Qifeng Chen

i got a wrong result, i'm confused. there are some details: dataset: face ffhq 512x512 (The public dataset you mentioned in readme.md) config: default file in project sr_sr3_64_512.json gpu: 2x32G...

> Hello, have you solved this problem yet? I have also encountered this situation Do you have any experience about these worse results? Could you share any suggestion? Thanks in...

> The code for the point cloud initialization and Chamfer loss is not released in this repo, as this trick is not really crucial as shown in figure 3, and...

> 透视投影的高斯不是标准高斯，因此不会用标准高斯的mean和 variance 计算 bbox。详细细节你可以参考 diff-surfel-rasterization 上的相关 [issue8](https://github.com/hbb1/diff-surfel-rasterization/issues/8) 和 [issue15](https://github.com/hbb1/diff-surfel-rasterization/issues/15) 作者你好，参考你在issue15内提供的“如何计算aabb”的文档，我理解了你的计算过程，但我不是很明白为什么aabb的半径在数值上就是|x1-x2|/2，还是我理解有误？ ``` // Compute AABB // ref issue15. a=sum(f*T3*T3) b=-2sum(f*T0*T3) c=sum(f*T0*T0) f=(R**2,R**2,-1) for pix R=3sigma=3 float3 temp_point = {cutoff *...

> `def load_pretrained_weights(self, path): w_dict = torch.load(path + "/backbone", map_location=lambda storage, loc: storage) self.backbone.load_state_dict(w_dict, strict=True) w_dict = torch.load(path + "/segmentation_decoder", map_location=lambda storage, loc: storage) self.decoder.load_state_dict(w_dict, strict=True)` > > thanks for...