Alun King alunkingusw

Lecturer in Computer Science at University of South Wales

alunkingusw / gist:75ac37c44ca5b00ed9fdd5f50b391afa

Created June 20, 2025 20:12

Whisper & Pyannote Transcription segment by segment

	import logging

	# Configure logging level and format
	logging.basicConfig(
	level=logging.INFO, # options: DEBUG, INFO, WARNING, ERROR, CRITICAL
	format='%(asctime)s [%(levelname)s] %(message)s',
	handlers=[
	logging.StreamHandler() # prints to console
	]
	)

alunkingusw / gist:2eb29682a98f94a714d10080ed0f4896

Last active June 28, 2025 06:09

Speaker diarisation with Whisper & Pyannote using pyannote-whisper. Then speaker labelling using pyannote/embedding

	import logging

	# Configure logging level and format
	logging.basicConfig(
	level=logging.INFO, # options: DEBUG, INFO, WARNING, ERROR, CRITICAL
	format='%(asctime)s [%(levelname)s] %(message)s',
	handlers=[
	logging.StreamHandler() # prints to console
	]
	)

alunkingusw / gist:4275a26ba79cefd49f5a2e91d91c4da2

Last active June 16, 2025 08:42

Performing diarization on an audio file and comparing embeddings of known speakers.

	!pip install -qq pyannote.audio
	from datetime import datetime
	from pyannote.core import Segment
	from subprocess import CalledProcessError, run
	from pyannote.audio import Pipeline
	from pyannote.audio.core.model import Model # Import Model
	import os
	import torch # Import torch
	from pyannote.audio import Inference
	from sklearn.preprocessing import normalize