tlatkowski/inpainting-gmcnn-keras

Keras implementation of "Image Inpainting via Generative Multi-column Convolutional Neural Networks" paper published at NIPS 2018

👋 Hi there, I'm Tomek

I'm an Experienced AI Engineer with over 10 years of hands-on experience in applying deep and shallow machine learning methods for solving tasks in the fields of:

• Time Series
• Computer Vision
• Natural Language Processing
• Data Mining

inpainting-gmcnn-keras's Issues

Unable to run colab notebook

Hi,

I'm trying to run the colab notebook but I keep getting the following error when running this line:

Actually it first gives an error saying --experiment_name is required, but after passing that arg in I get the error below

!python inpainting-gmcnn-keras/runner.py --train_path images --mask_path mask -warm_up_generator --experiment_name tester

Traceback (most recent call last): File "inpainting-gmcnn-keras/runner.py", line 105, in <module> main() File "inpainting-gmcnn-keras/runner.py", line 52, in main config = main_config.MainConfig(MAIN_CONFIG_FILE) File "/content/inpainting-gmcnn-keras/config/main_config.py", line 22, in __init__ self.model = ModelConfig(model_config['MODEL']) File "/content/inpainting-gmcnn-keras/config/main_config.py", line 41, in __init__ self.add_mask_as_generator_input = parse_bool(model_section['ADD_MASK_AS_GENERATOR_INPUT']) File "/usr/lib/python3.6/configparser.py", line 1233, in __getitem__ raise KeyError(key) KeyError: 'ADD_MASK_AS_GENERATOR_INPUT'

Only one step for training

Hi,

I'm trying to test this inpainting alternative. I'm using code from branch tensorflow-1.15.2 and when execute runner.py this run only one step of training.

In the outputs folder, I see only step_000.png

Do you know why?

Thanks

Cannot predict image when I do some changes to mask dataset

Hi,
I'm trying to use opencv to erode the mask dataset you provide, however when execute runner.py, it cannot predict images.

error comes out like this.
Do you know why?
Thanks

Code for model testing

Hi,

Thank you so much for publishing the Keras implementation of gmcnn on Github.
Can you also provide the code for testing the model?

Error

Hello,
I'm trying to reproduce your algorithm, but I will get this error. What am I doing wrong?

Wrong results after Tensorflow 2x model conversion #70

Hi,

I hope you still give support for this repository. I tried converting this repository model to Tensorflow 2x step by step (also following the original repository by shepnerd). I think everything was fine, but when training and testing the model the results were not very satisfactory. So, maybe there is a problem with my conversion that I can't find it. I tried to train the model on a small subset of OpenImages V6. The principal parameters that I used (sorry, because names can be slightly different than yours):

--img_size 256x256x3
--batch_size 4
--learning_rate 1e-4
--gaussian_steps 7
--gaussian_kernel_size 32
--gaussian_kernel_std 20.0
--reconstruction_loss_weight 1.2
--adversarial_loss_weight 0.001
--gradient_penalty_loss_weight 10
--id_mrf_loss_weight' 0.03
--nn_stretch_sigma 0.5
--id_mrf_style_weight 1.0
--id_mrf_content_weight 1.0

I pretrained the model with only confidence reconstruction loss for the recommended steps, and results for this phase seem fine.

![imagen](https://

user-images.githubusercontent.com/39574343/111080637-a3871d00-84ff-11eb-90cb-c6de09d587f6.png)

However, in the training phase, after some steps, the results do not seem to converge and eliminate the mask.

I would like to know if you have any idea of what could be happening, or you have experienced any similar issue. I reviewed the code, networks, and losses many times, but could not find any solution.

Thank you very much.

outputs file is not being made, not able to get output

outputs file is not being made, not able to get output

Please help me in that

Link to dataset is dead

Hi.
When running notebook in colab, the link to Places356 dataset is dead:
404 Not Found