Update image to the latest version of faster whisper

[fishcakeday]

• Use CUDA 12.6 and Ubuntu 24.04
• Use venv
• Update to Faster Whisper 1.1.1
• Remove other than large-v3 and distil-large-v3 models to minimize container size
• Add batch_size and allow for faster parallel processing
• Make VAD used by default
• Add multilingual switch to allow for multilingual text and transcription
• Make model use large-v3 as default model
• Fix schema data types to allow None for some string parameters

[ef0xa]

Thanks for your submission! I'm in the middle of reworking this significantly, so I'm not going to accept it as is, but i will definitely look this over and try and use the parts I can.