VCTK_British_English_Females / prepare_model.py

updating prepend filenames

3b1bb15 about 1 year ago

3.63 kB

	import json
	import os
	import subprocess

	# Load the data from the provided dictionary
	data = {
	"VCTK_p225": {"age": 23, "gender": "F", "accents": "English", "region": "Southern England", "comments": ""},
	"VCTK_p228": {"age": 22, "gender": "F", "accents": "English", "region": "Southern England", "comments": ""},
	"VCTK_p229": {"age": 23, "gender": "F", "accents": "English", "region": "Southern England", "comments": ""},
	"VCTK_p230": {"age": 22, "gender": "F", "accents": "English", "region": "Stockton-on-tees", "comments": ""},
	"VCTK_p231": {"age": 23, "gender": "F", "accents": "English", "region": "Southern England", "comments": ""},
	"VCTK_p233": {"age": 23, "gender": "F", "accents": "English", "region": "Staffordshire", "comments": ""},
	"VCTK_p236": {"age": 23, "gender": "F", "accents": "English", "region": "Manchester", "comments": ""},
	"VCTK_p239": {"age": 22, "gender": "F", "accents": "English", "region": "Southwest England", "comments": ""},
	"VCTK_p240": {"age": 21, "gender": "F", "accents": "English", "region": "Southern England", "comments": ""},
	"VCTK_p244": {"age": 22, "gender": "F", "accents": "English", "region": "Manchester", "comments": ""},
	"VCTK_p250": {"age": 22, "gender": "F", "accents": "English", "region": "Southeast England", "comments": ""},
	"VCTK_p257": {"age": 24, "gender": "F", "accents": "English", "region": "Southern England", "comments": ""},
	"VCTK_p267": {"age": 23, "gender": "F", "accents": "English", "region": "Yorkshire", "comments": ""},
	"VCTK_p268": {"age": 23, "gender": "F", "accents": "English", "region": "Southern England", "comments": ""},
	"VCTK_p269": {"age": 20, "gender": "F", "accents": "English", "region": "Newcastle", "comments": ""},
	"VCTK_p276": {"age": 24, "gender": "F", "accents": "English", "region": "Oxford", "comments": ""},
	"VCTK_p277": {"age": 23, "gender": "F", "accents": "English", "region": "Northeast England", "comments": ""},
	"VCTK_p282": {"age": 23, "gender": "F", "accents": "English", "region": "Newcastle", "comments": ""}
	}

	# Convert the data to JSON format
	json_data = json.dumps(data, indent=2)

	# Save the JSON data to a file
	with open('speakers-log.json', 'w') as file:
	file.write(json_data)

	# Run the TTS command to get the speaker indices
	command = "tts --model_path checkpoint_85000.pth --config_path config.json --list_speaker_idxs \| grep -vE '^(\s\\|\|\s>\|\s*$)'"
	output = subprocess.check_output(command, shell=True, text=True)

	# Parse the JSON output into a Python dictionary
	speaker_indices = eval(output)

	# Load the speaker IDs from speakers.json
	with open('speakers-log.json', 'r') as file:
	speaker_ids = json.load(file)


	for speaker_idx in speaker_indices:
	# # Remove the 'VCTK_' prefix
	speaker_id = speaker_idx
	# speaker_id = speaker_idx.replace('VCTK_', '')

	# Lookup the speaker ID in the loaded speaker IDs
	if speaker_id in speaker_ids:
	speaker_id_json = speaker_ids[speaker_id]
	else:
	continue

	# # Generate the TTS command to create the audio file
	# text = f"Hello, I am from {speaker_id_json['region']}. I hope that you will select my voice for your project. Thank you."
	# # make samples directory if it doesn't exist
	# if not os.path.exists("samples"):
	# os.makedirs("samples")

	# out_path = f"samples/{speaker_id}.wav"
	# tts_command = f"tts --text \"{text}\" --model_path checkpoint_85000.pth --language_idx en --config_path config.json --speaker_idx \"VCTK_{speaker_id}\" --out_path {out_path}"

	# Execute the TTS command
	# os.system(tts_command)