Aborted (core dumped) in `tf.raw_ops.ResourceApplyRMSProp/tf.raw_ops.ResourceSparseApplyRMSProp`

[x0w3n]

Issue type

Bug

Have you reproduced the bug with TensorFlow Nightly?

Yes

Source

source

TensorFlow version

tf 2.16

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.ResourceApplyRMSProp/tf.raw_ops.ResourceSparseApplyRMSProp triggers "Aborted (core dumped)". We ran the code on colab's latest TensorFlow Nightly, which also triggers the crash. The cause of the crash may be the grad parameter.

Standalone code to reproduce the issue

import tensorflow as tf

var = tf.Variable([1.0, 2.0, 3.0])
accum = tf.Variable([0.1, 0.2, 0.3], dtype=tf.complex64) 
accum2 = tf.Variable([0, 1.2, -3.3], dtype=tf.complex64) 
lr = 0.01
grad = tf.constant([0.1, 0.2, 0.3])
momentum = 0.9

# crash
tf.raw_ops.ResourceApplyRMSProp(
    var=var.handle,
    ms=accum.handle,
    mom=accum2.handle,
    lr=lr,
    rho=lr,
    momentum=momentum,
    epsilon=0.9,
    grad=grad,
    use_locking=False,
)

# tf.raw_ops.ResourceSparseApplyRMSProp(
#     var=var.handle,
#     ms=accum.handle,
#     mom=accum2.handle,
#     lr=lr,
#     rho=lr,
#     momentum=momentum,
#     indices=tf.constant([1,2,2]),
#     epsilon=0.9,
#     grad=grad,
#     use_locking=False,
# )

Relevant log output

2024-06-06 01:07:00.757682: I tensorflow/core/platform/cpu_feature_guard.cc:210] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2024-06-06 01:07:01.835541: F tensorflow/core/framework/tensor.cc:844] Check failed: dtype() == expected_dtype (8 vs. 1) float expected, got complex64
Aborted (core dumped)

[sushreebarsa]

@x0w3n I was able to replicate the issue reported here using TF v2.16 and tf-nightly. Thank you!