search

SheffieldML / GPy

Gaussian processes framework in python
BSD 3-Clause "New" or "Revised" License
2.01k stars 557 forks source link

The input of GPCoregionalizedRegression #1024

Open BinglinW opened 1 year ago

BinglinW commented 1 year ago

Hi,

When I use the GPCoregionalizedRegression, the train data has some wrong. I will input X_0 and Y_0 as two training sets, X_0 is an array format, it has 20 rows, 3 columns, so it has 3 dimensional input, 20 sample points, and Y_0 has 30 output, so it is a 20 by 30 array. I test my input format with a minimal example of two dots, and it keeps getting errors

K = GPy.kern.Matern32(input_dim=dim)
 icm = GPy.util.multioutput.ICM(input_dim=dim ,num_outputs=t_max,kernel=K)

 m = GPy.models.GPCoregionalizedRegression(X_list=[X_0[0, :].reshape(-1,1),X_0[1, :].reshape(-1,1)],
Y_list=[Y_0[0, :].reshape(-1,),Y_0[1, :].reshape(-1,)],
kernel=icm)

Then the error is 

---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
File [~/anaconda3/envs/wbl_gpy_py39/lib/python3.9/site-packages/GPy/kern/src/kern.py:115](https://vscode-remote+ssh-002dremote-002b127-002e0-002e0-002e1.vscode-resource.vscode-cdn.net/home/dell/wbl/~/anaconda3/envs/wbl_gpy_py39/lib/python3.9/site-packages/GPy/kern/src/kern.py:115), in Kern._slice_X(self, X)
    [114](file:///home/dell/anaconda3/envs/wbl_gpy_py39/lib/python3.9/site-packages/GPy/kern/src/kern.py?line=113) try:
--> [115](file:///home/dell/anaconda3/envs/wbl_gpy_py39/lib/python3.9/site-packages/GPy/kern/src/kern.py?line=114)     return X[:, self._all_dims_active].astype('float')
    [116](file:///home/dell/anaconda3/envs/wbl_gpy_py39/lib/python3.9/site-packages/GPy/kern/src/kern.py?line=115) except:

IndexError: index 2 is out of bounds for axis 1 with size 2

During handling of the above exception, another exception occurred:

IndexError                                Traceback (most recent call last)
/home/dell/wbl/main_compare_multioutputGP_gpy.py in line 7
      [73](file:///home/dell/wbl/main_compare_multioutputGP_gpy.py?line=72) X_0_list = [list(row) for row in X_0]
      [74](file:///home/dell/wbl/main_compare_multioutputGP_gpy.py?line=73) Y_0_list = [list(row) for row in Y_0]
----> [75](file:///home/dell/wbl/main_compare_multioutputGP_gpy.py?line=74) m = GPy.models.GPCoregionalizedRegression(X_list=[X_0[0, :].reshape(-1,1),X_0[1, :].reshape(-1,1)],
      [76](file:///home/dell/wbl/main_compare_multioutputGP_gpy.py?line=75)                                           Y_list=[Y_0[0, :].reshape(-1,),Y_0[1, :].reshape(-1,)],
      [77](file:///home/dell/wbl/main_compare_multioutputGP_gpy.py?line=76)                                           kernel=icm)

File [~/anaconda3/envs/wbl_gpy_py39/lib/python3.9/site-packages/paramz/parameterized.py:58](https://vscode-remote+ssh-002dremote-002b127-002e0-002e0-002e1.vscode-resource.vscode-cdn.net/home/dell/wbl/~/anaconda3/envs/wbl_gpy_py39/lib/python3.9/site-packages/paramz/parameterized.py:58), in ParametersChangedMeta.__call__(self, *args, **kw)
     [56](file:///home/dell/anaconda3/envs/wbl_gpy_py39/lib/python3.9/site-packages/paramz/parameterized.py?line=55) self._model_initialized_ = False
     [57](file:///home/dell/anaconda3/envs/wbl_gpy_py39/lib/python3.9/site-packages/paramz/parameterized.py?line=56) if initialize:
---> [58](file:///home/dell/anaconda3/envs/wbl_gpy_py39/lib/python3.9/site-packages/paramz/parameterized.py?line=57)     self.initialize_parameter()
     [59](file:///home/dell/anaconda3/envs/wbl_gpy_py39/lib/python3.9/site-packages/paramz/parameterized.py?line=58) else:
     [60](file:///home/dell/anaconda3/envs/wbl_gpy_py39/lib/python3.9/site-packages/paramz/parameterized.py?line=59)     import warnings
...
    [115](file:///home/dell/anaconda3/envs/wbl_gpy_py39/lib/python3.9/site-packages/GPy/kern/src/kern.py?line=114)     return X[:, self._all_dims_active].astype('float')
    [116](file:///home/dell/anaconda3/envs/wbl_gpy_py39/lib/python3.9/site-packages/GPy/kern/src/kern.py?line=115) except:
--> [117](file:///home/dell/anaconda3/envs/wbl_gpy_py39/lib/python3.9/site-packages/GPy/kern/src/kern.py?line=116)     return X[:, self._all_dims_active]

IndexError: index 2 is out of bounds for axis 1 with size 2

I read the instructions carefully, but still couldn't find the problem.The environment is Ubuntu, and I got a prompt when loading gpy-1.10.0

warning in coregionalize: failed to import cython module: falling back to numpy
warning in choleskies: failed to import cython module: falling back to numpy

I don't know if these warnings are related to the error

MartinBubel commented 11 months ago

Hi @BinglinW could you please add a script reproducing the error? Or give hints on the values of the variables dim, t_max, X_0, Y_0 so I can dig into this?