Please tell me how to solve this problem

[WangchukTsering]

raise KeyError(f"No object named '{name}' found in '{self._name}' registry!") KeyError: "No object named 'TextDegradationDataset' found in 'dataset' registry!"

[csxmli2016]

raise KeyError(f"No object named '{name}' found in '{self._name}' registry!") KeyError: "No object named 'TextDegradationDataset' found in 'dataset' registry!"

Did you follow the installation steps? You can solve it by running python setup.py develop

[WangchukTsering]

According to the Readme file, after Python setup. py, it displays: Finished processing dependencies for marconet=1.1.0 But in the end, this error will still be reported!

[WangchukTsering]

I was unable to train, please ask the author for a favor.

[csxmli2016]

I was unable to train, please ask the author for a favor.

I can successfully train the model with the given instruction. Please double-check the following steps: 1) All the training steps are performed in cd ./Train path 2) Remove the environment and run the following steps again:

conda create -n marcotrain python=3.8 -y
conda activate marcotrain
pip install -r requirements.txt
BASICSR_EXT=True pip install basicsr
python setup.py develop

3) python scripts/download.py 4) CUDA_VISIBLE_DEVICES=0 python tspgan/train.py -opt options/train.yml

With the above steps, I can successfully run the training code. Could you show me the output by running pip list | grep marconet after you run python setup.py develop?

[WangchukTsering]

Thanks especially for your reply. I just got out of the lab. I'll try to re-match this environment tomorrow.

---Original--- From: @.> Date: Tue, Jan 2, 2024 22:00 PM To: @.>; Cc: @.**@.>; Subject: Re: [csxmli2016/MARCONet] Please tell me how to solve this problem (Issue #16)

I was unable to train, please ask the author for a favor.

I can successfully train the model with the given instruction. Please double-check the following steps:

All the training steps are performed in cd ./Train path

Remove the environment and run the following steps again: conda create -n marcotrain python=3.8 -y conda activate marcotrain pip install -r requirements.txt BASICSR_EXT=True pip install basicsr python setup.py develop
python scripts/download.py

CUDA_VISIBLE_DEVICES=0 python tspgan/train.py -opt options/train.yml

With the above steps, I can successfully run the training code. Could you show me the output by running pip list | grep marconet after you run python setup.py develop?

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

[WangchukTsering]

I reinstalled the environment today and this error still occurs. After running python setup.py develop, it shows success, but cannot be trained.

[csxmli2016]

I reinstalled the environment today and this error still occurs. After running python setup.py develop, it shows success, but cannot be trained.

what is the output when you run pip list | grep marconet

[WangchukTsering]

display after list | grep marconet

[csxmli2016]

Author

Could you show me the files in Text_Image_SR/MARCONet/Train/tspgan/data

[WangchukTsering]

Or can you help me debug it remotely? I have been trying to solve this problem for three weeks!

[csxmli2016]

Or can you help me debug it remotely? I have been trying to solve this problem for three weeks!

Sure. You can send me an email (csxmli@gmail.com) about connecting to the server.