semchan
HI, I use DBG to produce the proposals from the feature's data which you provided, and then use the proposals and the TSN features directly to do the temporal detection,...
> Hi, could you please provide more detail? I directly use the bounded T&T organized by [NSVF](https://github.com/facebookresearch/NSVF) and use the unbounded T&T organized by [nerf++](https://github.com/Kai-46/nerfplusplus). Thanks very much for your...
It is not support the dataset of Unbounded inward-facing: [T&T](https://drive.google.com/file/d/11KRfN91W1AxAW6lOFs4EeYDbeoQZCi87/view?usp=sharing), not the T&T (masked)](https://dl.fbaipublicfiles.com/nsvf/dataset/TanksAndTemple.zip)
> hi，did you find the way to achieve this? thanks a lot.
I also meet the same issue. could you please help. thanks
hi, since the speed is too slow, is that can be modified the implementation from " self._depthwise_conv = Conv2dSamePadding( in_channels=oup, out_channels=oup, groups=oup, # groups makes it depthwise kernel_size=k, stride=s, bias=False)"...
防止出现负数
嗨，非常感谢你的贡献．我看到你代码里做了一些调整，比如不能使用batch进行训练，是否考虑今后添加这个功能？另外，你测试过训练后的精度吗？与官方相比的话，有什么差距吗？谢谢您
> @semchan Hi 您好，目前没有添加batch的计划，以后有时间会试一下的。我最多只跑过3个epoch，精度比官方数据低，如果跑的时间长一点可能会得到更好的结果 非常感谢您的快速回复．我在自己的数据集上进行训练并测试到epoch=11，发现比起官方发布的预训练模型的要差,正在查找原因
hi, would you please guide me? thanks a lot