ValueError: test_size=0 should be either positive and smaller than the number of samples 0 or a float in the (0,1) range

Asked Nov 20 '19 at 21:17

Active Nov 20 '19 at 21:30

Viewed 6,689 times

X= data
X_data= 'trained.h5'
#Labels
Y_data = X_data[:,(X_data.shape[1]-1)]
#Training Data
X_data = X_data[:,0:(X_data.shape[1]-1)]
x_tr,x_ts,y_tr,y_ts = train_test_split(X_data, Y_data, test_size=0.3,random_state=10)

Error comes '(0,1) range'.format(test_size, n_samples))

ValueError: test_size=0 should be either positive and smaller than the number of samples 0 or a float in the (0,1) range

edited Nov 20 '19 at 21:30

asked Nov 20 '19 at 21:17

Sanaa

can you try printing the shapes of Y_data and X_data ? – venkata krishnan Nov 21 '19 at 01:49
basically i just want to split the data. and ```x_tr,x_ts,y_tr,y_ts = train_test_split(X_data, Y_data, test_size=0.3,random_state=10)```` error occur in this line – Sanaa Nov 21 '19 at 02:37
yes. I get that. for that, if we know, what is the shape of X_data and Y_data, we will have a better idea of why it gives this error. all you need to do is print(X_data.shape) and print(Y_data.shape) and see what it gives, and add it to your question, which will be of nice use for us to debug. Thanks – venkata krishnan Nov 21 '19 at 03:16
i try to run but this error comes````Y_data = X_data[:,(X_data.shape[1]-1)] AttributeError: 'str' object has no attribute 'shape'``` – Sanaa Nov 21 '19 at 03:25
there you go. that's the problem. you need to know how to read data from .h5 files. there are lot of methods, mentioned in this thread, which will be helpful for you. https://stackoverflow.com/questions/28170623/how-to-read-hdf5-files-in-python – venkata krishnan Nov 21 '19 at 03:28
or is it a typo ? instead of using X, you are using X_data ? – venkata krishnan Nov 21 '19 at 03:29
yes its typo...and i read the file using ```filename = "trained.h5" dataset = h5py.File(filename,'r')``` – Sanaa Nov 21 '19 at 03:48
shall i close this question marking it as typo ? – venkata krishnan Nov 21 '19 at 03:54
@venkatakrishnan but my error is solved and still occur ``ValueError: test_size=0 should be either positive and smaller than the number of samples 0 or a float in the (0,1) range – Sanaa Nov 21 '19 at 04:00
what is the shape of your Y ? coz it clearly says " the number of samples 0 " – venkata krishnan Nov 21 '19 at 04:01
Do you mean dataset? its images – Sanaa Nov 21 '19 at 04:06
yea. but the size is 0. what output you get when you print the size ? – venkata krishnan Nov 21 '19 at 05:29
when I print, its come is (0,) – Sanaa Nov 21 '19 at 06:21
1

Thats why. You dont have any data. Something wrong in reading h5 file – venkata krishnan Nov 21 '19 at 16:08
Thanks, @venkatakrishnan, can you guide me further reading .h5 , because with I print .h5 gives an error. – Sanaa Nov 21 '19 at 18:27
https://stackoverflow.com/questions/28170623/how-to-read-hdf5-files-in-python this link must help you. – venkata krishnan Nov 22 '19 at 00:51

ValueError: test_size=0 should be either positive and smaller than the number of samples 0 or a float in the (0,1) range

0 Answers0