libpng error: IDAT: CRC error Exception in thread Thread-1: TypeError: 'NoneType' object has no attribute '__getitem__' AttributeError: 'NoneType' object has no attribute 'shape'
Opened this issue · 2 comments
Firstly, my terminal outputs are like this:
libpng error: IDAT: CRC error
Exception in thread Thread-1:
Traceback (most recent call last):
File "/home/wh/anaconda2/envs/dy/lib/python2.7/threading.py", line 801, in __bootstrap_inner
self.run()
File "/home/wh/anaconda2/envs/dy/lib/python2.7/threading.py", line 754, in run
self.__target(*self.__args, **self.__kwargs)
File "experiments/faster_rcnn/../../faster_rcnn/../lib/utils/PrefetchingIter.py", line 60, in prefetch_func
self.next_batch[i] = self.iters[i].next()
File "experiments/faster_rcnn/../../faster_rcnn/core/loader.py", line 701, in next
self.get_batch_individual()
File "experiments/faster_rcnn/../../faster_rcnn/core/loader.py", line 808, in get_batch_individual
rst.append(self.parfetch(iroidb))
File "experiments/faster_rcnn/../../faster_rcnn/core/loader.py", line 816, in parfetch
data, label = get_rpn_batch_quadrangle(iroidb, self.cfg)
File "experiments/faster_rcnn/../../faster_rcnn/../lib/rpn/rpn.py", line 91, in get_rpn_batch_quadrangle
imgs, roidb = get_image_quadrangle_bboxes(roidb, cfg)
File "experiments/faster_rcnn/../../faster_rcnn/../lib/utils/image.py", line 61, in get_image_quadrangle_bboxes
im = im[:, ::-1, :]
TypeError: 'NoneType' object has no attribute 'getitem'
and sometimes like this:
Exception in thread Thread-1:
Traceback (most recent call last):
File "/home/wh/anaconda2/envs/dy/lib/python2.7/threading.py", line 801, in __bootstrap_inner
self.run()
File "/home/wh/anaconda2/envs/dy/lib/python2.7/threading.py", line 754, in run
self.__target(*self.__args, **self.__kwargs)
File "experiments/faster_rcnn/../../faster_rcnn/../lib/utils/PrefetchingIter.py", line 61, in prefetch_func
self.next_batch[i] = self.iters[i].next()
File "experiments/faster_rcnn/../../faster_rcnn/core/loader.py", line 701, in next
self.get_batch_individual()
File "experiments/faster_rcnn/../../faster_rcnn/core/loader.py", line 808, in get_batch_individual
rst.append(self.parfetch(iroidb))
File "experiments/faster_rcnn/../../faster_rcnn/core/loader.py", line 816, in parfetch
data, label = get_rpn_batch_quadrangle(iroidb, self.cfg)
File "experiments/faster_rcnn/../../faster_rcnn/../lib/rpn/rpn.py", line 91, in get_rpn_batch_quadrangle
imgs, roidb = get_image_quadrangle_bboxes(roidb, cfg)
File "experiments/faster_rcnn/../../faster_rcnn/../lib/utils/image.py", line 66, in get_image_quadrangle_bboxes
im, im_scale = resize(im, target_size, max_size, stride=config.network.IMAGE_STRIDE)
File "experiments/faster_rcnn/../../faster_rcnn/../lib/utils/image.py", line 219, in resize
im_shape = im.shape
AttributeError: 'NoneType' object has no attribute 'shape'
I'm sure that my images' format is .png . And my log is as followed:
2018-10-12 20:21:08,725 training config:{'CLASS_AGNOSTIC': False,
'MXNET_VERSION': 'mxnet',
'RESIZE_TO_FIX_SIZE': True,
'SCALES': [(1024, 1024)],
'TEST': {'BATCH_IMAGES': 1,
'CXX_PROPOSAL': False,
'DO_MULTISCALE_TEST': False,
'HAS_RPN': True,
'MULTISCALE': [1.0, 1.2, 1.4, 1.6],
'NMS': 0.3,
'PROPOSAL_MIN_SIZE': 0,
'PROPOSAL_NMS_THRESH': 0.7,
'PROPOSAL_POST_NMS_TOP_N': 2000,
'PROPOSAL_PRE_NMS_TOP_N': 20000,
'RPN_MIN_SIZE': 0,
'RPN_NMS_THRESH': 0.7,
'RPN_POST_NMS_TOP_N': 300,
'RPN_PRE_NMS_TOP_N': 6000,
'max_per_image': 300,
'save_img_path': '/home/wh/Faster_RCNN_for_DOTA/data/vis',
'test_epoch': 59},
'TRAIN': {'ALTERNATE': {'RCNN_BATCH_IMAGES': 0,
'RPN_BATCH_IMAGES': 0,
'rfcn1_epoch': 0,
'rfcn1_lr': 0,
'rfcn1_lr_step': '',
'rfcn2_epoch': 0,
'rfcn2_lr': 0,
'rfcn2_lr_step': '',
'rpn1_epoch': 0,
'rpn1_lr': 0,
'rpn1_lr_step': '',
'rpn2_epoch': 0,
'rpn2_lr': 0,
'rpn2_lr_step': '',
'rpn3_epoch': 0,
'rpn3_lr': 0,
'rpn3_lr_step': ''},
'ASPECT_GROUPING': True,
'BATCH_IMAGES': 1,
'BATCH_ROIS': 128,
'BATCH_ROIS_OHEM': 128,
'BBOX_MEANS': [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
'BBOX_NORMALIZATION_PRECOMPUTED': False,
'BBOX_REGRESSION_THRESH': 0.5,
'BBOX_STDS': [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0],
'BBOX_WEIGHTS': array([1., 1., 1., 1., 1., 1., 1., 1.]),
'BG_THRESH_HI': 0.5,
'BG_THRESH_LO': 0.1,
'CXX_PROPOSAL': False,
'ENABLE_OHEM': True,
'END2END': True,
'FG_FRACTION': 0.25,
'FG_THRESH': 0.5,
'FLIP': True,
'RESUME': False,
'RPN_BATCH_SIZE': 256,
'RPN_BBOX_WEIGHTS': [1.0, 1.0, 1.0, 1.0],
'RPN_CLOBBER_POSITIVES': False,
'RPN_FG_FRACTION': 0.5,
'RPN_MIN_SIZE': 0,
'RPN_NEGATIVE_OVERLAP': 0.3,
'RPN_NMS_THRESH': 0.7,
'RPN_POSITIVE_OVERLAP': 0.7,
'RPN_POSITIVE_WEIGHT': -1.0,
'RPN_POST_NMS_TOP_N': 300,
'RPN_PRE_NMS_TOP_N': 6000,
'SHUFFLE': True,
'begin_epoch': 0,
'end_epoch': 60,
'lr': 0.0005,
'lr_factor': 0.1,
'lr_step': '45,52',
'model_prefix': 'rcnn_DOTA_quadrangle',
'momentum': 0.9,
'warmup': True,
'warmup_lr': 5e-05,
'warmup_step': 1000,
'wd': 0.0005},
'dataset': {'NUM_CLASSES': 16,
'dataset': 'DOTA_oriented',
'dataset_path': '/home/wh/Faster_RCNN_for_DOTA/data',
'image_set': 'train',
'proposal': 'rpn',
'root_path': '/home/wh/Faster_RCNN_for_DOTA/data',
'test_image_set': 'test'},
'default': {'frequent': 100, 'kvstore': 'device'},
'gpus': '0',
'network': {'ANCHOR_RATIOS': [0.5, 1, 2],
'ANCHOR_SCALES': [8, 16, 32],
'FIXED_PARAMS': ['conv1',
'bn_conv1',
'res2',
'bn2',
'gamma',
'beta'],
'FIXED_PARAMS_SHARED': ['conv1',
'bn_conv1',
'res2',
'bn2',
'res3',
'bn3',
'res4',
'bn4',
'gamma',
'beta'],
'IMAGE_STRIDE': 0,
'NUM_ANCHORS': 9,
'PIXEL_MEANS': array([103.06, 115.9 , 123.15]),
'RCNN_FEAT_STRIDE': 16,
'RPN_FEAT_STRIDE': 16,
'pretrained': './model/pretrained_model/resnet_v1_101',
'pretrained_epoch': 0},
'output_path': './output/rcnn/DOTA_quadrangle',
'symbol': 'resnet_v1_101_rcnn_quadrangle'}
2018-10-12 20:21:11,505 bucketing: data "gt_boxes" has a shape (1L, 386L, 9L), which is larger than already allocated shape (1L, 100L, 9L). Need to re-allocate. Consider putting default_bucket_key to be the bucket taking the largest input for better memory sharing.
2018-10-12 20:21:19,930 bucketing: data "gt_boxes" has a shape (1L, 597L, 9L), which is larger than already allocated shape (1L, 386L, 9L). Need to re-allocate. Consider putting default_bucket_key to be the bucket taking the largest input for better memory sharing.
2018-10-12 20:21:50,888 training config:{'CLASS_AGNOSTIC': False,
'MXNET_VERSION': 'mxnet',
'RESIZE_TO_FIX_SIZE': True,
'SCALES': [(1024, 1024)],
'TEST': {'BATCH_IMAGES': 1,
'CXX_PROPOSAL': False,
'DO_MULTISCALE_TEST': False,
'HAS_RPN': True,
'MULTISCALE': [1.0, 1.2, 1.4, 1.6],
'NMS': 0.3,
'PROPOSAL_MIN_SIZE': 0,
'PROPOSAL_NMS_THRESH': 0.7,
'PROPOSAL_POST_NMS_TOP_N': 2000,
'PROPOSAL_PRE_NMS_TOP_N': 20000,
'RPN_MIN_SIZE': 0,
'RPN_NMS_THRESH': 0.7,
'RPN_POST_NMS_TOP_N': 300,
'RPN_PRE_NMS_TOP_N': 6000,
'max_per_image': 300,
'save_img_path': '/home/wh/Faster_RCNN_for_DOTA/data/vis',
'test_epoch': 59},
'TRAIN': {'ALTERNATE': {'RCNN_BATCH_IMAGES': 0,
'RPN_BATCH_IMAGES': 0,
'rfcn1_epoch': 0,
'rfcn1_lr': 0,
'rfcn1_lr_step': '',
'rfcn2_epoch': 0,
'rfcn2_lr': 0,
'rfcn2_lr_step': '',
'rpn1_epoch': 0,
'rpn1_lr': 0,
'rpn1_lr_step': '',
'rpn2_epoch': 0,
'rpn2_lr': 0,
'rpn2_lr_step': '',
'rpn3_epoch': 0,
'rpn3_lr': 0,
'rpn3_lr_step': ''},
'ASPECT_GROUPING': True,
'BATCH_IMAGES': 1,
'BATCH_ROIS': 128,
'BATCH_ROIS_OHEM': 128,
'BBOX_MEANS': [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
'BBOX_NORMALIZATION_PRECOMPUTED': False,
'BBOX_REGRESSION_THRESH': 0.5,
'BBOX_STDS': [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0],
'BBOX_WEIGHTS': array([1., 1., 1., 1., 1., 1., 1., 1.]),
'BG_THRESH_HI': 0.5,
'BG_THRESH_LO': 0.1,
'CXX_PROPOSAL': False,
'ENABLE_OHEM': True,
'END2END': True,
'FG_FRACTION': 0.25,
'FG_THRESH': 0.5,
'FLIP': True,
'RESUME': False,
'RPN_BATCH_SIZE': 256,
'RPN_BBOX_WEIGHTS': [1.0, 1.0, 1.0, 1.0],
'RPN_CLOBBER_POSITIVES': False,
'RPN_FG_FRACTION': 0.5,
'RPN_MIN_SIZE': 0,
'RPN_NEGATIVE_OVERLAP': 0.3,
'RPN_NMS_THRESH': 0.7,
'RPN_POSITIVE_OVERLAP': 0.7,
'RPN_POSITIVE_WEIGHT': -1.0,
'RPN_POST_NMS_TOP_N': 300,
'RPN_PRE_NMS_TOP_N': 6000,
'SHUFFLE': True,
'begin_epoch': 0,
'end_epoch': 60,
'lr': 0.0005,
'lr_factor': 0.1,
'lr_step': '45,52',
'model_prefix': 'rcnn_DOTA_quadrangle',
'momentum': 0.9,
'warmup': True,
'warmup_lr': 5e-05,
'warmup_step': 1000,
'wd': 0.0005},
'dataset': {'NUM_CLASSES': 16,
'dataset': 'DOTA_oriented',
'dataset_path': '/home/wh/Faster_RCNN_for_DOTA/data',
'image_set': 'train',
'proposal': 'rpn',
'root_path': '/home/wh/Faster_RCNN_for_DOTA/data',
'test_image_set': 'test'},
'default': {'frequent': 100, 'kvstore': 'device'},
'gpus': '0',
'network': {'ANCHOR_RATIOS': [0.5, 1, 2],
'ANCHOR_SCALES': [8, 16, 32],
'FIXED_PARAMS': ['conv1',
'bn_conv1',
'res2',
'bn2',
'gamma',
'beta'],
'FIXED_PARAMS_SHARED': ['conv1',
'bn_conv1',
'res2',
'bn2',
'res3',
'bn3',
'res4',
'bn4',
'gamma',
'beta'],
'IMAGE_STRIDE': 0,
'NUM_ANCHORS': 9,
'PIXEL_MEANS': array([103.06, 115.9 , 123.15]),
'RCNN_FEAT_STRIDE': 16,
'RPN_FEAT_STRIDE': 16,
'pretrained': './model/pretrained_model/resnet_v1_101',
'pretrained_epoch': 0},
'output_path': './output/rcnn/DOTA_quadrangle',
'symbol': 'resnet_v1_101_rcnn_quadrangle'}
2018-10-12 20:21:55,379 bucketing: data "gt_boxes" has a shape (1L, 263L, 9L), which is larger than already allocated shape (1L, 100L, 9L). Need to re-allocate. Consider putting default_bucket_key to be the bucket taking the largest input for better memory sharing.
2018-10-12 20:21:56,718 bucketing: data "gt_boxes" has a shape (1L, 295L, 9L), which is larger than already allocated shape (1L, 263L, 9L). Need to re-allocate. Consider putting default_bucket_key to be the bucket taking the largest input for better memory sharing.
2018-10-12 20:22:08,772 bucketing: data "gt_boxes" has a shape (1L, 1028L, 9L), which is larger than already allocated shape (1L, 295L, 9L). Need to re-allocate. Consider putting default_bucket_key to be the bucket taking the largest input for better memory sharing.
2018-10-12 20:22:19,004 Epoch[0] Batch [100] Speed: 3.89 samples/sec Train-RPNAcc=0.838219, RPNLogLoss=0.442614, RPNL1Loss=0.823888, RCNNAcc=0.750155, RCNNLogLoss=2.288598, RCNNL1Loss=0.208822,
So, what's wrong?
I think the problem should lie in data processing. Have a look. It might not be the format problem. Maybe some of your images are corrupted or the image path is not pointing to an existing image. I don't quite know. I guess it should be the data problem.
There are some wrong images, but I check the other images
and find no problem and there are still the same problems when training:
TypeError: 'NoneType' object has no attribute 'getitem'
The image name was output in the screen and I check that it can be opened.