-
Notifications
You must be signed in to change notification settings - Fork 5
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
I encountered the following error while attempting to reproduce the steps you provided. Could you please tell me why this error occurred? #3
Comments
Hello! |
Thanks for the answer, I have deleted the models file as requested, but the following error appears, please answer again, thank you very much Traceback (most recent call last): |
Hi, |
Thanks. The environment has been reconfigured as required, and the code shows that there is a problem with the clip package, and I didn't find the ModifiedResNet module it needs, please let me know, thank you. At the same time I'm not very clear about the meaning of these two in your training code, -m torch.distributed.launch --nproc_per_node=4. |
For the first problem, did you install clip package correctly? Would you please run 'pip list' and check the version of clip package? We do not meet such problem. Make sure you do not meet network problem during installing of clip package. |
Thanks for the answer, the model has run successfully. I also want to ask, I see that this model does not have an evaluation index, how to judge the effect of super-score, and there is no log record to show the training progress, will these be considered to be added? Thank you |
In our implementation, we will evaluate the performance of the trained model every 5000 iterations on the Set5 benchmark, as set in yml file. If you want to evaluate it on another dataset, you may change it by yourself~ If I haven't misunderstood your meaning, our "evaluation index" is PSNR and SSIM. |
Hello, I found that there are three evaluation indicators in your paper: LPIPS, PSNR, and SSIM, but there is no LPIPS in the code, can you add it? |
Hi, you can use the powerful python package pyiqa to calculate LPIPS, we will add these to our code, thanks for your advice! |
I see that the generator type in your code is currently RRDB, I want to use the SwinIR you provided, please how can I do it, only modify the model_type will give an error, thanks for the answer |
What's your error message? |
I first modified the weight file as follows: |
This is because official SwinIR code stores their weights using an extra key called "params". You can modify the line 107 of train.py into "model.load_state_dict(ckpt['params'])" |
Thanks for the answer, pls if there is a problem with my data_root and data_test_root: |
The dataset used to run me can run through。 |
It seems nothing wrong. |
Traceback (most recent call last):
File "F:\pythonProject1\SeD-main\train.py", line 15, in
from models import model_rrdb, model_swinir, sed
File "F:\pythonProject1\venv\Lib\site-packages\models_init_.py", line 37, in
import project
ModuleNotFoundError: No module named 'project'
Traceback (most recent call last):
File "F:\pythonProject1\SeD-main\train.py", line 15, in
from models import model_rrdb, model_swinir, sed
File "F:\pythonProject1\venv\Lib\site-packages\models_init_.py", line 37, in
import project
ModuleNotFoundError: No module named 'project'
The text was updated successfully, but these errors were encountered: