Bianjiang Yang

What do you mean by replacing SFT layer by direct concatenation

1

In table 1, you proposed the ablation study. Could you give the sample code for direct concatenation in the intermediate layers? I am not sure how to concatenate the kernel_code...

Hi, in your paper, you said "It should be noted that we only learn the color/tone enhancement part of the imaging pipeline rather than learning the entire pipeline from raw...

Have you reproduced the results shown in the paper?

8

Hi, have you reproduced the results on the FiveK dataset?

How to adjust the resolution of the saved dot image?

2

How to adjust the resolution of the saved dot image? My computational graph is too long...then the image seems not clear enough. 

Hi, What img size (`928x1600` or smaller), bev_embed size (seems `100x100` and upsampled to `200x200`), and other different settings you have for small UniAD? Can you share more information about...

https://github.com/DerryHub/BEVFormer_tensorrt/blob/303d3140c14016047c07f9db73312af364f0dd7c/det2trt/models/modules/encoder.py#L256C10-L258C84 ``` bev_mask = (1 - (1 - bev_mask).prod(0)).view(6, -1, 1) bev_mask = bev_mask / torch.clamp(bev_mask.sum(0, keepdims=True), min=1e-4) ``` What's the purpose of this, it makes `bev_mask` different from original...