Demo is not working, help required

[ghost]

Hi there,
First of all I would like to congratulate you for this great work and thank you for saharing the this awsome content publicly.
I am trying to run the demo.py after following all the steps described in the demo paragraph but unfortunately I am encountering problems and no visual results are generated in results folder except an empty results.txt file. Following is the terminal output that I am getting. The "Datalist len" is 0. I don't know why it is so. I have provided a total of 778 frames but I am not getting any output. Is there a minimum nunber of frames to be provided for detection?
Please help, Thanks.

python3 demo.py
Warning: If you want to use fp16, please apex with cuda support (https://github.com/NVIDIA/apex) and update pytorch to 1.0
Warning: If you want to use fp16, please apex with cuda support (https://github.com/NVIDIA/apex) and update pytorch to 1.0
Warning: If you want to use fp16, please apex with cuda support (https://github.com/NVIDIA/apex) and update pytorch to 1.0
Loading pretrain model from pretrained/ava_step.pth
Building I3D model...
Building I3D head for global branch...
Building I3D head for global branch...
Building I3D head for global branch...
Building I3D head for context branch...
Datalist len: 0

[xyang35]

Thanks for your interest in our work! It seems that the dataloader cannot find the frames you provided. Could you please double check the data folder you've created? Make sure to follow the structure mentioned in the Demo section. It may be also helpful to take a quick look at the code snippet for processing the input frames.

[ghost]

Hi there, thanks for reply. Unfortunately I am still not able to run the demo. I am getting following error when i try to run customize.py
Traceback (most recent call last): File "customize.py", line 14, in <module> from utils.tube_utils import scale_tubes, scale_tubes_abs ModuleNotFoundError: No module named 'utils.tube_utils'

Althoug utils package is installed but still Iam getting this error.
Also Could please eloberate further on the placement of the input frames and video. At the moment I have placed video frames at two places.

|datasets/
| |-- frames/
| |-- <video_id1>/
| |-- frame0000.jpg
| |-- frame0001.jpg
| |-- ...

and

|--data/
| |--data_root/
| |-- frames/
| |-- <video_id1>/
| |-- frame0000.jpg
| |-- frame0001.jpg
| |...

Still getting same results:
Warning: If you want to use fp16, please apex with cuda support (https://github.com/NVIDIA/apex) and update pytorch to 1.0
Warning: If you want to use fp16, please apex with cuda support (https://github.com/NVIDIA/apex) and update pytorch to 1.0
Warning: If you want to use fp16, please apex with cuda support (https://github.com/NVIDIA/apex) and update pytorch to 1.0
Loading pretrain model from pretrained/ava_step.pth
Building I3D model...
Building I3D head for global branch...
Building I3D head for global branch...
Building I3D head for global branch...
Building I3D head for context branch...
Datalist len: 0

Please help me solve this use thanks :)

[ronghui19]

args.data_root = 'datasets/demo/frames/' not args.data_root = '/datasets/demo/frames/'

[AhmadManzoor]

@IronHide99 please try to give absolute path

[ghost]

@AhmadManzoor can we get in touch?

[AhmadManzoor]

@IronHide99 sure , how may i help you. [email protected]

[EllisHui]

hello ,Have you solve your problems? I met the same Questions

[ronghui19]

CHANGE args.data_root = '/datasets/demo/frames/' TO args.data_root = 'datasets/demo/frames/'

[ghost]

hello ,Have you solve your problems? I met the same Questions

After multiple attempts I haven't been able to run this code. I don't know what is the problem here even though I have followed the demo tutorial down to detail.

[EllisHui]

| | 17600591193 | | ***@***.*** | 签名由网易邮箱大师定制 On 3/30/2021 ***@***.***> wrote： CHANGE args.data_root = '/datasets/demo/frames/' TO args.data_root = 'datasets/demo/frames/' — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.