search

phil-bergmann / tracking_wo_bnw

Implementation of "Tracking without bells and whistles” and the multi-object tracking "Tracktor"
https://arxiv.org/abs/1903.05625
GNU General Public License v3.0
818 stars 196 forks source link

RuntimeError when run test_tracktor.py #143

Open sfwang20 opened 3 years ago

sfwang20 commented 3 years ago

Hi Philipp, Thanks for your excellent work! I followed README.md to build the environment and when I run test_tracktor.py, it shows the RuntimeError as following:

INFO - main - Initializing object detector(s). INFO - main - Initializing reID network(s). INFO - main - Tracking: MOT17-02-FRCNN 0%| | 0/600 [00:02<?, ?it/s] ERROR - test_tracktor - Failed after 0:00:27! Traceback (most recent calls WITHOUT Sacred internals): File "experiments/scripts/test_tracktor.py", line 143, in main tracker.step(frame_data) File "d:\research\projects\tracking_wo_bnw\src\tracktor\tracker.py", line 261, in step self.obj_detect.load_image(blob['img']) File "d:\research\projects\tracking_wo_bnw\src\tracktor\frcnn_fpn.py", line 75, in load_image self.features = self.backbone(preprocessed_images.tensors) File "C:\Users\shinfu\anaconda3\envs\tracking_wo_bnw\lib\site-packages\torch\nn\modules\module.py", line 722, in _call_impl result = self.forward(*input, kwargs) File "C:\Users\shinfu\anaconda3\envs\tracking_wo_bnw\lib\site-packages\torchvision\models\detection\backbone_utils.py", line 39, in forward x = self.body(x) File "C:\Users\shinfu\anaconda3\envs\tracking_wo_bnw\lib\site-packages\torch\nn\modules\module.py", line 722, in _call_impl result = self.forward(*input, *kwargs) File "C:\Users\shinfu\anaconda3\envs\tracking_wo_bnw\lib\site-packages\torchvision\models_utils.py", line 63, in forward x = module(x) File "C:\Users\shinfu\anaconda3\envs\tracking_wo_bnw\lib\site-packages\torch\nn\modules\module.py", line 722, in _call_impl result = self.forward(input, kwargs) File "C:\Users\shinfu\anaconda3\envs\tracking_wo_bnw\lib\site-packages\torch\nn\modules\conv.py", line 419, in forward return self._conv_forward(input, self.weight) File "C:\Users\shinfu\anaconda3\envs\tracking_wo_bnw\lib\site-packages\torch\nn\modules\conv.py", line 416, in _conv_forward self.padding, self.dilation, self.groups) RuntimeError: cuDNN error: CUDNN_STATUS_ALLOC_FAILED

I search by google if there are solutions about the error and it seems like the memory of my GPU is not enough. My GPU is Nvidia GTX 1660Ti with 6 GB memory. Besides, I also tried change the batch size (to 32) in res50.yml but it didn't work (it seems like testing won't use this file, I'm not sure about that) Is there any other way for me to get smaller GPU memory usage? Thanks for your help! Best regards.

timmeinhardt commented 3 years ago

Yes, the file you changed is just for the training of the reid model. You could run the tracking evaluation on the CPU or reduce the resolution of the images. The latter might hurt the performance though.