Can't reproduce the result for BurstSR dataset(Track2)

[Egkang-Luis]

Hi, thanks for your wonderful work. I am also working on BurstSR and your codes have helped me a lot.

While I have some problems when I try to fine-tune an SR model on the BurstSR dataset. I firstly train the SR model on the Synthetic dataset and everything is OK. Then I fine-tune this model on the BurstSR dataset. The training loss keeps decreasing while the validation PSNR only grows at the beginning and drops gradually after it achieves the best result.
Below image is PSNR result per epoch. X axis is epoch and Y axis is PSNR.

Changed points were as below.

1. spatial_color_alignment.py : Change lstsq from "c = torch.lstsq(ir.t(), iq.t())" to "c = torch.linalg.lstsq(iq.t(), ir.t())" due to Pytorch version.
2. correlation.py : Change "@cupy.util.memoize(for_each_device=True)" to "@cupy.memoize(for_each_device=True)" due to version.
3. PWC pre-train : I used download.bash in pwcnet folder.
4. Train epoch setting is 25 epoch because your original code was 25 epoch.
5. Train patch size is 24 and validation patch size is 80 as like your released code.

Have you met the same problem? I would appreciate it a lot if you could help me figure this out.

Thanks.

[Egkang-Luis]

I'm waiting your response :)

[cvkaiming]

I'm waiting your response :)

Hi, did you encounter any errors while running the Track 2 code, I found that when loading samsung_00\im_raw.png, it reported a path error, can you please tell me what you did in the Burst Super-resolution\data_sets\burstsr_dataset.py ? Thank you, bro.

[Egkang-Luis]

I'm waiting your response :)

Hi, did you encounter any errors while running the Track 2 code, I found that when loading samsung_00\im_raw.png, it reported a path error, can you please tell me what you did in the Burst Super-resolution\data_sets\burstsr_dataset.py ? Thank you, bro.

Could you share your including your error comment?

[cvkaiming]

I'm waiting your response :)

Hi, did you encounter any errors while running the Track 2 code, I found that when loading samsung_00\im_raw.png, it reported a path error, can you please tell me what you did in the Burst Super-resolution\data_sets\burstsr_dataset.py ? Thank you, bro.

Could you share your including your error comment?

When I run Burstormer_Track_2_training.py directly I get the error in the image, after debugging I found that at SamsungRAWImage.load('{}/{}/samsung_{:02d}'.format(self.root, self.burst _list[burst_id], im_id)) This line of code is reporting an error, the path does not include the name of the subfolder under the train dataset train00 folder for example. I assume you were able to run it successfully, so that should have solved the problem

thanks for your reply

[Egkang-Luis]

I'm waiting your response :)

Hi, did you encounter any errors while running the Track 2 code, I found that when loading samsung_00\im_raw.png, it reported a path error, can you please tell me what you did in the Burst Super-resolution\data_sets\burstsr_dataset.py ? Thank you, bro.

Could you share your including your error comment?

When I run Burstormer_Track_2_training.py directly I get the error in the image, after debugging I found that at SamsungRAWImage.load('{}/{}/samsung_{:02d}'.format(self.root, self.burst _list[burst_id], im_id)) This line of code is reporting an error, the path does not include the name of the subfolder under the train dataset train00 folder for example. I assume you were able to run it successfully, so that should have solved the problem thanks for your reply

Honestly, I don't remember clearly. But I didn't have any error that position as I remember.

Could you check your dataset folder?

[cvkaiming]

I'm waiting your response :)

Hi, did you encounter any errors while running the Track 2 code, I found that when loading samsung_00\im_raw.png, it reported a path error, can you please tell me what you did in the Burst Super-resolution\data_sets\burstsr_dataset.py ? Thank you, bro.

Could you share your including your error comment?

When I run Burstormer_Track_2_training.py directly I get the error in the image, after debugging I found that at SamsungRAWImage.load('{}/{}/samsung_{:02d}'.format(self.root, self.burst _list[burst_id], im_id)) This line of code is reporting an error, the path does not include the name of the subfolder under the train dataset train00 folder for example. I assume you were able to run it successfully, so that should have solved the problem thanks for your reply

Honestly, I don't remember clearly. But I didn't have any error that position as I remember.

Could you check your dataset folder?

May I ask, did you put all the files in these folders under one folder?

[cvkaiming]

When testing at SyntheticBurst, I get strange images，can you help me？