You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Traceback (most recent call last):
File "/home/project/Hagrid/run.py", line 88, in <module>
run(args)
File "/home/project/Hagrid/run.py", line 54, in run
train_dataloader, val_dataloader, test_dataloader, model = load_train_objects(config, args.command, args.n_gpu)
File "/home/project/Hagrid/custom_utils/train_utils.py", line 89, in load_train_objects
train_dataset[0]
File "/home/project/Hagrid/dataset/dataset.py", line 175, in __getitem__
for i in range(len(row["united_bbox"])):
TypeError: object of type 'NoneType' has no len()
It seems to be a bug in both branch master and hagridv2-1m.
Because the united_bbox row in .json file usually is None.
The code works well with the following adjustments. But I'm not sure if this will train correctly, because I'm not sure what the united_bbox does
The text was updated successfully, but these errors were encountered:
I have reviewed the training process for SSDLiteMobileNetV3Large and did not encounter the errors you mentioned. Could you please verify that you have downloaded the correct version of the annotation files (from the master branch)?
I have reviewed the training process for SSDLiteMobileNetV3Large and did not encounter the errors you mentioned. Could you please verify that you have downloaded the correct version of the annotation files (from the master branch)?
Hello! Thank you for your replay.
Yes, I'm sure the annotation file is downloaded from the master branch. And I tried it agin, it does have some gestures with NULL united_bbox.
Finally, I'm guessing this could be a problem caused by a different python version. My conda environment with Python 3.8, which is lower than master branch. It is possible that higher versions of Python will make additional judgments about len(null)
When I train the code with following command, the program reports an error.
I found that the error occurs in the dataset.py
It seems to be a bug in both branch master and hagridv2-1m.
Because the
united_bbox
row in .json file usually is None.The code works well with the following adjustments. But I'm not sure if this will train correctly, because I'm not sure what the
united_bbox
doesThe text was updated successfully, but these errors were encountered: