dvlp-cmaterju / rectinet Goto Github PK

Hi, can you please add scripts for generating a synthetic dataset. It would be really helpful. Thanks.

您好，我根据您提供的脚本生成了自己的数据集，但在运行的时候，他报了以下错误。
raise KeyError("%s is not a file in the archive"% key)
keyError:''grid is not a file in the archive'
是我数据集格式存在问题吗，希望您能指点一下

使用提供的预训练模型，效果不好，请问大神用了多少数据量训练的

I have trained 100 epochs according to your code, but the result is not ideal. How many epochs have you trained and got the best result?

thank you very much for sharing your code.

How can I run the predict.py with cpu only?

When trying to run train.py I am getting the error that cannot import initialize_weights. On checking the utils_model.py file in utils folder I am finding that there is no function or class named initialize_weights. Is this an error in the branch or something?

I I have downloaded the data you provided, but in the dataset.py In, label is. Npz format, but I didn't find this format in the data you provided. Can you tell me how to apply the downloaded data to the code you provided?

Can you add some sample datasets and how it is arranged?

我想了解门控网络的作用，使用预训练模型，发现门控网络输出为黑色图片，这里有什么错误吗？还是门控网络根本不起作用？

the pretrained model weight name is different from the model define, so the model can not load the weight

I am having trouble understanding the directory structure. I have generated the images using the AGUN repository you have linked. However, I am just trying to understand how to use those images in the directory structure. Like which images go in the image folder of datagen ? And which is the label .npz file required.

Hi!
I have tried to run your code, but after training the data, I have some questions about the Models.
The project provieded two model, the first one is in the model.py and the second one is in the model_pred.py. When I tried to run the predict.py, it asserted missing keys in state_dict. I checked the code and found different models used in train.py and predict.py.
I guess the code following is trying to import the intersection of two state dict;
File "predict.py", line 108, "state_dict = {k: v for k, v in state_dict.items() if k in model_dict}"
I can predict with the pre_trained model provided, but failed in the model trained by myself. This Block may not work well.

I want to know the difference between the two models and why the project use two model for train and prediect respectively?