search

ratschlab / GP-VAE

TensorFlow implementation for the GP-VAE model described in https://arxiv.org/abs/1907.04155
MIT License
124 stars 27 forks source link

Trying to reproduce the code for custom dataset #18

Open Niharikajo opened 2 years ago

Niharikajo commented 2 years ago

Hello, I was trying to reproduce your code using a custom data set (attached below) 10.csv

I have made changes to the values of some flags in train.py (file attached below) train.txt

I ran the following python script !CUDA_VISIBLE_DEVICES=* python train.py --model_type gp-vae --data_type 10

Here I'm getting the following error image

Please help on the following error.

Thank you

mayavy commented 1 year ago

"data" np.ndarray doesn't accepts "string" types 'x_train_full' as index. Wrong syntax basically in the source. You have pass indexes of the types given in the error-message. Or try to load "data" as dictionary instead of array.

ugggggggg commented 1 year ago

Hello

I am trying to reproduce your code and I would like to know some issues with the 'physionet. npz' dataset

  1. '’m_ train_artificial' What does mean? 2.'x_train_full', 'x_train_miss'Which is the original data ?
  2. What is the relationship between 'x_test_full' and the original dataset?
  3. I didn't understand this code either image image