Sorry, I can not understand Japanese language. I think your work looks promising :+1: I've also written some codes trying to achieve the results as Xu did in her article, but what I got is pretty messy as in the following image(70 chars a line) :disappointed_relieved: Hope you can write an article explaining the details about how you train the model :grin:

Large file size so file wasn't uploaded: https://github.com/OsciiArt/DeepAA/blob/master/model/weight.hdf5

Possibly supply file in releases or upload elsewhere.

Executing output.py returns the following error:

$ python output.py                                                                                              v8.8.1 | 01:08 
Using TensorFlow backend.
/usr/lib/python3.6/importlib/_bootstrap.py:219: RuntimeWarning: compiletime version 3.5 of module 'tensorflow.python.framework.fast_tensor_util' does not match runtime version 3.6
  return f(*args, **kwds)
2017-11-13 13:11:08.607633: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA
Traceback (most recent call last):
  File "output.py", line 38, in <module>
    model.load_weights(weight_path)
  File "/usr/lib/python3.6/site-packages/keras/engine/topology.py", line 2615, in load_weights
    raise ImportError('`load_weights` requires h5py.')
ImportError: `load_weights` requires h5py.

Adding h5py (2.7.1) to requirements list fix the issue

pip でtensorflowとKerasはインストールしましたー。

#スタックトレース

Using TensorFlow backend. Traceback (most recent call last): File "C:/Users/(ユーザー名)/Downloads/DeepAA-master/deepaa_output.py", line 24, in model = model_from_json(json_string) File "C:\Program Files\Anaconda3\lib\site-packages\keras\models.py", line 325, in model_from_json return layer_module.deserialize(config, custom_objects=custom_objects) File "C:\Program Files\Anaconda3\lib\site-packages\keras\layers_init_.py", line 46, in deserialize printable_module_name='layer') File "C:\Program Files\Anaconda3\lib\site-packages\keras\utils\generic_utils.py", line 140, in deserialize_keras_object list(custom_objects.items()))) File "C:\Program Files\Anaconda3\lib\site-packages\keras\engine\topology.py", line 2370, in from_config process_layer(layer_data) File "C:\Program Files\Anaconda3\lib\site-packages\keras\engine\topology.py", line 2339, in process_layer custom_objects=custom_objects) File "C:\Program Files\Anaconda3\lib\site-packages\keras\layers_init_.py", line 46, in deserialize printable_module_name='layer') File "C:\Program Files\Anaconda3\lib\site-packages\keras\utils\generic_utils.py", line 141, in deserialize_keras_object return cls.from_config(config['config']) File "C:\Program Files\Anaconda3\lib\site-packages\keras\engine\topology.py", line 1202, in from_config return cls(**config) TypeError: init() got an unexpected keyword argument 'input_dtype'

#環境 Python 3.5.2 |Anaconda custom (64-bit)| (default, Jul 5 2016, 11:41:13) [MSC v.1900 64 bit (AMD64)] on win32 依存ライブラリのバージョン

• tensorflow (1.0.1)
• Keras (2.0.1)
• numpy (1.12.0)
• Pillow (3.3.1)

I was wondering which font would be best for output. I generated ascii art someting like this:

　  ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,　　　
　　|　　　　　　　　　　　　　　　　　　　　　 | 　　　
　　|　　'|''''''''''''''''''''''''''''''''''''''''''''''''''''''''''|　 　 |　　　
　　|　　:|　　,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,-,,,,　　 |　 　 |　　　
　　|　　:|　　 |　　　　　　　　　　]|　　|　 　 |　　　
　　|　　:|　　 |　　　　　　　　　　:|　　:|　 　 |　　　
　　|　　:|　　 |　　　　　　　　　　:|　　:|　 　 |　　　
　　|　　:|　　 |　　　　　　　 　　 l|　　 |　 　 |　　　
　　|　　:| 　 └――――――-┘　 |　 　 |　　　
　　|　　.|＿＿＿＿＿＿＿＿＿＿__,|　 　 |　　　
　  |　　　　　　　　　　　　　　　　　　　　　 :|　　　
　　|-----------------------------┘

and I think fixed width font is not proper for this.

The bullet point "training data (additional, download it from [here], extract it and place the extracted directory in dir data.) )" in the readme seems like it should link to a training data file but it actually links to the same weights.hdf5 file as in the previous bullet point.

I was trying to download the weights file over google drive and it asked for me to request access. I believe there is a way to set it to not make people request access, but still not be able to edit the file. That will make it so you don't have to give access to anyone who wants to download in the future.

Thanks for the cool program!

as_matrix is deprecated since 0.23.0, to_numpy should be used instead. Also fixed some grammar in readme. .idea/ folder is not needed. It is automatically generated by Jetbrains IDE.