vision4robotics / tctrack Goto Github PK

Who have the .json file in the UAVDT dataset ? I have only .txt file in the UAVDT dataset !

请问在训练时，每个数据集的NUM_USE是取的100000嘛，以及总的VIDEOS_PER_EPOCH = 600000？是否与公开的配置文件相同

作者你好，训练时出现了这个报错：

File "/private/TX/anaconda3/envs/TCTrack/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 1229, in _process_data
    data.reraise()
  File "/private/TX/anaconda3/envs/TCTrack/lib/python3.8/site-packages/torch/_utils.py", line 434, in reraise
    raise exception
AttributeError: Caught AttributeError in DataLoader worker process 0.
Original Traceback (most recent call last):
  File "/private/TX/anaconda3/envs/TCTrack/lib/python3.8/site-packages/torch/utils/data/_utils/worker.py", line 287, in _worker_loop
    data = fetcher.fetch(index)
  File "/private/TX/anaconda3/envs/TCTrack/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py", line 49, in fetch
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "/private/TX/anaconda3/envs/TCTrack/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py", line 49, in <listcomp>
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "/private/TX/TCTrack-main/pysot/datasets/dataset.py", line 306, in __getitem__
    current_templatebox = self._get_bbox(current_templateimage, current[0][1])
  File "/private/TX/TCTrack-main/pysot/datasets/dataset.py", line 260, in _get_bbox
    imh, imw = image.shape[:2]
AttributeError: 'NoneType' object has no attribute 'shape'

经过debug后发现在pysot/datasets/dataset.py中

        current_templateimage=cv2.imread(current[0][0])
        current_searchimg=cv2.imread(current[1][0])

中current_templateimage和current_searchimg获取到的信息为None 但是我单独写了一个test 代入同样的image路径和以上的代码 发现是可以读取到数据的（非None） test代码如下：

import cv2
import numpy as np

current=(('./train_dataset/got10k/crop511/GOT-10k_Train/GOT-10k_Train_004614/000003.00.x.jpg', [864, 526, 911, 561]), ('./train_dataset/got10k/crop511/GOT-10k_Train/GOT-10k_Train_004614/000010.00.x.jpg', [855, 515, 910, 555]))
current_templateimage=cv2.imread(current[0][0])
current_searchimg=cv2.imread(current[1][0])
cv2.imshow("test", current_templateimage)
cv2.waitKey(0)

请问该如何解决呢

Where is the Online TAdaConv code?

Thank you for your open source work in the field of UAV Tracking. We want to generate the results of figure 6 in your paper, and can you provide the result files for other comparison algorithms? Very Grateful!

Could you provide the code for the preprocessing of the UAV123 dataset and other test dataset? I can't find the .json file in the raw UAV123 dataset, so maybe you have preprocessed it.

Thank you for your contributions!
Could you tell me:
def forward(self, xset): B,L, _,_,_ = xset.size()
1.What does the 'B' above denote?
2.Does the 'L' denotes the number of frames' features?(Is the 'L' above as same as the 'L' in the paper's Figure.3?)
&
`class TemporalAlexNet(nn.Module):
configs = [3, 96, 256, 384, 384, 256]

#input (B*L)*C*W*H, A1,A2,A3,A4,B1,B2,B3,B4...`

Could you tell me more details about the structure of the inputs and outputs?

When I want to test TCTrack, I encountered this error.
I test this on Ubuntu pytorch 1.6 CUDA 10.2

Traceback (most recent call last):
File "./tools/test.py", line 14, in
from pysot.core.config import cfg
ModuleNotFoundError: No module named 'pysot'

The following dataset OTB100 can be provided？

已解决

您好，在test.py读取UAV123_10fps数据集时，图像文件夹与标注文本文件并不一一对应，如person4对应person4_1.txt和person4_2.txt，导致无法找到person4.txt报错，请问应该如何处理呢？谢谢！

Thanks

您好，请问现在TCTrack有通过TensorRT算法加速的版本了吗

Hello, may I ask which module is extracting timing features and where in the code?

您好，您在github里发的TCTrack++的UAV123的那几个评估结果的文件里怎么是空的呢，可以再发一下么

What is the meaning of L in BLCWH?

thanks for you great work

I am confused why different codes were used in the training and prediction phases (AT-Trans)

Dear Ziang, first of all thanks for releasing the code for this promising pproach and making it publicly available. Is it possible to upload the pretrained model file to a more accessible platform, such as google drive or dropbox? That would be great, as it is not that straigthforward to download files from Baidu outside China, as far as I know. At least, my attempts were not successful so far. Thanks in advance!

请问作者以下问题怎么解决？
Traceback (most recent call last):
File "E:/liuxinhao/python_project/TCTrack-main/tools/train_tctrack.py", line 286, in
main()
File "E:/liuxinhao/python_project/TCTrack-main/tools/train_tctrack.py", line 281, in main
train(train_loader, dist_model, optimizer, lr_scheduler, tb_writer)
File "E:/liuxinhao/python_project/TCTrack-main/tools/train_tctrack.py", line 169, in train
outputs = model(data)
File "D:\anaconda\envs\siam_track\lib\site-packages\torch\nn\modules\module.py", line 550, in call
result = self.forward(*input, **kwargs)
File "D:\anaconda\envs\siam_track\lib\site-packages\torch\nn\parallel\data_parallel.py", line 155, in forward
outputs = self.parallel_apply(replicas, inputs, kwargs)
File "D:\anaconda\envs\siam_track\lib\site-packages\torch\nn\parallel\data_parallel.py", line 165, in parallel_apply
return parallel_apply(replicas, inputs, kwargs, self.device_ids[:len(replicas)])
File "D:\anaconda\envs\siam_track\lib\site-packages\torch\nn\parallel\parallel_apply.py", line 85, in parallel_apply
output.reraise()
File "D:\anaconda\envs\siam_track\lib\site-packages\torch_utils.py", line 395, in reraise
raise self.exc_type(msg)
RuntimeError: Caught RuntimeError in replica 0 on device 0.
Original Traceback (most recent call last):
File "D:\anaconda\envs\siam_track\lib\site-packages\torch\nn\parallel\parallel_apply.py", line 60, in _worker
output = module(*input, **kwargs)
File "D:\anaconda\envs\siam_track\lib\site-packages\torch\nn\modules\module.py", line 550, in call
result = self.forward(*input, **kwargs)
File "E:\liuxinhao\python_project\TCTrack-main\pysot\models\utile_tctrack\model_builder.py", line 135, in forward
xf.size(-1))z
**RuntimeError: shape '[70, 4, 256, 26, 26]' is invalid for input of size 72683520

how to generate coco training datasets?

Is it possible to use nn.Conv2d to perform the feature after Online TAdaConv?

在两个Nvidia 1080TiGPU上训练，需要10天左右正常吗，为什么需要训练这么久呢

Traceback (most recent call last):
File "tctrack_original/tools/train_tctrackpp.py", line 303, in
main()
File "tctrack_original/tools/train_tctrackpp.py", line 298, in main
train(train_loader, dist_model, optimizer, lr_scheduler, tb_writer)
File "tctrack_original/tools/train_tctrackpp.py", line 187, in train
outputs = model(data,videorange)
File "tctrack/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "tctrack/lib/python3.9/site-packages/torch/nn/parallel/data_parallel.py", line 168, in forward
outputs = self.parallel_apply(replicas, inputs, kwargs)
File "tctrack/lib/python3.9/site-packages/torch/nn/parallel/data_parallel.py", line 178, in parallel_apply
return parallel_apply(replicas, inputs, kwargs, self.device_ids[:len(replicas)])
File "tctrack/lib/python3.9/site-packages/torch/nn/parallel/parallel_apply.py", line 86, in parallel_apply
output.reraise()
File "tctrack/lib/python3.9/site-packages/torch/_utils.py", line 461, in reraise
raise exception
RuntimeError: Caught RuntimeError in replica 0 on device 0.
Original Traceback (most recent call last):
File "tctrack/lib/python3.9/site-packages/torch/nn/parallel/parallel_apply.py", line 61, in _worker
output = module(*input, **kwargs)
File "tctrack/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "tctrack_original/pysot/models/utile_tctrackplus/model_builder.py", line 142, in forward
loc,cls2,cls3=self.grader(xf[:,-1,:,:,:],zf,xf[:,:-1,:,:,:].permute(1,0,2,3,4))
File "tctrack/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(input, **kwargs)
File "tctrack_original/pysot/models/utile_tctrackplus/utile.py", line 99, in forward
ppres=self.conv1(self.xcorr_depthwise(px[0],z))
File "tctrack_original/pysot/models/utile_tctrackplus/utile.py", line 88, in xcorr_depthwise
x = x.reshape(1, batchchannel, x.size(2), x.size(3))
RuntimeError: shape '[1, 8960, 26, 26]' is invalid for input of size 12113920

Thank you for your open source work in the field of UAV Tracking! Can you provide a pre-trained configuration file on Imagenet?

Traceback (most recent call last):
File "train_tctrack.py", line 288, in
main()
File "train_tctrack.py", line 283, in main
train(train_loader, dist_model, optimizer, lr_scheduler, tb_writer)
File "train_tctrack.py", line 132, in train
for idx, data in enumerate(train_loader):
File "/usr/local/lib/python3.6/dist-packages/torch/utils/data/dataloader.py", line 363, in next
data = self._next_data()
File "/usr/local/lib/python3.6/dist-packages/torch/utils/data/dataloader.py", line 989, in _next_data
return self._process_data(data)
File "/usr/local/lib/python3.6/dist-packages/torch/utils/data/dataloader.py", line 1014, in _process_data
data.reraise()
File "/usr/local/lib/python3.6/dist-packages/torch/_utils.py", line 395, in reraise
raise self.exc_type(msg)
AttributeError: Caught AttributeError in DataLoader worker process 0.
Original Traceback (most recent call last):
File "/usr/local/lib/python3.6/dist-packages/torch/utils/data/_utils/worker.py", line 185, in _worker_loop
data = fetcher.fetch(index)
File "/usr/local/lib/python3.6/dist-packages/torch/utils/data/_utils/fetch.py", line 44, in fetch
data = [self.dataset[idx] for idx in possibly_batched_index]
File "/usr/local/lib/python3.6/dist-packages/torch/utils/data/_utils/fetch.py", line 44, in
data = [self.dataset[idx] for idx in possibly_batched_index]
File "/mnt/TCTrack-main/pysot/datasets/dataset.py", line 323, in getitem
current_templatebox = self._get_bbox(current_templateimage, current[0][1])
File "/mnt/TCTrack-main/pysot/datasets/dataset.py", line 277, in _get_bbox
imh, imw = image.shape[:2]
AttributeError: 'NoneType' object has no attribute 'shape'

您好，关注你们很多的工作了，从siamapn开始到++再到TCTrack，请问下方便分享下，您在论文中展示的其他的算法的结果嘛，就是除了SiamAPN++等你们开源的能找到，很多你们论文中提到的都找不到，我也是做无人机跟踪这块的最近写论文，苦于UAV数据集的对比太少，你的论文中非常详细，如果可以的话，非常感谢。

你好，感谢你的工作！
调试eval.py的时候出现了如图的bug

我检查了Tiger1的results和对应的真值的txt文件，发现results中是有349行结果的，但是评估的时候只有344行？我应该排查哪里的代码呢？
非常感谢您的解答。

Hi,

Congrats on achieving this amazing result. I was so impressed by the speed of this tracker. However, I'm facing a serious accuracy loss when the object becomes occluded (fully, not partial). As can be seen in this video, when the bus goes under the bridge, TCTrack cannot track it correctly anymore. I'm getting the same result with some other aerial videos, while Stark can still handle all of those.

I'm thinking of improving this by trading a bit of performance for accuracy and extending the number of frames that the model saves. Doing so could help save more temporal information and potentially catch the correct object. Is it possible to extend the number L? If so, can you provide some hints on where should I modify?

Thank you.

在下载VID的时候，代码中提供的“wget http://bvisionweb1.cs.unc.edu/ilsvrc2015/ILSVRC2015_VID.tar.gz”访问不了，哪位大佬可以提供一下VID，got10k，lasot的数据集吗

您好，我发现提供的模型文件在测试是存在无法载入的问题，想请教下您这是因为模型文件和代码不匹配的问题造成的吗？

What is the torch.size() of xet in def generateweight(self, xet)?

您好，看了你们的论文，从里面受益很多，我看到SiamAPN++、SiamAPN、SiamAPN也是你们团队的论文，可以分享一下在DTB70和UAV123数据集上其他算法的跟踪结果么，我将非常感谢

1. What does 'DATASET.VIDEOS_PER_EPOCH = 600000' in config.py mean? the number of videos or the number picture pairs?If the dataset is small, can I reduce this param to reduce the time of training?

您好，我看了您的论文很受启发，请问您那有DTB70、UAV12310fps的json文件和UAVTrack112L的数据集以及json文件么，可不可以分享一下，十分感谢，我没有找到，期待您的回复

Hello,

Thanks for sharing your great work! I see that the dataset preparation scripts write an "x" and a "z" image to disk, but here it looks like the dataset uses the "x" image for everything, rather than making use of the "z" image for template, which I believe may have been intended. Is that correct? If not, what is the "z" crop used for?

Thank you!

作者您好，发现您论文中提到学习率是从0.005下降至0.0005，而在代码config文件中显示学习率设置为0.005下降至0.00005，请问实际是以哪个为准？望回复

GOT10K par_crop.py中的sub_sets里面是‘GOT-10k_Train_split_01’意味着什么呢？我下载下来是GOT-10k_Train_000001-009335

您好，请问下，提取时序特征的是哪个模块，以及在代码中的哪里？

Hi, have you used TensorRT acceleration when deploying projects with jetson at frame rates over 27FPS?

您好，我想请教一下我按照您论文里的描述将UAV123数据集下载至tools目录下后运行test.py仍报错找不到json文件，请问如何解决呢或者您是否有相应的下载链接，不胜感激。

Can you share the source or how to make pretrained wights of backbone?

hi,professor:
like the topic said
please help