Getting error when trying to use Speech to Text on CUDA container #2395

New Issue

2025-11-11T15:06:20-06:00

GiteaMirror commented

2025-11-11 15:06:20 -06:00

Originally created by @nikito on GitHub (Oct 16, 2024).

Bug Report

Installation Method

Pod in k8s

Environment

Open WebUI Version: 0.3.32
Ollama (if applicable): 0.3.13 (running on separate pod)
Operating System: k8s in Ubuntu 22.04
Browser (if applicable): Chrome

Confirmation:

[ x] I have read and followed all the instructions provided in the README.md.
[x ] I am on the latest version of both Open WebUI and Ollama.
I have included the browser console logs.
I have included the Docker container logs.
I have provided the exact steps to reproduce the bug in the "Steps to Reproduce" section below.

Expected Behavior:

Speech to Text should successfully process the speech.

Actual Behavior:

Speech to text fails, throwing error in backend, causing app to crash and restart.

Description

Bug Summary:

When using the CUDA enabled container, upon trying to perform speech to text, the app crashes, and I see this error in my logs:
Could not load library libcudnn_ops_infer.so.8. Error: libcudnn_ops_infer.so.8: cannot open shared object file: No such file or directory

Looking in the container, it appears all my libraries are the .9 version, not .8. This is likely due to my cluster using a higher version of CUDA. Is it possible to make the app compatible with the later versions of CUDA as well?

Reproduction Details

Steps to Reproduce:
Install the cuda variant of the container, with latest CUDA enabled container.

Logs and Screenshots

Browser Console Logs:

Docker Container Logs:

<html> 2024-10-16 08:19:39.141 | Could not load library libcudnn_ops_infer.so.8. Error: libcudnn_ops_infer.so.8: cannot open shared object file: No such file or directory | -- | -- | -- | | 2024-10-16 08:19:38.594 | transcribe /app/backend/data/cache/audio/transcriptions/153c6178-0da2-4c9d-9687-68590fda567a.wav </html>

Originally created by @nikito on GitHub (Oct 16, 2024). # Bug Report ## Installation Method Pod in k8s ## Environment - **Open WebUI Version:** 0.3.32 - **Ollama (if applicable):** 0.3.13 (running on separate pod) - **Operating System:** k8s in Ubuntu 22.04 - **Browser (if applicable):** Chrome **Confirmation:** - [ x] I have read and followed all the instructions provided in the README.md. - [x ] I am on the latest version of both Open WebUI and Ollama. - [ ] I have included the browser console logs. - [ ] I have included the Docker container logs. - [ ] I have provided the exact steps to reproduce the bug in the "Steps to Reproduce" section below. ## Expected Behavior: Speech to Text should successfully process the speech. ## Actual Behavior: Speech to text fails, throwing error in backend, causing app to crash and restart. ## Description **Bug Summary:** When using the CUDA enabled container, upon trying to perform speech to text, the app crashes, and I see this error in my logs: Could not load library libcudnn_ops_infer.so.8. Error: libcudnn_ops_infer.so.8: cannot open shared object file: No such file or directory Looking in the container, it appears all my libraries are the .9 version, not .8. This is likely due to my cluster using a higher version of CUDA. Is it possible to make the app compatible with the later versions of CUDA as well? ## Reproduction Details **Steps to Reproduce:** Install the cuda variant of the container, with latest CUDA enabled container. ## Logs and Screenshots **Browser Console Logs:** **Docker Container Logs:** <html> <body>  2024-10-16 08:19:39.141 | Could not load library libcudnn_ops_infer.so.8. Error: libcudnn_ops_infer.so.8: cannot open shared object file: No such file or directory | -- | -- | -- | | 2024-10-16 08:19:38.594 | transcribe /app/backend/data/cache/audio/transcriptions/153c6178-0da2-4c9d-9687-68590fda567a.wav  </body> </html>

GiteaMirror closed this issue