Crash in tf.raw_ops.SparseCountSparseOutput

[x0w3n]

Issue type

Bug

Have you reproduced the bug with TensorFlow Nightly?

Yes

Source

source

TensorFlow version

TensorFlow Nightly

Custom code

Yes

OS platform and distribution

Linux Ubuntu 22.04.3 LTS (x86_64)

Mobile device

No response

Python version

3.9.13

Bazel version

No response

GCC/compiler version

No response

CUDA/cuDNN version

No response

GPU model and memory

No response

Current behavior?

On specific input, tf.raw_ops.SparseCountSparseOutput will output "The session crashed because it took up all available RAM."

Standalone code to reproduce the issue

import tensorflow as tf

binary_output = True
indices = tf.constant(0, shape=[3456,2], dtype=tf.int64)
values = tf.constant(536870912, shape=[3456], dtype=tf.int32)
dense_shape = tf.constant([125099989676412,125099989676412], shape=[2], dtype=tf.int64)
weights = tf.constant(51, shape=[3456], dtype=tf.int32)


tf.raw_ops.SparseCountSparseOutput(
   indices=indices, values=values, dense_shape=dense_shape, weights=weights, binary_output=binary_output,
    minlength=0,
    maxlength=0,
    name=None
)

Relevant log output

Jun 10, 2024, 11:31:57 AM	WARNING	WARNING:root:kernel 7c5d2f95-9ab7-4b7c-99c0-d8949477c9f1 restarted
Jun 10, 2024, 11:31:57 AM	INFO	KernelRestarter: restarting kernel (1/5), keep random ports
Jun 10, 2024, 11:31:29 AM	WARNING	2024-06-10 03:31:29.145870: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Jun 10, 2024, 11:31:26 AM	WARNING	To enable the following instructions: AVX2 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
Jun 10, 2024, 11:31:26 AM	WARNING	2024-06-10 03:31:26.481147: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
Jun 10, 2024, 11:31:26 AM	WARNING	2024-06-10 03:31:26.469832: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
Jun 10, 2024, 11:31:26 AM	WARNING	2024-06-10 03:31:26.467591: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
Jun 10, 2024, 11:31:26 AM	WARNING	2024-06-10 03:31:26.467524: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
Jun 10, 2024, 11:31:18 AM	INFO	Kernel started: 7c5d2f95-9ab7-4b7c-99c0-d8949477c9f1, name: python3
Jun 10, 2024, 11:30:19 AM	INFO	Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).

[tilakrayal]

I was able to reproduce the issue on TensorFlow v2.15, v2.16 and tf-nightly. Kindly find the gist and also the screenshot for the reference.