search

zhang-can / ECO-pytorch

PyTorch implementation for "ECO: Efficient Convolutional Network for Online Video Understanding", ECCV 2018
BSD 2-Clause "Simplified" License
188 stars 93 forks source link

Trained ECO models on Kinetics dataset #16

Open Tord-Zhang opened 6 years ago

Tord-Zhang commented 6 years ago

As reported in the paper of ECO, pretrained 2D BNInception and 3D resnet-18 models on Kinetics dataset are not enough to get a good result, training the ECO model for another 10 epochs on Kinectics would promise a better result. However, I am really not able to train the model on Kinectics(GPU and memory limitation). Since you mentioned that you are training the models on Kinetics, would you please share the trained model? I am going to use the trained weights to initialize the model and train it on something something dataset. I can report the testing result and share it in this repository.

Tord-Zhang commented 6 years ago

and I am wondering if all your modification available on this repository?https://github.com/mzolfaghari/ECO-efficient-video-understanding

zhang-can commented 6 years ago

Hi @mangdian ,

The pretrained model on Kinetics can be automatically downloaded, see as:

https://github.com/zhang-can/ECO-pytorch/blob/master/tf_model_zoo/ECO/pytorch_load.py#L11

jangho2001us commented 5 years ago

@zhang-can Hi Can,

Your link does not work (https://s3.us-east-2.amazonaws.com/zhangcan/kin_ECO_epoch-12-c89e9dc0.pth.tar)

Is there any pretrained model with kinetics?

chen-ming2019 commented 5 years ago

@mangdian can you share the something-something dataset with me? My downloaded something dataset from it's official website couldn't use.