sauradip / fewshotQAT

[BMVC 2021]: Official PyTorch implementation of : "Few Shot Temporal Action Localization using Query Adaptive Transformers"

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

GTAD checkpoint

lucky-23 opened this issue · comments

commented

Hello, I am unable to run align.py due to incompatibility issue with my CUDA/CUDNN version and I have been stuck with this problem for a long time. Can you please share the pre-trained weights of GTAD model on ActivityNet dataset?
Thank you very much!

commented

And also in the paper you compared with BMN(Boundary matching) as an embedding module. I would be grateful if you could kindly let me know from which layer did you extract the features from the pre-trained BMN model?
Thank you!.

Hello, I am unable to run align.py due to incompatibility issue with my CUDA/CUDNN version and I have been stuck with this problem for a long time. Can you please share the pre-trained weights of GTAD model on ActivityNet dataset? Thank you very much!

Here is the checkpoint for your convenience

And also in the paper you compared with BMN(Boundary matching) as an embedding module. I would be grateful if you could kindly let me know from which layer did you extract the features from the pre-trained BMN model? Thank you!.

We took the output of this 3d conv block https://github.com/JJBOY/BMN-Boundary-Matching-Network/blob/a92c1d79c19d88b1d57b5abfae5a0be33f3002eb/models.py#L71

commented

Thank you very much!