, nan, nan, nan, nan, nan],
  1787. [nan, nan, nan, nan, nan, nan, nan, nan],
  1788. [nan, nan, nan, nan, nan, nan, nan, nan],
  1789. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1790. grad_fn=<AddmmBackward>)
  1791. landmarks are: tensor([[501.4312, 241.7457, 680.0000, 161.0000, 315.0000, 210.0000, 548.7982,
  1792. 317.8241],
  1793. [494.1500, 245.1300, 699.0000, 237.0000, 302.0000, 336.0000, 498.0000,
  1794. 342.0000],
  1795. [494.6500, 244.5700, 707.0000, 271.0000, 305.0000, 269.0000, 462.0000,
  1796. 338.0000],
  1797. [497.5500, 246.8200, 654.0000, 169.0000, 314.0000, 167.0000, 472.0000,
  1798. 321.0000],
  1799. [502.0016, 241.4330, 617.7352, 124.7889, 351.3434, 134.0229, 514.8118,
  1800. 317.3988],
  1801. [502.6900, 241.7500, 707.0000, 227.0000, 318.0000, 171.0000, 506.7784,
  1802. 305.4101],
  1803. [500.7600, 247.1900, 641.0000, 141.0000, 391.0000, 92.0000, 502.5046,
  1804. 293.6965],
  1805. [501.8700, 246.5800, 712.0000, 229.0000, 335.0000, 130.0000, 468.6673,
  1806. 290.0746]], device='cuda:0')
  1807. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1808. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1809. loss_train: nan
  1810. step: 60
  1811. running loss: nan
  1812. Train Steps: 60/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1813. [nan, nan, nan, nan, nan, nan, nan, nan],
  1814. [nan, nan, nan, nan, nan, nan, nan, nan],
  1815. [nan, nan, nan, nan, nan, nan, nan, nan],
  1816. [nan, nan, nan, nan, nan, nan, nan, nan],
  1817. [nan, nan, nan, nan, nan, nan, nan, nan],
  1818. [nan, nan, nan, nan, nan, nan, nan, nan],
  1819. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1820. grad_fn=<AddmmBackward>)
  1821. landmarks are: tensor([[495.6700, 245.2700, 711.0000, 275.0000, 360.0000, 341.0000, 491.0000,
  1822. 353.0000],
  1823. [496.3500, 242.9600, 638.0000, 162.0000, 315.0000, 154.0000, 456.0000,
  1824. 311.0000],
  1825. [496.1100, 241.6200, 642.3539, 163.6537, 323.3445, 138.6042, 478.4196,
  1826. 323.4764],
  1827. [501.7164, 244.8729, 726.5198, 293.6544, 296.0000, 257.0000, 532.8420,
  1828. 316.4105],
  1829. [503.7131, 242.6839, 731.0000, 246.0000, 338.5284, 254.5401, 593.7525,
  1830. 317.8241],
  1831. [490.7800, 239.6200, 633.0000, 183.0000, 290.0000, 183.0000, 467.0000,
  1832. 303.0000],
  1833. [503.7131, 243.6220, 728.0000, 196.0000, 378.0822, 202.0042, 595.6520,
  1834. 321.2947],
  1835. [501.7164, 242.3712, 731.0000, 225.0000, 370.0000, 157.0000, 578.5567,
  1836. 324.7654]], device='cuda:0')
  1837. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1838. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1839. loss_train: nan
  1840. step: 61
  1841. running loss: nan
  1842. Train Steps: 61/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1843. [nan, nan, nan, nan, nan, nan, nan, nan],
  1844. [nan, nan, nan, nan, nan, nan, nan, nan],
  1845. [nan, nan, nan, nan, nan, nan, nan, nan],
  1846. [nan, nan, nan, nan, nan, nan, nan, nan],
  1847. [nan, nan, nan, nan, nan, nan, nan, nan],
  1848. [nan, nan, nan, nan, nan, nan, nan, nan],
  1849. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1850. grad_fn=<AddmmBackward>)
  1851. landmarks are: tensor([[500.7400, 249.3600, 705.0000, 191.0000, 382.0000, 112.0000, 497.5185,
  1852. 297.6010],
  1853. [490.0800, 243.9900, 691.0000, 323.0000, 335.0000, 291.0000, 401.3054,
  1854. 323.5448],
  1855. [501.6600, 240.3000, 676.0000, 364.0000, 308.0000, 299.0000, 485.4824,
  1856. 278.9525],
  1857. [499.1300, 250.4500, 674.0000, 344.0000, 386.0000, 270.0000, 450.0000,
  1858. 356.0000],
  1859. [487.8500, 239.4600, 691.0000, 283.0000, 341.0000, 298.0000, 417.0000,
  1860. 339.0000],
  1861. [497.4400, 245.9000, 579.0000, 111.0000, 339.0000, 151.0000, 503.0000,
  1862. 321.0000],
  1863. [501.0700, 243.4400, 704.0000, 230.0000, 292.0000, 223.0000, 509.9838,
  1864. 288.2302],
  1865. [496.3500, 242.9600, 638.0000, 162.0000, 315.0000, 154.0000, 456.0000,
  1866. 311.0000]], device='cuda:0')
  1867. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1868. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1869. loss_train: nan
  1870. step: 62
  1871. running loss: nan
  1872. Train Steps: 62/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1873. [nan, nan, nan, nan, nan, nan, nan, nan],
  1874. [nan, nan, nan, nan, nan, nan, nan, nan],
  1875. [nan, nan, nan, nan, nan, nan, nan, nan],
  1876. [nan, nan, nan, nan, nan, nan, nan, nan],
  1877. [nan, nan, nan, nan, nan, nan, nan, nan],
  1878. [nan, nan, nan, nan, nan, nan, nan, nan],
  1879. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1880. grad_fn=<AddmmBackward>)
  1881. landmarks are: tensor([[490.9200, 242.2600, 685.0000, 243.0000, 305.0000, 153.0000, 408.4806,
  1882. 297.2106],
  1883. [ nan, nan, 548.7455, 131.6165, 332.0000, 112.0000, 412.2570,
  1884. 343.7506],
  1885. [ nan, nan, 682.0000, 133.0000, 433.0000, 142.0000, 589.3204,
  1886. 328.9302],
  1887. [498.0082, 251.4398, 680.0000, 250.0000, 295.0000, 194.0000, 452.0000,
  1888. 339.0000],
  1889. [494.8000, 245.8500, 707.0000, 294.0000, 363.0000, 348.0000, 503.0000,
  1890. 324.0000],
  1891. [486.4500, 236.9900, 683.0000, 280.0000, 308.0000, 295.0000, 427.3566,
  1892. 297.2106],
  1893. [497.9000, 243.6700, 719.0000, 258.0000, 307.0000, 285.0000, 489.0000,
  1894. 329.0000],
  1895. [496.1000, 243.8400, 695.0000, 303.0000, 338.0000, 306.0000, 491.0000,
  1896. 330.0000]], device='cuda:0')
  1897. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1898. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1899. loss_train: nan
  1900. step: 63
  1901. running loss: nan
  1902. Train Steps: 63/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1903. [nan, nan, nan, nan, nan, nan, nan, nan],
  1904. [nan, nan, nan, nan, nan, nan, nan, nan],
  1905. [nan, nan, nan, nan, nan, nan, nan, nan],
  1906. [nan, nan, nan, nan, nan, nan, nan, nan],
  1907. [nan, nan, nan, nan, nan, nan, nan, nan],
  1908. [nan, nan, nan, nan, nan, nan, nan, nan],
  1909. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1910. grad_fn=<AddmmBackward>)
  1911. landmarks are: tensor([[491.6300, 240.4200, 702.0000, 272.0000, 365.0000, 332.0000, 487.0000,
  1912. 332.0000],
  1913. [488.5952, 242.9966, 696.0000, 291.0000, 357.5752, 290.8813, 403.4423,
  1914. 325.8875],
  1915. [492.5886, 246.1237, 681.0000, 229.0000, 289.0000, 220.0000, 407.7161,
  1916. 353.4143],
  1917. [489.6200, 240.8100, 549.0000, 169.0000, 296.0000, 167.0000, 441.0000,
  1918. 340.0000],
  1919. [492.3300, 242.9000, 568.0000, 124.0000, 347.0000, 100.0000, 433.0551,
  1920. 313.2191],
  1921. [498.8640, 244.8729, 686.0000, 180.0000, 297.0000, 182.0000, 444.0000,
  1922. 338.0000],
  1923. [502.0200, 242.8900, 679.0000, 173.0000, 357.0000, 122.0000, 505.7100,
  1924. 309.3146],
  1925. [490.7200, 245.4200, 554.0701, 169.1422, 283.0000, 194.0000, 445.0000,
  1926. 340.0000]], device='cuda:0')
  1927. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1928. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1929. loss_train: nan
  1930. step: 64
  1931. running loss: nan
  1932. Train Steps: 64/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1933. [nan, nan, nan, nan, nan, nan, nan, nan],
  1934. [nan, nan, nan, nan, nan, nan, nan, nan],
  1935. [nan, nan, nan, nan, nan, nan, nan, nan],
  1936. [nan, nan, nan, nan, nan, nan, nan, nan],
  1937. [nan, nan, nan, nan, nan, nan, nan, nan],
  1938. [nan, nan, nan, nan, nan, nan, nan, nan],
  1939. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1940. grad_fn=<AddmmBackward>)
  1941. landmarks are: tensor([[497.1300, 238.9300, 715.0000, 271.0000, 293.0000, 245.0000, 469.0000,
  1942. 303.0000],
  1943. [497.5400, 245.8000, 699.0000, 204.0000, 285.0000, 247.0000, 478.0000,
  1944. 341.0000],
  1945. [491.7329, 244.8729, 683.0000, 204.0000, 293.0000, 189.0000, 411.3298,
  1946. 292.5252],
  1947. [490.0100, 240.6100, 692.0000, 274.0000, 293.0000, 234.0000, 448.0000,
  1948. 334.0000],
  1949. [496.6100, 244.9400, 683.0000, 184.0000, 287.0000, 223.0000, 489.0000,
  1950. 331.0000],
  1951. [496.5300, 244.8400, 613.0000, 124.0000, 317.0000, 192.0000, 505.0000,
  1952. 318.0000],
  1953. [494.0900, 241.0100, 703.0000, 306.0000, 326.0000, 315.0000, 473.0000,
  1954. 302.0000],
  1955. [494.0100, 245.5700, 704.0000, 266.0000, 326.0000, 262.0000, 410.2613,
  1956. 294.0870]], device='cuda:0')
  1957. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1958.  
  1959. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1960. loss_train: nan
  1961. step: 65
  1962. running loss: nan
  1963. Train Steps: 65/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1964. [nan, nan, nan, nan, nan, nan, nan, nan],
  1965. [nan, nan, nan, nan, nan, nan, nan, nan],
  1966. [nan, nan, nan, nan, nan, nan, nan, nan],
  1967. [nan, nan, nan, nan, nan, nan, nan, nan],
  1968. [nan, nan, nan, nan, nan, nan, nan, nan],
  1969. [nan, nan, nan, nan, nan, nan, nan, nan],
  1970. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  1971. grad_fn=<AddmmBackward>)
  1972. landmarks are: tensor([[504.1500, 240.4100, 708.0000, 330.0000, 289.0000, 271.0000, 506.7784,
  1973. 300.7246],
  1974. [492.8600, 245.9100, 699.0000, 263.0000, 303.0000, 329.0000, 448.3696,
  1975. 301.1151],
  1976. [495.4411, 249.8763, 707.0000, 282.0000, 332.0000, 292.0000, 434.1604,
  1977. 315.6382],
  1978. [495.4411, 242.0584, 620.5229, 140.8099, 298.6643, 175.1677, 474.4262,
  1979. 295.5407],
  1980. [498.4000, 246.7900, 577.0000, 119.0000, 346.0000, 142.0000, 501.0000,
  1981. 324.0000],
  1982. [501.9300, 256.1800, 715.0000, 298.0000, 284.0000, 257.0000, 456.0000,
  1983. 344.0000],
  1984. [ nan, nan, 601.2303, 162.4709, 319.0000, 136.0000, 413.0000,
  1985. 334.0000],
  1986. [491.5200, 241.9000, 685.0000, 197.0000, 282.0000, 252.0000, 462.0000,
  1987. 335.0000]], device='cuda:0')
  1988. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1989. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  1990. loss_train: nan
  1991. step: 66
  1992. running loss: nan
  1993. Train Steps: 66/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  1994. [nan, nan, nan, nan, nan, nan, nan, nan],
  1995. [nan, nan, nan, nan, nan, nan, nan, nan],
  1996. [nan, nan, nan, nan, nan, nan, nan, nan],
  1997. [nan, nan, nan, nan, nan, nan, nan, nan],
  1998. [nan, nan, nan, nan, nan, nan, nan, nan],
  1999. [nan, nan, nan, nan, nan, nan, nan, nan],
  2000. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2001. grad_fn=<AddmmBackward>)
  2002. landmarks are: tensor([[488.8600, 241.9700, 664.0000, 196.0000, 287.0000, 200.0000, 435.5481,
  2003. 338.2080],
  2004. [496.2200, 243.3300, 715.3384, 245.1700, 291.4701, 217.0026, 473.3853,
  2005. 328.9417],
  2006. [492.3034, 246.7491, 563.0000, 139.0000, 339.0000, 110.0000, 428.0000,
  2007. 336.0000],
  2008. [501.0800, 242.3100, 720.0000, 264.0000, 290.0000, 280.0000, 513.9015,
  2009. 288.2302],
  2010. [492.2300, 247.1400, 677.0000, 230.0000, 288.0000, 192.0000, 408.5175,
  2011. 333.7941],
  2012. [489.7400, 240.3700, 708.0000, 253.0000, 327.0000, 331.0000, 485.0000,
  2013. 331.0000],
  2014. [500.0100, 246.6100, 579.0000, 124.0000, 341.0000, 113.0000, 450.0000,
  2015. 338.0000],
  2016. [495.0900, 241.9900, 692.0000, 316.0000, 359.0000, 309.0000, 474.0000,
  2017. 303.0000]], device='cuda:0')
  2018. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2019. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2020. loss_train: nan
  2021. step: 67
  2022. running loss: nan
  2023. Train Steps: 67/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2024. [nan, nan, nan, nan, nan, nan, nan, nan],
  2025. [nan, nan, nan, nan, nan, nan, nan, nan],
  2026. [nan, nan, nan, nan, nan, nan, nan, nan],
  2027. [nan, nan, nan, nan, nan, nan, nan, nan],
  2028. [nan, nan, nan, nan, nan, nan, nan, nan],
  2029. [nan, nan, nan, nan, nan, nan, nan, nan],
  2030. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2031. grad_fn=<AddmmBackward>)
  2032. landmarks are: tensor([[492.3300, 242.9000, 568.0000, 124.0000, 347.0000, 100.0000, 433.0551,
  2033. 313.2191],
  2034. [483.0400, 236.7800, 673.0000, 293.0000, 285.0000, 273.0000, 421.3020,
  2035. 281.5925],
  2036. [501.7164, 244.8729, 726.5198, 293.6544, 296.0000, 257.0000, 532.8420,
  2037. 316.4105],
  2038. [496.7200, 235.1800, 692.0000, 322.0000, 352.0000, 304.0000, 482.0000,
  2039. 297.0000],
  2040. [486.3200, 237.8300, 593.6238, 177.4812, 285.0000, 175.0000, 428.0689,
  2041. 298.7724],
  2042. [498.6900, 241.4000, 711.0000, 278.0000, 318.0000, 346.0000, 512.0000,
  2043. 311.0000],
  2044. [492.9800, 236.9300, 707.0000, 271.0000, 340.0000, 311.0000, 467.0000,
  2045. 330.0000],
  2046. [494.3001, 240.4949, 688.0120, 240.8839, 314.5486, 150.9156, 456.8577,
  2047. 326.3061]], device='cuda:0')
  2048. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2049. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2050. loss_train: nan
  2051. step: 68
  2052. running loss: nan
  2053. Train Steps: 68/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2054. [nan, nan, nan, nan, nan, nan, nan, nan],
  2055. [nan, nan, nan, nan, nan, nan, nan, nan],
  2056. [nan, nan, nan, nan, nan, nan, nan, nan],
  2057. [nan, nan, nan, nan, nan, nan, nan, nan],
  2058. [nan, nan, nan, nan, nan, nan, nan, nan],
  2059. [nan, nan, nan, nan, nan, nan, nan, nan],
  2060. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2061. grad_fn=<AddmmBackward>)
  2062. landmarks are: tensor([[488.1900, 240.0300, 619.0000, 215.0000, 277.0000, 228.0000, 441.9588,
  2063. 341.3316],
  2064. [501.1200, 242.0800, 711.0000, 293.0000, 324.0000, 313.0000, 508.9154,
  2065. 287.4493],
  2066. [492.8600, 245.9100, 699.0000, 263.0000, 303.0000, 329.0000, 448.3696,
  2067. 301.1151],
  2068. [489.3200, 241.1600, 683.0000, 244.0000, 281.0000, 215.0000, 453.0000,
  2069. 308.0000],
  2070. [498.0082, 251.7526, 703.0000, 208.0000, 300.0000, 204.0000, 433.0000,
  2071. 326.0000],
  2072. [507.9918, 247.0619, 669.0000, 163.0000, 388.0000, 102.0000, 515.2408,
  2073. 310.1886],
  2074. [489.7400, 240.3700, 708.0000, 253.0000, 327.0000, 331.0000, 485.0000,
  2075. 331.0000],
  2076. [486.6400, 237.4500, 691.0000, 297.0000, 349.0000, 305.0000, 427.7128,
  2077. 298.7724]], device='cuda:0')
  2078. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2079. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2080. loss_train: nan
  2081. step: 69
  2082. running loss: nan
  2083. Train Steps: 69/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2084. [nan, nan, nan, nan, nan, nan, nan, nan],
  2085. [nan, nan, nan, nan, nan, nan, nan, nan],
  2086. [nan, nan, nan, nan, nan, nan, nan, nan],
  2087. [nan, nan, nan, nan, nan, nan, nan, nan],
  2088. [nan, nan, nan, nan, nan, nan, nan, nan],
  2089. [nan, nan, nan, nan, nan, nan, nan, nan],
  2090. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2091. grad_fn=<AddmmBackward>)
  2092. landmarks are: tensor([[ nan, nan, 677.0000, 153.0000, 468.0000, 128.0000, 570.3256,
  2093. 364.3312],
  2094. [491.1100, 241.1500, 571.0000, 129.0000, 350.0000, 92.0000, 423.4389,
  2095. 300.3342],
  2096. [501.0800, 241.7300, 712.0000, 296.0000, 326.0000, 311.0000, 512.4769,
  2097. 285.4970],
  2098. [503.9984, 240.8076, 715.0000, 321.0000, 294.0000, 276.0000, 516.5071,
  2099. 298.3883],
  2100. [496.1000, 242.9900, 620.5920, 134.0372, 356.6960, 107.8909, 478.0393,
  2101. 325.5612],
  2102. [501.7164, 244.8729, 726.5198, 293.6544, 296.0000, 257.0000, 532.8420,
  2103. 316.4105],
  2104. [489.3200, 241.0900, 525.0000, 118.0000, 299.0000, 153.0000, 422.3705,
  2105. 306.1910],
  2106. [496.8600, 244.1200, 700.0000, 307.0000, 332.0000, 294.0000, 470.0000,
  2107. 310.0000]], device='cuda:0')
  2108. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2109. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2110. loss_train: nan
  2111. step: 70
  2112. running loss: nan
  2113. Train Steps: 70/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2114. [nan, nan, nan, nan, nan, nan, nan, nan],
  2115. [nan, nan, nan, nan, nan, nan, nan, nan],
  2116. [nan, nan, nan, nan, nan, nan, nan, nan],
  2117. [nan, nan, nan, nan, nan, nan, nan, nan],
  2118. [nan, nan, nan, nan, nan, nan, nan, nan],
  2119. [nan, nan, nan, nan, nan, nan, nan, nan],
  2120. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2121. grad_fn=<AddmmBackward>)
  2122. landmarks are: tensor([[ nan, nan, 617.9645, 156.6336, 294.0000, 164.0000, 433.0000,
  2123. 310.0000],
  2124. [498.8640, 238.6186, 718.8183, 288.3773, 306.7708, 312.9795, 529.0106,
  2125. 309.9863],
  2126. [496.8600, 244.1200, 700.0000, 307.0000, 332.0000, 294.0000, 470.0000,
  2127. 310.0000],
  2128. [497.0900, 250.5100, 584.0000, 173.0000, 347.0000, 130.0000, 455.8488,
  2129. 346.4075],
  2130. [496.0800, 242.1700, 687.7038, 230.9800, 279.3204, 227.0880, 478.2295,
  2131. 307.8407],
  2132. [495.4600, 246.4600, 595.0000, 162.0000, 292.0000, 221.0000, 499.0000,
  2133. 343.0000],
  2134. [502.0016, 240.1822, 728.0000, 227.0000, 351.0000, 188.0000, 564.6272,
  2135. 320.6006],
  2136. [492.9600, 240.8300, 704.0000, 320.0000, 300.0000, 289.0000, 479.0000,
  2137. 317.0000]], device='cuda:0')
  2138. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2139.  
  2140. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2141. loss_train: nan
  2142. step: 71
  2143. running loss: nan
  2144. Train Steps: 71/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2145. [nan, nan, nan, nan, nan, nan, nan, nan],
  2146. [nan, nan, nan, nan, nan, nan, nan, nan],
  2147. [nan, nan, nan, nan, nan, nan, nan, nan],
  2148. [nan, nan, nan, nan, nan, nan, nan, nan],
  2149. [nan, nan, nan, nan, nan, nan, nan, nan],
  2150. [nan, nan, nan, nan, nan, nan, nan, nan],
  2151. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2152. grad_fn=<AddmmBackward>)
  2153. landmarks are: tensor([[504.8541, 239.8694, 638.0000, 346.0000, 307.0000, 291.0000, 586.1546,
  2154. 320.6006],
  2155. [497.8700, 247.7900, 648.0000, 179.0000, 282.0000, 234.0000, 455.4926,
  2156. 306.1910],
  2157. [501.0800, 241.7700, 720.0000, 286.0000, 304.0000, 310.0000, 513.1892,
  2158. 286.2780],
  2159. [484.2800, 242.0800, 551.7881, 114.9384, 320.0000, 127.0000, 435.1920,
  2160. 310.0955],
  2161. [499.1492, 246.4364, 652.9544, 165.8065, 290.0000, 216.0000, 479.0000,
  2162. 342.0000],
  2163. [491.4477, 242.3712, 659.0000, 200.0000, 326.0000, 127.0000, 410.9736,
  2164. 298.3820],
  2165. [487.7395, 241.4330, 679.0000, 223.0000, 310.0000, 331.0000, 466.8895,
  2166. 335.4748],
  2167. [507.1360, 248.9381, 704.0000, 300.0000, 312.0000, 317.0000, 603.2499,
  2168. 325.4596]], device='cuda:0')
  2169. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2170. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2171. loss_train: nan
  2172. step: 72
  2173. running loss: nan
  2174. Train Steps: 72/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2175. [nan, nan, nan, nan, nan, nan, nan, nan],
  2176. [nan, nan, nan, nan, nan, nan, nan, nan],
  2177. [nan, nan, nan, nan, nan, nan, nan, nan],
  2178. [nan, nan, nan, nan, nan, nan, nan, nan],
  2179. [nan, nan, nan, nan, nan, nan, nan, nan],
  2180. [nan, nan, nan, nan, nan, nan, nan, nan],
  2181. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2182. grad_fn=<AddmmBackward>)
  2183. landmarks are: tensor([[501.8900, 245.2000, 670.0000, 342.0000, 322.0000, 288.0000, 456.0000,
  2184. 367.0000],
  2185. [498.7100, 250.7200, 626.0000, 207.0000, 305.0000, 172.0000, 454.0000,
  2186. 337.0000],
  2187. [507.7065, 249.8763, 731.0000, 239.0000, 310.0000, 259.0000, 597.5515,
  2188. 328.2361],
  2189. [497.9900, 246.9700, 693.0000, 211.0000, 293.0000, 194.0000, 467.0000,
  2190. 319.0000],
  2191. [504.8541, 240.4949, 634.8398, 344.7328, 312.0000, 302.0000, 556.3961,
  2192. 321.9889],
  2193. [482.4700, 239.1800, 597.0000, 170.0000, 291.0000, 163.0000, 420.2336,
  2194. 283.5448],
  2195. [493.1591, 237.3677, 700.5271, 305.2639, 343.9919, 319.1815, 481.7994,
  2196. 312.1400],
  2197. [494.7600, 244.7600, 707.0000, 277.0000, 387.0000, 339.0000, 494.0000,
  2198. 351.0000]], device='cuda:0')
  2199. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2200. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2201. loss_train: nan
  2202. step: 73
  2203. running loss: nan
  2204. Train Steps: 73/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2205. [nan, nan, nan, nan, nan, nan, nan, nan],
  2206. [nan, nan, nan, nan, nan, nan, nan, nan],
  2207. [nan, nan, nan, nan, nan, nan, nan, nan],
  2208. [nan, nan, nan, nan, nan, nan, nan, nan],
  2209. [nan, nan, nan, nan, nan, nan, nan, nan],
  2210. [nan, nan, nan, nan, nan, nan, nan, nan],
  2211. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2212. grad_fn=<AddmmBackward>)
  2213. landmarks are: tensor([[493.2300, 246.3500, 606.0000, 104.0000, 307.0000, 159.0000, 454.4241,
  2214. 306.9719],
  2215. [496.3600, 244.8300, 716.0000, 241.0000, 303.0000, 282.0000, 477.0000,
  2216. 340.0000],
  2217. [ nan, nan, 601.0000, 127.0000, 343.0000, 120.0000, 448.0000,
  2218. 337.0000],
  2219. [493.4200, 246.6600, 521.3622, 172.4778, 295.0000, 169.0000, 418.2301,
  2220. 350.2430],
  2221. [ nan, nan, 669.0000, 199.0000, 285.0000, 202.0000, 426.2881,
  2222. 308.5337],
  2223. [490.3400, 244.1100, 700.0000, 304.0000, 310.0000, 254.0000, 418.8311,
  2224. 352.8785],
  2225. [491.6200, 238.9700, 696.0000, 301.0000, 352.0000, 288.0000, 430.0000,
  2226. 345.0000],
  2227. [494.0900, 241.0100, 703.0000, 306.0000, 326.0000, 315.0000, 473.0000,
  2228. 302.0000]], device='cuda:0')
  2229. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2230. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2231. loss_train: nan
  2232. step: 74
  2233. running loss: nan
  2234. Train Steps: 74/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2235. [nan, nan, nan, nan, nan, nan, nan, nan],
  2236. [nan, nan, nan, nan, nan, nan, nan, nan],
  2237. [nan, nan, nan, nan, nan, nan, nan, nan],
  2238. [nan, nan, nan, nan, nan, nan, nan, nan],
  2239. [nan, nan, nan, nan, nan, nan, nan, nan],
  2240. [nan, nan, nan, nan, nan, nan, nan, nan],
  2241. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2242. grad_fn=<AddmmBackward>)
  2243. landmarks are: tensor([[495.9500, 244.2800, 608.0000, 127.0000, 323.0000, 166.0000, 491.0000,
  2244. 333.0000],
  2245. [501.3800, 245.6600, 697.0000, 185.0000, 352.0000, 136.0000, 500.0000,
  2246. 312.0000],
  2247. [507.1360, 246.1237, 687.0000, 188.0000, 354.0000, 127.0000, 513.3414,
  2248. 305.3296],
  2249. [495.8900, 246.1100, 716.0000, 265.0000, 321.0000, 322.0000, 489.0000,
  2250. 358.0000],
  2251. [496.4300, 240.2100, 715.0000, 293.0000, 293.0000, 300.0000, 508.5592,
  2252. 296.8201],
  2253. [494.3001, 240.4949, 688.0120, 240.8839, 314.5486, 150.9156, 456.8577,
  2254. 326.3061],
  2255. [485.1400, 241.1400, 692.0000, 271.0000, 323.0000, 322.0000, 456.2567,
  2256. 336.5189],
  2257. [491.8400, 239.7800, 679.0000, 232.0000, 279.0000, 244.0000, 469.0000,
  2258. 300.0000]], device='cuda:0')
  2259. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2260. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2261. loss_train: nan
  2262. step: 75
  2263. running loss: nan
  2264. Train Steps: 75/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2265. [nan, nan, nan, nan, nan, nan, nan, nan],
  2266. [nan, nan, nan, nan, nan, nan, nan, nan],
  2267. [nan, nan, nan, nan, nan, nan, nan, nan],
  2268. [nan, nan, nan, nan, nan, nan, nan, nan],
  2269. [nan, nan, nan, nan, nan, nan, nan, nan],
  2270. [nan, nan, nan, nan, nan, nan, nan, nan],
  2271. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2272. grad_fn=<AddmmBackward>)
  2273. landmarks are: tensor([[501.8400, 245.9700, 571.0000, 128.0000, 320.0000, 159.0000, 486.0000,
  2274. 338.0000],
  2275. [503.7131, 244.5602, 672.0000, 335.0000, 296.0000, 262.0000, 550.0646,
  2276. 329.6244],
  2277. [492.0700, 247.4900, 699.0000, 265.0000, 286.0000, 227.0000, 411.0000,
  2278. 329.0000],
  2279. [500.8000, 252.8600, 683.0000, 330.0000, 293.0000, 226.0000, 447.0000,
  2280. 359.0000],
  2281. [502.4100, 243.2900, 688.0000, 319.0000, 304.0000, 282.0000, 502.0000,
  2282. 308.0000],
  2283. [506.8508, 245.1856, 712.0000, 237.0000, 312.0000, 177.0000, 520.3061,
  2284. 303.9413],
  2285. [495.8300, 246.2900, 636.0000, 196.0000, 294.0000, 226.0000, 483.0000,
  2286. 370.0000],
  2287. [502.0016, 242.9966, 723.0000, 226.0000, 307.0000, 212.0000, 565.8936,
  2288. 334.4833]], device='cuda:0')
  2289. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2290. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2291. loss_train: nan
  2292. step: 76
  2293. running loss: nan
  2294. Train Steps: 76/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2295. [nan, nan, nan, nan, nan, nan, nan, nan],
  2296. [nan, nan, nan, nan, nan, nan, nan, nan],
  2297. [nan, nan, nan, nan, nan, nan, nan, nan],
  2298. [nan, nan, nan, nan, nan, nan, nan, nan],
  2299. [nan, nan, nan, nan, nan, nan, nan, nan],
  2300. [nan, nan, nan, nan, nan, nan, nan, nan],
  2301. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2302. grad_fn=<AddmmBackward>)
  2303. landmarks are: tensor([[ nan, nan, 593.0000, 132.0000, 356.0000, 91.0000, 425.0000,
  2304. 299.0000],
  2305. [496.1400, 243.1500, 691.0000, 317.0000, 363.6628, 306.4949, 472.0000,
  2306. 308.0000],
  2307. [ nan, nan, 567.7618, 140.7894, 340.0000, 111.0000, 414.0000,
  2308. 335.0000],
  2309. [498.7100, 250.7200, 626.0000, 207.0000, 305.0000, 172.0000, 454.0000,
  2310. 337.0000],
  2311. [494.2245, 240.7296, 712.0364, 269.7266, 311.2985, 337.0000, 477.7910,
  2312. 292.7526],
  2313. [492.8600, 245.9100, 699.0000, 263.0000, 303.0000, 329.0000, 448.3696,
  2314. 301.1151],
  2315. [503.9400, 258.1900, 637.0000, 236.0000, 388.0000, 137.0000, 438.3973,
  2316. 373.3486],
  2317. [504.3800, 238.9900, 716.0000, 290.0000, 295.0000, 281.0000, 510.0000,
  2318. 307.0000]], device='cuda:0')
  2319. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2320.  
  2321. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2322. loss_train: nan
  2323. step: 77
  2324. running loss: nan
  2325. Train Steps: 77/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2326. [nan, nan, nan, nan, nan, nan, nan, nan],
  2327. [nan, nan, nan, nan, nan, nan, nan, nan],
  2328. [nan, nan, nan, nan, nan, nan, nan, nan],
  2329. [nan, nan, nan, nan, nan, nan, nan, nan],
  2330. [nan, nan, nan, nan, nan, nan, nan, nan],
  2331. [nan, nan, nan, nan, nan, nan, nan, nan],
  2332. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2333. grad_fn=<AddmmBackward>)
  2334. landmarks are: tensor([[488.8600, 241.9700, 664.0000, 196.0000, 287.0000, 200.0000, 435.5481,
  2335. 338.2080],
  2336. [490.7800, 239.6200, 633.0000, 183.0000, 290.0000, 183.0000, 467.0000,
  2337. 303.0000],
  2338. [500.5754, 241.4330, 693.7882, 325.3166, 335.6755, 311.8676, 535.5466,
  2339. 312.2101],
  2340. [500.0049, 237.6805, 693.7882, 295.7652, 335.9166, 298.3375, 504.9382,
  2341. 318.0700],
  2342. [499.1500, 243.6600, 708.0000, 276.0000, 338.0000, 312.0000, 491.0000,
  2343. 327.0000],
  2344. [488.7100, 240.8900, 613.4007, 218.3425, 281.0000, 220.0000, 415.9966,
  2345. 338.4796],
  2346. [502.8574, 243.3093, 720.0000, 283.0000, 301.0000, 281.0000, 561.4614,
  2347. 329.6244],
  2348. [493.0100, 246.8500, 612.0000, 121.0000, 301.0000, 172.0000, 450.5065,
  2349. 304.2387]], device='cuda:0')
  2350. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2351. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2352. loss_train: nan
  2353. step: 78
  2354. running loss: nan
  2355. Train Steps: 78/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2356. [nan, nan, nan, nan, nan, nan, nan, nan],
  2357. [nan, nan, nan, nan, nan, nan, nan, nan],
  2358. [nan, nan, nan, nan, nan, nan, nan, nan],
  2359. [nan, nan, nan, nan, nan, nan, nan, nan],
  2360. [nan, nan, nan, nan, nan, nan, nan, nan],
  2361. [nan, nan, nan, nan, nan, nan, nan, nan],
  2362. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2363. grad_fn=<AddmmBackward>)
  2364. landmarks are: tensor([[498.0700, 246.1700, 686.0000, 207.0000, 351.0000, 124.0000, 463.0000,
  2365. 323.0000],
  2366. [494.0148, 240.8076, 618.6979, 167.0053, 357.9991, 107.1851, 463.1682,
  2367. 321.0350],
  2368. [491.4477, 242.0584, 704.0000, 290.0000, 361.0000, 322.0000, 423.0828,
  2369. 305.8005],
  2370. [495.0900, 242.9100, 711.0000, 265.0000, 337.0000, 312.0000, 479.0000,
  2371. 338.0000],
  2372. [496.0300, 237.6400, 676.8727, 329.8349, 331.3488, 274.6205, 471.3836,
  2373. 324.1019],
  2374. [494.5853, 238.3059, 684.1613, 354.8681, 294.6041, 250.8712, 455.0547,
  2375. 322.6822],
  2376. [494.5853, 237.9932, 661.0566, 183.8918, 282.0930, 249.6721, 495.3221,
  2377. 317.4111],
  2378. [490.3600, 243.8000, 699.0000, 315.0000, 345.0000, 284.0000, 418.4304,
  2379. 352.4393]], device='cuda:0')
  2380. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2381. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2382. loss_train: nan
  2383. step: 79
  2384. running loss: nan
  2385. Train Steps: 79/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2386. [nan, nan, nan, nan, nan, nan, nan, nan],
  2387. [nan, nan, nan, nan, nan, nan, nan, nan],
  2388. [nan, nan, nan, nan, nan, nan, nan, nan],
  2389. [nan, nan, nan, nan, nan, nan, nan, nan],
  2390. [nan, nan, nan, nan, nan, nan, nan, nan],
  2391. [nan, nan, nan, nan, nan, nan, nan, nan],
  2392. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2393. grad_fn=<AddmmBackward>)
  2394. landmarks are: tensor([[496.3600, 244.8300, 716.0000, 241.0000, 303.0000, 282.0000, 477.0000,
  2395. 340.0000],
  2396. [494.3001, 237.6805, 667.7954, 361.2005, 310.9871, 277.4219, 457.1582,
  2397. 322.3528],
  2398. [487.5800, 238.2200, 695.0000, 286.0000, 388.7642, 292.7355, 415.2475,
  2399. 296.4297],
  2400. [501.7164, 244.8729, 726.5198, 293.6544, 296.0000, 257.0000, 532.8420,
  2401. 316.4105],
  2402. [493.4444, 243.3093, 606.0000, 177.0000, 324.0000, 163.0000, 475.0000,
  2403. 370.0000],
  2404. [494.7800, 244.0100, 707.0000, 267.0000, 323.0000, 284.0000, 417.0282,
  2405. 308.5337],
  2406. [500.0100, 246.6100, 579.0000, 124.0000, 341.0000, 113.0000, 450.0000,
  2407. 338.0000],
  2408. [ nan, nan, 677.0000, 153.0000, 468.0000, 128.0000, 570.3256,
  2409. 364.3312]], device='cuda:0')
  2410. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2411. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2412. loss_train: nan
  2413. step: 80
  2414. running loss: nan
  2415. Train Steps: 80/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2416. [nan, nan, nan, nan, nan, nan, nan, nan],
  2417. [nan, nan, nan, nan, nan, nan, nan, nan],
  2418. [nan, nan, nan, nan, nan, nan, nan, nan],
  2419. [nan, nan, nan, nan, nan, nan, nan, nan],
  2420. [nan, nan, nan, nan, nan, nan, nan, nan],
  2421. [nan, nan, nan, nan, nan, nan, nan, nan],
  2422. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2423. grad_fn=<AddmmBackward>)
  2424. landmarks are: tensor([[ nan, nan, 513.0000, 114.0000, 339.0000, 113.0000, 439.0000,
  2425. 342.0000],
  2426. [490.1500, 245.0200, 696.0000, 268.0000, 319.0000, 259.0000, 401.0383,
  2427. 328.2302],
  2428. [486.6900, 238.8800, 687.0000, 314.0000, 368.0000, 322.0000, 454.4241,
  2429. 300.3342],
  2430. [492.8739, 241.4330, 707.0000, 275.0000, 295.0000, 224.0000, 424.8636,
  2431. 320.6377],
  2432. [496.9300, 241.9800, 715.0000, 250.0000, 305.0000, 256.0000, 449.0000,
  2433. 335.0000],
  2434. [496.1400, 243.1500, 691.0000, 317.0000, 363.6628, 306.4949, 472.0000,
  2435. 308.0000],
  2436. [501.8300, 248.5600, 700.0000, 342.0000, 319.0000, 283.0000, 481.0000,
  2437. 328.0000],
  2438. [489.7400, 240.3700, 708.0000, 253.0000, 327.0000, 331.0000, 485.0000,
  2439. 331.0000]], device='cuda:0')
  2440. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2441. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2442. loss_train: nan
  2443. step: 81
  2444. running loss: nan
  2445. Train Steps: 81/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2446. [nan, nan, nan, nan, nan, nan, nan, nan],
  2447. [nan, nan, nan, nan, nan, nan, nan, nan],
  2448. [nan, nan, nan, nan, nan, nan, nan, nan],
  2449. [nan, nan, nan, nan, nan, nan, nan, nan],
  2450. [nan, nan, nan, nan, nan, nan, nan, nan],
  2451. [nan, nan, nan, nan, nan, nan, nan, nan],
  2452. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2453. grad_fn=<AddmmBackward>)
  2454. landmarks are: tensor([[501.2500, 244.2100, 697.0000, 336.0000, 297.0000, 287.0000, 462.0000,
  2455. 366.0000],
  2456. [496.1500, 243.9800, 699.6677, 202.5880, 297.3302, 185.4055, 467.3902,
  2457. 309.9255],
  2458. [502.0016, 241.4330, 617.7352, 124.7889, 351.3434, 134.0229, 514.8118,
  2459. 317.3988],
  2460. [501.6600, 240.3000, 676.0000, 364.0000, 308.0000, 299.0000, 485.4824,
  2461. 278.9525],
  2462. [499.7000, 245.3900, 557.1127, 121.6097, 314.0000, 161.0000, 487.0000,
  2463. 335.0000],
  2464. [500.0400, 246.9800, 696.0000, 291.0000, 372.0000, 334.0000, 487.0000,
  2465. 311.0000],
  2466. [501.4300, 246.1100, 715.0000, 220.0000, 322.0000, 170.0000, 502.0000,
  2467. 311.0000],
  2468. [502.6300, 256.9900, 598.1877, 212.5051, 410.0000, 115.0000, 440.0000,
  2469. 370.0000]], device='cuda:0')
  2470. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2471. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2472. loss_train: nan
  2473. step: 82
  2474. running loss: nan
  2475. Train Steps: 82/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2476. [nan, nan, nan, nan, nan, nan, nan, nan],
  2477. [nan, nan, nan, nan, nan, nan, nan, nan],
  2478. [nan, nan, nan, nan, nan, nan, nan, nan],
  2479. [nan, nan, nan, nan, nan, nan, nan, nan],
  2480. [nan, nan, nan, nan, nan, nan, nan, nan],
  2481. [nan, nan, nan, nan, nan, nan, nan, nan],
  2482. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2483. grad_fn=<AddmmBackward>)
  2484. landmarks are: tensor([[490.8700, 247.0400, 696.0000, 310.0000, 335.0000, 305.0000, 411.7228,
  2485. 329.6944],
  2486. [490.0100, 240.6100, 692.0000, 274.0000, 293.0000, 234.0000, 448.0000,
  2487. 334.0000],
  2488. [496.2100, 244.5800, 688.8793, 172.7032, 324.0115, 153.2296, 472.5629,
  2489. 329.7797],
  2490. [ nan, nan, 727.0000, 227.0000, 365.0000, 157.0000, 539.3008,
  2491. 334.4833],
  2492. [493.1591, 237.3677, 700.5271, 305.2639, 343.9919, 319.1815, 481.7994,
  2493. 312.1400],
  2494. [ nan, nan, 601.0000, 127.0000, 343.0000, 120.0000, 448.0000,
  2495. 337.0000],
  2496. [490.9100, 237.3000, 672.0000, 196.0000, 280.0000, 252.0000, 469.0000,
  2497. 328.0000],
  2498. [503.1100, 241.8900, 673.4919, 326.7499, 326.0000, 301.0000, 505.0000,
  2499. 307.0000]], device='cuda:0')
  2500. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2501.  
  2502. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2503. loss_train: nan
  2504. step: 83
  2505. running loss: nan
  2506. Train Steps: 83/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2507. [nan, nan, nan, nan, nan, nan, nan, nan],
  2508. [nan, nan, nan, nan, nan, nan, nan, nan],
  2509. [nan, nan, nan, nan, nan, nan, nan, nan],
  2510. [nan, nan, nan, nan, nan, nan, nan, nan],
  2511. [nan, nan, nan, nan, nan, nan, nan, nan],
  2512. [nan, nan, nan, nan, nan, nan, nan, nan],
  2513. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2514. grad_fn=<AddmmBackward>)
  2515. landmarks are: tensor([[490.0100, 240.6100, 692.0000, 274.0000, 293.0000, 234.0000, 448.0000,
  2516. 334.0000],
  2517. [501.0800, 241.7300, 712.0000, 296.0000, 326.0000, 311.0000, 512.4769,
  2518. 285.4970],
  2519. [490.1500, 245.0200, 696.0000, 268.0000, 319.0000, 259.0000, 401.0383,
  2520. 328.2302],
  2521. [490.1700, 246.8700, 573.0000, 173.0000, 290.0000, 177.0000, 426.1470,
  2522. 329.6944],
  2523. [501.7164, 242.3712, 720.0000, 195.0000, 395.0000, 138.0000, 575.3909,
  2524. 324.7654],
  2525. [495.8100, 239.8400, 686.5782, 321.6804, 329.3477, 300.9463, 475.3770,
  2526. 308.0492],
  2527. [502.8574, 242.3712, 695.7136, 182.8364, 313.9077, 173.2258, 504.0367,
  2528. 322.0233],
  2529. [490.1900, 243.9500, 684.0000, 334.0000, 372.9783, 308.4714, 405.5792,
  2530. 324.7162]], device='cuda:0')
  2531. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2532. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2533. loss_train: nan
  2534. step: 84
  2535. running loss: nan
  2536. Train Steps: 84/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2537. [nan, nan, nan, nan, nan, nan, nan, nan],
  2538. [nan, nan, nan, nan, nan, nan, nan, nan],
  2539. [nan, nan, nan, nan, nan, nan, nan, nan],
  2540. [nan, nan, nan, nan, nan, nan, nan, nan],
  2541. [nan, nan, nan, nan, nan, nan, nan, nan],
  2542. [nan, nan, nan, nan, nan, nan, nan, nan],
  2543. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2544. grad_fn=<AddmmBackward>)
  2545. landmarks are: tensor([[494.0148, 240.8076, 618.6979, 167.0053, 357.9991, 107.1851, 463.1682,
  2546. 321.0350],
  2547. [492.0181, 236.1169, 695.7136, 309.4855, 371.7718, 319.7672, 483.3020,
  2548. 309.1750],
  2549. [501.0600, 243.8700, 723.0000, 259.0000, 287.0000, 273.0000, 506.0000,
  2550. 315.0000],
  2551. [494.0148, 240.8076, 712.0000, 270.0000, 350.0000, 351.0000, 494.0000,
  2552. 323.0000],
  2553. [493.0700, 240.3800, 703.0000, 281.0000, 293.0000, 293.0000, 471.0000,
  2554. 301.0000],
  2555. [496.0800, 240.2600, 702.8997, 302.5531, 306.6687, 308.2590, 479.7508,
  2556. 305.5475],
  2557. [495.6700, 245.2700, 711.0000, 275.0000, 360.0000, 341.0000, 491.0000,
  2558. 353.0000],
  2559. [501.6600, 241.1700, 670.0000, 365.0000, 314.0000, 292.0000, 482.9497,
  2560. 277.5642]], device='cuda:0')
  2561. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2562. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2563. loss_train: nan
  2564. step: 85
  2565. running loss: nan
  2566. Train Steps: 85/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2567. [nan, nan, nan, nan, nan, nan, nan, nan],
  2568. [nan, nan, nan, nan, nan, nan, nan, nan],
  2569. [nan, nan, nan, nan, nan, nan, nan, nan],
  2570. [nan, nan, nan, nan, nan, nan, nan, nan],
  2571. [nan, nan, nan, nan, nan, nan, nan, nan],
  2572. [nan, nan, nan, nan, nan, nan, nan, nan],
  2573. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2574. grad_fn=<AddmmBackward>)
  2575. landmarks are: tensor([[483.8400, 239.2300, 609.5974, 231.6849, 278.0000, 250.0000, 410.9215,
  2576. 327.9374],
  2577. [501.7700, 247.2200, 723.0000, 247.0000, 298.0000, 192.0000, 494.0000,
  2578. 315.0000],
  2579. [488.0600, 238.6400, 684.0000, 340.0000, 309.0000, 265.0000, 410.2613,
  2580. 292.1347],
  2581. [501.0800, 241.7700, 720.0000, 286.0000, 304.0000, 310.0000, 513.1892,
  2582. 286.2780],
  2583. [492.2300, 247.1400, 677.0000, 230.0000, 288.0000, 192.0000, 408.5175,
  2584. 333.7941],
  2585. [502.6300, 256.9900, 598.1877, 212.5051, 410.0000, 115.0000, 440.0000,
  2586. 370.0000],
  2587. [491.7329, 244.8729, 683.0000, 204.0000, 293.0000, 189.0000, 411.3298,
  2588. 292.5252],
  2589. [502.1600, 246.1500, 647.0000, 343.0000, 335.0000, 285.0000, 453.0000,
  2590. 365.0000]], device='cuda:0')
  2591. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2592. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2593. loss_train: nan
  2594. step: 86
  2595. running loss: nan
  2596. Train Steps: 86/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2597. [nan, nan, nan, nan, nan, nan, nan, nan],
  2598. [nan, nan, nan, nan, nan, nan, nan, nan],
  2599. [nan, nan, nan, nan, nan, nan, nan, nan],
  2600. [nan, nan, nan, nan, nan, nan, nan, nan],
  2601. [nan, nan, nan, nan, nan, nan, nan, nan],
  2602. [nan, nan, nan, nan, nan, nan, nan, nan],
  2603. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2604. grad_fn=<AddmmBackward>)
  2605. landmarks are: tensor([[507.1360, 249.5636, 672.0000, 337.0000, 306.0000, 249.0000, 587.4209,
  2606. 344.8954],
  2607. [488.6600, 240.6300, 643.0000, 204.0000, 296.0000, 176.0000, 452.6434,
  2608. 337.0366],
  2609. [491.8600, 244.5800, 693.2687, 289.2242, 331.0000, 304.0000, 420.0000,
  2610. 346.0000],
  2611. [502.8574, 256.4433, 680.0000, 270.0000, 362.0000, 155.0000, 435.1920,
  2612. 372.5677],
  2613. [ nan, nan, 593.0000, 132.0000, 356.0000, 91.0000, 425.0000,
  2614. 299.0000],
  2615. [503.4279, 245.8110, 704.0000, 151.0000, 421.0000, 156.0000, 594.3857,
  2616. 322.6830],
  2617. [495.7300, 239.1300, 704.0000, 277.0000, 335.0000, 287.0000, 455.0000,
  2618. 333.0000],
  2619. [508.8475, 246.1237, 692.0000, 179.0000, 391.0000, 120.0000, 536.1351,
  2620. 327.5420]], device='cuda:0')
  2621. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2622. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2623. loss_train: nan
  2624. step: 87
  2625. running loss: nan
  2626. Train Steps: 87/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2627. [nan, nan, nan, nan, nan, nan, nan, nan],
  2628. [nan, nan, nan, nan, nan, nan, nan, nan],
  2629. [nan, nan, nan, nan, nan, nan, nan, nan],
  2630. [nan, nan, nan, nan, nan, nan, nan, nan],
  2631. [nan, nan, nan, nan, nan, nan, nan, nan],
  2632. [nan, nan, nan, nan, nan, nan, nan, nan],
  2633. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2634. grad_fn=<AddmmBackward>)
  2635. landmarks are: tensor([[486.9100, 241.1400, 622.5284, 256.7020, 290.0000, 261.0000, 411.9899,
  2636. 317.1024],
  2637. [ nan, nan, 712.0000, 193.0000, 403.0000, 134.0000, 535.5019,
  2638. 336.5657],
  2639. [493.1591, 244.5602, 707.0000, 247.0000, 297.0000, 333.0000, 499.0000,
  2640. 321.0000],
  2641. [504.5688, 241.7457, 719.0000, 289.0000, 315.0000, 210.0000, 584.8883,
  2642. 322.6830],
  2643. [502.0600, 247.2100, 699.0000, 188.0000, 338.0000, 133.0000, 496.2461,
  2644. 293.5293],
  2645. [503.7131, 244.5602, 672.0000, 335.0000, 296.0000, 262.0000, 550.0646,
  2646. 329.6244],
  2647. [490.2700, 247.0800, 691.0000, 320.0000, 370.0000, 316.0000, 415.4624,
  2648. 328.5230],
  2649. [495.9800, 239.5800, 691.1306, 221.2412, 292.6610, 188.3306, 480.1311,
  2650. 322.4341]], device='cuda:0')
  2651. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2652. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2653. loss_train: nan
  2654. step: 88
  2655. running loss: nan
  2656. Train Steps: 88/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2657. [nan, nan, nan, nan, nan, nan, nan, nan],
  2658. [nan, nan, nan, nan, nan, nan, nan, nan],
  2659. [nan, nan, nan, nan, nan, nan, nan, nan],
  2660. [nan, nan, nan, nan, nan, nan, nan, nan],
  2661. [nan, nan, nan, nan, nan, nan, nan, nan],
  2662. [nan, nan, nan, nan, nan, nan, nan, nan],
  2663. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2664. grad_fn=<AddmmBackward>)
  2665. landmarks are: tensor([[491.1800, 244.2000, 696.0000, 300.0000, 369.0000, 294.0000, 420.8344,
  2666. 351.1215],
  2667. [491.6300, 240.4700, 700.0000, 323.0000, 318.0000, 279.0000, 445.0000,
  2668. 332.0000],
  2669. [488.0700, 242.5300, 622.0000, 157.0000, 297.0000, 169.0000, 435.1920,
  2670. 338.9889],
  2671. [488.5952, 240.8076, 696.0000, 279.0000, 403.9442, 310.5674, 468.0000,
  2672. 333.0000],
  2673. [506.5656, 247.0619, 739.0000, 256.0000, 321.0000, 284.0000, 601.9836,
  2674. 326.1537],
  2675. [489.3200, 241.1600, 683.0000, 244.0000, 281.0000, 215.0000, 453.0000,
  2676. 308.0000],
  2677. [504.7900, 241.0400, 685.0000, 348.0000, 295.0000, 285.0000, 506.0662,
  2678. 300.3342],
  2679. [499.9900, 243.2300, 701.6359, 283.3869, 373.0000, 322.0000, 493.0000,
  2680. 326.0000]], device='cuda:0')
  2681. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2682.  
  2683. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2684. loss_train: nan
  2685. step: 89
  2686. running loss: nan
  2687. Train Steps: 89/90 Loss: nan predictions are: tensor([[nan, nan, nan, nan, nan, nan, nan, nan],
  2688. [nan, nan, nan, nan, nan, nan, nan, nan],
  2689. [nan, nan, nan, nan, nan, nan, nan, nan],
  2690. [nan, nan, nan, nan, nan, nan, nan, nan],
  2691. [nan, nan, nan, nan, nan, nan, nan, nan],
  2692. [nan, nan, nan, nan, nan, nan, nan, nan],
  2693. [nan, nan, nan, nan, nan, nan, nan, nan],
  2694. [nan, nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
  2695. grad_fn=<AddmmBackward>)
  2696. landmarks are: tensor([[486.4500, 236.9900, 683.0000, 280.0000, 308.0000, 295.0000, 427.3566,
  2697. 297.2106],
  2698. [505.9951, 249.8763, 720.0000, 277.0000, 305.0000, 294.0000, 598.8177,
  2699. 326.8478],
  2700. [497.4600, 248.2400, 581.0000, 134.0000, 326.0000, 159.0000, 497.0000,
  2701. 347.0000],
  2702. [500.0400, 246.9800, 696.0000, 291.0000, 372.0000, 334.0000, 487.0000,
  2703. 311.0000],
  2704. [500.5700, 242.0600, 663.0000, 140.0000, 314.0000, 163.0000, 506.4223,
  2705. 294.0870],
  2706. [494.3001, 239.5567, 714.0048, 287.3219, 310.3356, 293.9974, 483.3020,
  2707. 316.7523],
  2708. [ nan, nan, 682.0000, 133.0000, 433.0000, 142.0000, 589.3204,
  2709. 328.9302],
  2710. [495.9400, 237.1000, 685.8777, 322.4308, 326.0126, 281.2020, 475.3770,
  2711. 322.6425]], device='cuda:0')
  2712. loss_train_step before backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2713. loss_train_step after backward: tensor(nan, device='cuda:0', grad_fn=<MseLossBackward>)
  2714. loss_train: nan
  2715. step: 90
  2716. running loss: nan
  2717. Valid Steps: 10/10 Loss: nan
  2718. --------------------------------------------------
  2719. Epoch: 1 Train Loss: nan Valid Loss: nan
  2720. --------------------------------------------------
  2721. Training Complete
  2722. Total Elapsed Time : 95.45278215408325 s
  2723.  
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement
Public Pastes
Advertisement
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy.  OK, I Understand
Not a member of Pastebin yet?
Sign Up, it unlocks many cool features!