ops.copy makes object (arrays only?) immutable
Mrutyunjay01 opened this issue · 1 comments
Mrutyunjay01 commented
Issue:
When trying to overwrite an object after making a copy using ops.copy
, it makes the object immutable. Even after explicitly converting to numpy
using ops.convert_to_numpy
, the copied object remains immutable.
How to Reproduce:
import numpy as np
from keras import ops
arr = np.arange(10)
ops_arr = ops.copy(arr)
print(type(ops_arr)) # now it's EagerTensor
ops_arr_np = ops.convert_to_numpy(ops_arr)
print(ops_arr_np) # of type ndarray
ops_arr_np[0] = 10
And the error that follows:
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
# in <cell line: 9>()
7 print(ops_arr_np) # of type ndarray
8
----> 9 ops_arr_np[0] = 10
ValueError: assignment destination is read-only
fchollet commented
Thanks for the report, this is now fixed at HEAD.