kkahatapitiya / coarse-fine-networks Goto Github PK

how to train train_coarse_fineFEAT.py？

Thank you very much for your work.
I want to inference the wild videos. Can you help me with how to inference my own videos? or test the pre-trained models? I only found the training codes.
Thanks

how to test a custom video？

Could you please mention the hardware setup (number of GPUs and type of GPUs) you used to train/test the models and the time taken to train the models (per epoch)?

Thanks in advance!

Hello,
I recently read your paper and it states that for "Charades dataset, we configure our network to use T = 64,
T= 128 and α = 1/4."
Could you point me to where in the code the number of frames are set to 128?

I've only seen the frames parameter being set to 80 in:

• train_fine.py on line 57 (80x4)
• extract_fineFEAT.py on line 61
• train_coarse_fineFEAT.py line 60 (80x4)

Would you mind clarifying why the value is 80?
I'm trying to train my own dataset and would like to make sure I adjust the number of frames at the correct points in the code. I would greatly appreciate it if you could point me in the right direction :)

How to prepare training data?
@kkahatapitiya

Hi. It's me again.
In your paper and code, mAP was used for performance comparison.

1. Is it frame-mAP? or video-mAP?
2. Do you have any plans to add another one?

Thanks.

Hi @kkahatapitiya ,

Thank you in advance for your excellent work!
As you mentioned in your README.md: "use this evaluation script to reproduce our reported numbers."

I want to know where I can get your results (such as this "test_submission_localize.txt", which is required by "Charades_v1_localize.m"), because I want to reproduce the numbers you reported.

I'm looking forward to your reply!

Thanks

Hi, again. :)
I'm still in python train_fine.py -gpu {gpu_num}

dataset size:0
Traceback (most recent call last):
File "train_fine.py", line 269, in
run()
File "train_fine.py", line 86, in run
dataloader = torch.utils.data.DataLoader(dataset, batch_size=batch_size, shuffle=True,
File "/opt/conda/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 262, in __init__
sampler = RandomSampler(dataset, generator=generator) # type: ignore
File "/opt/conda/lib/python3.8/site-packages/torch/utils/data/sampler.py", line 103, in __init__
raise ValueError("num_samples should be a positive integer "
ValueError: num_samples should be a positive integer value, but got num_samples=0

I downloaded Charades dataset(Charades_v1_rgb.tar 76GB) and unzipped it. However, dataset size is still zero.

What can I do?

+) Plus, extract_fineFEAT.py also has import videotransforms.

Hi
I'm just trying to run this code.
However, I faced to some module error from the first train code,
python train_fine.py -gpu {gpu_num}

File "train_fine.py", line 22, in <module>
   import videotransforms
ModuleNotFoundError: No module named 'videotransforms'

File "/workspace/Coarse-Fine-Networks/x3d_fine.py", line 7, in <module>
    import loss_utils
ModuleNotFoundError: No module named 'loss_utils'

Would you explain how to solve the error?

Dear author:
thanks for sharing the insightful work! it helps!
I wonder which tool you used to draw these diagram in your paper:

it looks very vivid. congratulations!