ZiMo Chen comments

Results 21 comments of


                                            ZiMo Chen

Can you please provide the GFLOPs calculation code

> ``` > from fvcore.nn import FlopCountAnalysis > from fvcore.nn import flop_count_table > from fvcore.nn import flop_count_str > > # Similar codes in main.py are omitted here for simplicity >...

assert (boxes1[:, 2:] >= boxes1[:, :2]).all()

> I am also following these kind of errors in detr architectures in recent weeks. I do not come across in this repo but one of my architectural modification to...

assert (boxes1[:, 2:] >= boxes1[:, :2]).all()

> 嘿，这意味着您对代码的修改无法使模型工作并产生非法的框预测。例如，这里的错误表示框的左下点坐标值大于右上点。 It is strange that this problem occurs in the middle of the training, and the results of the previous training are ok

cannot import name '_LinearWithBias' from 'torch.nn.modules.linear'

> It is possibly caused by incompatible PyTorch version. Which PyTorch version do you use? '1.10.1

Sizes of tensors must match except in dimension 1.

> It seems you set the wrong dimension. Please check your code. I didn't change the code, I ran the code with --use_dn --amp --save_results The code works fine if...

Visual result

> @cxq1 thanks! relative log amplitudes are plotted with https://colab.research.google.com/github/xxxnell/how-do-vits-work/blob/transformer/fourier_analysis.ipynb. Fourier spectrum of feature maps is visualized with the function np. fft.fft2() Thank you for your contribution. Could you also...

ZiMo Chen

Can you please provide the GFLOPs calculation code

assert (boxes1[:, 2:] >= boxes1[:, :2]).all()

assert (boxes1[:, 2:] >= boxes1[:, :2]).all()

cannot import name '_LinearWithBias' from 'torch.nn.modules.linear'

Sizes of tensors must match except in dimension 1.

Visual result

Visual result

Could you please provide the evaluation code of CPR?

[CPR] An error occurred while running code

[CPR] An error occurred while running code