argsort does not support complex128 tensors on CPU backend #99164
Description
Issue type
Bug
Have you reproduced the bug with TensorFlow Nightly?
Yes
Source
source
TensorFlow version
2.17
Custom code
Yes
OS platform and distribution
Ubuntu 20.04
Mobile device
No response
Python version
No response
Bazel version
No response
GCC/compiler version
No response
CUDA/cuDNN version
No response
GPU model and memory
No response
Current behavior?
The operation fails with the following error:
tensorflow.python.framework.errors_impl.InvalidArgumentError: Value for attr 'T' of complex128 is not in the list of allowed values: float, double, int32, uint8, int16, int8, int64, bfloat16, uint16, half, uint32, uint64
; NodeDef: {{node TopKV2}}; Op<name=TopKV2>; Op<name=TopKV2>
Expected behavior
argsort should either support complex64/complex128 tensors on CPU, or raise a clear error indicating that the operation is not supported for complex types.
Standalone code to reproduce the issue
import os
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3'
import tensorflow as tf
import tensorflow.experimental.numpy as tnp
# Generate random complex input data
real_part = tnp.random.randn(3, 4)
imag_part = tnp.random.randn(3, 4)
complex_data = tf.complex(real_part, imag_part)
# Attempt to apply argsort
sorted_indices = tnp.argsort(a=complex_data, kind='stable')Relevant log output
Input complex data:
[[-0.56195893-1.64089445j -1.54584942-0.36713052j -1.3005634 +1.6114282j
0.55223823-0.18570225j]
[ 1.14189741+0.47845684j 0.23988903-1.09895554j -0.87619739+0.73307129j
1.04041324-0.80997881j]
[ 1.14454671-0.10186608j 0.14604414+0.00396092j 1.10731772-0.09357997j
-0.61830007-0.10900073j]]
tensorflow.python.framework.errors_impl.InvalidArgumentError: Value for attr 'T' of complex128 is not in the list of allowed values: float, double, int32, uint8, int16, int8, int64, bfloat16, uint16, half, uint32, uint64
; NodeDef: {{node TopKV2}}; Op<name=TopKV2>