Closed maybeLee closed 2 days ago
I notice that this issue also occurs when using the tf.signal.irfft
import tensorflow as tf
input = tf.constant([0.27], dtype='complex64')
tf.signal.irfft(input, fft_length=[0])
error:
Skipping registering GPU devices...
DUCC FFT c2r failed:
bazel-out/k8-opt/bin/external/ducc/_virtual_includes/fft/ducc/src/ducc0/fft/fft1d_impl.h: 2948 (static Trpass<Tfs> ducc0::detail_fft::rfftpass<float>::make_pass(size_t, size_t, size_t, const Troots<Tfs> &, bool) [Tfs = float]):
Assertion failure
no zero-sized FFTs
Aborted (core dumped)
Hi @maybeLee , Apologize for the delay. I tried to run your code on Colab using TensorFlow version 2.17.0, nightly and encountered the same issue. As an alternative, I have provided a Gist for your reference. This issue is already being tracked, so it is a duplicate. Could you please check and let me know if i am wrong. Thank you!
This issue is stale because it has been open for 7 days with no activity. It will be closed if no further activity occurs. Thank you.
Hi @Venkat6871 ,
Yes I think this issue is the duplicate of previous issue. I am closing this one and I will keep tracking the thread on previous issue. It is much appreciated if this issue can be fixed.
Issue type
Bug
Have you reproduced the bug with TensorFlow Nightly?
Yes
Source
source
TensorFlow version
tf 2.17.0
Custom code
Yes
OS platform and distribution
No response
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?
When setting the fft_length to [0], the
tf.signal.rfft
will raises a program abort. If this parameter is invalid, raising an invalid argument error looks more clear to me instead of a program abort.Standalone code to reproduce the issue