tulip-lab / open-code Goto Github PK

to be released

please also update the overall readme, with links.

You can use pull request

Hi, I opened this issue because executing Eval.py gives me the following error:

Traceback (most recent call last):
File "Eval.py", line 117, in
valuelist = walk_forward(train_X,train_Y,test_X, test_Y,grid_search,scaler)
File "Eval.py", line 30, in walk_forward
model = build_model(history_x,history_y,lstm_unit=x[0],drop_rate=x[1],dense_unit=x[2],batch_size=x[3],epochs=x[4])
File "/mnt/c/Users/User/Desktop/UniUD/LabAI/open-code-master/DLM/DLM.py", line 45, in build_model
dense_unit)
File "/mnt/c/Users/User/Desktop/UniUD/LabAI/open-code-master/DLM/DLM.py", line 27, in model_attention_applied_before_lstm
attention_mul = attention_block(inputs,TIME_STEPS,INPUT_DIM)
File "/mnt/c/Users/User/Desktop/UniUD/LabAI/open-code-master/DLM/DLM.py", line 21, in attention_block
output_attention_mul = merge([inputs, a_probs], name='attention_mul', mode='mul')
TypeError: 'module' object is not callable

Upon a closer code inspection I figured out what the problem was:
line 21 of DLM.py

output_attention_mul = merge([inputs, a_probs], name='attention_mul', mode='mul')

constains merge, a layer not implemented in more recent versions of Keras.

I tried changing that line to

output_attention_mul = Multiply()([inputs, a_probs])

and now the code seems to work. Could you please confirm whether my fix is valid or not?

I also have a question concerning Setting.py: since I'm trying to recreate the results of your paper "Tourism demand forecasting - A deep learning approach" can I please know which were the actual settings used to get the values in the tables 2,3,4 and 6,7,8 or the article?

Finalize the information for the data or code, corresponding to the published paper.

老师您好 您preprocess.py文件中 引用了minepy的一个方法 但是包中好像没有这个，请问是怎么回事儿呢？
from minepy import MINE --第十一行

I looked at the code of GANBLR and GANBLR_Extend, but didn't understand how to generate numerical tabular data on the sample code. >_<