Skip to content

Instantly share code, notes, and snippets.

@fakerybakery
Created December 6, 2025 02:04
Show Gist options
  • Select an option

  • Save fakerybakery/38c3ee0deaa5d3ee270b7aad387e5e3c to your computer and use it in GitHub Desktop.

Select an option

Save fakerybakery/38c3ee0deaa5d3ee270b7aad387e5e3c to your computer and use it in GitHub Desktop.
import json
from glob import glob
from tqdm import tqdm
import os
from datasets import Dataset, Audio
files = [f for f in glob('tagged/output_gemini/*.json') if not f.endswith('.timestamps.json') and os.path.exists(f.replace('.json', '.timestamps.json'))]
datas = []
for f in files:
with open(f) as fp:
data = json.load(fp)
with open(f.replace('.json', '.timestamps.json')) as tfp:
tdata = json.load(tfp)
voice = data['voice'].strip()
text = data['text'].strip()
caption = tdata['items'][0]['caption'].replace('S1:', '').strip()
path_mp3 = f.replace('.json', '.mp3')
path_wav = f.replace('.json', '.wav')
path = path_mp3 if os.path.exists(path_mp3) else path_wav
datas.append({
'voice': voice,
'text': text,
'caption': caption,
'audio': path
})
ds = Dataset.from_list(datas).cast_column('audio', Audio())
ds.push_to_hub('mrfakename/emoact_prompts')
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment