Open heboyong opened 5 years ago
Hi, I think the reason will be either the data loader too slow or your BATCH_SIZE too high. You can use more threads to load the data, or lower the BATCH_SIZE. Also, I noticed that you might set the TIMEOUT_FOR_WAIT_QUEUE=30. You can also enlarge the timeout to see if the data loader can provide the video on time.
yeah,I have tried to run this model on another computer and it worked well ,so I think the reason of this error is small memory of GPU or CPU.Any way,thanks for your reply andyour job
Hi @heboyong , I think it is the memory of the CPU. It consume nearly 25GB/32GB main memory in my computer. Your post remind me that I should consider the hardware capability of the ordinary devices while I set the default settings. Thanks!
@JoshuaPiinRueyPan First of all, thank you so much for putting in so much effort. Secondly, can you please tell me how I can change these settings? I have 16GB RAM and it uses almost 8-9 GB. I faced same error.
Also, this project is not using my GPU. I have tried most of the solutions available online.
Lastly, do you have checkpoints trained on multiple datasets? For example on BEHAVE, Crowd Violence Detection & Real Life Violence Detection (Kaggle).
how can i solve this error:
Start Training...
Training terminate at epoch: 0 , steps: 0 Queue info = listOfData.len() = 61; WaitingQueue.len() = 40; LoadedQueue.len() = 0; Pause = False
Traceback (most recent call last): File "D:\Desktop\ViolenceDetection\src\data\DataManager.py", line 271, in AssignBatchData timeout=dataSettings.TIMEOUT_FOR_WAIT_QUEUE) File "D:\python\lib\queue.py", line 172, in get raise Empty queue.Empty
During handling of the above exception, another exception occurred:
Traceback (most recent call last): File "D:/Desktop/ViolenceDetection/Train.py", line 197, in
main.Run()
File "D:/Desktop/ViolenceDetection/Train.py", line 64, in Run
self.trainer.PrepareNewBatchData()
File "D:\Desktop\ViolenceDetection\src\Trainer.py", line 77, in PrepareNewBatchData
raise error
File "D:\Desktop\ViolenceDetection\src\Trainer.py", line 71, in PrepareNewBatchData
self._dataManager.AssignBatchData(self._batchData)
File "D:\Desktop\ViolenceDetection\src\data\DataManager.py", line 279, in AssignBatchData
raise TimeoutError(errorMessage)
TimeoutError: In TrainDataManager: Unable to get batch data in duration: 30(s)
TrainQueue info:
listOfData.len() = 61; WaitingQueue.len() = 40; LoadedQueue.len() = 0; Pause = False