One tool for the full dataset pipeline — resize, caption, align, generate, mask, filter, degrade, train LoRA adapters, train TTS voices, upload to HuggingFace, and automate with workflows.
pip install datasety # core
pip install datasety[all] # everything# 1. Resize raw photos
datasety resize -i ./raw -o ./dataset -r 1024x1024
# 2. Generate captions with a trigger word
datasety caption -i ./dataset -o ./dataset --trigger-word "[trigger]"
# 3. Generate face masks for focused training
datasety mask -i ./dataset -o ./masks -k "face,hair"Or define it as a workflow:
# datasety.yaml
steps:
- command: resize
args: { input: ./raw, output: ./dataset, resolution: 1024x1024 }
- command: caption
args: { input: ./dataset, output: ./dataset, trigger-word: "[trigger]" }
- command: mask
args: { input: ./dataset, output: ./masks, keywords: "face,hair" }datasety workflow --dry-run # validate
datasety workflow # execute