Jia-Chang Feng

About the mean and std of I3D (inception-v1)?

1

At test_configs/I3D_RGB/i3d_kinetics400_3d_rgb_inception_v1_seg1_f64s1.py, I find out the mean and std is ` img_norm_cfg = dict( mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)` But in the source repo [kinetics-i3d](https://github.com/deepmind/kinetics-i3d), whose ReadMe.md tells...

The padding of I3D model should be symmetrical

1

The Module of MaxPool3dTFPadding with kernel_size=(1,3,3), stride(1,2,2) can lead to asymmetrical padding. It would influence the output feature map, as the bottom right would be usually higher than other part...

Pretrained models on Baidu Cloud Were Deleted

1

The pretrained models on baidu cloud have been deleted by the website. Could you re-upload the models? ( If you use validation code, The models may survived. )

Bug in Sub-graph loss

1

https://github.com/frostinassiky/gtad/blob/6deb5b1bc6883b48bd22e0cc593069643c953e3d/gtad_lib/loss_function.py#L23 ``` def subgraph_loss_func(pred_bm, gt_iou_map, bm_mask, lambda1=10): pred_bm_wcs = pred_bm[:, 0].contiguous() pred_bm_mse = pred_bm[:, 1].contiguous() -> wce_loss = pem_reg_loss_func(pred_bm_wcs, gt_iou_map, bm_mask) -> mse_loss = pem_cls_loss_func(pred_bm_mse, gt_iou_map, bm_mask) subgraph_loss = wce_loss...

Can you offer some samples about manual attribute?

9

I am wondering how you annote mannual attribute on the video, can you give me some hints?