AssertionError: size of input tensor and input format are different.

[luluning]

I want to train my own model,and I try ./experiments/scripts/train_faster_rcnn.sh 0 pascal_voc vgg16.I got an error message
Loading initial model weights from data/imagenet_weights/vgg16.pth
Loaded.
Traceback (most recent call last):
File "./tools/trainval_net.py", line 164, in
max_iters=args.max_iters)
File "/XXX/pytorch-faster-rcnn-master/tools/../lib/model/train_val.py", line 393, in train_net
sw.train_model(max_iters)
File "/XXX/pytorch-faster-rcnn-master/tools/../lib/model/train_val.py", line 286, in train_model
self.net.train_step_with_summary(blobs, self.optimizer)
File "/XXX/pytorch-faster-rcnn-master/tools/../lib/nets/network.py", line 501, in train_step_with_summary
summary = self._run_summary_op()
File "/XXX/pytorch-faster-rcnn-master/tools/../lib/nets/network.py", line 348, in _run_summary_op
summaries.append(self._add_gt_image_summary())
File "/XXX/pytorch-faster-rcnn-master/tools/../lib/nets/network.py", line 66, in _add_gt_image_summary
image[0].astype('float32') / 255.0)
File "/XXX/lib/python3.7/site-packages/tensorboardX/summary.py", line 251, in image
tensor = convert_to_HWC(tensor, dataformats)
File "/XXX/lib/python3.7/site-packages/tensorboardX/utils.py", line 103, in convert_to_HWC
tensor shape: {}, input_format: {}".format(tensor.shape, input_format)
AssertionError: size of input tensor and input format are different. tensor shape: (375, 500, 3), input_format: NCHW
Command exited with non-zero status 1
10.72user 3.29system 0:39.37elapsed 35%CPU (0avgtext+0avgdata 3427892maxresident)k
3184968inputs+0outputs (398major+817987minor)pagefaults 0swaps

[Rongtao-Xu]

I get the same problem，did you solve it？

[yongyixiao]

I get the same problem，did you solve it

I also get the same problem，did you solve it

[luluning]

I think there's something wrong with the tensorboardX,but I still don't solve it.You can search tensorboard and see some similar problems,maybe they can help you.

[yongyixiao]

ok，thank you。But instead i use the res101 net，i get the new problem ：KeyError: 'resnet.conv1.weight'。

…

------------------ 原始邮件 ------------------ 发件人: "luluning"<[email protected]>; 发送时间: 2019年5月10日(星期五) 下午3:51 收件人: "ruotianluo/pytorch-faster-rcnn"<[email protected]>; 抄送: "Ｍvp_FerryMan"<[email protected]>; "Comment"<[email protected]>; 主题: Re: [ruotianluo/pytorch-faster-rcnn] AssertionError: size of inputtensor and input format are different. (#139) I think there's something wrong with the tensorboardX,but I still don't solve it.You can search tensorboard and see some similar problems,maybe they can help you. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.

[adityaarun1]

See #126

[Spark001]

Maybe you can use version like these:

torch==0.4.1
torchvision==0.2.2
tensorboard==1.11.0
tensorboardX==1.4