Shangqian Gao comments

Results 19 comments of


                                            Shangqian Gao

choose Only Testing , but still train data

Didn't notice this new issue, what are the results you got? like screen shot, etc.

choose Only Testing , but still train data

Yeah, @jmessou is right. you can refer to line 87 in spatial_cnn.py ` if os.path.isfile(self.resume) ` This line of code will verify whether provided path is a file or not...

choose Only Testing , but still train data

The code from this repository reproduce the training procedure from temporal segment network(TSN). In TSN, during training, they sample three times from a video, during testing, they sample 20 times...

Improvement on motion-cnn result: 84.1% on split-1, with VGG-16

You should use this data augmentation during training to get desired results. zhujian 于2018年9月13日周四03:15写道： > I have some problem about the accuracy , when i only use the center...

Improvement on motion-cnn result: 84.1% on split-1, with VGG-16

You can refer to the related paper, during training there are extensive data augmentation used such as multi-scale, corner crops, etc. The author of this project only used very simple...

Improvement on motion-cnn result: 84.1% on split-1, with VGG-16

That's weird, you can try these two models I converted from the project of their paper [https://github.com/feichtenhofer/twostreamfusion](url), the link for the models https://drive.google.com/file/d/1JydxdPMEHU7uJnRyi8A8uF82jSgE9FGe/view?usp=sharing. They are VGG-16 models.

Improvement on motion-cnn result: 84.1% on split-1, with VGG-16

What is the results you got? You'd better open an new issue to discuss about this.

Improvement on motion-cnn result: 84.1% on split-1, with VGG-16

I think you need to use lambda expression, for example: ``` transforms.Compose([ transforms.Resize(256), transforms.FiveCrop([224, 224]), transforms.Lambda(lambda crops: torch.stack([transforms.Normalize( mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])(crop) for crop in crops])) ]) ```

Improvement on motion-cnn result: 84.1% on split-1, with VGG-16

I used trained model from other projects, I provided converted pytorch model link in previous comments in this issue. When test with 5-crops/center-crop, I can achieve around 82%/78% accuracy with...

Improvement on motion-cnn result: 84.1% on split-1, with VGG-16

Right now, I may not have time for sharing my code. But after CVPR deadline, I will refine the code concerning this project and make it public available. Regarding two-stream...