Add an experimental implementation of Attention VAD

Add a small cleanup on how to update `ctx.textarea`

Add initial cut at annotating meta speaker information

Enable TimelinePlugin()

Rename clear_active_region() -> unset_active_region()

Enable users to upload their audio files

Filter out a region if it has no length

Add textarea to display the annotation result

Add a simple favicon to 'voice-web'

Add first cut at 'voice-web' application

Import wavesurfer.js v7.10.1

Document 'silero-vad/train.py' in README.md

Reduce the output tensor dimension from 3 to 1

Use safetensors instead of torch.save

Add the initial revision of 'silero-vad/train.py'