All setups
Set up this box
A plain box with Whisper on disk
A GPU and a shell, with the Whisper large-v3-turbo weights already pulled down. Nothing else installed.
Set up this boxWorkload
Plain boxes
Preinstalled
Nothing preinstalled
Weights on disk
openai/whisper-large-v3-turbo
What you get
- A GPU and a shell. No template, no service started for you.
- The Whisper large-v3-turbo weights downloaded into the box before you open it.
- You install your own transcription stack: whisper.cpp, faster-whisper, transformers, whatever you already use.
- That install, and your audio, stay on the box's disk between sessions.
The weights are downloaded onto the box before you open it. Nothing is running them for you. You bring your own code, or use whatever the box has installed. For what card this needs and what it costs an hour right now, see the GPU requirements for openai/whisper-large-v3-turbo.
Why this one is here
Its GPU-requirements page took 21 impressions at average position 23.5 over 28 days, and the three openai/whisper pages took 56 between them.
