Riva ASR issue on transcribing demo audio

AI & Data Science Deep Learning (Training & Inference) Riva
1

Please provide the following information when requesting support.

Hardware - GPU Tesla P100
Hardware - CPU Intel(R) Xeon(R) CPU @ 2.30GHz
Operating System Debian GNU/Linux 11
Riva Version v 2.10.0
How to reproduce the issue ? (This is for errors. Please share the command and the detailed log here)

I follow the Riva Quick Start Guide to deploy the ASR models on the sever. When running “Offline recognition for English”, the issue, as shown in the attached snapshot, pops up. All previous steps are executed without error.

Specific configs in the Riva server config.sh file are

service_enabled_asr=true
service_enabled_nlp=false
service_enabled_tts=false
service_enabled_nmt=false

language_code=("en-US")

asr_acoustic_model=("conformer_xl")

Could you please troubleshoot the issue?

Screenshot 2023-04-19 at 4.13.43 PM
Screenshot 2023-04-19 at 4.13.43 PM1524×886 68.8 KB

2

Hi @xin.zhang3

Thanks for your interest in Riva

I will reproduce from my end and check this further with the team,

request to kindly provide

  1. config.sh used attached as file in this thread
  2. complete log output of docker logs riva-speech as file
  3. Client python script/notebook used (Added in your screenshot as file)

Thanks

3

Hi there,

Thanks for the quick response!

I have attached 3 files below as requested. Can you please take a look?

asr-basics.ipynb (192.2 KB)
config.sh (12.6 KB)
log1.txt (22.3 KB)

4

Hi @rvinobha

All required files have been uploaded. Could you please take a look?

Thanks!