# Automatic-speech-recognition problems

**URL:** <https://discuss.huggingface.co/t/automatic-speech-recognition-problems/42037>\
**Category:** Models\
**Created:** [June 4, 2023, 3:00pm UTC](https://discuss.huggingface.co/t/automatic-speech-recognition-problems/42037 "2023-06-04T15:00:40Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Superchik](https://avatars.discourse-cdn.com/v4/letter/s/34f0e0/32.png) [@Superchik](https://discuss.huggingface.co/u/Superchik)\
**Post date:** [June 4, 2023, 3:00pm UTC](https://discuss.huggingface.co/t/automatic-speech-recognition-problems/42037/1 "2023-06-04T15:00:40Z")

</div>

Infinity response of {‘error’: ‘Model \< automatic-speech-recognition model name\>is currently loading’, ‘estimated\_time’: 20.0}  
I tried to wait estimated\_time. Rarely it helps, but almost always not.

Why the input language is Russian, it understands it and translates it into English and sends it to me.  
Words are lost when translated into English  
Can I get speech recognition without translation?

My suggestions to `automatic-speech-recognition`:

1. Add `wait_for_model`  
This will allow you to avoid sending an endless number of queries to find out if the model is turned on or not.

2. Send the correct time before starting the model (`'estimated_time'`)  
This will let you know if you need to wait for the model to start or if you can skip this task.

3. Add `output_language`  
This will allow you to select the output language instead of English