BSRT

BSRT copied to clipboard

验证集和测试集一样？

1

请问作者，从代码看，你们训练用的验证集和测试用的测试集是同一个，这样测试集就是已经见过的数据，其测试结果如何能证明模型真实的泛化能力呢？

githublei-min

只在Real data上训练

1

作者您好，请问您有没有试过只在Real data上训练呢？结果怎么样？是否会出现过拟合的现象？

greatlog

Improving Model for Text Super-Resolution

1

Dear author, first and foremost thank you very much for your great work and your great effort towards improving burst super-resolution. I am currently myself training and testing your model...

raphischwarz

About "FlowGuidedDCN: Offset mean is 276.96575927734375, larger than 100."

1

Has anyone encountered this issue before?

fanjunkai1

Question about magnitude in FLOWDCN

5

Dear author, I see you set the value of offset in 10*tanh(offset), what the difference between original offset and [-10,10] offset? It seems that the magnitude is a super parameter,...

K1NSA

Get flow between frames ref and other

5

flows_list = [flow.view(b, n, 2, h // (2 ** (i)), w // (2 ** (i))) for flow, i in zip(flows, range(3))] RuntimeError: shape '[4, 7, 2, 64, 64]' is invalid...

qtct

输入raw图的个数是恒定的？

1

输入必须是 batch,14,4, Height, Width,否则会报错： `RuntimeError: Given groups=1, weight of size [60, 896, 1, 1], expected input[1, 512, 160, 160] to have 896 channels, but got 512 channels instead` 请问怎么修改，比如我要输入8张raw序列，谢谢

anguoyang

function mosaic with mode == 'grbg'

1

code/synthetic/bsrt/data_processing/camera_pipeline.py line140： blue = image[:, 2, 0::2, 1::2] => blue = image[:, 2, 1::2, 0::2] ?

nuaalyh

The real-world dataset

3

Excuse me, how to download the dataset which named real-world dataset in your paper? Thank you.

ZXXY09