SHARE
TWEET

torchvision

Reikriam Aug 12th, 2019 114 Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
  1. ========================================================================================== FAILURES ===========================================================================================
  2. _____________________________________________________________________________ Tester.test_fasterrcnn_resnet50_fpn _____________________________________________________________________________
  3.  
  4. self = <test_models.Tester testMethod=test_fasterrcnn_resnet50_fpn>, model_name = 'fasterrcnn_resnet50_fpn'
  5.  
  6.     def do_test(self, model_name=model_name):
  7. >       self._test_detection_model(model_name)
  8.  
  9. test/test_models.py:144:
  10. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  11. test/test_models.py:54: in _test_detection_model
  12.     out = model(model_input)
  13. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  14.     result = self.forward(*input, **kwargs)
  15. build/lib.linux-x86_64-3.7/torchvision/models/detection/generalized_rcnn.py:51: in forward
  16.     proposals, proposal_losses = self.rpn(images, features, targets)
  17. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  18.     result = self.forward(*input, **kwargs)
  19. build/lib.linux-x86_64-3.7/torchvision/models/detection/rpn.py:411: in forward
  20.     boxes, scores = self.filter_proposals(proposals, objectness, images.image_sizes, num_anchors_per_level)
  21. build/lib.linux-x86_64-3.7/torchvision/models/detection/rpn.py:336: in filter_proposals
  22.     keep = box_ops.batched_nms(boxes, scores, lvl, self.nms_thresh)
  23. build/lib.linux-x86_64-3.7/torchvision/ops/boxes.py:72: in batched_nms
  24.     keep = nms(boxes_for_nms, scores, iou_threshold)
  25. build/lib.linux-x86_64-3.7/torchvision/ops/boxes.py:32: in nms
  26.     _C = _lazy_import()
  27. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  28.  
  29.     def _lazy_import():
  30.         """
  31.         Make sure that CUDA versions match between the pytorch install and torchvision install
  32.         """
  33.         global _C
  34.         if _C is not None:
  35.             return _C
  36.         import torch
  37. >       from torchvision import _C as C
  38. E       ImportError: dynamic module does not define module export function (PyInit__C)
  39.  
  40. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  41. ____________________________________________________________________________ Tester.test_keypointrcnn_resnet50_fpn ____________________________________________________________________________
  42.  
  43. self = <test_models.Tester testMethod=test_keypointrcnn_resnet50_fpn>, model_name = 'keypointrcnn_resnet50_fpn'
  44.  
  45.     def do_test(self, model_name=model_name):
  46. >       self._test_detection_model(model_name)
  47.  
  48. test/test_models.py:144:
  49. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  50. test/test_models.py:54: in _test_detection_model
  51.     out = model(model_input)
  52. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  53.     result = self.forward(*input, **kwargs)
  54. build/lib.linux-x86_64-3.7/torchvision/models/detection/generalized_rcnn.py:51: in forward
  55.     proposals, proposal_losses = self.rpn(images, features, targets)
  56. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  57.     result = self.forward(*input, **kwargs)
  58. build/lib.linux-x86_64-3.7/torchvision/models/detection/rpn.py:411: in forward
  59.     boxes, scores = self.filter_proposals(proposals, objectness, images.image_sizes, num_anchors_per_level)
  60. build/lib.linux-x86_64-3.7/torchvision/models/detection/rpn.py:336: in filter_proposals
  61.     keep = box_ops.batched_nms(boxes, scores, lvl, self.nms_thresh)
  62. build/lib.linux-x86_64-3.7/torchvision/ops/boxes.py:72: in batched_nms
  63.     keep = nms(boxes_for_nms, scores, iou_threshold)
  64. build/lib.linux-x86_64-3.7/torchvision/ops/boxes.py:32: in nms
  65.     _C = _lazy_import()
  66. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  67.  
  68.     def _lazy_import():
  69.         """
  70.         Make sure that CUDA versions match between the pytorch install and torchvision install
  71.         """
  72.         global _C
  73.         if _C is not None:
  74.             return _C
  75.         import torch
  76. >       from torchvision import _C as C
  77. E       ImportError: dynamic module does not define module export function (PyInit__C)
  78.  
  79. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  80. ______________________________________________________________________________ Tester.test_maskrcnn_resnet50_fpn ______________________________________________________________________________
  81.  
  82. self = <test_models.Tester testMethod=test_maskrcnn_resnet50_fpn>, model_name = 'maskrcnn_resnet50_fpn'
  83.  
  84.     def do_test(self, model_name=model_name):
  85. >       self._test_detection_model(model_name)
  86.  
  87. test/test_models.py:144:
  88. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  89. test/test_models.py:54: in _test_detection_model
  90.     out = model(model_input)
  91. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  92.     result = self.forward(*input, **kwargs)
  93. build/lib.linux-x86_64-3.7/torchvision/models/detection/generalized_rcnn.py:51: in forward
  94.     proposals, proposal_losses = self.rpn(images, features, targets)
  95. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  96.     result = self.forward(*input, **kwargs)
  97. build/lib.linux-x86_64-3.7/torchvision/models/detection/rpn.py:411: in forward
  98.     boxes, scores = self.filter_proposals(proposals, objectness, images.image_sizes, num_anchors_per_level)
  99. build/lib.linux-x86_64-3.7/torchvision/models/detection/rpn.py:336: in filter_proposals
  100.     keep = box_ops.batched_nms(boxes, scores, lvl, self.nms_thresh)
  101. build/lib.linux-x86_64-3.7/torchvision/ops/boxes.py:72: in batched_nms
  102.     keep = nms(boxes_for_nms, scores, iou_threshold)
  103. build/lib.linux-x86_64-3.7/torchvision/ops/boxes.py:32: in nms
  104.     _C = _lazy_import()
  105. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  106.  
  107.     def _lazy_import():
  108.         """
  109.         Make sure that CUDA versions match between the pytorch install and torchvision install
  110.         """
  111.         global _C
  112.         if _C is not None:
  113.             return _C
  114.         import torch
  115. >       from torchvision import _C as C
  116. E       ImportError: dynamic module does not define module export function (PyInit__C)
  117.  
  118. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  119. _____________________________________________________________________ RoIPoolTester.test_roi_pool_align_non_cont_grad_cpu _____________________________________________________________________
  120.  
  121. self = <test_ops.RoIPoolTester testMethod=test_roi_pool_align_non_cont_grad_cpu>
  122.  
  123.     def test_roi_pool_align_non_cont_grad_cpu(self):
  124.         devices = ['cpu']
  125.         if torch.cuda.is_available():
  126.             devices.append('cuda')
  127.    
  128.         for d in devices:
  129.             device = torch.device(d)
  130.             rois = torch.tensor([
  131.                 [0, 0, 0, 9, 9],
  132.                 [0, 0, 5, 5, 9],
  133.                 [0, 5, 5, 9, 9]], dtype=self.dtype, device=device)
  134.    
  135.             grad_cont = torch.rand(3, 1, 5, 5, dtype=self.dtype, device=device)
  136.             grad = grad_cont.permute(2, 1, 3, 0).contiguous().permute(3, 1, 0, 2)
  137.    
  138.             for op in ['RoIPool', 'RoIAlign']:
  139.                 x = torch.rand(1, 1, 10, 10, dtype=self.dtype, device=device, requires_grad=True)
  140.                 kwargs = {}
  141.                 if op == 'RoIAlign':
  142.                     kwargs['sampling_ratio'] = 1
  143.                 m = getattr(ops, op)((5, 5), 1, **kwargs)
  144.    
  145. >               y = m(x, rois)
  146.  
  147. test/test_ops.py:162:
  148. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  149. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  150.     result = self.forward(*input, **kwargs)
  151. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:75: in forward
  152.     return roi_pool(input, rois, self.output_size, self.spatial_scale)
  153. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:62: in roi_pool
  154.     return _RoIPoolFunction.apply(input, rois, output_size, spatial_scale)
  155. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:19: in forward
  156.     _C = _lazy_import()
  157. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  158.  
  159.     def _lazy_import():
  160.         """
  161.         Make sure that CUDA versions match between the pytorch install and torchvision install
  162.         """
  163.         global _C
  164.         if _C is not None:
  165.             return _C
  166.         import torch
  167. >       from torchvision import _C as C
  168. E       ImportError: dynamic module does not define module export function (PyInit__C)
  169.  
  170. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  171. ____________________________________________________________________________ RoIPoolTester.test_roi_pool_basic_cpu ____________________________________________________________________________
  172.  
  173. self = <test_ops.RoIPoolTester testMethod=test_roi_pool_basic_cpu>
  174.  
  175.     def test_roi_pool_basic_cpu(self):
  176.         device = torch.device('cpu')
  177.         x = torch.rand(1, 1, 10, 10, dtype=self.dtype, device=device)
  178.         rois = torch.tensor([[0, 0, 0, 4, 4]],  # format is (xyxy)
  179.                             dtype=self.dtype, device=device)
  180.    
  181.         pool_h, pool_w = (5, 5)
  182.         roi_pool = ops.RoIPool((pool_h, pool_w), 1)
  183. >       y = roi_pool(x, rois)
  184.  
  185. test/test_ops.py:50:
  186. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  187. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  188.     result = self.forward(*input, **kwargs)
  189. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:75: in forward
  190.     return roi_pool(input, rois, self.output_size, self.spatial_scale)
  191. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:62: in roi_pool
  192.     return _RoIPoolFunction.apply(input, rois, output_size, spatial_scale)
  193. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:19: in forward
  194.     _C = _lazy_import()
  195. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  196.  
  197.     def _lazy_import():
  198.         """
  199.         Make sure that CUDA versions match between the pytorch install and torchvision install
  200.         """
  201.         global _C
  202.         if _C is not None:
  203.             return _C
  204.         import torch
  205. >       from torchvision import _C as C
  206. E       ImportError: dynamic module does not define module export function (PyInit__C)
  207.  
  208. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  209. ___________________________________________________________________________ RoIPoolTester.test_roi_pool_basic_cuda ____________________________________________________________________________
  210.  
  211. self = <test_ops.RoIPoolTester testMethod=test_roi_pool_basic_cuda>
  212.  
  213.     @unittest.skipIf(not torch.cuda.is_available(), "CUDA unavailable")
  214.     def test_roi_pool_basic_cuda(self):
  215.         device = torch.device('cuda')
  216.         x = torch.rand(1, 1, 10, 10, dtype=self.dtype, device=device)
  217.         rois = torch.tensor([[0, 0, 0, 4, 4]],  # format is (xyxy)
  218.                             dtype=self.dtype, device=device)
  219.    
  220.         pool_h, pool_w = (5, 5)
  221.         roi_pool = ops.RoIPool((pool_h, pool_w), 1)
  222. >       y = roi_pool(x, rois)
  223.  
  224. test/test_ops.py:200:
  225. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  226. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  227.     result = self.forward(*input, **kwargs)
  228. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:75: in forward
  229.     return roi_pool(input, rois, self.output_size, self.spatial_scale)
  230. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:62: in roi_pool
  231.     return _RoIPoolFunction.apply(input, rois, output_size, spatial_scale)
  232. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:19: in forward
  233.     _C = _lazy_import()
  234. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  235.  
  236.     def _lazy_import():
  237.         """
  238.         Make sure that CUDA versions match between the pytorch install and torchvision install
  239.         """
  240.         global _C
  241.         if _C is not None:
  242.             return _C
  243.         import torch
  244. >       from torchvision import _C as C
  245. E       ImportError: dynamic module does not define module export function (PyInit__C)
  246.  
  247. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  248. _______________________________________________________________________________ RoIPoolTester.test_roi_pool_cpu _______________________________________________________________________________
  249.  
  250. self = <test_ops.RoIPoolTester testMethod=test_roi_pool_cpu>
  251.  
  252.     def test_roi_pool_cpu(self):
  253.         device = torch.device('cpu')
  254.         x = torch.rand(2, 1, 10, 10, dtype=self.dtype, device=device)
  255.         rois = torch.tensor([[0, 0, 0, 9, 9],  # format is (xyxy)
  256.                              [0, 0, 5, 4, 9],
  257.                              [0, 5, 5, 9, 9],
  258.                              [1, 0, 0, 9, 9]],
  259.                             dtype=self.dtype, device=device)
  260.    
  261.         pool_h, pool_w = (5, 5)
  262.         roi_pool = ops.RoIPool((pool_h, pool_w), 1)
  263. >       y = roi_pool(x, rois)
  264.  
  265. test/test_ops.py:72:
  266. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  267. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  268.     result = self.forward(*input, **kwargs)
  269. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:75: in forward
  270.     return roi_pool(input, rois, self.output_size, self.spatial_scale)
  271. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:62: in roi_pool
  272.     return _RoIPoolFunction.apply(input, rois, output_size, spatial_scale)
  273. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:19: in forward
  274.     _C = _lazy_import()
  275. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  276.  
  277.     def _lazy_import():
  278.         """
  279.         Make sure that CUDA versions match between the pytorch install and torchvision install
  280.         """
  281.         global _C
  282.         if _C is not None:
  283.             return _C
  284.         import torch
  285. >       from torchvision import _C as C
  286. E       ImportError: dynamic module does not define module export function (PyInit__C)
  287.  
  288. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  289. _________________________________________________________________________ RoIPoolTester.test_roi_pool_cpu_empty_rois __________________________________________________________________________
  290.  
  291. self = <test_ops.RoIPoolTester testMethod=test_roi_pool_cpu_empty_rois>
  292.  
  293.     def test_roi_pool_cpu_empty_rois(self):
  294.         device = torch.device('cpu')
  295.         x = torch.tensor(
  296.             [[[[0.1767, 1.2851, 4.2325, 4.8645, 7.1496]],
  297.               [[2.5916, 4.3361, 3.8143, 6.1329, 2.0230]],
  298.               [[1.4492, 3.3384, 4.0816, 6.3116, 5.1068]]]],
  299.             dtype=self.dtype, device=device)
  300.         rois = torch.tensor(
  301.             [[0., 1., 0., 4., 0.],
  302.              [0., 2., 0., 3., 0.],
  303.              [0., 0., 0., 0., 0.],
  304.              [0., 0., 0., 0., 0.],
  305.              [0., 2., 0., 2., 0.]],
  306.             dtype=self.dtype, device=device)
  307.    
  308.         pool_h, pool_w = (1, 2)
  309.         roi_pool = ops.RoIPool((pool_h, pool_w), 1)
  310. >       y = roi_pool(x, rois)
  311.  
  312. test/test_ops.py:100:
  313. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  314. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  315.     result = self.forward(*input, **kwargs)
  316. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:75: in forward
  317.     return roi_pool(input, rois, self.output_size, self.spatial_scale)
  318. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:62: in roi_pool
  319.     return _RoIPoolFunction.apply(input, rois, output_size, spatial_scale)
  320. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:19: in forward
  321.     _C = _lazy_import()
  322. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  323.  
  324.     def _lazy_import():
  325.         """
  326.         Make sure that CUDA versions match between the pytorch install and torchvision install
  327.         """
  328.         global _C
  329.         if _C is not None:
  330.             return _C
  331.         import torch
  332. >       from torchvision import _C as C
  333. E       ImportError: dynamic module does not define module export function (PyInit__C)
  334.  
  335. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  336. ______________________________________________________________________________ RoIPoolTester.test_roi_pool_cuda _______________________________________________________________________________
  337.  
  338. self = <test_ops.RoIPoolTester testMethod=test_roi_pool_cuda>
  339.  
  340.     @unittest.skipIf(not torch.cuda.is_available(), "CUDA unavailable")
  341.     def test_roi_pool_cuda(self):
  342.         device = torch.device('cuda') if torch.cuda.is_available() else torch.device('cpu')
  343.         x = torch.rand(2, 1, 10, 10, dtype=self.dtype, device=device)
  344.         rois = torch.tensor([[0, 0, 0, 9, 9],  # format is (xyxy)
  345.                              [0, 0, 5, 4, 9],
  346.                              [0, 5, 5, 9, 9],
  347.                              [1, 0, 0, 9, 9]],
  348.                             dtype=self.dtype, device=device)
  349.    
  350.         pool_h, pool_w = (5, 5)
  351.         roi_pool = ops.RoIPool((pool_h, pool_w), 1)
  352. >       y = roi_pool(x, rois)
  353.  
  354. test/test_ops.py:222:
  355. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  356. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  357.     result = self.forward(*input, **kwargs)
  358. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:75: in forward
  359.     return roi_pool(input, rois, self.output_size, self.spatial_scale)
  360. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:62: in roi_pool
  361.     return _RoIPoolFunction.apply(input, rois, output_size, spatial_scale)
  362. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:19: in forward
  363.     _C = _lazy_import()
  364. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  365.  
  366.     def _lazy_import():
  367.         """
  368.         Make sure that CUDA versions match between the pytorch install and torchvision install
  369.         """
  370.         global _C
  371.         if _C is not None:
  372.             return _C
  373.         import torch
  374. >       from torchvision import _C as C
  375. E       ImportError: dynamic module does not define module export function (PyInit__C)
  376.  
  377. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  378. __________________________________________________________________________ RoIPoolTester.test_roi_pool_gradcheck_cpu __________________________________________________________________________
  379.  
  380. self = <test_ops.RoIPoolTester testMethod=test_roi_pool_gradcheck_cpu>
  381.  
  382.     def test_roi_pool_gradcheck_cpu(self):
  383.         device = torch.device('cpu')
  384.         x = torch.rand(1, 1, 10, 10, dtype=self.dtype, device=device, requires_grad=True)
  385.         rois = torch.tensor([
  386.             [0, 0, 0, 9, 9],
  387.             [0, 0, 5, 5, 9],
  388.             [0, 5, 5, 9, 9]], dtype=self.dtype, device=device)
  389.    
  390.         m = ops.RoIPool((5, 5), 1).to(dtype=self.dtype, device=device)
  391.    
  392.         def func(input):
  393.             return m(input, rois)
  394.    
  395. >       assert gradcheck(func, (x,)), 'gradcheck failed for roi_pool CPU'
  396.  
  397. test/test_ops.py:188:
  398. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  399. /usr/lib/python3.7/site-packages/torch/autograd/gradcheck.py:256: in gradcheck
  400.     func_out = func(*tupled_inputs)
  401. test/test_ops.py:186: in func
  402.     return m(input, rois)
  403. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  404.     result = self.forward(*input, **kwargs)
  405. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:75: in forward
  406.     return roi_pool(input, rois, self.output_size, self.spatial_scale)
  407. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:62: in roi_pool
  408.     return _RoIPoolFunction.apply(input, rois, output_size, spatial_scale)
  409. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:19: in forward
  410.     _C = _lazy_import()
  411. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  412.  
  413.     def _lazy_import():
  414.         """
  415.         Make sure that CUDA versions match between the pytorch install and torchvision install
  416.         """
  417.         global _C
  418.         if _C is not None:
  419.             return _C
  420.         import torch
  421. >       from torchvision import _C as C
  422. E       ImportError: dynamic module does not define module export function (PyInit__C)
  423.  
  424. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  425. _________________________________________________________________________ RoIPoolTester.test_roi_pool_gradcheck_cuda __________________________________________________________________________
  426.  
  427. self = <test_ops.RoIPoolTester testMethod=test_roi_pool_gradcheck_cuda>
  428.  
  429.     @unittest.skipIf(not torch.cuda.is_available(), "CUDA unavailable")
  430.     def test_roi_pool_gradcheck_cuda(self):
  431.         device = torch.device('cuda')
  432.         x = torch.rand(1, 1, 10, 10, dtype=self.dtype, device=device, requires_grad=True)
  433.         rois = torch.tensor([
  434.             [0, 0, 0, 9, 9],
  435.             [0, 0, 5, 5, 9],
  436.             [0, 5, 5, 9, 9]], dtype=self.dtype, device=device)
  437.    
  438.         m = ops.RoIPool((5, 5), 1).to(dtype=self.dtype, device=device)
  439.    
  440.         def func(input):
  441.             return m(input, rois)
  442.    
  443. >       assert gradcheck(func, (x,)), 'gradcheck failed for roi_pool CUDA'
  444.  
  445. test/test_ops.py:274:
  446. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  447. /usr/lib/python3.7/site-packages/torch/autograd/gradcheck.py:256: in gradcheck
  448.     func_out = func(*tupled_inputs)
  449. test/test_ops.py:272: in func
  450.     return m(input, rois)
  451. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  452.     result = self.forward(*input, **kwargs)
  453. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:75: in forward
  454.     return roi_pool(input, rois, self.output_size, self.spatial_scale)
  455. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:62: in roi_pool
  456.     return _RoIPoolFunction.apply(input, rois, output_size, spatial_scale)
  457. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:19: in forward
  458.     _C = _lazy_import()
  459. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  460.  
  461.     def _lazy_import():
  462.         """
  463.         Make sure that CUDA versions match between the pytorch install and torchvision install
  464.         """
  465.         global _C
  466.         if _C is not None:
  467.             return _C
  468.         import torch
  469. >       from torchvision import _C as C
  470. E       ImportError: dynamic module does not define module export function (PyInit__C)
  471.  
  472. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  473. __________________________________________________________________________ RoIPoolTester.test_roi_pool_gradient_cpu ___________________________________________________________________________
  474.  
  475. self = <test_ops.RoIPoolTester testMethod=test_roi_pool_gradient_cpu>
  476.  
  477.     def test_roi_pool_gradient_cpu(self):
  478.         device = torch.device('cpu')
  479.         x = torch.ones(1, 1, 10, 10, dtype=self.dtype, device=device, requires_grad=True)
  480.         rois = torch.tensor([
  481.             [0, 0, 0, 9, 9],
  482.             [0, 0, 5, 4, 9],
  483.             [0, 0, 0, 4, 4]],
  484.             dtype=self.dtype, device=device)
  485.    
  486.         layer = ops.RoIPool((5, 5), 1).to(dtype=self.dtype, device=device)
  487.    
  488. >       y = layer(x, rois)
  489.  
  490. test/test_ops.py:122:
  491. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  492. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  493.     result = self.forward(*input, **kwargs)
  494. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:75: in forward
  495.     return roi_pool(input, rois, self.output_size, self.spatial_scale)
  496. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:62: in roi_pool
  497.     return _RoIPoolFunction.apply(input, rois, output_size, spatial_scale)
  498. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:19: in forward
  499.     _C = _lazy_import()
  500. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  501.  
  502.     def _lazy_import():
  503.         """
  504.         Make sure that CUDA versions match between the pytorch install and torchvision install
  505.         """
  506.         global _C
  507.         if _C is not None:
  508.             return _C
  509.         import torch
  510. >       from torchvision import _C as C
  511. E       ImportError: dynamic module does not define module export function (PyInit__C)
  512.  
  513. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  514. __________________________________________________________________________ RoIPoolTester.test_roi_pool_gradient_cuda __________________________________________________________________________
  515.  
  516. self = <test_ops.RoIPoolTester testMethod=test_roi_pool_gradient_cuda>
  517.  
  518.     @unittest.skipIf(not torch.cuda.is_available(), "CUDA unavailable")
  519.     def test_roi_pool_gradient_cuda(self):
  520.         device = torch.device('cuda')
  521.         layer = ops.RoIPool((5, 5), 1).to(dtype=self.dtype, device=device)
  522.         x = torch.ones(1, 1, 10, 10, dtype=self.dtype, device=device, requires_grad=True)
  523.         rois = torch.tensor([
  524.             [0, 0, 0, 9, 9],
  525.             [0, 0, 5, 4, 9],
  526.             [0, 0, 0, 4, 4]],
  527.             dtype=self.dtype, device=device)
  528.    
  529. >       y = layer(x, rois)
  530.  
  531. test/test_ops.py:243:
  532. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  533. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  534.     result = self.forward(*input, **kwargs)
  535. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:75: in forward
  536.     return roi_pool(input, rois, self.output_size, self.spatial_scale)
  537. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:62: in roi_pool
  538.     return _RoIPoolFunction.apply(input, rois, output_size, spatial_scale)
  539. build/lib.linux-x86_64-3.7/torchvision/ops/roi_pool.py:19: in forward
  540.     _C = _lazy_import()
  541. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  542.  
  543.     def _lazy_import():
  544.         """
  545.         Make sure that CUDA versions match between the pytorch install and torchvision install
  546.         """
  547.         global _C
  548.         if _C is not None:
  549.             return _C
  550.         import torch
  551. >       from torchvision import _C as C
  552. E       ImportError: dynamic module does not define module export function (PyInit__C)
  553.  
  554. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  555. ___________________________________________________________________________ RoIAlignTester.test_roi_align_basic_cpu ___________________________________________________________________________
  556.  
  557. self = <test_ops.RoIAlignTester testMethod=test_roi_align_basic_cpu>
  558.  
  559.     def test_roi_align_basic_cpu(self):
  560.         device = torch.device('cpu')
  561.         x = self.x.to(device)
  562.         single_roi = self.single_roi.to(device)
  563.         gt_y_single = self.gt_y_single.to(device)
  564.    
  565.         pool_h, pool_w = (5, 5)
  566.         roi_align = ops.RoIAlign((pool_h, pool_w), spatial_scale=1, sampling_ratio=2).to(device=device)
  567. >       y = roi_align(x, single_roi)
  568.  
  569. test/test_ops.py:345:
  570. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  571. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  572.     result = self.forward(*input, **kwargs)
  573. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:83: in forward
  574.     return roi_align(input, rois, self.output_size, self.spatial_scale, self.sampling_ratio)
  575. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:69: in roi_align
  576.     return _RoIAlignFunction.apply(input, rois, output_size, spatial_scale, sampling_ratio)
  577. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:21: in forward
  578.     _C = _lazy_import()
  579. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  580.  
  581.     def _lazy_import():
  582.         """
  583.         Make sure that CUDA versions match between the pytorch install and torchvision install
  584.         """
  585.         global _C
  586.         if _C is not None:
  587.             return _C
  588.         import torch
  589. >       from torchvision import _C as C
  590. E       ImportError: dynamic module does not define module export function (PyInit__C)
  591.  
  592. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  593. __________________________________________________________________________ RoIAlignTester.test_roi_align_basic_cuda ___________________________________________________________________________
  594.  
  595. self = <test_ops.RoIAlignTester testMethod=test_roi_align_basic_cuda>
  596.  
  597.     @unittest.skipIf(not torch.cuda.is_available(), "CUDA unavailable")
  598.     def test_roi_align_basic_cuda(self):
  599.         device = torch.device('cuda')
  600.         x = self.x.to(device)
  601.         single_roi = self.single_roi.to(device)
  602.         gt_y_single = self.gt_y_single.to(device)
  603.    
  604.         pool_h, pool_w = (5, 5)
  605.         roi_align = ops.RoIAlign((pool_h, pool_w), spatial_scale=1, sampling_ratio=2).to(device=device)
  606. >       y = roi_align(x, single_roi)
  607.  
  608. test/test_ops.py:376:
  609. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  610. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  611.     result = self.forward(*input, **kwargs)
  612. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:83: in forward
  613.     return roi_align(input, rois, self.output_size, self.spatial_scale, self.sampling_ratio)
  614. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:69: in roi_align
  615.     return _RoIAlignFunction.apply(input, rois, output_size, spatial_scale, sampling_ratio)
  616. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:21: in forward
  617.     _C = _lazy_import()
  618. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  619.  
  620.     def _lazy_import():
  621.         """
  622.         Make sure that CUDA versions match between the pytorch install and torchvision install
  623.         """
  624.         global _C
  625.         if _C is not None:
  626.             return _C
  627.         import torch
  628. >       from torchvision import _C as C
  629. E       ImportError: dynamic module does not define module export function (PyInit__C)
  630.  
  631. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  632. ______________________________________________________________________________ RoIAlignTester.test_roi_align_cpu ______________________________________________________________________________
  633.  
  634. self = <test_ops.RoIAlignTester testMethod=test_roi_align_cpu>
  635.  
  636.     def test_roi_align_cpu(self):
  637.         device = torch.device('cpu')
  638.         x = self.x.to(device)
  639.         rois = self.rois.to(device)
  640.         gt_y_multiple = self.gt_y_multiple.to(device)
  641.    
  642.         pool_h, pool_w = (5, 5)
  643.         roi_align = ops.RoIAlign((pool_h, pool_w), spatial_scale=1, sampling_ratio=2).to(device=device)
  644. >       y = roi_align(x, rois)
  645.  
  646. test/test_ops.py:360:
  647. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  648. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  649.     result = self.forward(*input, **kwargs)
  650. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:83: in forward
  651.     return roi_align(input, rois, self.output_size, self.spatial_scale, self.sampling_ratio)
  652. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:69: in roi_align
  653.     return _RoIAlignFunction.apply(input, rois, output_size, spatial_scale, sampling_ratio)
  654. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:21: in forward
  655.     _C = _lazy_import()
  656. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  657.  
  658.     def _lazy_import():
  659.         """
  660.         Make sure that CUDA versions match between the pytorch install and torchvision install
  661.         """
  662.         global _C
  663.         if _C is not None:
  664.             return _C
  665.         import torch
  666. >       from torchvision import _C as C
  667. E       ImportError: dynamic module does not define module export function (PyInit__C)
  668.  
  669. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  670. _____________________________________________________________________________ RoIAlignTester.test_roi_align_cuda ______________________________________________________________________________
  671.  
  672. self = <test_ops.RoIAlignTester testMethod=test_roi_align_cuda>
  673.  
  674.     @unittest.skipIf(not torch.cuda.is_available(), "CUDA unavailable")
  675.     def test_roi_align_cuda(self):
  676.         device = torch.device('cuda')
  677.         x = self.x.to(device)
  678.         rois = self.rois.to(device)
  679.         gt_y_multiple = self.gt_y_multiple.to(device)
  680.    
  681.         pool_h, pool_w = (5, 5)
  682.         roi_align = ops.RoIAlign((pool_h, pool_w), spatial_scale=1, sampling_ratio=2).to(device=device)
  683. >       y = roi_align(x, rois)
  684.  
  685. test/test_ops.py:392:
  686. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  687. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  688.     result = self.forward(*input, **kwargs)
  689. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:83: in forward
  690.     return roi_align(input, rois, self.output_size, self.spatial_scale, self.sampling_ratio)
  691. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:69: in roi_align
  692.     return _RoIAlignFunction.apply(input, rois, output_size, spatial_scale, sampling_ratio)
  693. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:21: in forward
  694.     _C = _lazy_import()
  695. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  696.  
  697.     def _lazy_import():
  698.         """
  699.         Make sure that CUDA versions match between the pytorch install and torchvision install
  700.         """
  701.         global _C
  702.         if _C is not None:
  703.             return _C
  704.         import torch
  705. >       from torchvision import _C as C
  706. E       ImportError: dynamic module does not define module export function (PyInit__C)
  707.  
  708. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  709. _________________________________________________________________________ RoIAlignTester.test_roi_align_gradcheck_cpu _________________________________________________________________________
  710.  
  711. self = <test_ops.RoIAlignTester testMethod=test_roi_align_gradcheck_cpu>
  712.  
  713.     def test_roi_align_gradcheck_cpu(self):
  714.         dtype = torch.float64
  715.         device = torch.device('cpu')
  716.         m = ops.RoIAlign((5, 5), 0.5, 1).to(dtype=dtype, device=device)
  717.         x = torch.rand(1, 1, 10, 10, dtype=dtype, device=device, requires_grad=True)
  718.         rois = self.rois.to(device=device, dtype=dtype)
  719.    
  720.         def func(input):
  721.             return m(input, rois)
  722.    
  723. >       assert gradcheck(func, (x,)), 'gradcheck failed for RoIAlign CPU'
  724.  
  725. test/test_ops.py:428:
  726. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  727. /usr/lib/python3.7/site-packages/torch/autograd/gradcheck.py:256: in gradcheck
  728.     func_out = func(*tupled_inputs)
  729. test/test_ops.py:426: in func
  730.     return m(input, rois)
  731. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  732.     result = self.forward(*input, **kwargs)
  733. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:83: in forward
  734.     return roi_align(input, rois, self.output_size, self.spatial_scale, self.sampling_ratio)
  735. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:69: in roi_align
  736.     return _RoIAlignFunction.apply(input, rois, output_size, spatial_scale, sampling_ratio)
  737. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:21: in forward
  738.     _C = _lazy_import()
  739. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  740.  
  741.     def _lazy_import():
  742.         """
  743.         Make sure that CUDA versions match between the pytorch install and torchvision install
  744.         """
  745.         global _C
  746.         if _C is not None:
  747.             return _C
  748.         import torch
  749. >       from torchvision import _C as C
  750. E       ImportError: dynamic module does not define module export function (PyInit__C)
  751.  
  752. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  753. ________________________________________________________________________ RoIAlignTester.test_roi_align_gradcheck_cuda _________________________________________________________________________
  754.  
  755. self = <test_ops.RoIAlignTester testMethod=test_roi_align_gradcheck_cuda>
  756.  
  757.     @unittest.skipIf(not torch.cuda.is_available(), "CUDA unavailable")
  758.     def test_roi_align_gradcheck_cuda(self):
  759.         dtype = torch.float64
  760.         device = torch.device('cuda')
  761.         m = ops.RoIAlign((5, 5), 0.5, 1).to(dtype=dtype, device=device)
  762.         x = torch.rand(1, 1, 10, 10, dtype=dtype, device=device, requires_grad=True)
  763.         rois = self.rois.to(device=device, dtype=dtype)
  764.    
  765.         def func(input):
  766.             return m(input, rois)
  767.    
  768. >       assert gradcheck(func, (x,)), 'gradcheck failed for RoIAlign CUDA'
  769.  
  770. test/test_ops.py:462:
  771. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  772. /usr/lib/python3.7/site-packages/torch/autograd/gradcheck.py:256: in gradcheck
  773.     func_out = func(*tupled_inputs)
  774. test/test_ops.py:460: in func
  775.     return m(input, rois)
  776. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  777.     result = self.forward(*input, **kwargs)
  778. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:83: in forward
  779.     return roi_align(input, rois, self.output_size, self.spatial_scale, self.sampling_ratio)
  780. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:69: in roi_align
  781.     return _RoIAlignFunction.apply(input, rois, output_size, spatial_scale, sampling_ratio)
  782. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:21: in forward
  783.     _C = _lazy_import()
  784. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  785.  
  786.     def _lazy_import():
  787.         """
  788.         Make sure that CUDA versions match between the pytorch install and torchvision install
  789.         """
  790.         global _C
  791.         if _C is not None:
  792.             return _C
  793.         import torch
  794. >       from torchvision import _C as C
  795. E       ImportError: dynamic module does not define module export function (PyInit__C)
  796.  
  797. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  798. _________________________________________________________________________ RoIAlignTester.test_roi_align_gradient_cpu __________________________________________________________________________
  799.  
  800. self = <test_ops.RoIAlignTester testMethod=test_roi_align_gradient_cpu>
  801.  
  802.     def test_roi_align_gradient_cpu(self):
  803.         """
  804.         Compute gradients for RoIAlign with multiple bounding boxes on CPU
  805.         """
  806.         device = torch.device('cpu')
  807.         pool_h, pool_w = (5, 5)
  808.         roi_align = ops.RoIAlign((pool_h, pool_w), spatial_scale=1, sampling_ratio=2).to(device=device)
  809.    
  810.         x = self.x.to(device).clone()
  811.         rois = self.rois.to(device)
  812.         gt_grad = self.x_grad.to(device)
  813.    
  814.         x.requires_grad = True
  815. >       y = roi_align(x, rois)
  816.  
  817. test/test_ops.py:412:
  818. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  819. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  820.     result = self.forward(*input, **kwargs)
  821. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:83: in forward
  822.     return roi_align(input, rois, self.output_size, self.spatial_scale, self.sampling_ratio)
  823. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:69: in roi_align
  824.     return _RoIAlignFunction.apply(input, rois, output_size, spatial_scale, sampling_ratio)
  825. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:21: in forward
  826.     _C = _lazy_import()
  827. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  828.  
  829.     def _lazy_import():
  830.         """
  831.         Make sure that CUDA versions match between the pytorch install and torchvision install
  832.         """
  833.         global _C
  834.         if _C is not None:
  835.             return _C
  836.         import torch
  837. >       from torchvision import _C as C
  838. E       ImportError: dynamic module does not define module export function (PyInit__C)
  839.  
  840. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  841. _________________________________________________________________________ RoIAlignTester.test_roi_align_gradient_cuda _________________________________________________________________________
  842.  
  843. self = <test_ops.RoIAlignTester testMethod=test_roi_align_gradient_cuda>
  844.  
  845.     @unittest.skipIf(not torch.cuda.is_available(), "CUDA unavailable")
  846.     def test_roi_align_gradient_cuda(self):
  847.         """
  848.         Compute gradients for RoIAlign with multiple bounding boxes on the GPU
  849.         """
  850.         device = torch.device('cuda')
  851.         pool_h, pool_w = (5, 5)
  852.         roi_align = ops.RoIAlign((pool_h, pool_w), spatial_scale=1, sampling_ratio=2).to(device=device)
  853.    
  854.         x = self.x.to(device).clone()
  855.         rois = self.rois.to(device)
  856.         gt_grad = self.x_grad.to(device)
  857.    
  858.         x.requires_grad = True
  859. >       y = roi_align(x, rois)
  860.  
  861. test/test_ops.py:445:
  862. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  863. /usr/lib/python3.7/site-packages/torch/nn/modules/module.py:493: in __call__
  864.     result = self.forward(*input, **kwargs)
  865. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:83: in forward
  866.     return roi_align(input, rois, self.output_size, self.spatial_scale, self.sampling_ratio)
  867. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:69: in roi_align
  868.     return _RoIAlignFunction.apply(input, rois, output_size, spatial_scale, sampling_ratio)
  869. build/lib.linux-x86_64-3.7/torchvision/ops/roi_align.py:21: in forward
  870.     _C = _lazy_import()
  871. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  872.  
  873.     def _lazy_import():
  874.         """
  875.         Make sure that CUDA versions match between the pytorch install and torchvision install
  876.         """
  877.         global _C
  878.         if _C is not None:
  879.             return _C
  880.         import torch
  881. >       from torchvision import _C as C
  882. E       ImportError: dynamic module does not define module export function (PyInit__C)
  883.  
  884. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  885. _____________________________________________________________________________________ NMSTester.test_nms ______________________________________________________________________________________
  886.  
  887. self = <test_ops.NMSTester testMethod=test_nms>
  888.  
  889.     def test_nms(self):
  890.         boxes, scores = self._create_tensors(1000)
  891.         err_msg = 'NMS incompatible between CPU and reference implementation for IoU={}'
  892.         for iou in [0.2, 0.5, 0.8]:
  893.             keep_ref = self.reference_nms(boxes, scores, iou)
  894. >           keep = ops.nms(boxes, scores, iou)
  895.  
  896. test/test_ops.py:501:
  897. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  898. build/lib.linux-x86_64-3.7/torchvision/ops/boxes.py:32: in nms
  899.     _C = _lazy_import()
  900. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  901.  
  902.     def _lazy_import():
  903.         """
  904.         Make sure that CUDA versions match between the pytorch install and torchvision install
  905.         """
  906.         global _C
  907.         if _C is not None:
  908.             return _C
  909.         import torch
  910. >       from torchvision import _C as C
  911. E       ImportError: dynamic module does not define module export function (PyInit__C)
  912.  
  913. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  914. ___________________________________________________________________________________ NMSTester.test_nms_cuda ___________________________________________________________________________________
  915.  
  916. self = <test_ops.NMSTester testMethod=test_nms_cuda>
  917.  
  918.     @unittest.skipIf(not torch.cuda.is_available(), "CUDA unavailable")
  919.     def test_nms_cuda(self):
  920.         boxes, scores = self._create_tensors(1000)
  921.         err_msg = 'NMS incompatible between CPU and CUDA for IoU={}'
  922.    
  923.         for iou in [0.2, 0.5, 0.8]:
  924. >           r_cpu = ops.nms(boxes, scores, iou)
  925.  
  926. test/test_ops.py:510:
  927. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  928. build/lib.linux-x86_64-3.7/torchvision/ops/boxes.py:32: in nms
  929.     _C = _lazy_import()
  930. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
  931.  
  932.     def _lazy_import():
  933.         """
  934.         Make sure that CUDA versions match between the pytorch install and torchvision install
  935.         """
  936.         global _C
  937.         if _C is not None:
  938.             return _C
  939.         import torch
  940. >       from torchvision import _C as C
  941. E       ImportError: dynamic module does not define module export function (PyInit__C)
  942.  
  943. build/lib.linux-x86_64-3.7/torchvision/extension.py:12: ImportError
  944. ====================================================================================== warnings summary =======================================================================================
  945. build/lib.linux-x86_64-3.7/torchvision/datasets/lsun.py:8
  946.   /home/reik/.cache/yay/python-torchvision/src/vision-0.4.0/build/lib.linux-x86_64-3.7/torchvision/datasets/lsun.py:8: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
  947.     from collections import Iterable
  948.  
  949. /usr/lib/python3.7/site-packages/av/container/__init__.py:1
  950.   /usr/lib/python3.7/site-packages/av/container/__init__.py:1: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
  951.     from .core import Container, open
  952.  
  953. test/test_cpp_models.py::Tester::test_resnet101
  954.   /usr/lib/python3.7/site-packages/torch/jit/__init__.py:702: TracerWarning: Output nr 1. of the traced function does not match the corresponding output of the Python function. Detailed error:
  955.   Not within tolerance rtol=1e-05 atol=1e-05 at input[0, 929] (-2.8711626529693604 vs. -2.87127685546875) and 1 other locations (0.00%)
  956.     _check_trace([example_inputs], func, executor_options, traced, check_tolerance, _force_outplace)
  957.  
  958. test/test_cpp_models.py::Tester::test_wide_resnet101_2
  959.   /usr/lib/python3.7/site-packages/torch/jit/__init__.py:702: TracerWarning: Output nr 1. of the traced function does not match the corresponding output of the Python function. Detailed error:
  960.   Not within tolerance rtol=1e-05 atol=1e-05 at input[0, 402] (-1.1914129257202148 vs. -1.1916732788085938) and 1 other locations (0.00%)
  961.     _check_trace([example_inputs], func, executor_options, traced, check_tolerance, _force_outplace)
  962.  
  963. -- Docs: https://docs.pytest.org/en/latest/warnings.html
  964. =============================================================== 23 failed, 159 passed, 5 skipped, 4 warnings in 1421.04 seconds ===============================================================
  965. ==> ERROR: A failure occurred in check().
  966.     Aborting...
  967. Error making: python-torchvision
RAW Paste Data
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
 
Top