mzolfaghari / ECO-efficient-video-understanding

Code and models of paper " ECO: Efficient Convolutional Network for Online Video Understanding", ECCV 2018
MIT License
437 stars 96 forks source link

Training ECO PyTorch on the Moments in Time dataset #21

Closed sophia-wright-blue closed 6 years ago

sophia-wright-blue commented 6 years ago

Hello, I intend to use the ECO PyTorch code to train the model on the moments in time dataset. I have the frames extracted and the training and validation lists prepared.

I'd greatly appreciate any guidance on the changes to be made to the code (main.py) to train the model on the moments dataset (or any dataset other than kinetics).

Thank you,

mzolfaghari commented 6 years ago

You need to add the number of classes and image name format between lines 40-50 in main.py. In addition, add your dataset name (which you used in main.py) to opts.py line 3.