Problem with train_loader

[siyalsonarkar18]

When I'm running the main.py file, I'm getting this error: File "main.py", line 88, in pin_memory=True) File "/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 213, in init sampler = RandomSampler(dataset) File "/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/torch/utils/data/sampler.py", line 94, in init "value, but got num_samples={}".format(self.num_samples)) ValueError: num_samples should be a positive integer value, but got num_samples=0

How is the num_samples = 0?

[lrhjw]

have you solve this problem? i meet same error please tell me

[sudonto]

have you solve this problem? i meet same error please tell me

check your path of jpg files

[lrhjw]

@sudonto I can create the json file ,so i think my path of jpg is true

[lrhjw]

my file like this: root_path/data/ jpg/ action1/ action2/ ucf101_01.json
and my code to trian is: python main.py --root_path /home/zc/coode/3D-ResNets-PyTorch --video_path data/jpg --annotation_path data/ucf101_01.json --result_path results --dataset ucf101 --model resnet --model_depth 34 --n_classes 4 --batch_size 128 --n_threads 4 --checkpoint 5 @sudonto

[sudonto]

I think you should pay more attention here: https://github.com/kenshohara/3D-ResNets-PyTorch/blob/05a8a5e7f4feff0c6e0c0f1a498f191433c8e37c/datasets/loader.py#L32 and https://github.com/kenshohara/3D-ResNets-PyTorch/blob/05a8a5e7f4feff0c6e0c0f1a498f191433c8e37c/datasets/videodataset.py#L38

Those files are resposinsble to load images, if it is fail then ValueError will raise. Note, I assume you have correctly created json file.

[lrhjw]

@sudonto thank you i will try

[Chaoran-F]

@lrhjw did you solve it? i got same problem

[guilhermesurek]

Is there any possibility of having data described in your annotation json that doesn't exist in your disk? or videos without fps information in the annotations? like empty or zero.