NVIDIA RIVA ASR going down frequently for en-US

Hardware - GPU Tesla T4 (16GB)
Operating System - ubuntu
Riva Version - 2.6.0
model version: citrinet-1024-en-US-asr-streaming

Source of audio input: System mic
Audio metadata:
sampleRateHz: 16000,
format: ‘raw’,
language: ‘en-US’,
type: ‘start’,
encoding: ‘LINEAR16’

During conversion of Speech to text the riva service is stopping frequently with logs below:

Log1----------------------------------------------------------------
I0203 15:57:50.533809 205 grpc_riva_asr.cc:976] Using model citrinet-1024-en-US-asr-streaming for inference
I0203 15:57:50.533874 205 grpc_riva_asr.cc:992] Model sample rate= 16000 for inference
I0203 15:57:50.664098 205 riva_asr_stream.cc:214] Detected format: encoding = 1 RAW numchannels = 1 samplerate = 16000 bitspersample = 16
munmap_chunk(): invalid pointer
/opt/riva/bin/start-riva: line 55: 200 Aborted (core dumped) riva_server --triton_uri=localhost:8001 $ngc_args $@
One of the processes has exited unexpectedly. Stopping container.
Signal (15) received.
I0203 15:58:08.164999 102 server.cc:250] Waiting for in-flight requests to complete.
I0203 15:58:08.165020 102 server.cc:266] Timeout 30: Found 3 model versions that have in-flight inferences

Log2---------------------------------------------------------------------
I0203 16:22:08.551542 192 grpc_riva_asr.cc:976] Using model citrinet-1024-en-US-asr-streaming for inference
I0203 16:22:08.551601 192 grpc_riva_asr.cc:992] Model sample rate= 16000 for inference
I0203 16:22:08.648134 192 riva_asr_stream.cc:214] Detected format: encoding = 1 RAW numchannels = 1 samplerate = 16000 bitspersample = 16
malloc(): invalid next size (unsorted)
/opt/riva/bin/start-riva: line 55: 187 Aborted (core dumped) riva_server --triton_uri=localhost:8001 $ngc_args $@
One of the processes has exited unexpectedly. Stopping container.
Signal (15) received.
I0203 16:22:12.283945 91 server.cc:250] Waiting for in-flight requests to complete.
I0203 16:22:12.283969 91 server.cc:266] Timeout 30: Found 0 model versions that have in-flight inferences

Hi @scsourav18

Thanks for your interest in Riva,

Apologies you are facing issues

I will check with the internal team and get back soon

Thanks