Originally created by @PPTG on GitHub (Aug 8, 2024).
Why doesn't whisper(local) use my GPU processors, while ollama uses them correctly. nvidia-smi doesn't show any process during speech processing :( My docker file: services: open-webui: image: ghcr.io/open-webui/open-webui:ollama container_name: open-webui ports: - "3000:8080" - "11434:11434" volumes: - ollama:/root/.ollama environment: - WHISPER_MODEL=large - AUDIO_TTS_ENGINE=openai - AUDIO_TTS_VOICE=gosia - AUDIO_TTS_OPENAI_API_BASE_URL=http://openedai-speech:8111/v1 - AUDIO_TTS_OPENAI_API_KEY=sk-111111111 deploy: resources: reservations: devices: - driver: nvidia count: all capabilities: [gpu] restart: always
No dependencies set.
The note is not visible to the blocked user.
Originally created by @PPTG on GitHub (Aug 8, 2024).
Why doesn't whisper(local) use my GPU processors, while ollama uses them correctly. nvidia-smi doesn't show any process during speech processing :( My docker file:
services:
open-webui:
image: ghcr.io/open-webui/open-webui:ollama
container_name: open-webui
ports:
- "3000:8080"
- "11434:11434"
volumes:
- ollama:/root/.ollama
environment:
- WHISPER_MODEL=large
- AUDIO_TTS_ENGINE=openai
- AUDIO_TTS_VOICE=gosia
- AUDIO_TTS_OPENAI_API_BASE_URL=http://openedai-speech:8111/v1
- AUDIO_TTS_OPENAI_API_KEY=sk-111111111
deploy:
resources:
reservations:
devices:
- driver: nvidia
count: all
capabilities: [gpu]
restart: always