Question about generating validation dataset.

Question

Question about generating validation dataset.

Closed this issue 2 years ago · 5 comments

SingCheng commented 2 years ago

According to step two which is used to generate ten levels' foggy training images and val images, respectively

However, when we access data_make.py. We can observe that the img_name for the validation set has been commented.

The VOC structure is exactly the same as it is required.

is there anything I have missed?

Answer 1 · 2023-11-28T11:56:28.000Z

According to step two which is used to generate ten levels' foggy training images and val images, respectively

However, when we access data_make.py. We can observe that the img_name for the validation set has been commented.

The VOC structure is exactly the same as it is required.

is there anything I have missed?

I have the same question.Did you solve it?

Answer 2 · 2023-11-28T12:07:03.000Z

According to step two which is used to generate ten levels' foggy training images and val images, respectively
However, when we access data_make.py. We can observe that the img_name for the validation set has been commented.
The VOC structure is exactly the same as it is required.
is there anything I have missed?

I have the same question.Did you solve it?

Hi @lyq216398 i did solved it, if you are generating validation set, just uncomment it and comment the the line above.

Answer 3 · 2023-11-28T12:39:52.000Z

According to step two which is used to generate ten levels' foggy training images and val images, respectively
However, when we access data_make.py. We can observe that the img_name for the validation set has been commented.
The VOC structure is exactly the same as it is required.
is there anything I have missed?

I have the same question.Did you solve it?

Hi @lyq216398 i did solved it, if you are generating validation set, just uncomment it and comment the the line above.

Thank you for your reply!I want to know whether the validation set is for final testing or just is separated from the training set to verify the effectiveness during training.I am a little confused with the concept of "val and test".

Answer 4 · 2023-11-28T12:48:01.000Z

According to step two which is used to generate ten levels' foggy training images and val images, respectively
However, when we access data_make.py. We can observe that the img_name for the validation set has been commented.
The VOC structure is exactly the same as it is required.
is there anything I have missed?

I have the same question.Did you solve it?

Hi @lyq216398 i did solved it, if you are generating validation set, just uncomment it and comment the the line above.

Thank you for your reply!I want to know whether the validation set is for final testing or just is separated from the training set to verify the effectiveness during training.I am a little confused with the concept of "val and test".

The validation set is to verify the effectiveness during training. The author has mixing up the val and test in his code, if i'm remember correctly.

Answer 5 · 2023-11-28T12:51:45.000Z

According to step two which is used to generate ten levels' foggy training images and val images, respectively
However, when we access data_make.py. We can observe that the img_name for the validation set has been commented.
The VOC structure is exactly the same as it is required.
is there anything I have missed?

I have the same question.Did you solve it?

Hi @lyq216398 i did solved it, if you are generating validation set, just uncomment it and comment the the line above.

Thank you for your reply!I want to know whether the validation set is for final testing or just is separated from the training set to verify the effectiveness during training.I am a little confused with the concept of "val and test".

The validation set is to verify the effectiveness during training. The author has mixing up the val and test in his code, if i'm remember correctly.

Thank you !!!